Casting Design and Performance

Author: ASM International

335 downloads 975 Views 26MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Casting Design and Performance

Materials Park, Ohio 44073-0002 www.asminternational.org

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Copyright # 2009 by ASM InternationalW All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of the copyright owner. First printing, November 2009

Great care is taken in the compilation and production of this book, but it should be made clear that NO WARRANTIES, EXPRESS OR IMPLIED, INCLUDING, WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, ARE GIVEN IN CONNECTION WITH THIS PUBLICATION. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM’s control, ASM assumes no liability or obligation in connection with any use of this information. No claim of any kind, whether as to products or information in this publication, and whether or not based on negligence, shall be greater in amount than the purchase price of this product or publication in respect of which damages are claimed. THE REMEDY HEREBY PROVIDED SHALL BE THE EXCLUSIVE AND SOLE REMEDY OF BUYER, AND IN NO EVENT SHALL EITHER PARTY BE LIABLE FOR SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES WHETHER OR NOT CAUSED BY OR RESULTING FROM THE NEGLIGENCE OF SUCH PARTY. As with any material, evaluation of the material under end-use conditions prior to speciﬁcation is essential. Therefore, speciﬁc testing under actual conditions is recommended. Nothing contained in this book shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this book shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement. Comments, criticisms, and suggestions are invited, and should be forwarded to ASM International. Prepared under the direction of the ASM International Technical Book Committee (2008–2009), Lichun L. Chen, Chair. ASM International staff who worked on this project include Scott Henry, Senior Manager of Product and Service Development; Steven Lampman, Technical Editor; Ann Britton, Editorial Assistant; Bonnie Sanders, Manager of Production; Madrid Tramble, Senior Production Coordinator; and Diane Whitelaw, Production Coordinator. Library of Congress Control Number: 2009935431 ISBN-13: 978-0-87170-724-6 ISBN-10: 0-87170-724-1 SAN: 204-7586

ASM InternationalW Materials Park, OH 44073-0002 www.asminternational.org Printed in the United States of America

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Contents Design Problems Involving Junctions . . . . . . . . . . . . . . . . . . 147 Design Problems Involving Distortion . . . . . . . . . . . . . . . . . 155

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Part I: Casting Design Principles and Practices Casting Design Issues and Practices . . . . . . . . . . . . . . . . . . . . . 1 Casting Design and Processes . . . . . . . . . . . . . . . . . . . . . . . . . 9 Modeling of Casting and Solidiﬁcation Processing . . . . . . . . . 37 Part II: Process Design Riser Design . . . . . . . . . . . . . . . . . . Gating Design . . . . . . . . . . . . . . . . . Design for Economical Sand Molding Design for Economical Coring . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. 61 . 73 . 81 . 89

Part III: Design and Geometry Casting Design and Geometry . . . . . . . . . . . . Design Problems Involving Thin Sections. . . . Design Problems Involving Uniform Sections . Design Problems Involving Unequal Sections .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

101 121 133 139

Part IV: Casting Performance Corrosion of Cast Irons . . . . . . . . . . . . . . . . . . Corrosion of Cast Carbon and Low-Alloy Steels . Corrosion of Cast Stainless Steels . . . . . . . . . . . Fatigue and Fracture Properties of Cast Irons . . . Fatigue and Fracture Properties of Cast Steels. . . Fatigue and Fracture Properties of Aluminum Alloy Castings . . . . . . . . . . . . . . . . . . . . . . . Friction and Wear of Cast Irons . . . . . . . . . . . . Friction and Wear of Aluminum-Silicon Alloys . Failure Analysis of Castings . . . . . . . . . . . . . . . Inspection of Castings . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

163 171 175 185 199

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

209 219 227 237 247

Appendix: Classiﬁcation of Casting Defects . . . . . . . . . . . . . . 251 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

iii

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Preface Component geometry is a powerful aspect of casting design in terms of both effective production and the function of a cast part. Many tools have been developed for casting design, and the examples of past designs, even those of years ago, provide an important baseline in producing effective castings. Computer modeling and simulation has greatly facilitated the design process, and the article “Modeling of Casting and Solidification Processing” provides an extensive review of the subject. In addition, the complex aspects of configuration design are detailed in a series of articles in the sections on “Process Design” and “Design and Geometry.” Several of these articles are based on the ASM publication Casting Design Handbook (1962), which has been out-of-print for many years. Nonetheless, the lessons are still relevant today, as the basic fundamentals of geometry, metallurgy, and physics remain unchanged (even within the view of new modern perspectives and the advent of more powerful analytical or numerical tools). It is noted that the distinct sections on “Process Design” and “Design and Geometry” are a somewhat artificial division of topics, because really both process and design are intertwined in complex ways, especially for castings. In a sense, the “design” is like the fulcrum that leverages these two important aspects of castings into effective products. It is hoped that this collection of articles provides a useful reference on casting design. Finally, the performance of cast products is covered in a series of articles in the last section. Ultimately, the performance of product determines its success.

S. Lampman February 2009

v

ASM International is the society for materials

engineers and scientists, a worldwide network dedicated to advancing industry, technology, and applications of metals and materials. ASM International, Materials Park, Ohio, USA www.asminternational.org This publication is copyright © ASM International®. All rights reserved. Publication title

Product code

Casting Design and Performance

05263G

To order products from ASM International: Online Visit www.asminternational.org/bookstore Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US) Fax 1-440-338-4634 Mail

Customer Service, ASM International 9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA

Email [email protected]

American Technical Publishers Ltd. 27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX, In Europe United Kingdom Telephone: 01462 437933 (account holders), 01462 431525 (credit card)

www.ameritech.co.uk Neutrino Inc. In Japan Takahashi Bldg., 44-3 Fuda 1-chome, Chofu-Shi, Tokyo 182 Japan Telephone: 81 (0) 424 84 5550 Terms of Use. This publication is being made available in PDF format as a benefit to members and customers of ASM International. You may download and print a copy of this publication for your personal use only. Other use and distribution is prohibited without the express written permission of ASM International. No warranties, express or implied, including, without limitation, warranties of merchantability or fitness for a particular purpose, are given in connection with this publication. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this information. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this publication shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement.

Casting Design and Performance Pages 1–8

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Casting Design Issues and Practices* H.W. Stoll, Northwestern University

DESIGN is the critical ﬁrst step in the development of cost effective, high quality castings. Designing a successful casting requires an integrated, concurrent engineering approach. It also requires systematic and structured use of sophisticated computer-aided design software and casting analysis and simulation software. In this paper, we discuss these and other issues impacting good casting design. In particular, design decisions that drive casting performance and cost are identiﬁed and used as the basis for a proposed holistic approach to casting design. This new design philosophy and methodology is aimed at optimizing both the structural performance and producibility of the casting while minimizing design time and effort. In casting, the part being produced and the tooling used to produce the part interact in complex ways, which effect both the quality and cost of the casting. This suggests that the design of a successful casting requires an integrated approach that considers both functional and process requirements simultaneously. In the traditional setting, however, the design engineer typically ﬁrst decides the geometry of a casting and then a casting engineer independently develops the mold and process. This decoupled arrangement often results in more costly castings with larger safety factors than necessary. In addition, lead times for casting development using this approach can be excessive. As manufactures seek to reduce weight and cost of products, casting has re-emerged as the manufacturing process of choice in many situations. This is because casting offers the important advantage of being able to produce highly complex functional shapes quickly and easily. Cost is reduced because numerous parts and complex construction and processing typically associated with built up structures and weldments can be replaced by a single cast part. Weight is reduced because material can be distributed to where it is needed and because sections can be thinner since load does not transfer across part interfaces (i.e., through fasteners or welds). To take advantage of these unique capabilities, castings must be properly conceived and designed from the start. This means that design

engineers and casting engineers will need to change the casting design and development process from the historical sequential, decoupled approach to a more integrated and concurrent approach. This new approach must be based on the recognition that the part geometry not only affects the load carrying functionality of the casting, but also the mold construction, mold ﬁlling and material solidiﬁcation processes involved in producing the casting. These processes in turn affect cycle time, casting quality, and material properties such as yield strength, ultimate strength, and fatigue resistance. Casting geometry must therefore be determined based on both functional and processing requirements. Hence, design engineers must become more knowledgeable of the casting process, and casting engineers must have a better understanding of the functional requirements that drive the design. In essence, the new design philosophy must not only facilitate good part and process design, it must also teach it. This paper discusses the issues and practices associated with good casting design. The focus of the paper is on casting design in general, and on sand and permanent mold aluminum casting in particular. We begin by examining the casting design process from a variety of design and processing perspectives. Strategies for casting design process improvement are then proposed which provide the basis for a new casting design philosophy and methodology. Two possible implementations are presented. The ﬁrst is a structured team approach that is intended as a possible means for quickly improving traditional casting design practice by integrating the casting geometry and process design. The second is a longer-term approach involving the development of casting design guidelines for the design of lightweight, high quality, and high performance structural castings.

Geometry/Material/Process Interactions Carefully planned geometry is the secret to efﬁcient load carrying capacity and acceptable component performance. It is also the secret

*Reprinted from Proceedings from Materials Solutions Conference ’98 on Aluminum Casting Technology

for achieving high quality (i.e., uniform properties, soundness, etc.) and for avoiding costly and time consuming problems associated with pouring and solidiﬁcation of the alloy. Understanding how casting geometry, the material in both its liquid and solid phases, and the casting process interact provides the insight needed to specify the best casting geometry from a function, form, and fabrication point of view. In the following, we explore several geometry/ material/process interactions that are pivotal to good casting design. Much of this discussion is based on the paper “Cost Effective Casting Design — Developing a Conceptual Framework for Designing Metal Castings” by Michael Gwyn (Ref 1). The reader is referred to this paper for a more in-depth discussion of many of the topics brieﬂy referred to here. Fluid life is the ability of the molten alloy to ﬁll the mold cavity, ﬂow through thin narrow channels to form thin walls and sections, and conform to ﬁne surface detail. In addition to temperature of the molten metal, ﬂuid life also depends on chemical, metallurgical, and surface tension factors. Therefore, the ﬂuid life of each alloy is different. For example, aluminum 356 is considered to have excellent ﬂuid life whereas the ﬂuid life of aluminum 206 is only fair to good. Fluid life determines the minimum wall thickness and maximum length of a thin section. It also determines the ﬁneness of cosmetic detail that is possible. Hence, knowing that an alloy has limited ﬂuid life suggests that the part should feature softer shapes (i.e., generous radii, etc.), larger lettering, ﬁner detail in the bottom portion of the mold, coarser detail in the upper portions of the mold, more taper leading to thin sections, and so forth. Solidiﬁcation Shrinkage. Shrinkage occurs in three distinct stages: liquid shrinkage, liquid-to-solid shrinkage, and solid shrinkage. Liquid shrinkage is the contraction of the liquid before solidiﬁcation. Liquid-to-solid shrinkage or solidiﬁcation shrinkage is the shrinkage that occurs as the liquid’s disconnected atoms and molecules form into the crystals of atoms and chemical compounds that comprise the solid metal. Solid shrinkage is the shrinkage that occurs as the solid metal casting cools to

2 / Casting Design and Performance ambient temperature. Although liquid shrinkage is important to the metal caster, it is not an important design consideration. Solidiﬁcation shrinkage and solid shrinkage, on the other hand, are extremely important and must be carefully considered during casting design. Different alloys have differing amounts of liquid-to-solid shrinkage (e.g., aluminum 356 has little while aluminum 206 has moderate to large). Most importantly, there are three different types of solidiﬁcation shrinkage: directional, eutectic, and equiaxed. In alloys such as malleable iron and carbon steel, which solidify directionally, solidiﬁcation moves along predictable pathways determined by the casting geometry and thermal gradients in the mold. For example, solidiﬁcation will typically begin at the mold wall and move perpendicularly toward the center of the part. This is called progressive solidiﬁcation. Solidiﬁcation will also begin in cooler regions where the mold surface area to metal volume ratio is large and travel toward the hotter regions of the casting. This is called directional solidiﬁcation. The key is to conﬁgure the part geometry so that directional solidiﬁcation can occur before progressive solidiﬁcation shuts off the source of liquid metal supply (the riser). Without proper pathway geometry (e.g., risering and tapering), voids or pores due to isolated internal shrinkage can result. In eutectic-type solidiﬁcation, the liquid metal cools and then solidiﬁes very quickly all over. This behavior minimizes internal shrinkage and the need for risers and makes this type of alloy the most forgiving of the three. Eutectic-type materials that have very little solidiﬁcation shrinkage like gray iron often require no risering at all. The key geometric concern for eutectic-type solidifying alloys like aluminum 356 which have small but appreciable solidiﬁcation shrinkage is to ensure that the avenue of liquid metal supply stays open and functioning all the way to ﬁnal solidiﬁcation. In addition to solidifying both progressively and directionally from the mold walls, alloys that exhibit an equiaxed solidiﬁcation behavior also begin to solidify throughout the liquid, forming mushy regions consisting of equiaxed islands of solid. These equiaxed islands can block the avenues of liquid metal supply making these alloys difﬁcult to feed. To offset this tendency, regions that solidify in an equiaxed-type manner should be designed to have small thermal gradients, that is, to be as thermally neutral as possible. Therefore, thermal mass in these regions should be spread out and distributed uniformly throughout the region. This causes the shrinkage to be distributed as microscopic pores throughout the volume of the casting. Although the thought of having microscopic holes in the casting is disturbing, the effect on mechanical properties is greatly minimized by the small size, rounded shape, and uniform dispersion produced by using the proper geometry for this type of alloy. Also, a uniform dispersion of very small pores

is clearly preferable to large, irregular pores concentrated in possibly critical regions of the casting that could result from less appropriate part geometries. Solid shrinkage is often called pattern maker’s shrink because the tooling must be properly sized so that the part will shrink to the desired ﬁnal size and shape upon cooling. Solid shrinkage is critical for two important reasons. First, the shrinkage must be predicted and then built into the patterns/dies and corebox dimensions. If this is not done correctly, then the tooling will need to be modiﬁed iteratively to achieve an acceptable production casting. This adds time and cost to the design cycle and introduces quality risk in the ﬁnal product. Second, as the casting cools, it may not be able to shrink uniformly because some regions are stiffer than others. This can result in undesirable residual stresses and/or undesirable warpage. Creating geometry’s that make shrinkage predictable and that avoid residual stress and warping is therefore highly desirable. Slag/Dross Formation. Slag is typically composed of liquid nonmetallic compounds (usually ﬂuxed refractories), products of alloying, and products of oxidation in air. Dross refers to nonmetallic compounds produced primarily by the molten metal reacting with air. Some molten metal alloys are much more sensitive to slag/dross formation than others. Castings made from these alloys are much more likely to contain nonmetallic inclusions. In addition to process and quality control techniques, part geometry can be used to dramatically reduce the likelihood of nonmetallic inclusions. For example, for castings made from alloys that have buoyant slag/dross, the probability of having an inclusion in a critical machined surface can be reduced by designing the part so those surfaces are in the lower portion of the mold. Similarly, rigging design (i.e., conﬁguration of sprues, runners, and gates) can be designed to control the amount of oxidation that occurs due to turbulent ﬂow and entrained air. Pouring Temperature. Molds used in the casting process must withstand the extremely high temperature of molten metal. Often, proper casting geometry can help make the mold robust against this thermal abuse. Also, recognizing the undesirable effects of high pouring temperature on casting quality can help. For example, high pouring temperatures can lead to poor as-cast surface ﬁnish due to metal penetration into small sand cores. Therefore, when pouring temperature is high, it is often advisable to machine rather than core small holes and other small features. Fluid Flow. Another key geometry/process interaction involves the ﬂow of molten metal into the mold cavity. As mentioned previously, turbulent ﬂow through gates and other channels can effect the amount of oxidation and consequent dross formation that occurs. Another consideration is the force generated by the molten metal as it ﬂows into the mold cavity and the

turbulence of the ﬂow in the cavity since both of these effects can displace cores and erode mold walls, especially sharp edges and high detail features. Steep thermal gradients can also arise due to ﬂuid ﬂow. If the ﬂow separates to pass around cores and other features and the joins together again, weld lines, nonmetallic inclusions, and other ﬂaws can occur due to cooling and oxidation of the ﬂow front. In order to minimize undesirable effects of ﬂuid ﬂow, the casting must be poured slowly. Unfortunately, this gives the molten metal more time to oxidize and increases process cycle time. Undesirable interactions due to ﬂuid ﬂow effects can often be reduced or eliminated by designing the casting geometry and rigging as a system. Heat Transfer Considerations. The geometry must also be selected with an understanding of the heat transfer conditions involved. High pouring temperatures mean that large amounts of heat must be transferred into the mold. If the geometry is such that the heat cannot escape, a hot spot is likely to occur. For example, narrow peninsulas or tight corners of mold material surrounded by molten metal will get hot very quickly and as a result, solidiﬁcation of the molten metal in these regions will be slower than surrounding regions. This creates the possibility of hot tears or shrinkage pulls because the hotter material will have less tensile strength and is therefor less able to resist internal forces that develop due to solidiﬁcation and solid shrinkage. Voids can also form because liquid metal supply paths close off before the material in the region of the hot spot is fully solidiﬁed. Just the opposite situation occurs when sharp corners or narrow peninsulas of molten metal are surrounded by mold material. In these cases, the molten metal cools and solidiﬁes very quickly. This is generally a desirable situation. However, if cooling is too rapid, it can cause cold cracking due to stressing of the solidiﬁed skin or thin region by solidiﬁcation shrinkage occurring at a slower rate in more massive adjacent regions. Also, difﬁcult to machine or undesirable material properties may result from to rapid cooling of some alloys. Geometry/Alloy Interactions. In most cases, it is the combination of material properties possessed by a particular alloy that determines the most desirable casting geometry. For example, gray iron has a moderately high pouring temperature, excellent ﬂuid life, and very small, eutectic type solidiﬁcation shrinkage. As a result, accept for the danger of hot spots due to its relatively high pouring temperature, gray iron is very casting friendly. It’s excellent ﬂuid life permits ﬁne detail and thin sections and its low solidiﬁcation shrinkage provides considerable geometry latitude. Aluminum 356 has a low pouring temperature, excellent ﬂuid life, and more eutectic type than directional type solidiﬁcation. The low pouring temperature makes it an excellent candidate for precision castings. Also, its excellent

Casting Design Issues and Practices / 3 ﬂuid life permits ﬁne detail and thin walls everywhere. However, although still relatively small, solidiﬁcation shrinkage is signiﬁcant enough to warrant consideration especially with respect to risering, section size, and feeding pathways. Also of concern is the sensitivity of aluminum to dross formation. Dross can be a particular problem because the speciﬁc gravity of aluminum oxide is close to that of molten aluminum and hence buoyancy does not aid separation. Carbon steel is at the opposite end of the spectrum. Steel has a very high pouring temperature, poor ﬂuid life, and large, directional type solidiﬁcation shrinkage. This combination of material properties makes steel very casting unfriendly. As a result, careful attention to casting geometry is essential. Because of its poor ﬂuid life and high pouring temperature, ﬁne detail and thin sections are difﬁcult. Most importantly, because of its large solidiﬁcation shrinkage, feeding of the casting is a great concern. Risers need to be large and the geometry must be carefully designed to ensure proper feeding of the casting. Because of its unfavorable combination of properties, steel and materials like it, require softer shapes (i.e., large radii, rounded shape, large lettering, no sharp detail) compared to casting friendly materials such as gray iron.

Cost Drivers A second way to look at the design of a casting is to understand how design decisions regarding casting geometry drive total cost of the casting. By total cost, we mean the sum of all costs, both the direct and indirect, that result from the design, production, distribution, use, and salvage of the casting over its lifetime. Although all components of total cost are important, we are particularly concerned in this paper with how design decisions drive both the direct cost associated with production of the casting and the cost of designing, building, and proving out the tooling. This cost can be calculated on a per unit basis as follows, CT C0 tcycle Cost ¼ þ Cc þ V Cm þ þ Cs N Y

(Eq 1)

where: CT = total tooling cost ($) N = lifetime number of castings CC = cost of coring ($/unit) V = total casting volume (in3) Cm = alloy cost ($/in3) C0 = casting equipment and labor cost ($/hr) tcycle = total casting lead time (hr) Y = yield (useable castings/N) Cs = cost of secondary processing ($/unit) It is important to note that total tooling cost (CT) includes all cost associated with tooling including the cost of pattern and corebox construction, the cost of producing and inspecting the ﬁrst article, and the cost of iteratively modifying the tooling to meet speciﬁcations. Also, material volume (V)

includes not only the volume of the casting, but also the volume of the risers, runners, and sprues used to feed the casting. Total casting cycle time (tcycle) is given by the following, tcycle ¼ tnp þ tbuild þ tcast þ tcool þ ttrim

(Eq 2)

where: tnp = non-productive time (hr) tbuild = mold build time including core placement (hr) tcast = time to pour the casting (hr) tcool = time to cool to ambient temperature (hr) ttrim = time to remove gates, risers, etc. (hr) Since many castings involve more than one core, the per unit cost of coring (CC) is calculated as, Cc ¼

nc X C00 t0cycle 0 V 0 Cm þ 0 YC i i¼1

(Eq 3)

where: nC = number of cores V0 = volume of core material (in3) 0 = cost of core material ($/in3) Cm 0 C0 = core making equipment and labor cost ($/hr) 0 tcycle = core making cycle time (hr) YC0 = yield (useable cores/lifetime number of cores) Similarly, secondary processing may involve more than one process such as machining, heat treating, welding, painting, and plating. In addition, processes such as machining might involve several different operations (e.g., drilling, milling, grinding, etc.). The per unit cost of secondary processing is therefore calculated as Cs ¼

ns X C000 t00cycle CT00 þ YS00 i i¼1

(Eq 4)

where: ns = number of secondary processes CT00 = secondary process tooling cost ($/unit) C000 = secondary process equipment and labor cost ($/hr) 00 tcycle = secondary process cycle time (hr) YS00 = secondary process yield (useable castings/N) Equations 1 through 4 are very clear as to what should be done to reduce the per unit production and design cost:

Design to minimize tooling cost Design to minimize material cost Design to minimize process cycle time Design to maximize yield Minimize the number of cores and secondary processes

By looking at how geometry decisions effect the sources of cost in the above equations, it is possible to make geometry decisions that reduce cost. For example, both tooling cost and

production cost will be reduced by selecting the parting plane early in the design and then creating geometry that minimizes the number of undercuts and other features that must be cored. Locating riser and gate contacts at easy to access areas on the casting will reduce trimming time. Also, locating riser and gate contacts so that they don’t interfere with machining ﬁxture targets will reduce trimming time, the cost of the ﬁxture, and possibly machining cycle time because ﬁxturing will be easier and more consistent. Designing so that critical dimensions do not cross the parting line will decrease build time and increase yield since the positional accuracy required between the cope and drag is reduced.

Shape Optimization Metal casting offers two unique and very desirable design advantages: (1) metal mass can be located exactly where it is needed and (2) complex, three-dimensional geometry is readily created. By properly capitalizing on these advantages, part geometry can be created that minimize both weight and cost of the part. For example, the use of continuously varying section geometry that fully utilizes the material strength while also satisfying deﬂection requirements is readily achievable. In addition, many parts can be consolidated into one part, thereby eliminating piece part fabrication cost, assembly cost, and all the indirect costs, interfacing information, quality risk, and manufacturing complexity associated with built-up parts and weldments. Shape optimization is the design perspective that seeks to leverage these advantages. This practice has been greatly facilitated by the development of powerful engineering workstations and solid modeling software that signiﬁcantly enhances the engineer’s ability to visualize complex three-dimensional geometry and to analyze stress levels and deﬂections of complex three-dimensional shapes.

Rigging System Design The rigging system includes the system of sprues, runners, gates, risers, and chills that channel and control the ﬂow of liquid metal into the mold cavity, feed the casting as it solidiﬁes, and control the heat transfer and rate of solidiﬁcation in critical regions. Rigging system design speciﬁes the size, dimensions and location of sprues, runners, gates, risers, and chills that comprise the system. In the traditional approach, an expert casting engineer designs the rigging system, usually after the geometry of the casting has been speciﬁed. Rigging design decisions typically include selection of the following: orientation of the cast part, parting line, potential sites for chills and chill types, sprue height and location, runner types and conﬁguration, ingate sites, choke area (smallest cross-section area present in the ﬂow system), riser sites and conﬁguration, and pouring rate and temperature.

4 / Casting Design and Performance

Casting Process Simulation In casting process simulation, comprehensive modeling of the intended production process is performed in order to determine the size and shape of sprues, runners, gates, and risers. A variety of simulation software for performing this type of simulation is available. In addition, methodologies have been developed to understand and predict the size and location of process related defects (microporosity, etc.). See for example references [2–6]. Using these methodologies, the rigging system design can be varied in the foundry system simulation to evaluate how defect size and location are to be controlled and/or eliminated. Using computer simulation early in the design process can greatly reduce the amount of guess work involved in specifying cost effective and functionally acceptable casting geometry. Computer based casting process simulation offers two important advantages: (1) design iterations and what-if analysis are much easier to perform and (2) the physics engine underlying the simulation software provides a consistent and predictable science base for casting design. This allows the casting geometry and rigging system to be speciﬁed and optimized as a coordinated system. Most importantly, it allows evaluation of the overall design before tools are cut and the design is irreversibly committed to hardware. When used properly, the result is a substantial reduction in design time and tooling iterations. It is extremely important to note however, that the use of casting process simulation software is not, in itself, a viable substitute for early input of experienced tooling and foundry engineers. Rather, it is a very powerful tool that helps leverage and assist the team approach.

Casting Design Improvement Strategies Casting design is an iterative process (Fig. 1). The problem of design is typically formulated in terms of functional requirements and constraints

that must be satisﬁed. Functional requirements relate to the functions the part must provide while constraints relate to the form (shape, size, surface ﬁnish, precision, etc.) and processing (parting line, draft, section thickness, etc.) requirements that constrain the geometry that can be selected. Based on the problem formulation, an initial design is created. This design is then evaluated and modiﬁed iteratively until an acceptable design is achieved. Typically, the redesign is guided by the design information, insight, and understanding developed in the evaluation step. To be acceptable, the design must satisfy all functional requirements and constraints. When the traditional approach to casting design is examined, we see that the iterative process characterized by Fig. 1 is essentially repeated at least two and perhaps several times as shown in Fig. 2. First the design engineer goes through the iterative design process to specify the casting geometry. This geometry is then passed on to the casting engineer who repeats the iterative design process to specify the rigging system and apply pattern maker shrink to the casting dimensions. Problems discovered during rigging system design can generate additional iterations if casting geometry changes are required. Additional iterations to the casting geometry and rigging system design may also be required during tool fabrication and preparation for production of the ﬁrst article. Finally, iterative changes to the tooling and perhaps the casting and rigging system geometry may be necessary to tweak the design to meet production requirements. Excessive design iterations can adversely impact the casting design in two important ways. First, design iterations signiﬁcantly increase design cost and time. Second, design iterations, especially those performed late in the process, can lead to suboptimal design. The result is a casting that falls short of cost, weight, and performance targets. Such designs give casting a bad reputation and are a disappointment to all concerned. Several strategies for improving the traditional casting design process are possible based on the design perspectives discussed above. These are summarized as follows: 1. Design the casting geometry and casting process as a coordinated system by integrating shape optimization and rigging system design into one concurrent process. Consider geometry, material, and process interactions

Fig. 1

Iterative model of the design process

Fig. 2

Traditional casting design process

2.

3.

4. 5.

and design related cost drivers from the beginning as part of the process. Develop a thorough understanding of all customer needs including downstream processing constraints before beginning the design. Focus on creating an acceptable initial design. By spending the time up-front to create the best possible initial design, a large number of lengthy analyze-redesign iterations are avoided. The evaluation phase should conﬁrm the design rather than create it. Use casting process simulation and other modern computer-aided analysis and inspection methods to quickly optimize the design. Develop a consistent, well deﬁned science base for casting design in the form of casting design guidelines and structured methodologies.

The goal of these strategies is to shorten the design cycle and help ensure that the best possible casting design is created. In the sections that follow, we propose some possible approaches for implementing these strategies.

Structured Team Approach Strategies 1 through 3, and to some extent, strategy 4 can be implemented very easily and quickly by adopting a structured team approach. By a structured team approach, we mean that the casting design is performed using a multi-disciplinary team and a structured design methodology. The goal of the team approach is to have all required product and process knowledge available when the key early design decisions are being made. In a structured design methodology, the overall problem of design is broken down into a series of sequential, easier to perform steps that proceed from the general to the speciﬁc. Excessive iteration and long design times are avoided by performing each step in a thorough and disciplined manner. In general, each step in the process can be further subdivided into steps to create a hierarchy of structured methodologies. To illustrate the structured team approach, we propose the simple methodology shown schematically in Fig. 3. This approach recognizes that not all members of the team can be available for designing the casting on a continuous basis. Team meetings are therefore scheduled at which all salient aspects of the design are reviewed and discussed. All members of

Casting Design Issues and Practices / 5 a casting engineer, a tooling design engineer, and perhaps one or more specialists who are familiar with casting process simulation software, ﬁnite element analysis, fracture mechanics, non-destructive evaluation (NDE) techniques, and so forth. In addition, it is essential that the end customer, the foundry that is to make the casting, and others concerned with secondary processing that is to be farmed out be properly represented on the team. This not only helps ensure that all customer and processing needs are appropriately considered, it also makes it possible to rapidly negotiate changes to the design speciﬁcation when necessary, and to quickly assess cost consequences of design decisions. Meeting 1: Clarify the Design Problem. Clarifying the problem consists of developing a general understanding of the cost, performance, and manufacturing goals and constraints of the design. A typical agenda for this meeting might include the following:

Fig. 3

Simple structured team approach

the team must be present at these meetings. The purpose of the meetings is to establish design direction, make key design decisions that require input and consensus from all team members, make sure that all process constraints and requirements are being properly considered, and resolve conﬂicts and impediments to the proposed design. The outcome of each meeting is a set of action steps to be implemented by individual team members. In this way, all team members are kept informed and participate in the design decision making process. At the same time, the actual detail work of creating the design is delegated to speciﬁc team members according to the skills and knowledge required. In the following, we brieﬂy discuss each step of the methodology and imagine how a casting would be designed using this methodology. Step 1: Form Team. This is the pivotal ﬁrst step in the methodology. Unless the arrangement is formalized in some way, it often is difﬁcult to get effective collaboration between the design and casting engineer early in the casting design process, especially before the casting geometry is deﬁned. By being formally assigned to a team, each individual team member takes personal responsibility for the design from the beginning. This fosters and facilitates the kind of collaborative attitude that is essential for good casting design. All stakeholders who have an interest in the casting should be represented on the team. A typical team might include a design engineer,

1. Review product background. 2. Customer requirements and design objectives. 3. Expected annual production volumes and target costs. 4. Geometry concepts and alternatives. 5. Material and processing options. 6. Foundry and secondary processing locations. 7. Potential geometry/material/process interactions. 8. Develop a preliminary conﬁguration design. 9. Make assignments to team members to create the initial design. As mentioned above, it is essential that all team members be present at each team meeting. For example, although the analyst may not be actively involved with the design before the casting geometry is fully deﬁned, it is extremely important that he or she participate in the early design decisions that lead to the proposed geometry. In this way, the analyst knows all the needs of the design problem and is familiar with the reasoning behind the particular geometry that will eventually be analyzed and optimized. Step 2: Create the Initial Design. The initial design establishes the detail layout of the casting geometry. It includes the conﬁguration and parametric design of the part together with the rigging design required to ﬁll the mold cavity and feed the solidifying casting. Conﬁguration design involves the determination of what features such as walls, holes, ribs, etc. will be present and how these features will be connected to provide the desired form, ﬁt, function, and manufacturability (e.g., parting line, coring, and draft for low tooling cost, short cycle time, and minimal trimming and secondary processing). Parametric design involves the determination of dimensions, tolerances, and exact material speciﬁcations needed to meet durability, stiffness, and/or natural frequency targets. As discussed previously, rigging design involves the location and sizing of the sprues, runners, gates, and risers.

As a general approach, a preliminary conﬁguration and rigging design might be proposed in Meeting 1 by the team as a whole. Using this as a starting point, the design engineer and casting engineer would work together to develop the details of the conﬁguration design, seeking input and consensus from various team members as necessary. Once the initial casting and rigging conﬁguration has been ﬁrmed up, a preliminary parametric design would be performed. The goal of this task is to quickly determine section dimensions, secondary processing requirements, and material property requirements using simple strength of materials methods or, if necessary, a rough ﬁnite element analysis. Once the approximate parametric design is complete, the overall design is evaluated and modiﬁed to minimize cost. This is easy and straightforward to perform with a minimum of analysis and iteration because the casting geometry, secondary processing, and rigging system have been conceived and developed as a coordinated system with input from all team members. Hence, all of the information needed to make quality design decisions is readily available. Meeting 2: Reﬁne and Approve the Initial Design. The goal of this meeting is to react as a team to the initial design and to make any adjustments or modiﬁcations deemed necessary by general consensus of the team. This is the time when all design and processing issues should be discussed and resolved. If there are signiﬁcant impediments to the design as proposed, these should be resolved before proceeding to Step 3. A typical agenda for this meeting might include the following: 1. Review the initial design. 2. Identify impediments, potential undesirable interactions and performance and processing concerns. 3. Discuss all design-related costs to ensure that the best casting geometry from a total cost standpoint has been identiﬁed. 4. Make assignments to team members to work out solutions to various impediments and schedule a follow on meeting, or 5. Approve the proposed initial design and make assignments to team members to reﬁne and optimize the design. Step 3: Reﬁne and Optimize the Design. Once the team is conﬁdent that the initial casting geometry and rigging design is the best solution possible, the effort required to optimize details of the casting geometry and rigging system by computer analysis can be justiﬁed. The goal of this step is therefore to computer model the design and iteratively improve it until all aspects have been appropriately optimized. The amount of effort expended on this step will depend on how important it is to optimize the casting. For example, if weight and/or material cost is critical, extensive effort to minimize the amount of material used can be justiﬁed. Similarly, if safety is an important issue, comprehensive analysis to

6 / Casting Design and Performance ensure acceptable fatigue life and reliable detection of ﬂaws can be justiﬁed. The key to this step is to start with a casting geometry that is close to the optimum. This will minimize the time, analysis effort, and number of iterations required to converge to the optimum design. Meeting 3: Approve the Final Design. The result of Step 3 will be a fully speciﬁed casting design including the detailed casting geometry, parting plane, draft, coring, riser locations and sizes, and plumbing system design. In addition, the ﬁnished part design including machining, heat treating, and so forth will be fully speciﬁed. The purpose of Meeting 3 is to formally review the ﬁnished design as a team and approve the design for release to manufacturing. When the structured team approach is performed properly, the ﬁnal design will almost always be approved. However, if the team decides that the design is not ready to be released, then appropriate action plans for correcting design deﬁciencies must be developed and implemented. One or more follow-on meetings may then be required before the design is released. Although painful at times, not releasing the part until it is ready helps insure a minimum number of tooling changes and tweaks and, in the long run, is the most cost effective policy. By strictly adhering to this policy, the casting design should proceed quickly and smoothly to ﬁrst article and production with little or no modiﬁcation. When this is the case, the team knows that it has done its job well.

facilitate design for damage tolerance, design for manufacture, and design for inspection. 2. Capture and reduce to practice the latest research results quantifying the relationships between casting processes, microstructure characteristics, cyclic (fatigue) properties, and component function for cast aluminum and magnesium alloys. 3. Facilitate the development and implementation of design strategies that seek to leverage part family concepts (i.e., group technology) and to standardize casting geometry, rigging, and tooling features. 4. Help the design engineer make better casting design decisions when input from experienced casting engineers is not readily available. This goal recognizes that many companies may not have deep in-house casting expertise. As shown in Fig. 4, the guidelines would be organized in a manner similar to the structured team approach discussed previously. The guidelines associated with each step in the process are brieﬂy discussed as follows. Step 1: Clarify the Problem. The purpose of these guidelines would be to help the casting design team to properly understand the problem of design by considering all customer needs

including downstream tooling and processing needs. These guidelines might consist of a stepby-step structured methodology for gathering and evaluating needs together with one or more check lists formulated to ensure that all design requirements and constraints are considered. They would also include details regarding design strategy and standardization objectives. Ideally, the design process would not continue until all items have been checked off and appropriately evaluated. Step 2: Create the Initial Design. This step includes the iterative development of the conﬁguration design, parametric design, and rigging design. Guidelines for conﬁguration design would include high level design for light metal casting guidelines and recommended practices. Such guidelines will recognize the important metallurgical characteristics of the foundry alloy and mechanical engineering structural characteristics that drive the geometry of good casting design. These will include the geometry, material, and process interactions discussed earlier as well as best practices for communicating design information to the foundry and design for guidelines pertaining to downstream considerations such as casting inspection, heat treating, and machining. Guidelines for parametric design would include best dimensioning and tolerancing

Casting Design Guidelines Casting design guidelines make needed alloy and process information readily available to the design and casting engineer in a timely fshion and in a form that is easy to understand and apply. Such guidelines typically consist of design rules suggested best practices, structured design methods, check lists look-up tables, graphs and charts, and computerized design tools that can be used by the casting design team to perform each step of the design process. In this section, we present a vision for the development of comprehensive casting design guidelines. The goal of the envisioned guidelines is to facilitate the design of high performance castings that are also easily manufactured in large production quantities for a reasonable cost. Speciﬁc development objectives include the following. 1. Provide the sophisticated design information needed to design light metal cast structural component systems for use in safety critical automotive applications. Such castings must be as light as possible and also manufacturable in large production quantities for a reasonable cost. This implies that every section must be optimized to use an absolute minimum amount of material while satisfying durability, stiffness, natural frequency, and other structural and performance targets. The design guidelines must therefore

Fig. 4

Structure of guidelines for light metal casting design

Casting Design Issues and Practices / 7 practices, material properties including those relevant to the geometry, material, and process interactions discussed earlier as well as structural properties needed for proper consideration of damage tolerance, etc. The guidelines would follow existing best practices such as those given in standard texts on automotive component design and machine design (Ref 7 and 8). The major difference would be in the way material properties such as ﬁnite life fatigue strength would be estimated. Rather than using the traditional casting factor, the design guidelines would provide charts and property estimation procedures based on recent and developing research results (Fig. 5 and Ref 9 and 10). Rigging design would include the approximate conﬁguration and parametric design of the risers, gates, runners, chills, insulation, etc. to feed the casting, prevent porosity formation, and minimize cycle time. As discussed previously, rigging design plays an extremely important role in the development of a successful casting design. As such, it must be considered early in the design process to achieve the goals outlined above. In addition, many of the computer simulations to be applied in the computer analysis step (see Fig. 4) require that the rigging system be modeled along with the part. The inclusion of as a part of the casting design represents a major departure from traditional casting design practice. Hence, the development of comprehensive guidelines that both teach and ensure good practice is essential. One approach that might be considered is use of computer-aided rigging design tools such as that described in Ref (11). Once an approximate rigging system design has been developed, it becomes possible to consider the casting geometry and rigging system as a coordinated system. This represents an important opportunity to improve the overall initial design quickly with a minimum of analysis and iteration. Development of guidelines for this step might include extensive use of a computer automated rapid prototyping (CARP) environment such as that shown schematically in Fig. 6.

This graph illustrates how material properties might be expressed as a function of section geometry. Developing such graphs will obviously require a focused research effort.

Computer Analysis. This step is necessary to meet the stringent weight, performance, and cost goals described above. It would be performed using computer based analysis and redesign methods to facilitate rapid convergence to the optimum coordinated casting geometry and rigging system design. As discussed in Ref 12 the computer analysis might begin with a casting process evaluation to estimate local material properties and to determine if and where ﬂaws (porosity, etc.) will form. This would be followed by a casting inspectability evaluation to determine the smallest ﬂaw that can be reliably detected in critical using modern NDE inspection techniques. Results of these evaluations would then be used in a damage tolerance evaluation to predict component durability and fatigue characteristics. Using component weight as the objective function and performance, cost, and quality targets as the constraints, these evaluations would be incorporated in an iterative optimization procedure that would modify component and rigging system geometry until an acceptable minimum weight design is achieved. It is expected that the guidelines for this step would consist of a procedure for using standard, commercially available ﬁnite element and casting process analysis software together with special computer programs and math based simulation techniques to perform the optimization see Ref 12 and 14.

Summary A broad range of casting design issues and practices has been reviewed. Based on this, a new casting design philosophy and methodology is proposed. The structured team approach is intended as a possible means for quickly improving traditional casting design practice by including early consideration of downstream tooling and processing constraints early in the design process. The development of casting design guidelines represents a longer-term approach for reducing design cycle time and

Fig. 5

Fig. 6

Computer automated rapid prototyping environment (CARP) Source: Ref 13

improving the quality and performance of lightweight structural castings. The methodology used for both approaches is to comprehensively understand the problem of design, quickly deﬁne an initial casting geometry and rigging system design that is close to target requirements, and then reﬁne the total design iteratively as a coordinated system using computer based analysis tools. Both approaches are predicted on the underlying assumption that, given appropriate design and processing information and guidance at each step, the design will converge quickly to the best possible result. It is important to note that development of design guidelines such as those suggested in this paper require a substantial research and development undertaking. It is also important to point out that the proposed methodologies represent starting points for casting design process improvement. It is expected that the effectiveness and usefulness of these methods will increase as they are used and improved over time by casting design professionals and the foundry industry.

REFERENCES 1. M.A. Gwyn, “Cost Effective Casting Design,” presented at the AFS 101 Casting Congress, Preprint No. 97–147, American Foundrymen’s Society, Inc., Des Plaines, IL, 1997 2. E. Niyama, et.al., “A Method of Shrinkage Prediction and Its Application to Steel Casting Practice,” 49th International Foundry Congress, 1982 3. K. Tynelius, J. Major, and D. Apelian, “A Parametric Study of Microporosity in the A356 Casting Alloy System,” AFS Transactions, 1993, p 401–413 4. H. Huang, and J. Berry, “Evaluation Criteria Functions to Minimize Microporosity Formation in Long-Freezing Range Alloys,’ AFS Transactions, 1993, p 669–675 5. N. Tsumagari, et.al., “Construction and Application of Solidiﬁcation Maps for A356 and D357 Aluminum Alloys,” AFS Transactions, 1993, p 335–341 6. J. Huang, “Study of Criteria Function for Porosity Prediction in A356 Castings,” Masters Thesis, Mechanical Engineering, Northwestern University, Dec. 1995 7. Fatigue Design Handbook, 2nd Ed., SAE International. Warrendale, PA, 1988 8. J. Shigley, and C. Mischke, Mechanical Engineering Design, 5th Ed., McGraw-Hill, 1989 9. A.J. Hinkle, J.R. Brockenbrough, and J.T. Burg, “Microstructural Material Models for Fatigue Design of Castings,” SAE Paper 960161, SAE International, Warrendale, PA, 1996 10. J.M. Boileau, J.W. Zindel, and J.E. Allison, “The Effect of Solidiﬁcation Time on the Mechanical Properties in a Cast A356-T6

8 / Casting Design and Performance Aluminum Alloy,” SAE Paper 970019, SAE International. Warrendale, PA, 1997 11. N. Nanda, et.al., “Feature-Based Design of Gates and Risers in a Casting,” p 75–84, Knowledge-Based Applications in Materials Science and Engineering, Ed. by J.K. McDowell and K.J. Meltsner, The

Minerals, Metals, and Materials Society, 1994 12. J. Conley, B. Moran, and J. Gray, “A New Paradigm for the Design of Safety Critical Castings,” SAE Paper 980455, SAE International, Warrendale, PA, 1998 13. J. Conley, Private communication.

14. N. Palle, S. Singh, S. Mahadeva, R.J. Yang, and J.A. Dantzig, 1997, “An Optimization Based Methodology for Casting Process Design,” presented at the McNU ’97 Joint ASME, ASCE, and SES Summer Meeting, Northwestern University, June 29–July 2, 1997

Casting Design and Performance Pages 9–36

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Casting Design and Processes SHAPE CASTING, as a metal forming process, offers considerable ﬂexibility and wide scope in producing parts. Almost all metals and alloys can be cast with few restrictions on part weight or size, and casting is capable of producing highly reliable, cost-effective components ranging from low-volume, single-part prototype production runs to economies of scale for millions of parts. In terms of component design, metal casting is very ﬂexible in terms of conﬁguration design. Casting permits the formation of streamlined, intricate, integral parts of strength and rigidity that are not obtainable by other methods of fabrication. For example, the ﬂexibility of metal casting, particularly sand molding, may permit the use of difﬁcult design techniques, such as undercuts and curved or reﬂex contours that are not possible with other high-production processes. Tapered sections with thickened areas for bosses and generous ﬁllets are routine. The inherent design freedom of metal casting allows the designer to combine what would otherwise be several parts of a fabrication into a single piece. This is signiﬁcant when exact alignment must be held, as in highspeed machinery, machine tool parts, or engine end plates and housings that carry shafts. Nonetheless, there are a number of challenges in the design of castings. To begin, component geometry and properties are closely interrelated in casting design. Most castings involve complex conﬁgurations with sections of varying thicknesses, which inﬂuences solidiﬁcation rates and thus properties within various sections of a casting. For example, concave sections, or reentrant angles, solidify more slowly than ﬁns or protrusions, thus affecting the resultant local structure and the properties. In other words, property variations occur within a casting due to differences in cooling rates. These variations are predictable and need to be taken into account during component design. Designers also need to be aware that different casting alloys have different levels of “castability,” meaning that development of reliable casting geometries and properties may be more difﬁcult in some alloys than others. A suitable casting geometry for ductile iron may not be the same as that of cast aluminum bronze depending on key casting parameters such as pouring temperature, liquid-metal ﬂuid life; solidiﬁcation shrinkage, and formation of inclusions, slag or dross, in the melt. Design engineers thus

need to know how suitable casting geometries vary for the wide variety of casting alloys. Moreover, designers need to know how differences in suitable casting geometries can be utilized to achieve more optimum design conﬁgurations in terms of strength and rigidity. Traditionally, for sand-casting analysis, the use of geometric methods has been known as the section modulus approach. The technique uses the well-known Chvorinov’s rule, which is commonly used to compare the solidiﬁcation times of simple casting shapes. Although originally developed for the solidiﬁcation of pure metals and alloys solidifying over a very narrow temperature interval, the concept is more broadly applicable and states that the total solidiﬁcation time of a casting (or casting section) is proportional to the square of the volume-to-area ratio of the casting (or casting section). This rule can be stated mathematically as follows:

Vc Ac

tf ¼ K

2 (Eq 1)

where tf is the total solidiﬁcation time for the casting or casting section, Vc is the volume of the casting or casting section, Ac is the surface area of the casting or casting section, and K is a constant for a given metal and mold combination. For simple shapes, the modulus in Eq 1 can be calculated from the ratio of volume and surface area involved in cooling. However, for complex shapes discretized in a three-dimensional grid, the continuous distribution of modulus can be determined using the concept of distance from the mold. The modulus at each point in the casting is determined by the relation: M¼

2 N P

(Eq 2)

1=di

i¼1

The concept of the section modulus approach has been extended to other casting processes. The modulus approach is an approximate analysis scheme that uses geometry-based considerations to provide valuable insights into the solidiﬁcation times and, therefore, the propensity for defect formation during solidiﬁcation.

The next stage is to design the rigging system for the casting, which includes the design of the gate, risers, downsprue, and so forth. This has been based on the rules of thumb of foundry experts and empirical charts. Once the rigging design is established, the solidiﬁcation behavior can be evaluated. Here, heat, mass, and momentum transfer determine the cooling history of the casting. The factors ultimately inﬂuence microstructure evolution and the development of stresses in the casting This article provides a general introduction on casting processes and design. The next chapter describes computer modeling and simulation, which continues to be an important tool in solidiﬁcation modeling and casting design. The remaining chapters describe in more detail the role of design in meeting speciﬁc process or component requirements. Some of these chapters reﬂect more traditional or conventional solutions from the rules of thumb developed by foundries over years past. This retrospective view of casting design solutions may provide perspective in this age of computer–based engineering and perhaps an appreciation of the continuity in the basic principles of casting as both a science and a processing art.

Casting Methods As one of the oldest manufacturing methods, casting involves pouring molten metal into a mold cavity that is conﬁgured to the shapes and dimensions of the ﬁnished form. The methods of shape casting can be divided into several broad categories, as illustrated in Fig. 1. The main categories are: Expendable molds with permanent patterns Expendable molds with expendable patterns Metal or permanent mold processes

In the case of an expendable mold, made with bonded sand or other loose granular mold material, the patterns may be permanent as is typical in sand casting or expendable as in lost foam and investment casting. When patterns are permanent, the mold must be separable into two or more parts in order to permit withdrawal of the permanent pattern (Fig. 2). The tapered ends of the pattern permit it to be removed from the sand mold without restriction. Cores are separate

10 / Casting Design and Performance

Fig. 1

Chart of shape casting process

shapes that are placed in the mold to provide castings with contours, cavities, and passages that are not practical or obtainable with molds. Permanent molds must be separable into two or more parts in order to permit withdrawal of the raw casting from the mold or die. With expendable patterns, the limitations of two or more separable parts of the mold is not necessary, the molding media must only surround the expendable pattern and maintain its shape during molten metal pouring and solidiﬁcation. After the shape has solidiﬁed, the sand or other molding media is shaken off and out of part. Casting parts using expendable mold processes with either permanent patterns or expendable patterns is a very versatile molding method that provides tremendous freedom of design in terms of size, shape, and product quality. Table 1 summarizes the differences in the steps of casting a part between the permanent-pattern versus the expendable-pattern methods. In terms of capability, each shape casting process has its niche, depending on the metal or metals being cast, productivity; component size, dimensional tolerances, and conﬁguration details. The choice of pattern materials can affect dimensional tolerances. Tables 2 and 3 brieﬂy compare some of the typical capabilities of shape casting processes. The ratings and data are necessarily typical and approximate, as many material and design factors can inﬂuence the economical feasibility of a casting process.

The tables only provide a brief overview only for general approximate comparison. Ferrous metals, steel, gray and ductile iron, account for nearly 75% of all metals cast. On the nonferrous side, high-pressure die casting (HPDC) is the dominant process, largely because it readily accommodates scrap-based secondary alloys, and among metal-mold processes it enjoys the highest productivity because of short cycle times. Success depends, to a large extent, on the skill of the mold designer in understanding the very strong inﬂuences of the metal alloy and mold material. Even with the best design, solidiﬁcation on the mold walls closes passage of melt at a certain distance, and this leads to a minimum allowable section thickness. There are large differences in the ﬂuidity of alloys, and therefore minimum allowable section thickness depends on the alloy being cast. Lower minimum thickness can be allowed with zinc, aluminum, and cast iron than with steel. Streamlined ﬂow is helpful; thus sharp corners are avoided. Thicker sections solidify last and must be fed adequately.

General Principles of Casting Design Regardless of the casting method, good casting designs generally follow the same basic guidelines. The basic objective of good casting

design is to concentrate liquid-solid contraction to the last portion(s) of the casting to solidify. Almost all molten alloys, with some important exceptions, undergo shrinkage during solidiﬁcation. One exception is cast iron, both gray and ductile, because graphite expands during solidiﬁcation. This expansion of graphite often, but not always, can compensate for the shrinkage of the iron. During alloy solidiﬁcation, small nuclei of solid grains or dendrites form and grow as the temperature in the casting falls. As these grains grow, a mush of liquid and solid occurs within the freezing range of the alloy. Because each small dendrite forms a site of shrinkage, the shrinkage allows metal to ﬂow between the dendrites to the location where solidiﬁcation and shrinkage is taking place. As the dendrites grow, however; the paths for molten-metal ﬂow become smaller, thus restricting complete ﬁlling of the cavity. Thus, an overarching principle of good casting design is that casting sections should freeze progressively, so that contraction results in a sound part. There are several ways to achieve the goal of progressive solidiﬁcation and liquid-solid contraction. One basic technique of metalcasters is to design reservoirs of molten metal risers that feed liquid metals to the last portion of the casting to solidify. The risers supply liquid metal to feed shrinkage that occurs during solidiﬁcation, so that casting sections freeze

Casting Design and Processes / 11 Table 1 Differences in casting process steps between the permanent pattern and expendable pattern methods of the expendable molding process Expendable mold process Steps in casting a part

Permanent patterns

Design the part and Pick the casting process to be used for the part Design of tooling to make the part Make the tooling Make the patterns for the part Make the part the ﬁrst time Assemble the pattern sections, negative or positive shape of the part, with its gating system into an expendable mold Cast (create/make) the part

Simultaneous engineering with the product designer is a key to making a quality part. Once designer and caster have reached consensus on the casting process, the design can be ﬁnalized to meet functional requirements of both the ﬁnal part and the chosen foundry process. Design tooling to make the negative Design tooling to make the positive shape shape of the part of the part Permanent pattern tooling to make Expendable pattern tooling to make the the negative shape of the part positive shape of the part Negative shape of the part Positive shape of the part Not applicable If necessary, assemble the positive-shaped patterns into the part to be cast. Although pattern sections may be either a negative or positive shape, they must be assembled into a ﬁnal expendable mold assembly ready for casting. Introduce molten metal to the part in the selected casting process

and slag/dross forming tendency. However, systems for feeding and ﬁlling have evolved largely from the experience and capabilities of individual foundries. New tools and computer simulation have lead to the design of improved ﬁlling systems, but ﬁlling system designs for castings is not always systematic. Key objectives and principles that should be addressed for gating and risering include (Ref 1 and 2):

Design runners and gates, if two or more

Establish nonturbulent metal ﬂow. Systematically ﬁll the mold cavity with

metal of minimally degraded quality.

In conjunction with the selection of an

Major components of a sand mold (a) pattern assembly for cope and drag sections. (b) Cross section of mold with core

Fig. 2

progressively. The need for risers depends on the alloy being cast and the geometry of the mold cavity. For example, cast iron often needs very little in the way of risers, because graphite expansion during solidiﬁcation may compensate for the shrinkage of the iron. Adequate ﬁlling depends on the conﬁguration of the mold assembly and alloy castability factors such as ﬂuidity, freezing range, and shrinkage during solidiﬁcation. Thin sections, for example, may require risers, so that all portions of the mold cavity are properly ﬁlled with molten metal. Carefully planned geometry also can offset alloy limitations in terms of ﬂuid life, solidiﬁcation shrinkage, pouring temperature

Expendable patterns

appropriate pouring temperature, provide conditions for mold ﬁlling consistent with misrun avoidance. Establish thermal gradients within the cavity to promote directional solidiﬁcation and to enhance riser effectiveness. Design riser size and geometry, and locate risers and riser inlets to minimize the ratio of gross weight to net weight. Minimize to the extent possible the vertical distance the metal must travel from the lowest position of metal entry to the base of the sprue. Taper the sprue or use a sprue geometry other than cylindrical to minimize vortexing and aspiration. Keep the sprue continuously ﬁlled during pouring. Avoid abrupt changes in the direction of metal ﬂow; gate and runner passages should be streamlined for minimum induced turbulence at angles or points of divergence in the system. Provide contoured transitions in gate, runner, and infeed cross sections at points of cross-sectional area changes. Employ multiple gates to improve thermal distribution and to reduce metal velocity at entry points. Avoid molten metal impingement on mold surfaces or cores by appropriate gate location.

sprues are used, to prevent the turbulence associated with the collision of ﬂow patterns. Design risers to be of sufﬁcient size and effectiveness to compensate for volumetric shrinkage. Riser position, shape, and ﬁlling from the gating system relative to the casting cavity are interrelated and critical considerations. In general, risers should be placed to achieve the maximum pressure differential and, when possible, should be open to the mold surface. Blind or enclosed risers must be adequately vented. Observe the principles of directional solidiﬁcation. The use of chills, riser insulation, and casting design changes may be required. The effects of inadequate gating and riser design can in some cases be corrected only by complete redesign. Provide runner overruns, dross traps, or insystem ﬁltration to avoid the impact of degraded metal on casting quality. Locate the runners in the drag and locate the gates in the cope for horizontal mold orientation. This rule is subject to intelligent variation by the uniqueness of each part. Place the riser cavities in the gate path for maximum effectiveness whenever possible. Never place ﬁlters (if used) between riser and cavity. Design the gates so that metal entry occurs near the lowest surface of the casting cavity. Geometrically contour the runners to maintain uniform ﬂuid pressure throughout. Formulas applicable to all gravity casting methods have been developed for this purpose.

General Guidelines of Geometry in Casting Design. In terms of general geometry, good casting designs generally follow some basic guidelines: Limit drastic changes of thickness to mini-

mize stresses and aid in the casting process.

12 / Casting Design and Performance Table 2 Shape casting processes ratings chart Capabilities Mold method

Applicable metals

Expendable mold with permanent patterns Bonded sand processes Green sand CO2 sand Cold box Hot box Shell Slurry processes Plaster Ceramic Rammed graphite Nonbonded sand processes Vacuum molding Magnetic molding Expendable mold with expendable patterns Lost foam Investment Replicast Centrifugal casting Metal mold Sand mold Permanent mold and semipermanent mold Gravity Static top pour

Tilt pour Low-pressure permanent mold (LPPM) Standard LPPM Vacuum riserless casting/pressure riserless casting Counterpressure casting/pressure-counterpressure casting Countergravity casting High-pressure die casting (HPDC) Conventional HPDC Hot chamber Cold chamber Vacuum die casting Squeeze casting Indirect Direct Semisolid processing Thixocasting (billet) Rheocasting (slurry) Thixomolding (granular magnesium)

Productivity(a)

Disposable cores

Casting size

Dimensional control

All All All All All

5 2 2 2 3

Yes Yes Yes Yes Yes

Unlimited Unlimited Unlimited Unlimited Small to medium

Nonferrous All Reactive

1 1 1

Yes Yes Yes

Unlimited Unlimited Unlimited

Outstanding Outstanding ...

Aluminum Ferrous

1 1

NA NA

Unlimited Unlimited

... ...

Aluminum All All

3 2 1

NA NA NA

Medium to large Small to medium Medium to large

... Outstanding ...

Nonferrous All

3 2

Yes Yes

Small to medium Small to medium

... ...

Nonferrous and metal-matrix composites (MMCs) Nonferrous

3

Yes

Medium to large

...

3

Yes

Medium to large

...

Light metals Aluminum Aluminum All

3 3 3 3

Yes Yes Yes ...

Medium Medium Medium Small

... ... ... ...

Zn and Mg Nonferrous Aluminum

5 5 4

No No

Small Small to medium

Outstanding Outstanding

Al and MMCs Nonferrous and MMCs

4 1

No No

Small to medium Small to medium

... ...

Aluminum Nonferrous Magnesium

4 5 5

No No No

Small to medium Small to medium Small

Outstanding Outstanding Outstanding

... ... ... ... ...

NA, not applicable. (a) Productivity rankings: 1 = low productivity; 5 = high productivity

Taper sections as liberally as possible

Table 3 Comparison of several casting methods Approximate and depending on the metal

toward risers to help control directional solidiﬁcation. Do not interpose thin sections between thick sections and access to risers. Avoid extensive horizontal, ﬂat surfaces. Avoid isolated thick sections that create hot spots that are difﬁcult to feed. Critical dimensions should not cross parting lines in molds or injection or core dies.

Parameter

Green and casting

Permanent mold cast

Die casting

CO2-core casting

Investment casting

Relative cost in quantity Relative cost for small number Permissible weight of casting

Low Lowest

Low High

Lowest Highest

Medium-high Medium-high

Highest Medium

Up to approx. 900kg (1ton)

45kg (100lb)

40kg (85lb)

g(oz)–45kg (100lb)

Thinnest section castable, mm (in.) Typical dimensional tolerance, mm (in.) (not including parting lines) Relative surface ﬁnish

2.5(1/10)

3.0(1/8)

0.8(1/32)

Shell: g(oz)–115kg (250lb) CO2 0.22kg (½ lb)–Mg (tons) 2.5(1/10)

1.5(1/16)

0.3 (0.012)

0.75 (0.03)

0.25 (0.01)

0.25 (0.01)

0.25 (0.01)

Fair to good

Good

Best

Very good

Good

Good

Very good

Shell: good CO2: fair Good

Fair

Fair to good

Fair

Good

Good

Best

Best

Poor

Poorest

Fair

Fair

Designers also should have a good knowledge of the casting characteristics of available alloys, as this inﬂuences design approaches to work around limitations of alloy castability. See Chapter 8 “Casting Design and Geometry” in this book. In general, designing for progressive solidiﬁcation requires tapering walls so that they freeze from one end to the other (Fig. 3a, 3b), thus avoiding situations where two heavy sections are separated by a thin section. This is a poor design because metal must feed one heavy

Relative mechanical properties Relative ease of casting complex design Relative ease of changing design in production

Casting Design and Processes / 13 section through the thin section; when the thin section freezes before the heavy section, the ﬂow path will be cut off and shrinkage may form in the heavy section. Tapered conﬁgurations also facilitate removal of patterns from sand molds (Fig. 3c). Junctions also concentrate heat, leading to areas in the casting where heat is retained. These areas solidify more slowly than others, thus having a coarser structure and different properties from other sections, and solidify after the rest of the casting has solidiﬁed, so that shrinkage cannot be fed. Minimizing the concentration of heat in junctions, therefore, aids in improving casting properties. Examples of this are shown in Fig. 4. Concave corners concentrate heat, so they freeze later and more slowly than straight sections, while convex corners lose heat faster, and freeze sooner and more quickly than straight sections. In designing a casting, the designer should use properties from test bars that have solidiﬁed at the actual cooling rate in that section of the casting. This cooling rate can be determined by instrumenting a casting and measuring the cooling rate in various sections or by simulating its solidiﬁcation using a commercial solidiﬁcation simulation program. The hollow spaces in castings are formed by cores, refractory shapes placed in the molds and

around which the casting freezes. These cores are later removed from the casting, usually by thermal or mechanical means. However, each core requires tooling to form it and time to place it in the mold. Casting designs that minimize cores are preferable to minimize costs. Some examples are given in Fig. 5. The designer also must provide surfaces for the attachment of gates and risers. The casting must solidify toward the riser in order to be sound, and gates must be located so that the mold ﬁlls from the bottom to the top so that oxide ﬁlms that form are swept to the top surface of the casting or into risers where they will not affect casting properties. Gate and riser locations must be accessible for easy removal to minimize processing costs. The position of gate and riser contacts may also add costs if they are placed where subsequent machining will be required to remove gate stubs or riser pads in the ﬁnished component.

Expendable-Mold Casting Expendable mold processes are widely used in casting. The choice of molding method depends on several factors such as part size and shape, quantity, tooling, and the molten metal being poured into mold. The basic characteristics of foundry molds must address four basic requirements: Be formable into the desired shape around

some type of pattern

Be able to hold that shape while molten

metal is introduced

Be able to break down and become strippable

after the metal solidiﬁes The ﬁrst requirement is that the mold material must ﬂow or be pliable enough to encapsulate the surface around the shape of the pattern. Patterns can be either expendable; as in lost-foam casting and investment or lost-wax casting; or permanent. Permanent patterns are made from wood, steel, hard plastic, where numerous molds can be duplicated before signiﬁcant wear would alter the shape of the part being cast. Once the mold material is tightly bonded around the pattern, then the mold must be stripped and removed from that pattern, while not damaging either the permanent pattern material or the mold shape established. The mold shape, which is typically the reverse or mirror image of the part being cast, must then be strong enough to be handled and manipulated, so that it can be combined with other mold shapes such as copes, drags, cores, gatings, etc., that produce the empty cavity of the negative shape of the part being cast. Typically the plumbing system—the sprues, runners, gates—that allows the molten metal to be introduced into the cast part shape are also made from these same molding materials, and becomes part of the mold assembly. After complete assembly of this mold package, the molten metal is introduced in some manner into this mold assembly. The second requirement is the molding material must be able to withstand the erosion action of the molten metal as it ﬂows through the gating system and the part geometry without deteriorating or losing its shape. Any portion of the molding media that may break away could be

Be able to maintain that shape while the

molten metal solidiﬁes

Redesign with tapered sections. Tapered walls provided progressive solidiﬁcation in (a) elbow and (b) valve ﬁtting. (c) Taper allows easier removal of pattern from mold.

Fig. 3

Redesign of casting to minimize heat concentration. (a) Design has numerous hot spots (X junctions) that will cause the casting to distort. (b) Improved design using Y junctions.

Fig. 4

Redesign of castings to eliminate cores. (a) Casting redesigned to eliminate outside cores. (b) Simpliﬁcation of a base plate design to eliminate a core. (c) Redesign of a bracket to eliminate a core and to decrease stress problems.

Fig. 5

14 / Casting Design and Performance washed up in the molten metal, and end up as a defect within the solidiﬁed part being cast. The molten metal could also inﬁltrate the molding media at the mold/metal interface producing a nonhomogenous scab of solidiﬁed metal and molding media on the part being cast. The third requirement for molding media is to maintain shape while the molten metal solidiﬁes. This is a critical requirement to ensure that the cast part in the mold accurately reproduces the outline of the pattern used to make the mold. During solidiﬁcation, the thermal conductivity of the molding medium is also important in meeting the material properties of the metal being cast. Foundries must constantly balance the correct thermal conductivity of the molding media. If heat conducts too quickly from the mold cavity, freezing of the metal before ﬁlling the entire mold cavity could result. Conversely, slow heat conduction can result in slow cycle times or improper microstructure of the solidifying metal, which can lead to incorrect material properties for the part being cast. To assist solidiﬁcation, the mold may include regions with metal chill plates, which are added to assist in quicker solidiﬁcation in certain sections of cast parts. Most metals can be heat treated after casting to assist in meeting the material properties of any particular part. Of course this adds cost to the product, so any compromises in the heat conductance of the molding media that can be implemented would save this cost. Finally, the fourth basic attribute of the molding media is being able to be removed from the cast part. The expendable mold must be broken away and stripped from the solidiﬁed part. More cast parts, especially those with intricate internal passages, are rendered unusable during this shake out step of the casting process than in any other place in the foundry. This is why some of the expendable pattern practices, for example lost-foam and investment casting, may have an advantage over the permanent pattern methods of the expendable molding processes. Mold Materials. Sand is the most prevalent molding medium, and various types of binders are used to bond the sand into useable molds. Besides sand, other materials can meet the aforementioned requirements for molding media. Slurry molding uses a plaster and/or ceramic material to form the shape around either a permanent pattern or an expendable pattern. Graphite molding materials rammed around a pattern have also been utilized. Sand is an abundant raw material, although the sand for foundry use means something much more than a layman’s concept of sand. In casting, sand has numerous properties that need to be carefully controlled and monitored. In terms of sand molding, the basic requirements of the aggregate material include: Dimensional and thermal stability at eleva-

ted temperatures

Suitable particle size and shape

Chemically unreactive with molten metals Not readily wetted by molten metals Freedom from volatiles that produce gas

upon heating

Economical availability Consistent cleanliness, composition, and pH Compatibility with binder systems

Many minerals possess some of these features, but few have them all. The most prevalent mold base is silica sand, due to its wide availability and lower cost relative to other types of sands such as zircon, olivine, and chromite. Manufactured ceramics (such as mullite pellets) are also used as mold-base aggregates.

Expendable Molds Produced with Permanent Patterns Various methods are used to fabricate expendable molds from permanent patterns. The methods include: Molding of sand with a clay-water binder

and mechanical compaction green-sand molding Molding of sand with chemically thermal activated binders made from inorganic compounds such as silicates or organic resins Shell molding of sand with a thin resinbonded shell that is baked No-bond vacuum molding of sand, where molding media is held together with a vacuum source Plaster-mold casting Ceramic-mold casting Rammed graphite molding Magnetic or no-bond molding of ferrous shot

Green sand molding is perhaps the most popular molding process. Green sand mold; which are usually never green in color, but black; use natural or synthetic clays with other additives to bond the sand grains together. Green sand molding may be done either with ﬂasks or without ﬂasks. The resin-bonded mold processes called cold-box, hot-box, warm-box, and shell molding were originally based on organic resin binders, although lately even inorganic binders are mixed with the sand and then hardened or cured by chemical or thermal reactions to ﬁxate the shapes. Often these expendable resin bonded processes are used to produce cores that are placed in permanent molds. This combination of expendable cores with permanent molds is referred to as semipermanent molding. This is used extensively in the nonferrous metal casting area. Virtually all sand processes are suitable for casting both ferrous and nonferrous metals, but green sand excels over other sand processes because it is also the most productive. Green sand is by far the most often employed casting process, simply because it is the dominant process for casting ferrous metals. No other sand process can boast hundreds of molds per hour, with the potential for numerous cavities

per mold. Percentage of casting tonnage by molding process is All All All All

Sand Casting Permanent & Semi Permanent Mold Die Casting “other”

75% >5 19 <1

But every dominant process has its down side. Green sand does not provide the dimensional accuracy or surface ﬁnish of the other bondedsand processes, thus they ﬁnd their niche. The properties of many nonferrous alloys are sensitive to solidiﬁcation rate; when strength and ductility of those alloys is important, green sand gives way to metal mold processes. High speed green sand lines make the employment of chills somewhat difﬁcult, so when localized high cooling rates are required in a casting, CO2 or cold box sand processes may provide better local chill placement opportunity. High-pressure die casting (HPDC) is noted for spraying molten metal into the die at very high velocity, creating turbulence and entrapping cavity air and gasses in the solidiﬁed casting. Thus, HPDC is not suitable for castings that must be heat treated or welded or subjected to other elevated-temperature treatments like porcelain enameling. Variations on HPDC such as squeeze casting semisolid processing and high-vacuum die casting have emerged to overcome those shortcomings, but at a premium cost that renders them somewhat less competitive than HPDCs, so there must be a trade-off. The slurry processes provide exceptional dimensional accuracy and surface ﬁnish for precision parts having complex contours that would be difﬁcult and expensive to machine. High-pressure die casting and its variations also provide nearnet-shape castings but require much more expensive tooling so are generally warranted only for high annual volume requirements. The sand and permanent mold processes readily accommodate disposable internal cores, usually made from sand. The HPDC processes, on the other hand, have very limited ability to accommodate disposable cores. The high metal ﬂow velocities, and the high pressures applied during solidiﬁcation destroy most cores and those robust enough to withstand HPDC processing parameters cannot be easily removed from the solidiﬁed part. Some salt cores are routinely used in HPDC as are some low-melting-temperature metal cores, but HPDC must be considered, at this time, very limited with respect to application of disposable cores such as those applied to the sand and permanent mold processes. By assembling simple-to-mold shapes, the lost foam process has the ability to consolidate several complex cast conﬁgurations into one, sometimes enabling a cast part that would be impossible to machine. Lost foam requires no separate disposable internal cores to create internal shape that would require cores in other processes.

Casting Design and Processes / 15 While conventional high-pressure die casting dominates the nonferrous casting industry, it has been losing ground in recent years to other metal mold processes and to variations on itself that are better able to meet the demands for light weight structural castings such as those in the chassis and suspension systems of automobiles and light trucks. Squeeze casting was the ﬁrst variation on HPDC to prove capable of structural aluminum castings such as steering knuckles. Now the low pressure casting process (LPPM) and its variations such as VRC/PRC and CPC/PCPC are equally able to provide crash-worthy products and can achieve better cavity counts than squeeze casting because metal is not injected or solidiﬁed under pressure. But those processes are best suited for thicker-walled parts and semisolid processing and high-vacuum die casting are now emerging as the processes for large, very thin walled 2mm structural parts.

Expendable Molds Produced with Expendable Patterns. Casting with expendable molds is a very versatile metal-forming process, and the complexity and tolerance of cast products can be enhanced further when expendable patterns are used. Depending on the size and application, castings manufactured with the expendable mold process and with expendable patterns increase the tolerance from 1.5 to 3.5 times that of the permanent pattern methods. The two major expendable pattern methods are lost foam and investment or lost-wax casting. A hybrid of these two methods is the Replicast casting process, which involves patternmaking with polystyrene, like lost foam, but with a ceramic shell mold, like investments casting. One major difference between the permanent pattern casting methods and the expendable pattern methods is that the expendable pattern is typically always the positive shape of the part (Fig. 6). In contrast, permanent patterns are the

Fig. 6

negative or mirror image of the part to be cast. When using the expendable pattern method, the part is typically made twice—once in an expendable or disposable form of part, and then as the actual functional metal form of the part. Some sort of tooling is needed to replicate many expendable patterns for production. This extra set of tooling or patterns may not be needed if only one part or just a few parts are needed. If this is the case, then expendable patterns can be hand crafted or machined from some type of expendable pattern material. In contrast, higher production volumes of parts require repetitive fabrication of expendable patterns. In this case, tooling is needed to make the patterns. Investment casting uses an expendable/disposable pattern, typically made of wax and then coated with some type of slurry molding media. The wax positive shape of the part is usually made from tooling mounted in some type of wax injection machine. Again for very few parts, the wax could be hand crafted or shaped from a billet of the material. The positive shape of part, or more typically multiple parts, are combined with a gating system and made into a cluster (Fig. 7). The wax cluster is then typically coated with a ceramic slurry and dried. This process is repeated until a sufﬁciently thick enough shell is produced. This cluster with the wax disposable pattern is then ﬁred in an oven until all the wax is melted out, and the ceramic shell is rigid enough to support the introduction of the molten metal. Thus the term lost wax is used to refer to investment casting. The metal is then introduced into the empty cavity created by the wax shape that has now been melted out. The crux of the molding problem in investment casting involves withdrawal of the wax or plastic pattern from the metal die, and the withdrawal of cores if used from the pattern. See Chapter 7 “Design for Economical Coring” in this book. In other respects, the principles applicable to sand molding; as described in Chapter 6 “Design for Economical Sand Molding”, apply to investment casting.

Cast manifold with positive pattern for lost foam casting. Courtesy of the Vulcan Group

The lost foam casting process uses a disposable pattern typically made of expanded polystyrene (EPS). One previous name of this process was EPS casting. The EPS material is very similar to the material used for foam coffee cups, and plates, etc. The foam pattern embodies the positive shape of the part to be cast (Fig. 6). The foam pattern can be produced by hand crafting or machining the shape from a billet of the material. This method of making the expendable foam patterns is typically not the best way to produce a good lost-foam casting, but it is acceptable for prototypes or one-of-a-kind applications. More typically, especially for high volumes of parts, the foam pattern is produced from tooling installed on a foam blowing machine of some type. The lost foam foundries have discovered numerous EPS additives and foam molding techniques to produce the best foam pattern that results in the best casting with the least number of defects. Lost Foam is especially suitable for fairly complex castings with many internal passages. Cylinder blocks and cylinder heads for internal combustion engines are fairly good candidates for this process, where internal oil, water, and gas passages can be integrally cast as one piece. Typically these complex internal passages are made possible by gluing together numerous pattern sections and then attaching the ﬁnal gating system to this positive shape of part to be cast (Fig. 8). After the positive shape of part is assembled into a cluster, the cluster is coated with a thin ceramic slurry, dried, and placed into a ﬂask (Fig. 9) for investing of a casting mold consisting of unbonded sand or other molding media. Some fairly sophisticated sand investment and compaction methods for lost

Fig. 7

Wax cluster

16 / Casting Design and Performance decompositions to escape too quickly, allowing the unsupported sand to cave in and either mix with the molten metal stream, or loose the shape of the part being cast. To be successful, the lost foam foundryman must determine the key coating characteristics for making a particular part with a particular molten metal. Many of the coating / slurry suppliers to the industry have learned these lessons, and can be of invaluable service to the lost foam foundryman. A major attraction of lost foam casting is the next step—shakeout of the part from the expendable mold material (Fig. 10). In lost foam casting, the unbonded sand or other molding material just ﬂows out of part; no violent shaking and/or heating of the mold is needed to break up mold like in green-sand or resinset sand mold and cores. Replicast casting, a patented process exclusive to the Casting Technology International, UK; is a hybrid method between lost-foam and investment or lost-wax castings. The expendable pattern is made from some type of foam and/or plastic, but unlike lost foam, the pattern is removed from the mold cavity during ﬁring of the ceramic that surrounds the pattern. See the section “Replicast Molding” in this chapter. Replicast is; therefore, more like the investment casting process, where the disposable pattern material is removed from the mold shape before the introduction of the molten metal. Replicast was originally developed for steel casting; especially low-carbon steels; as the pattern removal before pouring eliminated carbon introduction in the metal as EPS is comprised mainly of carbon. The process is especially valuable for large one of a kind parts, where the caster can hand craft and/or machine the shape from a billet of foam material, and then mold it with some type of expendable molding media that will set-up and become the empty cavity for the molten metal to ﬁll after the plastic expendable pattern has been removed either by heat or some chemical means.

Lost foam pattern for cylinder head. (a) Four different sections glued together. (b) Foam cluster with its gating system and the casting after base cubing

Fig. 8

foam casting are utilized within the lost foam casting industry. The key sand compaction attribute is that the sand ﬂow into all the cavities of the part being cast and that it become fully dense behind the thin ceramic coating on the foam pattern. After the sand is fully compacted around this foam expendable pattern, the part is ready for molten metal introduction with the foam still in the mold assembly. The molten metal must now not only become the shape of the ﬁnal part, it must also decompose the foam pattern, thus the term lost foam. This process has also been referred to as full mold casting, because the mold cavity is not empty but rather ﬁlled with foam. A typical defect in lost foam casting is that this foam is not 100% lost, if it does not escape from the mold entirely, the liquid or gaseous remains can mix with the metal and become defects within the ﬁnal cast part. A key casting attribute of the ceramic slurry is that it acts as barrier and/or valve between the foam pattern and the sand during casting. The thin ceramic shell allows the foam decomposition materials to escape through it and into the sand at the proper rate. If it is too impermeable for the foam decomposition materials to pass through it, the molten metal could freeze off and the part would fail to ﬁll out. If the coating is too permeable, it would allow the foam

Permanent Mold Casting Permanent mold processes involve the production of castings by pouring molten metal into permanent metal molds using gravity, low pressure, vacuum or centrifugal pressure. Simple reusable cores are usually made of metal. More complex cores are made of sand, plaster, ceramics or even salt. When nonmetal destructible cores are used, the process is called semipermanent mold casting. the production of metal objects by placing molten metal in a durable mold of iron, steel, or solid graphite using gravity or low pressure. The general process of permanent mold casting process involves the following steps: Lost foam cluster being (a) dipped into ceramic slurry, (b) dried, and (c) inserted into sand-cast mold. Courtesy of the Vulcan Group

Fig. 9

1. A refractory wash or coating is applied onto the surfaces of the preheated mold that will

Casting Design and Processes / 17

Fig. 10

2. 3. 4. 5. 6.

Dumping and extraction of casting in lostfoam operation.

be in direct contact with the molten metal alloy. Cores, if applicable, are inserted, and the mold is closed either manually or mechanically. The alloy, heated to above its melting temperature, is introduced into the mold through the gating system. After the casting has solidiﬁed, metal cores and loose mold members are withdrawn, the mold is opened, and the casting removed. Steps 2 to 4 are repeated until repair of the refractory coating is required, at which time Step 1 is repeated to the extent necessary. The usual foundry practice is followed for trimming gates and risers from the castings.

The vast majority of permanent mold castings are of aluminum or its alloys. Other metals that can be cast in permanent molds include magnesium, zinc, copper alloys and hypereutectic gray iron. Permanent mold casting is particularly suitable for the high-volume production of castings with fairly uniform wall thickness and limited undercuts or intricate internal coring. The process can also be used to produce complex castings, but production quantities should be high enough to justify the cost of the molds. Practical sizes of permanent mold castings differ according to materials cast, part conﬁguration, and number of parts needed. Compared to sand casting, permanent mold casting permits the production of more uniform castings with closer dimensional tolerances and ﬁner surface ﬁnish (Fig. 11) and improved mechanical properties. In many applications,

Fig. 11

Approximate values of surface roughness and tolerance on dimensions typically obtained with different manufacturing processes. ECM, electrochemical machining; EDM, electrical discharge machining.

the higher mechanical properties can justify the cost of a permanent mold when production volume does not. Permanent mold casting has the following limitations:

Complicated and undercut internal sections

Not all alloys are suitable for permanent

mold casting

Because of relatively high tooling costs, the

process can be prohibitively expensive for low production quantities Some shapes cannot be made using permanent mold casting, because of parting line location, undercuts, or difﬁculties in removing the casting from the mold Coatings are required to protect the mold from attack by the molten metal Good designs of permanent mold casting generally follow the same guidelines as for any casting, as previously outlined in the section “General Principles of Casting Design” in this chapter. A number of factors must be considered when designing and building molds for cost-effective production of sound permanent mold castings. These factors include the following:

are usually made more easily with destructible cores than with metal cores, although collapsible steel cores or loose metal pieces can sometimes be used instead of expendable cores. Proper selection of mold material and wall thickness for economy and consistent mold temperature proﬁles. The gating system, which carries the metal into the mold cavity at the proper location and ﬁll rate to promote directional solidiﬁcation. The riser or feeding system, which contributes to proper temperature gradients and availability of molten metal to feed the solidifying metal front. Venting must be placed advantageously to allow air and gas to escape ahead of the poured metal. Chills which are sometimes inserted in the mold to initiate and accelerate solidiﬁcation in desired locations and directions must be properly sized and located. Antichills are occasionally needed. Use of both must be judicious.

Cavity dimensions must be compensated for

the shrinkage that occurs as the casting cools: Undercuts on the outside of a casting complicate mold design and increase casting cost because additional mold parts or expendable cores are needed.

Methods of Permanent Mold Casting Methods of permanent mold casting include: Gravity casting with static top pour or tilt

pour

18 / Casting Design and Performance Low-pressure casting with permanent mold Counter gravity casting High-pressure die casting

These methods are distinguished by the method of injecting molten metal into the mold cavity. Gravity casting is common, but the dynamics of gravity-fed systems such as turbulence with environmental reactions can introduce inclusion in the melt. The entrainment of inclusions such as oxides in molten aluminum is extremely difﬁcult to eliminate from gravity poured castings. Therefore, the latter three methods have advantages in reducing inclusions and improving properties. See ASM Handbook, Volume 15, Casting for more details on these and other casting methods. Low Pressure Casting with Permanent Molds. Low pressure metal casting has been utilized since the 1950s to produce high-volume, high-integrity castings in alloys ranging from aluminum to zinc. Low-pressure casting, as it is most widely practiced, is a variation of permanent mold casting, although it is used to a limited degree to ﬁll sand or plaster molds too. It might also be referred to as low pressure die casting, perhaps deriving from the European practice of referring to all metal mold processes as die casting. Low pressure casting is a process where molten metal is introduced to the mold by the application of pressure to a hermetically-sealed metal bath forcing the molten metal up through a narrow diameter ﬁll stalk tube from a furnace usually residing below the casting machine; although, there is a version using electromagnetic forces to lift metal into the mold and then the furnace might be an open hearth located beside the casting machine. The process can be considered for low to high volumes of castings from 5 to 100g and usually incorporates the use of iron or steel permanent molds. Recent developments in sand molding technology have made precision sand molds a viable choice for high-volume, low pressure casting as well. A wide range of casting core options such as expendable sand and shell cores and mechanical single or multipiece permanent cores are successfully used in the low pressure process. The large majority of castings produced in low pressure are aluminum, however iron, magnesium and copper-based alloys are successfully cast with this process also. The mechanical and fatigue properties of low pressure cast components are typically 5 to 6% greater than those of gravity cast components of a similar alloy. Controlled application of the pressure yields a smooth, nonturbulent cavity ﬁll from the bottom to the top which is highly desirable when casting light alloys such as aluminum and magnesium. The single point entry of metal to the casting and minimal or nonexistent risers minimizes subsequent processing to remove these items reducing overall manufacturing costs. There are three techniques on low-pressure casting:

to vertically open and close a casting mold situated over a ready supply of molten metal under low pressure Fig. 12 Counter-pressure casting, which is similar to low-pressure casting but with a die surrounded by an air tight mold chamber (Fig. 13) Vacuum riserless/pressure riserless casting (VRC/PRC), which incorporates the use of vacuum and pressure to bottom ﬁll multiple mold cavities from a hermetically sealed molten metal source More details on these processes are in ASM Handbook, Volume 15, Casting (Ref 3). These processes are used to producing high-integrity aluminum casting and have been used to replace aluminum forgings and squeeze castings. High-pressure die casting. (HPDC) is a fast (or perhaps the fastest) method for net shape manufacturing of parts from nonferrous alloys. It is ideal for the economical production of high volume metal parts with moderately accurate dimensions. The rapid cycle time of die casting provides for economical production with the capability for a high degree of automation. The process is adaptable to the use of metal saving designs and can also replace an assembly of separate parts with a single casting. One basic process limitation is that part’s

geometry must allow removal from the die cavity. Die casting is generally limited to metals with low melting points. Aluminum is the most commonly used, followed by zinc, magnesium, copper, tin and lead. Zinc, tin and lead alloys are considered to be low melting point alloys, while aluminum and magnesium alloys are considered as alloys with moderate melting points. Copper is considered as a high melting point alloy, and there are some copper-base alloys for die casting. Suitable high-temperature die materials are critical for economical die casting of copper alloys.

Mold Complexity Although casting solidiﬁcation, discussed in subsequent sections, is the ﬁrst major factor in casting design, the other major factor is mold complexity. This usually is related to a requirement for an excessive number of cores. That is, an important rule in economical casting design is to eliminate as many design features requiring cores as possible. The use of cores in the casting process provides a unique feature that is not available in most other methods of manufacture, yet each

Conventional low-pressure casting, which

incorporates a hydraulically operated machine

Fig. 12

Schematic of low-pressure casting machine with an electric resistance crucible furnace

Casting Design and Processes / 19

Fig. 13

Counterpressure casting process steps

core required adds to the ﬁnal casting cost. Thus, only those cores that are absolutely necessary for producing the desired shape should be used if design is to be directed toward lower cost castings. The ﬁrst principle that must be understood to eliminate cores is the method of mold manufacture. Even investment and expendable pattern or lost foam molds require that the molds or tooling used to make the patterns be fabricated as separate pieces or mold halves. For example, in sand processes, a boxlike device called a ﬂask is set over the pattern and ﬁlled with sand. The sand is hardened either through chemical or mechanical means, and the pattern is then removed. The pattern must be removed by drawing it away from the mold in a direction perpendicular to the parting line; therefore, the design factors related to this common practice must be considered by the design or casting engineers. As noted, taper or draft in the conﬁguration of a permanent pattern (Fig. 3c) facilitates withdrawal of the pattern from the mold. Although

pattern draft is usually not a problem, the requirement of removing the pattern by withdrawing it in a direction 90 from the parting line does restrict total design freedom if casting costs are to be minimized. Because the pattern is to be withdrawn straight out of the mold, no protrusions that restrict this movement can be allowed in the construction of the pattern. If the geometry of the part requires such protrusions, there are only two alternatives for the casting engineer: use a loose piece that remains in the mold or

core after the pattern is drawn and then is removed separately, or; provide a portion of the pattern that creates a cavity for the setting of a core to create the geometry. The ﬁrst alternative has limitations in that sufﬁcient room and access to remove the loose pieces must exist. Loose pieces can be used only in moderately low volume casting production because they limit productivity in high

production. Furthermore, because of the added handwork and potential for mold breakage during extraction, the use of loose pieces can be expected to increase casting costs even in low production. The second alternative involves the addition of a core; therefore, increases casting costs. The problem and some proposed solutions to undercuts are shown in Fig. 14. Finally, because part geometry can require more complex patterns due to irregular parting lines, the ﬁnal rule for reducing casting costs is to design straight parting lines whenever possible. This rule can be seen in practice in Fig. 15. Because much thicker patterns are required for offset parting lines, the cost of the pattern is also greater. The factors that affect the resulting cost of castings and that have sufﬁcient common traits to be independent of the casting process selected include: Requirements for surface ﬁnish and dimen-

sional tolerances

The number of castings to be produced

20 / Casting Design and Performance

Designing for a straight parting line to reduce pattern and casting costs. (a) Irregular parting line is a costly design. (b) Straight parting line is less expensive.

Fig. 15

Fig. 14

Restrictions to pattern removal, and some potential solutions

The material selected for the casting The pattern material selected Any special or unusual testing/inspection or

property requirements

The casting design The casting process selected

Although casting costs are affected by these seven points, casting cost is only one element in the equation that determines the shipping dock cost of the ﬁnal part headed to the consumer. If subsequent costs, for example, in machining, can be decreased by purchasing a more expensive casting, then the customer must weigh the potential cost savings against the cost of a less expensive casting. The way in which consideration of casting and pattern costs alone can lead to false economies can be easily demonstrated by an example of the manufacture of a simple cube. The choice of parting line is usually not even considered by a casting customer, but is left to the quoting foundry. Because of this, many casting customers miss opportunities for innovative engineering solutions to shipping dock cost reduction. For the cube example, there is a need to have the opposite sides of each face of the cube parallel and at right angles to one another. However, the need for draft angles on the tooling to allow removal of the pattern is in opposition to this. If this requirement is not conveyed to the casting supplier, he will undoubtedly quote the pattern equipment with a parting line as shown in Fig. 16(a). This parting line will be chosen by a casting supplier since it will produce the least costly pattern because it is the simplest pattern to make. Casting quotes are often rejected because of pattern prices; therefore, the casting supplier will generally opt for the lowest cost pattern. This choice of parting line will produce a cube in which only two of the opposing sides

will be parallel to each other. A different parting that results in a slightly more expensive pattern is one in which the diagonal of two sides of the cube is used as the parting line (Fig. 16b). This would produce a part in which four of the opposing sides of the cube are parallel and at right angles to one another. However, the complicated parting line shown in Fig. 16(c) eliminates the need for draft on the tooling. Thus, all six sides of the cube can be made parallel to their opposing sides, and all of the sides will be at right angles to each other. This will be the most expensive pattern because of the complexity of the geometry involved. However, based on the requirement for parallel sides, it is the least costly method of producing the required shape. The parting line should be considered in the design phase of a casting because of its importance in the ﬁnal cost of the part. If it is not speciﬁed on the drawing, it should be requested to be returned with any casting quote; only in this way can the customer be assured of comparing equivalent manufacturing methods on the returned quotes. The customer should also request information on cores and other related production tooling in a quote. Only with this information can the best alternative be found for very low shipping dock cost. Without such information, the potential of various tooling designs cannot be determined, let alone used to develop a unique manufacturing approach to reduce the shipping dock cost. However, foundries usually consider such information to be proprietary. If such manufacturing tooling design is of importance to the shipping dock cost of the ﬁnal component, then it should be speciﬁed by the casting customer. This approach is virtually nonexistent in the industrial design of casting components, but it is essential if truly low cost component design is to be achieved.

Parting line options for a cube that must have as many sides as possible parallel and at 90 to each other. (a) Cube parted with four drafted sides is the least expensive option but does not meet design requirements. (b) Parting along the diagonal is a moderate-cost solution that results in four faces in compliance with design requirements. (c) Irregular parting is the most costly option, but is the only one that allows all sides to be parallel and at 90 to each other.

Fig. 16

Fluidity and Solidiﬁcation Metals and alloys have different liquid-metal properties and solidiﬁcation characteristics that inﬂuence the castability of accurate and sound castings. Castability includes factors such as ﬂuidity, shrinkage, and resistance the ability of

Casting Design and Processes / 21 the alloy to withstand stresses developed by the contraction while cooling from the hot-short temperature range. Fluidity is the ability of liquid alloy to ﬂow readily in mold and to ﬁll thin sections. Fluidity is inversely related to the temperature range over which solidiﬁcation occurs. Thus, pure metals and eutectic alloys, with short freezing range, have greater ﬂuidity. Because the viscosity of molten metals is very low, ﬂuidity is determined more by the solidiﬁcation dynamics of the metal and the mold. A high surface tension of the liquid metal reduces ﬂuidity, and oxide ﬁlms forming on the surface of the molten metal thus have a signiﬁcantly adverse effect on ﬂuidity. Insoluble inclusions also reduce ﬂuidity. The solidiﬁcation of a casting can involve as many as three separate contractions as a result of cooling: Liquid-liquid contraction Solid-solid contraction Liquid-solid contraction

Shrinkage occurs during solidiﬁcation of most metals and alloys. Solid and liquid alloys usually have different densities, which means when solidiﬁcation is complete, the solid will occupy less space than the liquid in most alloys, as solid metal is typically denser than its liquid form. To compensate for shrinkage, the foundry adds risers or reservoir of molten metal, which supplies liquid metal to the solidifying casting and compensates for the shrinkage that occurs. If the riser is to function properly, the liquid metal that it contains must be able to ﬂow through the solidifying casting to reach the areas in the casting where solidiﬁcation and hence shrinkage is occurring. Liquid-liquid contraction occurs as a result of the liquid cooling from its pouring temperature; usually 110 to 165 C (200 to 300 F) above its melting point; down to the melting point or solidiﬁcation temperature. This particular factor is of little consequence to designers and is fairly easily dealt with by the foundry engineer. However, the design engineer must consider the other two contractions if costeffective designs and speciﬁcations are to be realized. Solid-solid contraction occurs after a casting has solidiﬁed and as it cools from the solidiﬁcation temperature to room temperature. The design engineer must be concerned with this contraction. To ensure that the dimensions of the casting are correct, the pattern used to produce a given casting usually must be made slightly larger than the casting dimensions at room temperature. The patternmaker compensates for this pattern enlargement for a particular alloy by using a shrink rule speciﬁcally for the alloy involved. Further, because the amount of solid contraction is a function of the particular metal to be cast, problems with dimensions can often occur when changing alloys if the same pattern equipment is used. Whenever such

a change is contemplated, the foundry engineer should be provided with this information because other factors such as gating and risering could also be involved. Therefore, the designers specifying the alloys should carefully consider any change in alloy to ensure that the cost of new equipment does not cancel out the beneﬁts to be achieved by such changes. Solid-solid contraction should also be considered in part design, because it is one of the primary causes of warped and cracked castings. A basic concept that governs the way castings cool is the casting modulus—the volume of a portion of a casting divided by the surface area of that portion of the casting. This relationship of geometry to cooling is easily understood by considering the effects of both volume and surface area on cooling rates. As volume increases, more hot metal will be contained within it, and the casting will therefore take longer to cool. Conversely, because all the heat within a casting must pass through a surface at the metal/ mold interface, the greater the surface area, the faster the casting will cool. Thus, as the volume-to-surface area ratio or casting modulus increases, the time required for cooling and solidiﬁcation is extended. Using this tool, one can easily see the problems created by the design shown in Fig. 17. Because the design contains a rather wide range of section thicknesses, the various sections will cool and solidify at different rates. The ﬁrst portion of the casting to solidify would be the sections having a modulus of 1. At some later time, the section with a modulus of 5 would solidify and begin cooling to room temperature. However, the thinner sections will have cooled and contracted by solid-solid contraction before the thicker section, and because of this the thicker section applies a compressive stress to the thinner sections as it cools. Such stresses have been measured to levels as high as 550 MPa (80ksi), depending on the alloy and section size variations. Therefore, a casting designed in this way will have a strong tendency toward warpage as a result of the imposed stresses. Although a casting may not be warped when taken from the mold, the internal stresses that develop as a result of design can appear at later stages as cracks or warping, often after heat treatment welding, and machining. If design features result in warpage, foundrymen often compensate for this by using tie bars that are usually small cast-on bars attached to various parts of the casting to brace the casting and therefore prevent warpage. Although such devices may result in castings that are not warped, other problems can result. For example, such devices cannot prevent the stresses that cause the castings to warp and, therefore, do nothing to prevent the cause of the problem. In some cases, warpage is eliminated at the expense of increased residual internal stresses. Because such stresses cannot be detected with nondestructive testing, they remain undetected, and upon application of heat or other stresses, warpage, cracking, or premature failure can

The effects of design on distortion of castings. (a) Top view of casting; numbers indicate moduli of the two sections. (b) Distortion caused by solidiﬁcation stresses

Fig. 17

result. In most cases, castings produced with the bars present no problems in service. However, tie bars do nothing to solve the root cause of the problem, namely, the stresses caused by design features; therefore, it is best to avoid their use by employing uniform section sizes whenever possible. The importance of solid-solid contraction as applied to design means that an attempt should be made to reduce dramatic variations in the section sizes of castings. Other factors in foundry practice that can lead to the distortion of castings include improper heat-treating practices, difﬁcult-to-collapse molds or cores, shakeout procedures, rigging, and the way in which the casting cools; these factors are under the control of the foundry engineer. Problems with warpage and cracks can often be eliminated through the cooperative effort of casting designers and foundry engineers. Liquid-solid contraction is by far the greatest difﬁculty due to solidiﬁcation that faces the foundry engineer. It is also one of the greatest opportunities for the design engineer to design for low cost, high quality, and timely delivery. Most metals contract as they pass from the liquid to the solid state. Certain compositions of gray and ductile iron are the exceptions to this rule in the major alloys produced by foundries. The entire founding process is possible only because volumetric contraction locates itself in solidifying castings in a systematic way.

Cooling Rates and Microstructure Casting microstructure is determined by cooling rate, that is, how fast each part of the casting freezes. The cooling rate is roughly proportional to the ratio of the square of the surface area of the casting to the square of its volume—a consequence of what metal casters know as Chvorinov’s law. In other words,

22 / Casting Design and Performance bulky castings freeze much slower than thin castings. For example, sphere of a given volume freezes more slowly than a thin plate of the same volume because the plate has much more surface area to transfer the same quantity of heat into the mold. One way to take account of this during the design phase is to apply a section size effect, a factor by which the local casting properties are adjusted based on local dimensions. While this technique is often effective, the designer should remember that cooling rate depends not only on section size but also on pouring temperature, mold material, gating system, the presence or absence of chills in the mold, mold coatings, and insulation. Thus, within limits, some very broad, the properties can be controlled by the metal caster to produce those which the designer desires. As the solidiﬁcation rate increases, the microstructure of the casting is reﬁned. That is, the grains are smaller, and the spacing between the arms of the dendrites that make up grains is ﬁner. Mechanical properties usually improve as the microstructure becomes ﬁner. Because the properties depend on the microstructure, which depends on the solidiﬁcation rate, which in turn depends on the processing variables used by the foundry and the casting design, designers have a major inﬂuence on the ﬁnal properties of the casting. Using solidiﬁcation simulation programs, foundries predict how fast each section of each casting will solidify and, therefore, what will be the properties of each section. The use of these programs combined with the application of statistical process control techniques has transformed casting into a high technology manufacturing process capable of reliably producing critical components for the most demanding applications.

The solidiﬁcation direction, as seen in Fig. 19 (b), is related to the casting modulus, which is the volume-to-surface-area ratio. Because a wedge casting has a natural solidiﬁcation direction, if the design is such that a riser cannot be placed at the wide end of the wedge, a shrinkage cavity will develop in the casting even though a riser was attached. As shown in Fig. 19(a), this occurs because solidiﬁcation begins somewhere near the middle of the wedge/riser combination. Thus, from a solidiﬁcation or thermal design point of view, there are actually two castings in this example, each having a last place to solidify. If the design allows for the placement of the riser at the proper end of the wedge, as in Fig. 19(b), a sound casting will result because of the unidirectional solidiﬁcation pattern. Shrinkage defects may not always represent a problem, but in cases where they jeopardize mechanical strength, appearance, or pressure tightness, the castings would have to be repaired or scrapped. Under such circumstances, the ﬁnal costs, including delays in delivery, can be severe. Very few castings are as simple as the wedge shown in Fig. 19; therefore, a technique capable of determining the solidiﬁcation sequence of complex shapes is needed if design engineers are to take advantage of the potential cost savings. Regardless of their apparent complexity, castings can usually be resolved into a series of rather simple shapes that make up the complex shape. These simple shapes are:

T-sections—plate intersections forming a T X-sections—plate intersections forming an X L-sections—plate intersections forming an L Plates Cylinders Cylinder/plate intersections

Most casting geometries can be covered by using these basic shapes. For example, a box shape can be simulated by a series of L-sections, and a hub casting can be simulated as a T-section revolved about a central axis. By solving the relationship between casting design parameters and the manner in which these simple shapes solidify, designers can obtain a useful guide to cost reduction and improved quality through casting design.

T-Sections Figure 20(a) shows a hub casting. There are four possible ways to position the risers, depending on the casting design. These four alternatives are shown in Fig. 20(b) and 20(c). Selection of one of these alternatives by the foundry engineer depends on the casting design. Figure 20(d) shows the dimensions and critical sections involved. In Fig. 20(d), as well as in each ﬁgure in this article that illustrates the dimensions of a given section type, a special notation is used to differentiate between the dimensions describing the size of the section components and the section itself. Although the same letter is used, for example, T and T, the roman T refers to the section, while the italic T refers to the dimension of the section. Factors Inﬂuencing the Solidiﬁcation Sequence. There are three basic design dimensions for a T-section: t, T, and R. If the cross plate dimension t is large with respect to the intersection plate dimension T, the last portion of the casting to solidify will be area t in Fig. 20(d). This will require the riser placement labeled Design 1 in Fig. 20(b). If the

Solidiﬁcation Sequence As noted, the objective of good casting design is to concentrate liquid-solid contraction to the last portion(s) of the casting to solidify. To achieve this goal, metalcasters attach reservoirs of molten metal or risers to help feed liquid metal into the last portion of the casting to solidify. This basic technique is illustrated in Fig. 18. Proper placement of risers on castings changes the way in which both casting and riser(s) solidify such that the riser is the last to solidify. When used properly, this produces a casting free of shrinkage because all the shrinkage for the entire mass of both casting and riser will be concentrated in the riser. However, in many cases, the design of the casting restricts the proper placement of risers, making the production of sound castings difﬁcult if not impossible. In this way, casting designers have a signiﬁcant impact on quality and cost. The technique of designing castings that are easily risered can be best understood by considering a simple shape such as a wedge (Fig. 19a).

Casting design and solidiﬁcation of a simple wedge. (a) Riser placed at narrow end of wedge; shrinkage occurs at wide end. The crosshatched region represents the approximate area of the casting where solidiﬁcation is ﬁrst complete, thus cutting off the feeding path of the casting. (b) Correct riser placement. V/SA is the volume-to-surface-area ratio (casting modulus).

Fig. 19

The function of risers. (a) Shrinkage in casting at last portion to solidify without a riser. (b) Shrinkage in riser allows the production of a sound casting.

Fig. 18

Casting Design and Processes / 23 intersecting plate dimension T is large compared to that of the cross plate dimension t, the last portion to solidify will be area T in Fig. 20(d). This type of design will require the riser placement labeled Design 2 in Fig. 20(b). Finally, if the dimensions t and T are nearly the same, either of the two possibilities shown in Fig. 20(c) can be used, because area J in Fig. 20(d) will tend to be the last area of the casting to solidify. In a solidiﬁcation sequence in which the last portion of a hub casting to solidify is the J portion of the T-section, as shown in Fig. 20(d), one of two difﬁculties usually arises. In many designs of this type, the hole cavity formed by the rotation of the T-section about its central axis is too small to allow a riser of sufﬁcient size to feed the section. When this cavity is too small, there is no alternative but to place the riser outside of the casting, as shown in Fig. 20(c), which necessitates the use of a core, thus adding cost. If the enclosed cavity should be large enough to contain a riser of sufﬁcient size, the problem then usually involves the removal of the riser

from within the cavity. In a high-production environment, such as automotive applications, special equipment to extract such internal risers can be economically feasible if the design allows placement in the internal cavity of the hub. Such a special case is the manner in which cast automotive connecting rods are produced. For this particular type of casting, a small wafer riser is placed within the crankshaft receiver end of the connecting rod, which adequately feeds this portion of the casting. However, the use of this casting design and the wafer riser is predicated on high production requirements, which allow the use of specially designed punch presses to remove the riser and the associated gating system. Without such special equipment, the internal riser placement would be highly undesirable. This case is an exception to the overall undesirable nature of solidiﬁcation sequences in which the J portion of a hub casting is the last section to solidify. This points out a very important concept; there are no deﬁnitive rules that always apply to casting design. It is this potential for alternative solutions based on

Cross section (a) of a hub casting showing that it is actually a T-section revolved around an axis. (b) and (c) Possible riser placements; X indicates the last portion of the casting to solidify. See text for discussion. (d) Principal design dimensions of T-section and component sections; see text for discussion.

Fig. 20

specialized conditions that makes the topic of casting design so interesting but very difﬁcult to conﬁne within a rigid set of guidelines. The goal is to design the casting with a particular solidiﬁcation sequence in mind so that the cast part can be risered easily and subsequent riser removal can be accomplished with ease. In general, however, either of the two riser placements shown in Fig. 20(c) can usually be considered undesirable. Because of the added costs imposed by a solidiﬁcation sequence of this type, such designs generally should be avoided. However, if such a design is unavoidable, some modiﬁcations can be made to facilitate manufacture. The key is to recognize the problem, namely, that area J of the T-section of the hub (Fig. 20d) is the last to solidify. Once this is known, appropriate measures can be taken to modify the solidiﬁcation sequence. One alternative would be to alter the casting modulus by changing the shape of the hub to reduce the volume contained and/or to increase the surface area. This would be an effective solution to the problem if the design already required a core to form the hole in the hub before the change in shape. If such a design change increases the number of cores required to produce the casting, this technique may not be cost effective, because additional cores usually necessitate increased casting costs. Another possible solution to the same solidiﬁcation sequence problem might be the use of a feed pad (Fig. 21). Once again, by recognizing the solidiﬁcation sequence problem imposed by the design, the designer could elect to add metal to the ﬂange of the hub to change the solidiﬁcation sequence. Foundrymen refer to this technique as padding. This is a good alternative solution if the pad can either remain on the casting in use or does not affect machining costs adversely. There are many other possible design alternatives, but the designer must ﬁrst be aware of the problem caused by the solidiﬁcation sequence.

Fig. 21 view

A hub casting with a pad to provide a feeding channel from the riser. (a) Top view. (b) Side

24 / Casting Design and Performance Either of the riser placements shown in Fig. 20(b) is an effective alternative if large production quantities are not required. Because of the need to extract the riser through the top of the cope half of the mold, Design 1 is less compatible with high-production molding techniques than the side riser used in Design 2. Therefore, if high production is expected, the preferred solidiﬁcation sequence would be that in which area T in Fig. 20(d) is the last portion of the casting to solidify. Development of Solidiﬁcation Sequence Graphs. A series of graphs has been developed to assist designers and foundry engineers in determining the solidiﬁcation sequence of the basic shapes listed earlier. These graphs show the solidiﬁcation sequences as a function of the design parameters constituting the sections. Considering the T-section shown in Fig. 20(d),

Fig. 22

one can easily see that there are three parameters governing the manner in which the section solidiﬁes, namely, dimensions T, t, and R. As each of these parameters changes, the casting modulus of each component making up the T-section changes. The solidiﬁcation sequence can be estimated by comparing the modulus of each component (t, T, and J) of the T-section. With the use of a computer, this process was employed to evaluate all the possible designs of the T-section from t = 0 to 10, and T = 0 to 10, while holding R = 0. The results of this analysis are shown in Fig. 22(a). The technique used in this analysis is based on Chvorinov’s rule (Eq 1) to compare solidiﬁcation times of simple casting shapes. The factors that relate to the constant K are mold temperature, metal temperature, the density of the solidifying metal, the speciﬁc heat of the

mold, the density of the mold material(s), the thermal conductivity of the metal, the thermal conductivity of the mold, and the speciﬁc heat of the metal. These factors involve the speciﬁcs of a given casting manufactured in a given mold. For a mold that consists of a single material, these factors can be considered a constant throughout the mold. For factors that are constant throughout the part of the mold containing a given portion of the casting, Chvorinov’s rule becomes a proportional equation in which the time for solidiﬁcation is proportional to the square of the volume divided by the surface area. In this form, the results of the equation are independent of the type of mold and metal being considered, provided they are the same over the casting section to which the results are to be applied. Furthermore, for the purposes of

Solidiﬁcation sequence curves for T-sections. Letters indicate the order of solidiﬁcation of the various parts of the casting (see Fig. 4d). (a) Fillet radius R = 0. (b) R = 2. (c) Composite of design curves for various ﬁllet radii

Casting Design and Processes / 25 providing a proper solidiﬁcation sequence, the actual time for a given portion of a casting to solidify is not needed. In this type of analysis, it is important only to determine the order in which each component solidiﬁes with respect to the other components being studied. By remembering these facts, the proportionality equation discussed previously, that is, solidiﬁcation time is proportional to the volume/surface area ratio, is all that is required to study the solidiﬁcation sequence of various section components relative to one another. Consider the components making up the Tsection in Fig. 20(d). There are three basic dimensions (T, t, and R) that describe the section, and the section consists of two plates labeled T and t whose intersection forms another section labeled J. Assuming from a thermal point of view that the formed section is inﬁnite, that is, the ends of plates in the zdirection perpendicular to the drawn section do not inﬂuence the solidiﬁcation of the section or the plates, the volume of the section and plates can be determined from the perimeter and enclosed cross-sectional areas of the sections. Such an assumption is valid if the thickness of the section in the z-direction is ﬁve times the thickness of the thickest section. For this case, the volume of the plate having a thickness T is equal to T2. The surface area is equal to 2T; therefore, the volume-to-surfacearea ratio or casting modulus of the plate is T/2. Similarly, it can be shown that the casting modulus for the other plate making up the section is t/2. To study the relationship of the solidiﬁcation sequence of all the components making up the T-junction, all that remains is to develop the casting modulus for the junction of the two plates referred to as J in Fig. 20(d). Assuming that the thermal effect of the junction extends one plate thickness beyond the tangent point of the radii, that is, the junction length along the t plate is considered to be 2t + 2R + T, the volume and surface area of the junction can be written as follows: Volume ¼ 2t2 þ tð2R þ T Þ þ RT þ T 2 þ

ð4R2 3:1416 R2 Þ 2

(Eq 3)

Surface area ¼ 4t þ 2R þ T þ 3:1416 R þ 2T (Eq 4)

Therefore, the casting moduli for the T-junction J and the plates making up the junction are as follows. For the t-plate, the casting modulus is t/2. For the T-plate, the casting modulus is T/2. For the J section (the T-junction itself ), the casting modulus is: 2t2 þ 2Rt þ tT þ RT þ T 2 þ 0:4292 R2 4t þ 3T þ 5:1416 R

(Eq 5)

Based on Eqs 3, 4, and 5 and the design parameters of R, t, and T, it is a simple matter to

determine the solidiﬁcation sequence of a Tsection. By entering the dimensions of t, T, and R into the respective equations and calculating the resulting casting moduli, one obtains the solidiﬁcation sequences based on the design parameters of the T-section and the plates making up the section. The results of such calculations for values of R = 0 and R = 2.0 for plate thicknesses from 0 to 10 are shown in Fig. 22(a) and 6(b). A series of T-section design curves for R values of 0, 0.25, 0.5, 1.0, and 2.0 for values of t and T from 0 to 2 are presented in Fig. 22 (c). For design curves not covered, one can easily use Eq 3, 4, and 5 for the junction and the respective plates of thickness t and T. Equations 3, 4, and 5 are not speciﬁc for the use of inches only, because the form of Eq 3, 4, and 5 and the results are purely numeric. Thus, one can use metric units such as centimeters or millimeters as the dimensions and use any of the graphs directly. Equations 3, 4, and 5 also allow the user to convert any of the graphs to any radius. All that is needed is to determine a ratio of the desired radius to a speciﬁc radius on the graph and then use this ratio as a scaling factor to adjust the scale of the two dimensional axes t and T. For example, assume a graph for a 1 in. or 1cm radius was needed but only Fig. 22(b) was available. Because the scaling ratio for this case is ½, the T dimension of the horizontal axis could be labeled 0, 1, 2, 3, 4, and 5, replacing the 0, 2, 4, 6, 8, and 10 currently on the graph. By following this same procedure for the t dimension on the vertical scale, one would produce the graph form necessary for the R = 1 problem. Thus, the user of these graphs is not limited to the cases presented, and any combination based on the required R value can be used. The following example further illustrates the use of such curves. Figure 20(a) shows the cross section of a hub casting. With reference to Fig. 20(d), it can be determined that the T-section dimensions are as follows: T = 4in. t = 6in. R=0

Note that T-sections with R = 0 are not recommended, because of the stress concentration and the potential for hot tears or cracks. By referring to Fig. 22(a) and plotting the T and t dimensions, it can be seen that for this design the T = 4, t = 6 point on the graph lies in the zone of the graph labeled T-t-J. Thus, based on casting modulus, a solidiﬁcation sequence of T-t-J would be expected. Because the J portion of the section is the last to solidify, the riser would have to be attached to this portion of the casting. Therefore, because of the solidiﬁcation sequence imposed, a riser placement scheme similar to that shown in Fig. 20(c) must be used. However, this type of solidiﬁcation sequence is generally undesirable, and design alternatives should be considered. Some metal can be removed from the junction, or the T

dimension should be increased somewhat above 8 inches as either a pad (Fig. 21) or a complete increase of the entire ﬂange of the casting. Yet another approach would be to increase or decrease the hub thickness dimension t. These are a few of the potential solutions to the problem that are available to the design engineer. The designer knows both the stresses and the environment in which the part will function; therefore, the designer alone must make the ﬁnal decision on any proposed changes. It is important to inform the design engineer of a potential manufacturing problem early in the designing process while ﬂexibility still exists, not at later stages when the ﬂexibility is lost. The previous example considered a T-section in which the radius R was 0. This case is somewhat academic because, in terms of well-known design principles, the design discussed leads to undue stress concentrations at the sharp corners. This is to be avoided from the standpoints of pure design and casting manufacture. Because of the contraction of the metal as it solidiﬁes and cools, such stress concentration can lead to either hot tearing or cold cracking of the casting. Therefore, the ﬁrst rule of economical casting design is to use generous ﬁllet radii whenever possible. Figure 22(b) was obtained by following the same procedure discussed for the development of Fig. 22(a). In Fig. 22(b), the radius dimension R of Fig. 20(a) was set to 2 inches. A comparison of Fig. 22(a) and 22(b) reveals the effects of adding a 2-inch ﬁllet radius. Very little difference in the solidiﬁcation sequence is observed at the heavy plate thicknesses, the upper right-hand corner of the graph. However, in those designs in which thin plates are used as in the lower left-hand corner of the graph, there is a much stronger tendency for the plate junction J to be the last portion of the T-section to solidify. It must be remembered that Fig. 22(b) represents all the design variations of the T-section from T and t = 0 to T and t = 10 to which a constant 2-inch ﬁllet radius is applied. A series of design curves for various ﬁllet radii are shown in Fig. 22(c). Whenever any ﬁllet radius is applied, an effect on the solidiﬁcation sequence must be expected because the ﬁllet radius causes an increase in volume and a decrease in surface area in the vicinity of the ﬁllet. The increase in volume and the decrease in surface area caused by the use of a ﬁllet are only functions of the ﬁllet radius; therefore, the effect of adding a given ﬁllet radius to the casting modulus of the t-junction is a constant based on the radius. Furthermore, because the casting modulus of the junction is a function of the plate thickness, the addition of a 2-inch ﬁllet to thin plate designs increases the casting modulus of the junction far more with respect to the modulus of the plate than for thick plates. Thus, the effect of the 2 inch. ﬁllet addition is far greater in thin plate designs than in heavier plate T-sections.

26 / Casting Design and Performance Using these graphs, the designer can recognize potential problem areas by remembering the second rule of casting design is that the last place of any casting to solidify must be accessible for riser placement. Ideal riser placement is at the parting line and at the outer perimeter of the casting. If this rule is not followed, weld repair, scrapped castings or sudden catastrophic failure could result, and these outcomes will cause increased casting costs, delivery delays, and a serious danger of premature failure.

X-Sections Another simple shape of importance to designers is the X-section. However, for X-sections, the designer must consider two conditions in order to cover all the possible design combinations that might be encountered.

Equal Opposite Legs. The ﬁrst type of Xsection design is that in which the opposite legs of the X are equal. The model for this case is shown in Fig. 23. The solidiﬁcation sequence curves of this model, which contain ﬁllet radii of R = 0 and R = 2, are presented in Fig. 24 (a) and Fig. 24(b). Design curves for various ﬁllet radii are shown in Fig. 24 (c). As before, the casting modulus equations for the plates making up the X-section are T/2 and t/2. For X-sections in which the opposite legs of the section are equal, the casting modulus for the J- or X-junction is as follows: Volume ¼ Surface area 2t2 þ 2T 2 þ 2RT þ 2Rt þ 0:8584 R2 4t þ 4T þ 6:2832 R

Fig. 23

Model of X-section with opposite legs equal. See Fig. 24 for solidiﬁcation sequence curves.

Fig. 24

Solidiﬁcation sequence curves for X-section with opposite legs equal (see Fig. 23). (a) R = 0. (b) R = 2. (c) Composite of design curves for various ﬁllet radii

(Eq 6)

Casting Design and Processes / 27 Because of the greater volume of metal at the junction, there are far more designs in which the J- or X-junction is the last to solidify compared to the T-junction curves discussed earlier. This is why many authors do not recommend the use of X-sections in casting design. However, this type of thinking can often cause excellent opportunities for cost reduction to be overlooked. Merely because the junction portion of an X, T, L, or boss section solidiﬁes last does not mean that the section is poorly designed. The most economically produced design will often take advantage of this very fact. The junction of a section solidifying last becomes a problem only when it violates the second rule of economical casting design; that is, it leaves that portion of the casting solidifying last in an inaccessible area for riser placement. Although accessibility for riser placement is essential, limiting the number of risers required can further improve economy by design. Thus, the third rule for economical design is to create designs requiring as few risers as possible. Solidiﬁcation proceeds through the casting in what is called the solidiﬁcation or feeding path. Each feeding path will require a riser, and each riser adds to the cost of the casting. Less expensive and better-quality castings will result from reducing the number of feeding paths and ensuring that the end of each feeding path is accessible. Therefore, designs in which the junctions solidify last can be advantageously used to combine feeding paths, thus reducing the overall cost while increasing the quality of the part. Three Legs Equal. The second type of Xsection is shown in Fig. 25. In this design, three of the plates constituting the section have the same thickness. The corresponding solidiﬁcation sequence curves for this case in which R = 0 and R = 2 are shown in Fig. 26(a) and Fig. 26(b), respectively. There is a difference in the curves across the isothickness line running diagonally from left to right on the curves. The upper left portions of the curves represent

those designs in which a thick plate of dimension t is being chilled by the three equal thickness plates whose dimensions are T. This portion of the curve is similar to the curves discussed previously and is fairly straightforward. The lower right-hand portion of the curve represents the case in which a T-section is to be chilled by the addition of a thin plate. In this case, if the thin plate is too small, it does not provide sufﬁcient surface area to chill the T-junction, thus causing it to solidify before the plates making up the T-junction. This is the t-T-J sequence portion near the horizontal axis of the curve. Only when the t plate provides sufﬁcient surface area to chill the T-junction can the t-J-T sequence be achieved. Finally, as the thinner plate continues to increase, the t-T-J sequence is again encountered. When one of the legs that make up the Xsection has a different dimension than the remaining three legs, two separate casting moduli equations must be used. For the case in which the odd plate labeled t on Fig. 25 is larger than the other three plates labeled T, the following casting modulus equation applies for the resulting J junction: Volume ¼ Surface area 2 2 t þ 3T þ 3RT þ Rt þ Tt þ 0:8584 R2 5T þ 3t þ 6:2832 R (Eq 7)

Equation 7 will apply to the upper left corner of the resulting design curves in which t is greater than T. For the case in which the dimension T is greater than t, the following casting modulus applies for the J junction: Volume 6t2 þ 3T 2 þ tR þ 3RT þ 0:8584 R2 ¼ Surface area 11t þ 7T þ 6:2832 R (Eq 8)

The effect of adding a 2-inch radius to this Xsection is identical to that seen on the T-section previously discussed. This can be demonstrated by comparing Fig. 26(a) and 26(b). Design curves for various ﬁllet radii are shown in Fig. 26(c) and 26(d).

L-Sections

Fig. 25

Model of X-section with three legs equal. See Fig. 26 for solidiﬁcation sequences.

The third basic shape of importance to design engineers is the L-section. The model of an Lsection is shown in Fig. 27. Two different radii are used to describe this section: internal radius r and external radius R. Because two different radii are required in describing this type of section, two design cases arise: one in which R is greater than r and one in which R is less than r. For the case in which the internal radius r is less than the external radius R, an L-section similar to that shown in Fig. 27 will result. For cases in which r is greater than R, L-sections of the form shown in Fig. 28 will result.

External Radius Greater than Internal Radius. The casting moduli equations for the Jor L-junction in which R > r are also a function of the design parameters being considered, and four equations must be used. For R > t + r and R < T + r, the casting modulus of the junction is: Volume ¼ Surface area 2 2 T þ t þ TR þ tr þ tT þ 0:2146 r2 0:2146 R2 t þ 3T þ 1:5707 r þ 1:5707 R (Eq 9)

For R > T + r and R > t + r, the casting modulus of the junction is: Volume ¼ Surface area 2 2 T þ r þ RT tT þ tR þ 0:2146 r2 þ 0:2146 R2 t þ T þ 2:8584 R 0:4292 r (Eq 10)

For R < T + r and R < t + r, the casting modulus of the junction is: Volume ¼ Surface area 2 2 T þ t þ Tt þ rt þ rT þ 0:2146 r2 0:2146 R2 3T þ 3t þ 2:8584 r 0:4292 R (Eq 11)

For R > T + r and R < t + r, the casting modulus of the junction is: Volume ¼ Surface area T 2 þ t2 þ Rt þ Tr þ tT þ 0:2146 r2 0:2146 R2 3t þ T þ 1:5707 R þ 1:5707 r (Eq 12)

Although the sequencing curves for R > r shown in Fig. 29(a) and Fig. 29(b) seem complex, the complexity is easily understood by remembering that both the internal and external radii, r and R respectively, are held constant on any given graph. The ﬁrst unusual feature of graphs of this type is the forbidden zone, that is, the dark area in the lower left corner of the graph. In this area, L-sections using the radii combinations given in the ﬁgure caption are impossible, and the internal radius punctures the external radius (Fig. 30a). Another unusual feature of these curves is the crossover point, which in Fig. 29(a) occurs at t = T = 0.25. At this point, the casting moduli of all the components of the L-section (t, T, and J) are equal. This occurs because a uniformly thick plate results from these particular design conditions, as shown in Fig. 30(b). The crossover point always occurs when t = T = (R r) for L-sections in which R> r. The effects of increasing the radii can be seen by comparing Fig. 29(a) with 29(d).

28 / Casting Design and Performance

Fig. 26

Solidiﬁcation sequence curves for X-section with three legs equal. (a) R = 0. (b) R = 2. (c) Composite of design curves for various ﬁllet radii. (d) One segment (R = 1) of the composite design curve shown in (c)

External Radius Less Than Internal Radius. If the external radius R of an L-section is designed to be less than the internal radius r, completely different sequencing curves result. The simpliﬁed geometry of L-sections in which r > R also considerably simpliﬁes the casting modulus of the resulting J- or L-junctions. Because of the geometry, only the following casting modulus equation is required:

Volume ¼ Surface area T 2 þ t2 þ rT þ rt þ Tt þ 0:2146 r2 0:2146 R2 3T þ 3t þ 3:5707 r 0:4292 R (Eq 13)

Curves for L-sections with two such R < r combinations are shown in Fig. 31. No crossover point occurs, nor does the curve contain

a forbidden zone. The reasons for this can be determined by noting the features of such L-section designs (Fig. 28).

Feeding To compensate for shrinkage during solidiﬁcation, molten metal is fed into the mold cavity

Casting Design and Processes / 29

Fig. 27 sequences.

Model of L-section for R > r. Compare with Fig. 28. See also Fig. 29 for solidiﬁcation

Two L-section designs that can occur when R < r. (a) L-section has a higher modulus than the plates when plate thicknesses are nearly equal. (b) Plate thicknesses are very unequal; the modulus of the L-section is less than that of the thick plate.

Fig. 28

from separate reservoirs containing molten metals. These reservoirs are called feeders or risers; although the term riser can refer to a speciﬁc type of feeder that is connected to a slot gate such that the metal rises in the reservoir while rising in the mold cavity (Ref 4). The ﬁrst question to answer is whether feeders or risers are even needed? If possible, the use of feeders should be reduced as much as possible. Just in terms of economics, feeders reduce yield; deﬁned the metals weight going into the foundry divided by the weight of good castings produced. Many small and medium-sized castings do not need to be fed, especially with thinner wall castings. Judicious uses of chills or cooling ﬁns may also reduce the need for feeders. If feeders are required, the next question is what is the required size of the feeder? Guidelines on the sizing and design of feeders are detailed by Campbell (Ref 4) in terms of six basic criteria: 1. Freezing time of the feeder must be as long as or longer than the freezing time of the casting. Freezing time depends on both the casting geometry and the thermal conductivity of the molten metal. 2. Feeder must contain a sufﬁcient amount of metal, for example, volume criterion 3. Freezing time in the junction between the feeder and casting should be greater than the freezing time of both the feeder and casting. If not, under feeding and shrinkage porosity may occur.

Fig. 29

Solidiﬁcation sequence curves for the L-section shown in Fig. 27 with R > r. (a) r = 0.25, R = 0.5. (b) r = 0.5, R = 1. (c) r = 1.0, R = 2.0. (d) r = 2.0, R = 4.0

4. The occurrence of a proper feeding path depends on geometric factors that should not be overlooked. 5. Sufﬁcient differential pressure must occur for the correct amount and direction of ﬂow. 6. Sufﬁcient pressure is needed at all points in the region of feeding so that cavities do not form and grow These guidelines are detailed by Campbell in Ref 4 and 5. Feeding is also discussed in other chapters of this volume along with following overviews on feeding of ﬂat plates and cylindrical sections. Computer modeling and simulation, as described in Chapter 3, have assisted in the design of improved ﬁlling systems. Solidiﬁcation of Flat Plates. The L-, X-, and T-sections discussed to this point consist of a combination of plates and the junctions themselves. Any discussion of casting design must also consider the solidiﬁcation of simple ﬂat plates as well as their more complex junctions.

The solidiﬁcation of metals does not occur as a ﬂat moving front, but rather as a series of protuberances ranging in shape from cone points to very complex, branched shapes called dendrites. The complexity of these shapes is a function of the speed of growth of the front, the cooling rates involved, and the composition of the alloy. For the purposes of design, the nature of these shapes concerns the ability of metal to develop a feeding path through uniform ﬂat plates or cylinders. Because these protuberances in most alloys of interest at commonly found growth rates and thermal gradients usually manifest themselves as dendrite-type shapes, they can provide a signiﬁcant barrier to the continued ﬂow of metal. This phenomenon can lead to shrinkage defects in castings because the feed metal supply cannot reach into the solidifying front at which the liquid-solid contraction is causing a feed metal demand. In particular, this effect inﬂuences casting design in the area of feeding distances.

30 / Casting Design and Performance

L-section designs that fall in the forbidden zones of Fig. 29. (a) Internal radius r punctures external radius R. (b) Crossover point occurs when moduli of all the components are equal—in this case at t = T = 0.25, as shown in Fig. 29(a).

Fig. 30

Example of a design using a solid cylinder (a) and the solidiﬁcation sequence curves for such a design (b)

Fig. 33

Feeding path design considerations. (a) Circular ﬂat plate with a single riser. (b) Addition of wedge-shaped ribs to ensure proper solidiﬁcation. (c) Branched ribs to overcome feeding problems at the circumference of the plate

Fig. 32

Solidiﬁcation sequence curves for L-sections with r > R. (a) r = 0.5, R = 0.25. (b) r = 1, R = 0.5. Compare with Fig. 29.

Fig. 31

In a simple plate having a constant thickness T, a set of solidiﬁcation fronts will move from each of its surfaces until they meet at its centerline. However, as these fronts move together, their dendritic character begins to cause a blockage of the metal ﬂow from a feeding path, which will result in some degree of shrinkage. This phenomenon has been quantiﬁed, and it has been determined that the ability of a riser to feed through a plate is very limited (Ref 6). Often, sound casting production can be achieved only for a few thicknesses along the plate even when chills are used. This is very important because long ﬂat plates will require extensive risering for this reason alone. This limited ability to feed can be overcome by simply reducing the tendency for the front closure to occur as a straight line. This is accomplished by using a plate whose cross section is not uniform in thickness but is continuously changing in thickness, such as the wedge casting discussed earlier. Another way of accomplishing the same objective without changing the geometry of the plate is to alter the thermal shape of the casting through the use of an insulating or chilling

material to change the rate of heat extraction. Still another way of accomplishing feeding through an extensive constant plate thickness is to create a series of feeding paths close to each other. Consider a circular ﬂat plate with a single riser at the center of the plate, as shown in Fig. 32(a). If the feeding distance creates a problem, multiple feeding paths could be considered as a solution. The design of feeding paths involves the development of shape geometries that permit solidiﬁcation to be directed to a portion of the casting that can be easily fed by a riser. For this example, it is necessary to create a feeding path to the center of the plate where it can be fed by the riser. If the addition of radial ribs presents no problem, wedge-shaped ribs could be added, with the taper increasing as the rib nears the center of the plate (Fig. 32b). Because feeding distance is a problem only for continuous-dimensioned ﬂat plates if the distance between these ribs does not exceed the feeding distance for the plate thickness, the ribs will ensure feeding of the plates to soundness. Because the ribs themselves are tapered to create a path to the riser, they should also be sound or free of shrink. If feeding distance again becomes a problem at the outer circumference of the circular plate, the ribs could be branched to ensure that the feeding distance for the ﬂat plate is maintained (Fig. 32c). If the geometry of the rib causes difﬁculty with the design of the resulting casting, the same idea could be applied through the use of insulating or chilling materials to create the same effect thermally without the addition of the added metal of the ribs. However, because the use of molding materials

Casting Design and Processes / 31 to create the needed feeding path would no doubt increase the cost and complexity of the solution, it should be avoided if possible. Solidiﬁcation of Cylinders. A restriction of feeding distances similar to that in plates also occurs in cylinders. Once again, this can be overcome through a change in the thermal shape by developing a cone shape either geometrically or through the use of chilling or insulating materials. Cylinders, however, do create a new condition. The new condition of concern is the junctions produced by the intersection of a cylinder and other cylinders or plates. If the cylinders are hollow, it is not necessary to consider them separately, because these conditions can be visualized as L- or T- sections merely revolved about an axis. However, a separate geometric case develops when the cylinder is solid. Figure 33 shows an example of such a case as well as the resulting design curve. On this curve, the solidiﬁcation sequence is either D-J-T or T-J-D, and no case exists in which the junction J is the last to freeze. For such a junction, either the cylinder acts as a chill for the junction or the plate tends to act as a chill dependent on the cylinder diameter D and the plate thickness T. Thus, such intersections are straightforward and should cause little problem for designers or casting engineers if the feeding distance relationships and other thermal shape factors for promoting directional design are followed.

Heat Transfer and Transport Phenomena Heat is transferred in any of three modes: conduction, convection, and radiation. One must consider which of these to include in the simulation. Conduction, except for some

Model for an L-section (a) cast using two different molding materials A and B. (b) Solidiﬁcation sequence model for L-section shown in (a) in which the insulating material is in the A position; r = 0.25, R = 0.5, diffusivity of mold material A = 2, diffusivity of mold material B = 1. (c) Baseline curve for comparison with Fig. 34(b), 35, 36, and 37 in which r = 0.25, R = 0.5, A = 1, and B = 1 where insulating materials are the same for positions A and B. See also Fig. 35 to 37.

approximation methods, is always included. Convection and radiation effects are often approximated as a single convective boundary condition—convection between the casting and mold, and between the mold and surroundings. The thermal resistance between the mold and casting greatly affects results for metallic mold casting simulations, such as high-pressure die-casting. It is also true that one must use an accurate convective heat transfer coefﬁcient between castings and metallic chills. Although literature values are useful, the generalized heat transfer methods must be carefully applied to speciﬁc casting designs. Knowledge about ﬂuid ﬂow during the ﬁlling of casting is important because it affects heat transfer both during and after ﬁlling. Heat transfer by forced convection predominates during the ﬁlling stages. The loss of liquid metal superheat in the casting cavity of the mold after the ﬁlling transients have died out occurs by buoyancy-generated convection currents. Buoyancygenerated natural convective heat and mass transfer occur before the phase change. Buoyancy-generated convection currents tend to redistribute the melt temperature and composition until solidiﬁcation begins. In the case of alloy melts, the difference in atomic weight of the constituent metals causes an additional convection pattern. There are several computational techniques to simulate ﬂuid ﬂow during mold ﬁlling. Mold-ﬁlling simulation, therefore, can be indispensable if a high-quality casting-solidiﬁcation analysis is desired. The information obtained could help avoid problems of cold shuts, where the melt solidiﬁes before ﬁlling a void, or where a molten front of liquid comes in contact with a solidiﬁed metal. In general, the transport of heat, mass, and momentum during solidiﬁcation processing controls such varied phenomena as solute macrosegregation,

Fig. 34

Another solidiﬁcation curve for the L-section shown in Fig. 18(a) in which a chill has been placed in the A position. In this case, r = 0.25, R = 0.5, A = 2, and B = 1. See also Fig. 34(b) and (c), 20, and 21.

Fig. 35

Another solidiﬁcation curve for the L-section shown in Fig. 34(a) in which the insulating material has been placed in the B position. r = 0.25, R = 0.5, A = 1, and B = 0.85. See also Fig. 34(b) and (c), 35, and 37.

Fig. 36

32 / Casting Design and Performance distribution of voids and porosity, shrinkage effects, and overall solidiﬁcation time. These parameters, in turn, result in a variation of the mechanical, thermophysical, and electrical properties of the solidiﬁed product. For example, macrosegregation primarily occurs due to the movement of solid phase by convection during solidiﬁcation. Governing equations for transport of mass, momentum, energy and solute are introduced by Danzig (Ref 7) for a moving solidiﬁcation front. Differences in density, enthalpy and composition between the liquid and solid are very important for the understanding of solidiﬁcation processes. A modeling process called scaling is used to understand how material properties and

Another solidiﬁcation curve for the L-section shown in Fig. 34(a) in which a chill has been placed in the B position. r = 0.25, R = 0.5, A = 1, and B = 2. Note that there is no crossover point in this ﬁgure. See also Fig. 34(b) and (c), 35, and 36.

geometry can be analyzed in the context of the hierarchal time and length scales associated with the effects of transport phenomena on important features of cast microstructures (Ref 7).

General Geometric Factors on Heat Transfer In addition to the transport phenomena during mold ﬁlling and solidiﬁcation, heat transfer is also inﬂuenced by geometric factors (Ref 5). For example, the designs of plates, cylinders, and T-, X-, and L-sections, as previously described, involve components with mold that extract heat uniformly. However, there are many instances in casting manufacture in which this is not the case. Such situations occur during the normal course of the casting process and in some cases are done intentionally to change the solidiﬁcation sequence with the use of a core or a chill. Use of Cores and Chills. Cores are often used in casting, but their effect on the solidiﬁcation sequence of a section depends on the relative heat extractive capability or heat diffusivity differences between the mold and core components. If the heat extractive character of the core components is greater than that of the mold components, the core will tend to act as a chill.

Fig. 37

Solidiﬁcation sequence curve for T-section shown in Fig. 38 with R = 0.5, A = 1, B = 1, and C = 0.85. Note that in one case (area marked tatb-T-J), ta and tb solidify at the same time. See also Fig. 40 and 41

Model of a T-section cast using three different mold materials. See Fig. 39 to 41 for possible solidiﬁcation sequences.

Fig. 38

This occurs because heat is removed more efﬁciently from the portions of the casting in contact with the chill than from those in contact with mold components. More commonly, the diffusivity of the core when made of resin bonded silica sand is lower than that of the green sand mold components, and the core acts as an insulator with respect to the mold components. It must be remembered that the relative heat extractive capacity of the various mold components determines the resulting solidiﬁcation sequence. The results shown here are independent of the actual mold and cores being considered because the solidiﬁcation sequence is important in casting design, not the time required for an individual portion of the shape to solidify. Only in a few rare cases, as in continuous casting and slush casting, the manufacture of hollow castings through decantation of unsolidiﬁed liquid, is the actual time for solidiﬁcation important. In these cases, the actual position of the solidiﬁcation front at a particular time is important. This is not true for most casting design situations, in which only the sequence of events is important because the order affects the formation of a feeding path. However, the actual time required for an individual component to solidify is usually relatively unimportant if its position in the overall solidiﬁcation sequence can be determined. Because of this comparative unimportance of the actual solidiﬁcation time of a given component, only the relative ratios of the thermal diffusivities of the two materials involved, rather than the actual diffusivity of each, need to be considered in the design curves. Thus, to understand the major effects on the design of T-, X-, and L-sections, only two cases must be considered—one in which one mold component is insulating with respect to the other, and a second in which one component has a chilling effect compared to the other. By convention,

Fig. 39

Solidiﬁcation sequence curve for T-section shown in Fig. 38 with R = 0.5, A = 1, B = 0.85, and C = 1. See also Fig. 39 and 41

Fig. 40

Solidiﬁcation sequence curve for T-section shown in Fig. 38 with R = 0.5, A = 0.85, B = 1, and C = 0.85. See also Fig. 39 and 40

Fig. 41

Casting Design and Processes / 33 the diffusivity ratio is always deﬁned as the ratio of a material divided by the diffusivity of the mold. For example, in a green sand mold, the ratio is the diffusivity of the nongreen sand component and the diffusivity of green sand, and for a permanent mold application, the ratio equals the diffusivity of a given component/diffusivity of die material. If one mold component is insulating with respect to the mold material, the diffusivity ratio would be less than one. A common condition of this type exists for the case in which a resin bonded solid sand core is placed in a green sand mold. The diffusivity ratio would be approximately 0.85 for this case. However, as stated above, the actual materials involved are relatively unimportant provided the same diffusivity ratio exists; the principal concern is the ratio of diffusivities rather than their actual values. L-Sections. Before discussing the design curves for L-sections, it is important to consider the geometric consequences of such placements. An L-section is shown in Fig. 34(a). In such a section, there are two possible positions, A and B, that might be occupied by the insulating material. The basic difference between these two possible placements has to do with the amount of casting surface area exposed to each. For placements of insulation material in the inner corner, the A position, less area at the Jjunction is exposed to the insulation material per unit volume than is exposed along the plates making up the section. This is important because one might expect the placement of insulation material along a junction to slow the solidiﬁcation of the section. However, when the resulting L-section design curve (Fig. 34b) for this situation is observed, the exact opposite seems to be the case; that is, there are far fewer designs in which the junction solidiﬁes last. This occurs because in position A the material extracts heat differently from the junction than the material along its position B outer area. Because a greater percentage of surface area is contacted by a material in position A along the plates making up the L-section than from the junction itself, the plates are insulated to a greater degree than the junction. This means that insulation material placed in the A position tends to act as a chill from the point of view of the junction with respect to the plates making up the section. This occurs because a greater effect of insulation is experienced by the plates due to the increase in surface area exposed to the insulation material than by the junction. This results in a slowing of solidiﬁcation of the plates more than its effect on the junction. Note that there are fewer designs in which J solidiﬁes last in Fig. 34(b) than in Fig. 34(c), in which the mold insulating materials are the same for the A and B positions. A similar apparent reversal of the expected effect when a chilling material is placed in the A position of an L-section can be seen in Fig. 35 for a relative diffusivity ratio of 2.0. Once

again, the chill has a greater effect on the plates making up the L-section than on the junction itself. Such chill materials, when placed in the A position, tend to act as insulation materials from the point of view of the L-section. That is, because the plates are chilled to a greater degree than the junction due to the exposed surface area to the chill, chill placement in the A position creates more designs in which the junction tends to be the last portion to solidify, as can be seen in Fig. 35. If material having an insulation character is placed in the B position of an L-section, it contacts a greater amount of surface area to volume in the junction than is contacted by the plates making up the section; the result of this is the reverse of that discussed for materials placed in the A position. This placement is illustrated in Fig. 36. Because the insulation effect is greater in this case on the junction due to the greater surface area contact, it creates more designs in which the L-junction solidiﬁes after the plates. The effect of a chill placed in the B position can be seen in Fig. 37. It is clear from Fig. 34 that, because the chill affects the junction to a greater degree than the plates, the resulting design curves contain more designs in which the junction is the ﬁrst to solidify. In addition, a radius in the A position placements tends to increase the area in contact with materials in the A position. However, external radii tend to lessen the surface area in contact with a material in the B position. T-Sections. In the case of a T-section, there are three potential positions in which insulating or chilling materials can be placed, namely, positions A, B, and C, as can be seen in Fig. 38. The T-shape causes several interesting effects. The ﬁrst of these is similar to that observed for the L-sections. Materials placed in either the A or B position have a greater inﬂuence on the plates making up the section than on the T-section itself. Similarly, material in the C position has a greater inﬂuence on the solidiﬁcation of the T-junction than on the plates making up the section. These effects are similar to those discussed for the L-sections, but because of the particular geometry involved in a T-section, several others also occur. For example, when the material in the C position is different from that in the A and B positions, the effects on the plates ta and tb are indistinguishable (Fig. 39). If the materials in the A and B positions are different, there will be a difference in the solidiﬁcation of the two halves of the cross plates ta and tb. Figure 40 illustrates this effect, in which each of the plates ta and tb are shown to provide an independent section on the design curve. This effect is a result of the imbalance of the heat extraction across the intersecting plate labeled T in Fig. 38, and it is evident in the use of chilling as well as insulating materials (Fig. 41). In addition, if A and C or B and C are similar materials, an imbalance still exists across the intersecting plate of the T-section, resulting in

a design curve in which the plates ta and tb separate into individual effects (Fig. 40 and 41). X-Sections. In the case of X-sections, a geometry-related effect exists. When an insulating or chilling material is placed in a single position, it inﬂuences two of the plates but not the other two plates. It also has an effect similar to placements on L-sections in their inner corner in that the material has a greater inﬂuence on the plates making up the section than on the X-section itself. This means that the effects are similar to those seen in L-sections discussed previously; chills tend to act as insulating materials from the point of view of the X-section. In addition, diagonal placements across the section cause a matched effect on all of the plates making up the section. Therefore, when mold materials of different diffusivities are employed, the results can be predicted, but geometric effects can often produce results that may not be obvious. Thus, to predict the results of a given set of conditions, one must consider the effects very carefully. However, the amount of contact area of each portion of a section in direct contact with a chill or insulator plays a signiﬁcant role in the eventual solidiﬁcation behavior.

Structure and Properties Properties of a material depend on its microstructure, which is inﬂuenced by the processing and the type of manufacturing method, for example, casting, forging, and machining. In many cases, the structure and properties of a material are assumed to be roughly uniform and isotropic, with some small inherent random scatter in properties. This is not always the case. For example, directionally solidiﬁed casting are purposely produced with a distinct orientation of structure for the directionally dependent properties of gas-turbine blades. Fiber-reinforced composites are another example of a nonuniform, anisotropic material. In castings, differences in cooling rate for different sections in a casting section, or from casting to casting result in variations in structure and thus properties. Discontinuities such as porosity or inclusions may form depending on variations in processing conditions. These variations, nonetheless, do have some underlying root causes that can be understood in predicting variations from point to point within a casting. For example, cast iron provides a dramatic example of the differences in properties caused by differences in structure resulting from cooling rate differences. When molten iron solidiﬁes as a solution of carbon and silicon in iron, the carbon can take different forms, depending on its composition and solidiﬁcation rate and the way the metal has been treated during melting. Chill cast or white iron has properties completely different from gray iron of the same composition, caused solely by the accelerated cooling rate in the chilled iron. White iron freezes so quickly that the carbon combines

34 / Casting Design and Performance with the iron to solidify as cementite (Fe3C), which is hard and brittle. In contrast, gray iron is produced by a slower solidiﬁcation process, such that the carbon solidiﬁes in the form of soft graphite ﬂakes. Ductile iron has properties signiﬁcantly different from gray iron; in this case the carbon solidiﬁes as tiny spheres in a steel matrix. Since the spheres of graphite are less effective as stress raisers, compared to the ﬂakes of carbon in gray iron, ductile iron has signiﬁcant ductility, whereas gray iron does not, even though they both may have nearly identical compositions. In this case, the difference in structure, which produces the property difference, is caused not by cooling rate, but by chemical treatment of the melt, which alters the undercooling at the beginning of solidiﬁcation, which in turn affects the way the carbon solidiﬁes. It is important here to realize that in castings, similar compositions, in similar geometries, can have very different properties, depending on the way the castings are made. It is critical that the casting designer specify the true design needs. Poor casting design can interfere with the ability of the foundry to use the best techniques to produce reliable castings. Defect-free castings can be produced at a price. The multitude of process variables such as molding mediums, binder, gating and feeding, melting and ladle practice, pouring technique, and heat treatment that must be controlled to produce sound castings requires a thorough understanding of the processes used and strict engineering supervision. The price of a casting often reﬂects the cost of such efforts. The attention to detail necessary to make good castings can reduce total cost of a manufactured part by signiﬁcantly reducing machining, repair, and ﬁt problems later in the assembly. Good casting design requires an understanding of how discontinuities affect casting performance and how casting design inﬂuences the tendency for discontinuities to occur during the casting process. Proper speciﬁcation, as opposed to over speciﬁcation or under speciﬁcation, is the key to the successful application. Over speciﬁcation causes needless expenses, and poor casting design can interfere with the ability of the foundry to use the best techniques to produce reliable castings. This can be avoided by understanding the effect of discontinuities on casting performance and the effect of casting design on the tendency for discontinuities to form during the casting process. Important types of casting discontinuities include porosity, inclusions, oxide ﬁlms, second phases, hot tears, metal penetration, and surface defects. See Appendix: Classiﬁcation of Casting Defects for illustration of casting defects. The existence of casting discontinuities does not, in and of itself, indicate that casting performance in service will be affected. Equally important are the size, location, and distribution of these discontinuities. Those discontinuities that are small and located near the center of the casting have little effect, while those

located at or near the surface of the casting are usually damaging. Clustered discontinuities and those that occur in a regular array have a greater effect on properties than those that are isolated and randomly distributed. In specifying acceptable levels of discontinuities, such as microporosity and inclusion sizes and distribution, the designer should determine the critical ﬂaw size that will deleteriously affect performance in service. This permits the foundry to design a casting practice that will eliminate such discontinuities at minimum cost.

Discontinuities and Defects Discontinuities are deﬁned as interruptions in the physical continuity of a casting, and the engineering community tends to make speciﬁc distinctions between the terms discontinuity and defect. By deﬁnition, a defect is a condition that must be removed or corrected. The word defect, therefore, should be carefully used because it implies that a casting is defective and requires corrective measures or rejection. An imperfection or ﬂaw may be a manufacturing defect, in the sense of being an unacceptable condition during an inspection regime, but not a functional defect. Thus, the term casting defect is sometimes used for general imperfections as well, even though the term discontinuity or imperfection is preferred when referring to deviations in less-than-perfect castings. Critical engineering assessment of discontinuities is necessary to deﬁne whether an imperfection is a defect or a harmless discontinuity that does not sacriﬁce the reliability or intended function of the casting. Unfavorable foundry practice can result in various casting imperfections that are detrimental in service and that may contribute to failure. However, many of the common causes of failures of castings occur from one or more aspects of design, materials selection, casting imperfections, faulty processing, improper assembly, or service conditions not initially anticipated. Some casting imperfections have no effect on the function or the service life of cast components, but will give an unsatisfactory appearance or will make further processing, such as machining, more costly. Many such imperfections can be easily corrected by shot-blast cleaning or grinding. Other imperfections or defects that can be more difﬁcult to remove can be acceptable in some locations. Moreover, many castings, especially iron and steel castings, generally lend themselves to being weld repaired, with proper attention to the details of weld repair procedures. Imperfections and defects can be removed and a suitable weld deposit substituted to give a ﬁnished product ﬁt for the intended service, although many failures of castings result from improper repair or rework of casting defects. Porosity is common in castings and takes many forms. Pores may be connected to the surface, where they can be detected by dye

penetrant techniques, or they may be wholly internal, where they require radiographic techniques to discover. Porosity and casting costs are minimized in casting designs that emphasize progressive solidiﬁcation toward a gate or riser, tapered walls, and the avoidance of hot spots. Porosity is not necessarily a defect or cause for rejection; it depends on product requirements of the designer. Macroporosity refers to pores that are large enough to see with the unaided eye on radiographic inspection, while microporosity refers to pores that are not visible without magniﬁcation. Both macroporosity and microporosity are caused by the combined action of metal shrinkage and gas evolution during solidiﬁcation. It has been shown (Ref 8 and 9) that nucleation of pores is difﬁcult in the absence of some sort of substrate, such as a nonmetallic inclusion, a grain reﬁner, or a second phase particle. This is why numerous investigations have shown that clean castings, those castings that are free from inclusions, have fewer pores than castings that contain inclusions. Microporosity is found not only in castings, but also in heavy section forgings that have not been worked sufﬁciently to close it up. When the shrinkage and the gas combine to form macroporosity, properties are deleteriously affected. Static properties are reduced at least by the portion of the cross-sectional area that is taken up with the pores, since there is no metal in the pores, there is no metal to support the load there, and the section acts as though its area was reduced. Because the pores may cause a stress concentration in the remaining material (Ref 10 and 11), static properties may be reduced by more than the percentage of cross-sectional area that is caused by the macroporosity. Liquid metal can dissolve much more gas in solution than solid metal. Therefore, when metal solidiﬁes, gas that is present in the liquid is rejected and forms bubbles. A commonly encountered example is that of hydrogen in aluminum. If these bubbles are trapped in the casting when it freezes, the result is a pore. Pores that result from gas may be spherical, indicating that they formed early in solidiﬁcation when the metal was mostly liquid, or they may be interdendritic in shape, showing that they formed late in solidiﬁcation, when the liquid that remained was present between dendrites. Microporosity is found between dendrites and, like macroporosity, is caused by the inability of feed metal to reach the interdendritic areas of the casting where shrinkage is occurring and where gas is being evolved. However, because this type of porosity occurs late in solidiﬁcation, particularly in long-range freezing or mushy-freezing alloys, it is particularly difﬁcult to eliminate. The most effective method is to increase the thermal gradient, often accomplished by increasing the solidiﬁcation rate, which decreases the length of the mushy zone. This technique may be limited

Casting Design and Processes / 35 by alloy and mold thermal properties, and by casting geometry, that is, the design of the casting. The shape of the micropore is as important as its size, with elongated pores having a greater effect than round pores. Hot Isostatic Pressing. Microporosity can be healed by hot isostatic pressing (HIP). Hot isostatic pressing treatment is relatively inexpensive. If the porosity is relatively large or a macroporosity, HIP will form small dimples on the surface of the casting that may require weld repair. Because HIP is a thermal as well as a pressure treatment, it can alter the microstructure and, therefore, the properties of components if not carefully designed. Hot isostatic pressing will not affect inclusions or oxide folds and, therefore, will not repair castings that have those types of discontinuities. It also has no effect on pores that are connected to the surface. Hot isostatic pressing has made possible the use of alloys that, because of their composition, do not solidify pore free and, without HIP, could not be used for high-integrity applications. The thermal proﬁle of a HIP cycle also is such that the process can eliminate the need for a stress relief or homogenization anneal. By adding time at elevated temperature to the overall thermal history of a casting, microstructural changes such as grain growth and precipitate coarsening can occur. For more information, see the article “Hot Isostatic Pressing of Castings” in Castings, Volume 15 of the ASM Handbook (Ref 12). Inclusions are nonmetallic particles that are found in the casting. They may form during solidiﬁcation as some elements, notably manganese and sulfur in steel, precipitate from solution in the liquid. More frequently, they are formed before solidiﬁcation begins. The former are sometimes called indigenous inclusions, and the latter are called exogenous inclusions. Inclusions are ceramic phases; they have little ductility. A crack may form in the inclusion and propagate from the inclusion into the metal, or a crack may form at the interface between the metal and the inclusion. In addition, because the inclusion and the metal have different coefﬁcients of thermal expansion, thermally induced stresses may appear in the metal surrounding the inclusion during solidiﬁcation (Ref 13). As a result, the inclusion acts as a stress concentration point and reduces dynamic properties. As in the case of microporosity, the size of the inclusion and its location determine its effect (Ref 14, and 15). Small inclusions that are located well within the center of the cross section of the casting have little effect, whereas larger inclusions and those located near the surface of the casting may be particularly detrimental to properties. Inclusions may also be a problem when machining surfaces, causing excessive tool wear and tool breakage. Exogenous inclusions are mostly oxides or mixtures of oxides and are primarily slag or dross particles, which are the oxides that result when the metal reacts with oxygen in the air during melting. These are removed from the

melt before pouring by ﬁltration. Most inclusions found in steel castings arise from the oxidation of metal during the pouring operation (Ref 16). This is known as reoxidation, and takes place when the turbulent ﬂow of the metal in the gating system causes the metal to break up into small droplets, which then react with the oxygen in the air in the gating system or casting cavity to form oxides. Metal casters use computer analysis of gating systems to indicate when reoxidation can be expected in a gating system and to eliminate them. However, casting designs that require molten metal to jet through a section of the casting to ﬁll other sections will recreate these inclusions and should be avoided. Oxide ﬁlms are similar to inclusions and have been found to reduce casting properties (Ref 17, 18, and 19). These form on the surface of the molten metal as it ﬁlls the mold. If this surface ﬁlm is trapped within the casting instead of being carried into a riser, it is a linear discontinuity and an obvious site for crack initiation. It has been shown (Ref 20, and 21) that elimination of oxide ﬁlms, in addition to substantially improving static properties, results in a ﬁve-fold improvement of fatigue life in axial tension-tension tests. Oxide ﬁlms are of particular concern in nonferrous castings, although they also must be controlled in steel and stainless steel castings; because of the high carbon content of cast iron, oxide ﬁlms do not form on that metal. If the ﬁlm folds over on itself as a result of turbulent ﬂow or waterfalling; when molten metal falls to a lower level in the casting during mold ﬁlling; the effects are particularly damaging. Casting design inﬂuences how the metal ﬁlls the mold, and features of the design that require the metal to fall from one level to another while the mold is ﬁlling should be avoided so that waterfalls are eliminated. Oxide ﬁlms are avoided by ﬁlling the casting from the bottom, in a controlled manner, by pumping the metal into the mold using pneumatic or electromagnetic pumps. If the casting is poorly designed, waterfalling will result. An example is given in Fig. 42. Second phases, which form during solidiﬁcation, may also nucleate cracks if they have the proper size and morphology (Ref 22). Example

are aluminum silicon alloys, where the silicon eutectic is present as large platelets, which nucleate cracks, and along which cracks propagate (Ref 23, and 24). The size of these platelets may be signiﬁcantly reduced by modifying the alloy with additions of sodium or strontium. However, such additions increase the size of micropores (Ref 25), and for this reason, many foundrymen rely on accelerated solidiﬁcation of the casting to reﬁne the silicon. As noted above, solidiﬁcation rates normally increase, and the structure is thus reﬁned, in thin sections. Heavy sections are to be avoided if a ﬁne structure is desired. Generally speaking, however, secondary phases in the structure of castings become important in limiting mechanical behavior of castings only in the absence of nonmetallic inclusions and microporosity (Ref 26). Hot tears form when casting sections are constrained by the mold from shrinking as they cool near the end of solidiﬁcation. These discontinuities are fairly large and are most often weld repaired. If not repaired, their effect is not readily predictable (Ref 27). While generally they are detrimental to casting properties, under some circumstances they do not affect them. Hot tears are caused by a combination of factors, including alloy type, metal cleanliness, and mold and core hardness. However, poor casting design is the primary cause. Castings should be designed so that solidifying sections are not subjected to tensile forces caused by shrinkage during solidiﬁcation, as the solidifying alloy has little strength before it solidiﬁes. An example is given in Fig. 43, and an extensive discussion on how to prevent hot tears through casting design is provided in Ref 29. Metal Penetration. Molten metal may penetrate the surface of the mold, forming a rough

Redesign of a casting to eliminate hot tears. Mold restraint coupled with nonuniform freezing of the various sections of this aluminum alloy 356 casting resulted in hot tears. Moving the wall and increasing its thickness corrected the problem. Part dimensions in inches. Source: Ref 28

Fig. 43 Redesign of a casting to avoid waterfalling. (a) In this design, waterfalling results when casting is ﬁlled from the bottom. (b) Improved design provides a path for the metal to follow as it ﬁlls the mold.

Fig. 42

36 / Casting Design and Performance surface or, in extreme cases, actually becoming intimately mixed with the sand in the mold. In iron castings, this is normally the result of the combination of metallostatic head, the pressure exerted on the molten iron at the bottom of the mold by the weight of the metal on top of it, and the surface tension relationships between the liquid iron and molding materials (Ref 30). In cast iron, it is frequently the result of the expansion of graphite at the end of solidiﬁcation, forcing liquid metal into the mold if the casting is not properly designed with a tapered wall to promote directional solidiﬁcation and avoid hot spots. ACKNOWLEDGMENTS

11.

12. 13. 14.

15.

Portions of this article were adapted from: T. Piwonka, Design for Casting, ASM Hand-

book Volume 20, 1997

16.

R. Kotschi, Casting Design, Metals Hand-

book, 9th Edition, Casting, 1986

REFERENCES 1. J.Campbell, Casting Practice—Guidelines for Effective Production of Reliable Castings, ASM Handbook, Vol 15, Casting, 2008, p 497 2. J. Campbell, Filling and Feeding System Concepts, ASM Handbook, Vol 15, Casting, 2008, p 506 3. G. Woycik and G. Peters, Low-Pressure Metal Casting, ASM Handbook, Vol 15, Casting, 2008, p 700 4. J. Campbell, Castings, ButterworthHeinemann, 1991 5. J. Campbell, Castings, 2nd Edition, Butterworth-Heinemann, 2003 6. R.A. Flinn, Fundamentals of Metal Casting, Addison Wesley, 1963 7. J. Danzig, Transport Phenomena During Solidiﬁcation, ASM Handbook, Vol 15, Casting, 2008, p 288 8. E.L. Rooy, Hydrogen: The One-Third Solution, AFS Trans., Vol 101, 1993, p 961 9. N. Roy, A.M. Samuel, and F.H. Samuel, Porosity Formation in Al-9Wt Pct Mg - 3 Wt Pct Cu Alloy Systems: Metallographic Observations, Met. Mater. Trans., Vol 27A, Feb 1996, p 415 10. M.K. Surappa, E. Blank, and J.C. Jaquet, Effect of Macro-porosity on the Strength

17.

18.

19.

20.

21.

22.

and Ductility of Cast Al-7Si-0.3Mg Alloy, Scr. Metall., Vol 20, 1986, p 1281 C.H. Ca´ceres, On the Effect of Macroporosity on the Tensile Properties of the Al7%Si-0.4%Mg Casting Alloy, submitted to Scr. Metall., 1994 S. Mashl, Hot Isostatic Pressing of Castings, ASM Handbook, Vol 15, Casting, 2008, p 408 I.P. Volchok, Non-Metallic Inclusions and the Failure of Ferritic-Pearlitic Cast Steel, Cast Metals, Vol 6 (No. 3), 1993, p 162 P. Heuler, C. Berger, and J. Motz, Fatigue Behaviour of Steel Casting Containing Near-Surface Defects, Fatigue Fract. Eng. Mater. Struct., Vol 16 (No. 1), 1992, p 115 J. Motz et al., Einﬂuss oberﬂa¨chener Fehlstellen im Stahlgub auf die Rib einleitung bei Schwingungsbeanspruchung, Geissereiforschung, Vol 43, 1991, p 37 C.E. Bates and C. Wanstall, Clean Steel Castings, in Metalcasting Competitiveness Research, Final Report, DOE/ID/13163-1 (DE95016652), Department of Energy, Aug 1994, p 51 J. Campbell, J. Runyoro, and S.M.A. Boutorabi, Critical Gate Velocities for FilmForming Alloys, AFS Trans., Vol 100, 1992, p 225 N.R. Green and J. Campbell, Proc. Spring 1993 Meeting (Strasbourg, France), European Division Materials Research Society, 4–7 May, 1993 N.R. Green and J. Campbell, The Inﬂuence of Oxide Film Filling Defects on the Strength of Al-7Si-Mg Alloy Castings, AFS Trans., Vol 102, 1994, p 341 C. Nyahumwa, N.R. Green, and J. Campbell, “The Effect of Oxide Film Filling Defects on the Fatigue Life Distributions of Al-7Si-Mg Alloy Castings,” presented at the International Symposium on Solidiﬁcation Science and Processing (Honolulu, HI), Japan Institute of Metals and TMS, Dec 1995 J. Campbell, The Mechanical Strength of Non-Ferrous Castings, Proc. 61st World Foundry Cong. (Beijing), 1995, p 104; available from the American Foundrymen’s Society, Des Plaines, IL K.E. Ho¨ner and J. Gross, Bruch verhalten und mechanische Eigenschafter von AluminiumSilicium-Gublegierungen in unterschiedlichen Behandlungszusta¨nden, Giessereiforschung, Vol 44 (No. 4), 1992, p 146

23. F.T. Lee, J.F. Major, and F.H. Samuel, Effect of Silicon Particles on the Fatigue Crack Growth Characteristics of Al-12Wt Pct - 0.35Wt Pct Mg - (0 to 0.02) Wt Pct Sr Casting Alloys, Met. Mater. Trans., Vol 26A (No. 6), June 1995, p 1553 24. J.F. Major, F.T. Lee, and F.H. Samuel, Fatigue Crack Growth and Fracture Behavior of Al-12 wt% Si-0.35 wt% Mg (0–0.02) % Sr Casting Alloys, Paper 96-027, AFS Trans., Vol 104, 1996 25. D. Argo and J.E. Gruzleski, Porosity in Modiﬁed Aluminum Alloy Castings, AFS Trans., Vol 96, 1988, p 65 26. T.L. Reinhart, “The Inﬂuence of Microstructure on the Fatigue and Fracture Properties of Aluminum Alloy Castings,” presented at Aeromat ’96 (Dayton, OH), ASM International, June 1996 27. M.J. Couper, A.E. Neeson, and J.R. Grifﬁths, Casting Defects and the Fatigue Behaviour of an Aluminium Casting Alloy, Fatigue Fract. Eng. Mater. Struct., Vol 13 (No. 3), 1990, p 213 28. Permanent Mold Casting, Forging and Casting, Vol 5, Metals Handbook, 8th ed., American Society for Metals, 1970, p 279 29. A.L. Kearney and J. Rafﬁn, Heat Tear Control Handbook for Aluminum Foundrymen and Casting Designers, American Foundrymen’s Society, 1987 30. D.M. Stefanescu et al., Cast Iron Penetration in Sand Molds—Part I: Physics of Penetration Defects and Penetration Model, Paper 96–206, AFS Trans., Vol 104, 1996

SELECTED REFERENCES J. Campbell, Castings, 2nd Edition, Butter-

worth-Heinemann, 2003

M.A. Gwyn, Cost-Effective Casting Design,

Modern Casting, Vol.88, No.5, 1998, pp.32–36. B. Ravi, Metal Casting—Computer Aided Design and Analysis, Prentice Hall (India), 2005 Hwang Weng-Sing and Kaohsiung, Ed., Modeling of Casting & Solidiﬁcation Processes, Taiwan, 2004 Kuang-O Yu, Modeling for Casting and Solidiﬁcation Processing, Marcel Dekker, Inc, 2002

Casting Design and Performance Pages 37–60

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Modeling of Casting and Solidiﬁcation Processing* Jianzheng Guo and Mark Samonds, ESI US R&D

CASTING AND SOLIDIFICATION PROCESSES are modeled and described in terms of thermodynamics, heat transfer, ﬂuid ﬂow, stress, defect formation, microstructure evolution, and thermophysical and mechanical properties (Ref 1). Simulation technologies are applied extensively in casting industries to understand the effects of alloy chemistry, heat transfer, ﬂuid transport phenomena, and their relationships to microstructure and the formation of defects. Thanks to the rapid progress of computer calculation capability and numerical modeling technologies, casting simulation can quickly answer some critical questions concerning ﬁlling, microstructure, defects, properties, and ﬁnal shape. Modeling can be used as a quality-assurance function and can also help to reduce the time for responding to customer inquiries. It shifts the trial-and-error procedure used on the shop ﬂoor to the computer, which does it faster, more easily, with greater transparency, and more economically. Casting simulation software has been available to the foundry industry for over 20 years. Simulation technology has come a long way since the early-to-mid-1980s when the design engineer could only work with two-dimensional models. The early days focused on identifying hot spots in the casting. As the computer-aided design and numerical simulation software packages evolved, foundry engineers could increasingly make quick changes to the feeding design to ﬁx these problems with relative ease. Most of the casting simulation packages in the market today can now handle solidiﬁcation and ﬂuid ﬂow in the mold. Today, the foundry industry wants to focus on more advanced predictions, such as stress and deformation, microstructure determination, defect formation, and mechanical properties. The primary phenomenon controlling casting is the heat transfer from the metal to the mold. The heat-transfer processes are complex. The cooling rates range from an order of one-tenth to thousands of degrees per second, and the corresponding length scales extend from several meters to a few micrometers. These various cooling rates produce different microstructures

and hence a variety of mechanical properties. Solidiﬁcation kinetics, including nucleation, growth, and coarsening, are now being investigated extensively. The incorporation of these principles into the more traditional thermal, ﬂuid ﬂow, and stress models enable quantitative predictions of microstructure and engineering properties such as tensile strength and elongation. The coupling of mechanical analysis with thermal analysis enables the prediction of residual stresses and distortion in the castings and molds. These predictions will enable design engineers to evaluate the effects of nonuniform properties and defects on life-cycle performance of components. Different casting processes are used to produce different kinds of casting components. Some of the most common processes are sand casting, investment casting, die casting, permanent mold casting, lost foam casting, centrifugal casting, continuous casting, and direct chill casting. Each casting process has its own features. For example, for pressure die casting machines, it is particularly important to optimize not only the casting quality but also the die behavior with respect to thermal stress and strains and life expectancy. In this chapter, the topic of computational thermodynamics is ﬁrst reviewed. The calculation of solidiﬁcation paths for casting alloys is introduced in which back diffusion is included so that the cooling condition can be accounted for. Then, a brief review of the calculation of thermophysical properties is presented. Fundamentals of the modeling of solidiﬁcation processes are discussed next. The modeling conservation equations are listed. Several commonly used microstructure simulation methods are presented. Ductile iron casting is chosen as an example to demonstrate the ability of microstructure simulation. Defect prediction is one of the main purposes for casting and solidiﬁcation simulation. The predictions for the major defects of a casting, such as porosity, hot tearing, and macrosegregation, are highlighted. In the end, several industry applications are presented.

*Originally written and subsequently published for Metals Process Simulation, Volume 22B, ASM Handbook.

Computational Thermodynamics Thermodynamic calculations are the foundation for performing basic materials research on the solidiﬁcation of metals. It is important to have proper phase information for an accurate prediction of casting solidiﬁcation (Ref 2). The method of phase diagram calculations was started by Van Laar, who computed a large number of prototype binary phase diagrams with different topological features using ideal and regular solution models. Since then, many researchers began to incorporate phase equilibrium data to evaluate the thermodynamic properties of alloys. It was not until the late 1980s that a number of phase diagram calculation software packages became available. Then, thermodynamic calculations for multicomponent systems became feasible, thanks to the rapid progress in the computer industry. Solidiﬁcation proceeds at various rates for castings. Thus, the microstructure and the composition are not homogeneous throughout the casting. The solidiﬁcation path determines the solidiﬁcation behavior of an alloy. For complex multicomponent alloys, the solidiﬁcation path is very complicated. Hence, the equilibrium of each phase at different temperatures must be calculated. The thermodynamic and the kinetic calculations are the base for the prediction of solidiﬁcation. The diffusion transport in the solid phase must be solved for each element. This requires knowledge of the diffusion coefﬁcient of the element, the length scale, and the cooling conditions. Thermodynamic modeling has recently become increasingly used to predict the equilibrium and phase relationships in multicomponent alloys, (Ref 3–5). Currently, several packages are able to simulate solidiﬁcation using the Scheil model and lever rule, such as Thermo-Calc, Pandat, and JMatPro. A modiﬁed Scheil model is applied in JMatPro. In their calculation, carbon and nitrogen are treated as completely diffused in the solid, which makes a great improvement, particularly for

38 / Casting Design and Performance iron-base alloy solidiﬁcation. In reality, there is always a ﬁnite back diffusion based on the cooling conditions (Ref 6). It is critical to have an accurate solidiﬁcation path for casting simulation (Ref 7–9). Obtaining the solidiﬁcation path is very important for understanding and controlling the solidiﬁcation process of the alloy. Normally, there are two ways to predict the solidiﬁcation path. One is the complete equilibrium approach, which can be calculated by the lever rule. The other is the Scheil model, which assumes that the solute diffusion in the solid phase is small enough to be considered negligible and that diffusion in the liquid is extremely fast, fast enough to assume that diffusion is complete. For almost all practical situations, the solidiﬁcation occurs under nonequilibrium conditions but does not follow the Scheil model. There is ﬁnite diffusion in the solid, or back diffusion, which is a function of the cooling rate. Back diffusion plays an important role in the calculation of segregation. There are many numerical and analytical models that attempt to handle such phenomena (Ref 10–16). For most of the models, constant partition coefﬁcients are assumed, which is a good approximation for many alloys. Unfortunately, sometimes the partition coefﬁcients can vary dramatically for some commercial alloys. The partition coefﬁcient of an element in an alloy can change from less than one to greater than one, or vice versa, during solidiﬁcation (Ref 17). For these cases, the analytical or earlier numerical models are not valid any more. The equilibrium of each phase at different temperatures should be calculated. This can be fulﬁlled by the coupling with the thermodynamic calculation. Recently, researchers have started to couple thermodynamic calculations with a modiﬁed Scheil model including back diffusion (Ref 18). Normally, the liquid is assumed to be completely mixed. The solid concentration is calculated by solving a one-dimensional diffusion equation for each element. The governing equations for conservation of species for multicomponent alloy solidiﬁcation are (Ref 19): Liquid species conservation: fl

@f @Clj j SDj j s ¼ Cl Cij þ Cs Cij @t @t L

(Ref 20) using the one-dimensional platelike dendrite geometry: L ¼ f6s l where l is the secondary dendrite arm spacing, which is a func

tion of cooling rate, l ¼ a T n . Here, a and n are constants determined by the alloy composi

tion, and T is the cooling rate. The interfacial area concentration S is related to the solid volume fraction and the secondary dendrite arm spacing, S ¼ 2=l. Combining Eq 1 and 2 and then discretizing provides: SDt j o ÞCi fs Csj Csj ¼ ðfs þ L SDt j ðfs þ ÞCs L

where the superscript o refers to the old value of the variable. Hence:

Csj ¼ Cij when o

¼ 0 Scheil model

SDt L

Based on mass conservation, the liquid concentration can be calculated from the solution of the solid concentration proﬁle accordingly. Yan (Ref 18) did an experiment for the solidiﬁcation of a quaternary Al-6.27Cu-0.22Si0.19Mg alloy with a cooling rate of 0.065 K/s. The microstructure of the solidiﬁed samples is dendritic. The calculated fraction of solid versus temperature relationship for this quaternary alloy is shown in Fig. 1. According to the Scheil model, the solidiﬁcation sequence for this alloy is L ! L+ face-centered cubic (fcc) ! L+fcc+theta ! L+fcc+theta+Al5Cu2Mg8 830

L-->fcc

820

where j is the species index, C is concentration, D is a diffusion coefﬁcient, f is the volume fraction of a phase, t is time, L is a diffusion length, and S is the interfacial area concentration. The subscript i refers to the solid-liquid interface, l is the liquid, and s is the solid. The diffusion length L can be determined using a model proposed by Wang and Beckermann (Ref 19) based on the work of Ohnaka

SDt ! 1Lever rule L

Csj ¼ Csj þ ðCij Cso Þfs =fs when

(Eq 1)

(Eq 2)

(Eq 4)

Equation 4 is used for calculating the concentration in the solid. Here, it is noticable that Eq 4 can automatically turn into the Scheil model or lever rule if the diffusion is zero or inﬁnity:

Temperature, K

@f @Csj j SDj s ¼ Ci Csj þ @t @t L

j ðfs fs ÞCsj þ ðfs þ SDt L ÞCi SDt fs þ L o

Csj ¼

Solid species conservation: fs

(Eq 3)

Si6(Q)! L+fcc+theta+Al5Cu2MG8Si6(Q)+Si. Experimentally, there were no Silicon and Al5Cu2Mg8Si6 phases formed according to metallographical examination and electron probe microanalysis. The back-diffusion model indicates that there are only fcc and theta phases formed during solidiﬁcation for this cooling condition. The results predicted by the back-diffusion mode are in agreement with the experimental quantitative image analysis program. The measured fractions of fcc phase were compared with the calculations from the Scheil model, lever rule, and the current back-diffusion model for three different cooling rates. The comparison is shown in Table 1. The fraction of fcc phase calculated the from the Scheil model is less than the measured values, and the fraction of fcc calculated from the lever rule is higher than that from the experiments for all three cooling rates. The back-diffusion model, which takes into account the cooling rate, gives good agreement with the experiments.

Thermophysical Properties The research on thermophysical properties is a very important part of materials science, particularly at the current time, because such data are a critical input for the simulation of metals processing. Yang and coworkers (Ref 21) investigated the sensitivity of investment casting simulations to the accuracy of thermophysical properties. They found that the temperature prediction and thermal gradient in the liquid are the most sensitive to the accuracy of the input values used for density and thermal conductivity in the solid. Thermal conductivity in the liquid, speciﬁc heat, and density have similar levels of inﬂuence on solidiﬁcation time; increasing their values increases the local solidiﬁcation time. Thermal conductivity in the solid has the opposite effect compared to all the other properties studied. Accurate thermophysical data are difﬁcult to obtain at high temperatures experimentally, especially for reactive alloys such as titanium and nickel-base superalloys. An extensive database for the calculation of thermophysical properties has been developed (Ref 8) that uses the phase fraction information predicted with the Gibb’s free energy minimization routines developed by (Ref 3) and extended by Ref 4.

810

800

Lever rule Back-diffusion model Scheil model

790

L-->fcc+Theta

Table 1 Comparison of experimental and calculated fraction of face-centered cubic phase (volume percent)

L-->fcc+Theta+Q Cooling rate, K/s

780 0.84

Fig. 1

L-->fcc+Theta+Q+(SI) 0.86

0.88

0.90 0.92 0.94 Fraction of solid

0.96

0.98

1.00

Solidiﬁcation path of a 2219 aluminum alloy

Lever rule 0.065 0.25 0.75 Scheil model

Area scan

Image analysis

Calculation

96.0 95.8 95.8

95.4 95.3 94.7

96.7 96.0 95.4 94.3 85.4

Modeling of Casting and Solidiﬁcation Processing / 39 These properties include density, speciﬁc heat, enthalpy, latent heat, electrical conductivity and resistivity, thermal conductivity, liquid viscosity, Young’s modulus, and Poisson’s ratio. The thermodynamic calculation is based on the thermodynamic databases from CompuTherm. A simple pairwise mixture model, which is similar to that used to model thermodynamic excess functions in multicomponent alloys, can be used to calculate the properties (Ref 5): X

XX

X

activation model. The former one is mainly based on the monatomic nature. There are some models available, but most of them are still under development and do not meet the technological need. The thermal activation method is applied here to predict the viscosity of alloys. The viscosity, Z, of pure liquid metals follows Andrade’s relationship (Ref 26): ZðT Þ ¼ Zo expðE=RT Þ

(Eq 7)

where P is the phase property, Pi is the property of the pure element in the phase, i is a binary interaction parameter, and xi and xj are the mole fractions of elements i and j in that phase.

where E is the activation energy, and R is the gas constant. Figure 4 shows an example of the calculated liquid viscosity of an IN718 alloy using the mixture model compared with experimental results. Figure 5 shows the comparison between experimental and calculated results for various alloys at different temperatures.

Thermal Conductivity

Density

The thermal conductivity mainly depends on the chemical composition of an alloy. It also depends to a lesser extent on the precipitates, bulk deformation, microstructures, and other factors (Ref 22, 23). These factors can usually be ignored in the calculation of conductivity for commercial alloys. The thermal conductivity of alloys is composed of two components: a lattice component and an electronic component. In well-conducting metals, the thermal conductivity is mainly electronic conductivity. The lattice conductivity is usually very small compared to the electronic one. Hence, only the electronic component is considered here. The thermal conductivity, l, and the electrical resistivity, r, are related according to the Wiedeman-Franz-Lorenz law (Ref 24, 25):

Currently, the casting simulation models have reached the stage where one of the limiting factors in their applicability is the accuracy of the thermophysical data for the materials to be modeled. Among all the thermophysical data, the temperature-dependent density is one of the most

v ðxi xj Þv

v

(Eq 5)

LT l¼ r

Liquid Viscosity Viscosity is an important property to be considered in dealing with ﬂuid ﬂow behavior. The liquid viscosity is a measure of resistance of the ﬂuid to ﬂow when subjected to an external force. There are two approaches to modeling of complex alloy viscosity. One is a fundamental molecular approach and the other is a thermal

Casting process modeling involves the simulation of mold ﬁlling, solidiﬁcation of the cast metal, microstructure formation, stress analysis on casting and mold, and so on. At the macroscopic scale, these processes are governed by basic equations that describe the conservation of mass,momentum, energy, and species. Heat transfer is perhaps the most important discipline in casting simulation.

200

14

180

12 160 140 Experiment Calculation

120 100

10 8 6

80

4

60

2

40 300

(Eq 6)

where the Lorentz constant, L; ¼ 2:44 1011 W K 2 , and T is the temperature. Based on this model, an example of the calculated thermal conductivity of an A356 alloy is shown in Fig. 2, with experimental results for comparison. The calculation can accurately predict the thermal conductivity variation with temperatures for this alloy in the liquid, solid, and mushy zone. Figure 3 shows the comparison with experimental results from Auburn University for various alloys at different temperatures. The agreement is good in general.

Fundamentals of the Modeling of Solidiﬁcation Processes

Viscosity, mPas

j>i

Fig. 2 alloy

400

500

600 700 800 Temperature, K

900

Comparison between experimental and calculated thermal conductivity for an A356

Experiment Calculation

0 1400

1000

1500

1600 Temperature, K

1700

1800

Comparison between experimental calculated viscosity for an IN718 alloy

Fig. 4

Thermal conductivity, W/mK

and

Liquid viscosity, cP

200

9 8 7

150

Calculation

i

xi xj

Thermal conductivity, Wm/K

xi Pi þ

Calculations

P¼

critical ones for the accurate simulation of solidiﬁcation microstructure and defect formation, such as shrinkage porosity (Ref 8). A database has been developed containing molar volume and thermal volume coefﬁcients of expansion of liquid, solid-solution elements, and intermetallic phases. This is linked to the thermodynamic calculations mentioned previousley. The densities of the liquid and solid phases of multicomponent systems are calculated by the simple mixture model (Ref 27). Figure 6 shows plots comparing experimental values with calculations for the density of various alloys at different temperatures. Figure 7 shows a comparison between the calculated and experimentally reported density for a CF8M stainless steel alloy.

100

6 5 4 3

50

2 1 0

0 0

Fig. 3 alloys

50

100 Experiments

150

200

Comparison between experimental and calculated thermal conductivity for various

0

1

2

3

4 5 Experiments

6

7

8

9

Comparison between experimental and calculated viscosity for various alloys at different temperatures

Fig. 5

40 / Casting Design and Performance fl is the fraction of liquid, and ui;l is the actual liquid velocity. Momentum equation is:

Density, g/cm3 9 8

Calculation

7

r

6 5 4

@ui @ui @ @ui m þ ruj þ pdij m ¼ rgi ui K @t @xj @xj @xj (Eq 10)

where dij is the Kronecker delta, p is pressure, gi is gravitational acceleration, and K is permeability.

3 2

Continuity equation is:

1 1

2

Fig. 6

3

4

5 6 Experiments

7

8

Comparison between experimental calculated density for various alloys

9

and

8.0 7.8

@r @ðrui Þ þ ¼0 @t @xi

(Eq 11)

In order to solve the previous equation, proper initial conditions and boundary conditions are needed. Basic initial conditions include temperature, velocity, pressure, and so on:

This is a simple version of radiation in which it is assumed that there is only one ambient temperature present. Thus, the view factor is equal to one. For investment casting modeling, the view factor must be calculated carefully. For that case, a view-factor radiation model can be used. View-Factor Radiation Model. A net ﬂux model can be used for more complex view-factor radiation. Rather than tracking the reﬂected radiant energy from surface to surface, an overall energy balance for each participating surface is considered. At a particular surface i, the radiant energy being received is denoted qin;i . The outgoing ﬂux is qout;i . The net radiative heat ﬂux is the difference of these two: qnet;i ¼ qout;i qin;i

(Eq 16)

Using the diffuse, gray-body approximation, the outgoing radiant energy can be expressed as:

Density, g/cm3

7.6 7.4

T ðx; 0Þ ¼ To ðxÞ

qout;i ¼ sei Ti4 þ ð1 ei Þqin;i

uðx; 0Þ ¼ uo ðxÞ

The ﬁrst term in the previous equation represents the radiant energy, which comes from direct emission. The second term is the portion of the incoming radiant energy, which is being reﬂected by surface i. The incoming radiant energy is a combination of the outgoing radiant energy from all participating surfaces being intercepted by surface i. Speciﬁcally, the view factor Fij is the fraction of the radiant energy leaving surface j that impinges on surface i. Thus:

7.0

Calculation Experiment

6.8

vðx; 0Þ ¼ vo ðxÞ

6.6 6.4 6.2

T ðx; 0Þ ¼ wo ðxÞ

6.0 0

Fig. 7

200

400

600 800 1000 1200 1400 1600 Temperature, °C

pðx; 0Þ ¼ po ðxÞ

Comparison between experimental and calculated density for a CF8M stainless steel alloy

The solidiﬁcation process depends on heat transfer from the part to the mold and from the mold to the environment. The heat can be transferred by conduction, convection, and/or radiation. Conduction refers to the heat transfer that occurs as a result of molecular interaction. Convection refers to the heat transfer that results from the movement of a liquid, such as liquid metal or air. Radiation refers to the heat transfer of electromagnetic energy between surfaces that do not require an intervening medium. Radiation is very important for investment casting processes, for instance. Basic mathematical formulations that govern the solidiﬁcation process can be summarized as follows (Ref 28): Energy equation is: r

(Eq 17)

7.2

@H @H þ rui rðkrT Þ qðxÞ ¼ 0 @t @xi

(Eq 8)

where r is density, k is thermal conductivity; H is enthalpy, a function of temperature that encompasses the effects of speciﬁc and latent heat; and ðT HðT Þ ¼ Cp dT þ L½1 fs ðT Þ

(Eq 9)

0

where L is latent heat, fs is the fraction of solid, ui ¼ fl ui;l is the component superﬁcial velocity,

Boundary Conditions. There are many kinds of boundary conditions. Some of the most common ones are listed as follows. The ﬁxed value or Dirichlet boundary condition is: Y ðxÞ ¼ Yd ðxÞfðtÞ on 1

(Eq 12)

where Y ðxÞ can be temperature, velocity, or pressure; 1 is some subset of the total boundary; Yd ðxÞ is speciﬁed variable vector; and f(t), the time function. The speciﬁed heat ﬂux boundary condition is: krT n_ ¼ qn fðtÞ on 2

(Eq 13)

where qn is the speciﬁed heat ﬂux, n_ is a unit vector normal to the surface, and 2 is some subset of the total boundary . The convective heat ﬂux boundary condition is: krT n_ ¼ qn fðtÞhfðtÞgðT Þ½T Ta ; on 2 (Eq 14)

where h is the convection coefﬁcient, g(T) is the temperature function, and Ta is the ambient or media temperature that could be a function of time. The radiation heat-boundary condition is: krT n_ ¼ segðT Þ½T 4

Ta4 ;

(Eq 15)

on 2 where s is the Stefan-Boltzmann constant, and e is emissivity.

qin;i ¼

N X

Fij qout;j

(Eq 18)

j¼1

where N is the total number of surfaces participating in the radiation model, and the view factors are calculated from the following integral: Fij ¼

1 Ai

ð ð

cos yj cos yi dAi dAj r2

(Eq 19)

Aj Ai

where Ai is the area of surface i, yi is the polar angle between the normal of surface i and the line between i and j, and r is the magnitude of the vector between surface i and j. Then, the vector of radiosities qout;i can be solved by: N X j¼1

Ai ei Ai dij Ai Fij qout;j ¼ sT 4 ð1 ei Þ ð1 ei Þ i (Eq 20)

Hence, the net radiant ﬂux is obtained by: qnet;i ¼

ei t sTi qout;i 1 ei

(Eq 21)

This heat ﬂux then appears as a boundary condition for the heat conduction analysis.

Modeling of Casting and Solidiﬁcation Processing / 41 Based on the previous equations, the basic heat-transfer and ﬂuid ﬂow problems can be solved with proper initial and boundary conditions.

Microstructure Simulation The purpose of casting solidiﬁcation micromodeling is to predict the microstructure of a casting, such as grain size. Understanding the casting solidiﬁcation process and microstructure formation will greatly facilitate casting design and quality control. Computer modeling also provides the basis for computer-aided manufacturing and product life-cycle management by being able to predict mechanical properties, quality, and useful life (Ref 29–34). There are several ways to simulate the microstructure formation during alloy solidiﬁcation, such as the deterministic method or the stochastic method. For the deterministic method, the density of grains that have nucleated in the bulk liquid at a given moment during solidiﬁcation is a deterministic function, for example, a function of undercooling. The stochastic method is a probabilistic means to predict the nucleation and growth of the grain, including the stochastic distribution of nucleation locations, the stochastic selection of the grain orientation, and so on. The stochastic method includes cellular automata and the phase-ﬁeld method.

Deterministic Micromodeling Modeling of solidiﬁcation processes and microstructural features has beneﬁted from the introduction of averaged conservation equations and the coupling of these equations with microscopic models of solidiﬁcation. When conservation equations are averaged over the liquid and solid phases, the interfacial continuity condition automatically vanishes and average entities (e.g., mean temperature or solute concentration) appear. Rappaz et al. (Ref 35, 36) proposed a model using averaging methods to predict the growth of equiaxed grains under isothermal conditions. The nucleation is based on the model proposed by Thevoz and co-workers (Ref 37), which is illustrated in Fig. 8. At a given undercooling, the grain density, n, is calculated by the integral

dn d (DT )

of the nucleation site distribution from zero undercooling to the current undercooling. The density of the grain nuclei is: nmax nðT ðtÞÞ ¼ pﬃﬃﬃﬃﬃﬃ 2 Td ! TððtÞ ðT ðtÞ Tn Þ2 exp dðT ðtÞÞ 2Ts2

(Eq 22)

where nmax is the maximum grain nuclei density, Td is the standard deviation undercooling, and Tn is the average undercooling. Rappaz and Boettinger (Ref 29) studied the growth of an equiaxed multicomponent dendrite. In their study, for each element, the supersaturation is cl;j co;j ¼ IvðPej Þ cl;j ð1 kj Þ

(Eq 23)

where J = 1, n is the solute element, cl;j is the tip liquid concentration, co;j is the nominal concentration, and kj is the partition coefﬁcient. The Peclet number is deﬁned as: Pej ¼

Rv 2Dj

where Dj is the diffusion coefﬁcient. The Ivanstsov function is deﬁned as IvðPeÞ ¼ Pe expðPeÞ E1 ðPeÞ, where E1 ðPeÞ is the ﬁrst exponential integral. Assuming growth at the marginal stability limit, the dendrite radius is calculated by: R¼P n j¼1

22 co;j ð1kj Þ mj Pej 1ð1k j ÞIvðPej Þ

(Eq 24)

where G is the Gibbs-Thomson coefﬁcient. Hence, the tip velocity is: v ¼ D1 Pe1

2 R

(Eq 25)

and the tip liquid concentration will be: cl;j ¼

co;j 1 ð1 kj ÞIvðPej Þ

(Eq 26)

Assume the liquid concentration in the interdendritic region is uniform. The solute proﬁles in the extradendritic liquid region can be obtained from an approximate model (Ref 29):

DTd DT

Jj ¼ Dj

n nmax

cl;j co;j dj =2

(Eq 27)

where the solute layer thickness is: DTn

Fig. 8

Nucleation model

DT

dj ¼

2Dj v

So the solute balance will become:

n X

(Eq 29)

mj Jj ¼ 0

j¼1

0

j ¼

n dfs X dT mj ðkj 1Þcl;j þ ðfg fs Þ dt dt j¼1

(Eq 28)

where fg is the envelope volume divided by the ﬁnal grain volume. Please refer to Ref 29 for details about the derivation of the previous equations. The secondary dendrite arm spacing is calculated by: l2 ¼ 5:5ðMtf Þ1=3

(Eq 30)

Where M¼P n j¼1

mj ð1 kj Þðce;j co;j Þ=Dj 0

1 m ð1 k Þc =D j j e;j j Bj¼1 C B C lnB P C @n A mj ð1 kj Þco;j =Dj n P

(Eq 31)

j¼1

Based on the previous equations, the solidiﬁcation of a multicomponent casting can be predicted. The averaged microstructure, such as grain size and secondary dendrite arm spacing, can be calculated based on the chemistry and cooling conditions. Some examples using such deterministic micromodeling are provided in a later section.

Cellular Automaton Models Cellular automaton (CA) models are algorithms that describe the discrete spatial and/or temporal evolution of complex systems by applying deterministic or probabilistic transformation rules to the sites of a lattice. In a CA model, the simulated domain is divided into a grid of cells, and each cell works as a small, independent automaton. Variables and state indices are attributed to each cell, and a neighborhood conﬁguration is also associated with it. The time is divided into ﬁnite steps. At a given time step, each cell automaton checks the variables and state indices of itself and its neighbors at the last time step and then decides the updated results at the present step according to the predeﬁned transition rules. By iterating this operation with each time step, the evolution of the variables and state indices of the whole system is obtained. The CA model is usually coupled with a ﬁnite-element (FE) heat ﬂow solver, such as the CAFE model developed by Rappaz and colleagues. The CA algorithm can be used to simulate nucleation and growth of grains. This model can be used to predict cellular-to-equiaxed transition (CET) in alloys. Several models have been developed over the years for the prediction of microstructure formation in casting (Ref 38–42). One example is shown in Fig. 9.

42 / Casting Design and Performance Cellular automaton-ﬁnite element models are particularly well suited to tracking the development of a columnar dendritic front in an undercooled liquid at the scale of the casting (Ref 40,41). Although these models do not directly describe the complicated nature of the solid/liquid interface that deﬁnes the dendritic microstructure, the crystallographic orientation of the grains as well as the effect of the ﬂuid ﬂow can be accounted for to calculate the undercooling of the mushy-zone growth front. Two- and three-dimensional CAFE models were successfully applied to predict features such as the columnar-to-equiaxed transition observed in aluminum-silicon alloys (Ref 40), the selection of a single grain and its crystallographic orientation due to the competition among columnar grains taking place while directionally solidifying a superalloy into a pigtail shape (Ref 41), as well as the effect of the ﬂuid ﬂow on the ﬁber texture selected during columnar growth (Ref 42). Coupling with macrosegregation has been developed (Ref 43), thus providing an advanced CAFE model to account for structure formation compared to purely macroscopic models developed previously, for example, Ref [44 to 46]. While both structure and segregation were predicted (Ref 43), comparison with experimental observation concerning structure formation was limited due to the lack of detailed data (Ref 47). Comparison is thus mainly conducted with the as-cast state. Wang et al. (Ref 48) investigated the effect of the direction of the temperature gradient on grain growth. As shown in Fig. 10(b), the temperature gradient was inclined at 45 relative to the macroscopic solidiﬁcation direction, and the magnitude of the gradient was 12 K/mm. It clearly shows that the direction of temperature gradient can affect both the macro- and microscale dendritic structures, as well as the maximum undercooling. While the CA methods produce realisticlooking dendritic growth patterns and have resulted in much insight into the CET, some questions remain regarding their accuracy. Independence of the results on the numerical grid size is rarely demonstrated. Furthermore, the CA techniques often rely on relatively arbitrary rules for incorporating the effects of crystallographic orientation while propagating the solid-liquid interface. It is now well accepted that dendritic growth of crystalline materials depends very sensitively on the surface energy anisotropy (Ref 49, 50).

Three-dimensional view of the ﬁnal grain structure calculated in the weak coupling mode for a directionally solidiﬁed turbine blade. The <100> pole ﬁgures are displayed for various cross sections perpendicular to the main blade axis. Source: Ref 41

Fig. 9

Phase-Field Model An alternative technique for investigating microstructure formation during solidiﬁcation is the phase-ﬁeld method. Phase-ﬁeld models were ﬁrst developed for simulating equiaxed growth under isothermal conditions (Ref 51, 52). A desirable extension of the model was to study the effect of heat ﬂow due to the release of latent heat. A simpliﬁed approach was proposed in which the temperature was assumed

(a) Predicted dendritic structure and (b) solutal adjusted undercooling distribution under thermal conditions of 45 inclined isotherms with respect to the growth direction moving at a constant velocity of 150 mm/s. Source: Ref 48

Fig. 10

Modeling of Casting and Solidiﬁcation Processing / 43 to remain spatially uniform at each instant, and a global cooling rate was imposed with consideration of the heat extraction rate and increase of the fraction of solid (Ref 53). The attempt to model nonisothermal dendritic solidiﬁcation of a binary alloy was made by Loginova et al. by solving both the solute and heat diffusion equations and considering the release of latent heat as well (Ref 54). Besides equiaxed growth in the supersaturated liquid, the phase-ﬁeld model was also applied to the simulation of directional solidiﬁcation, under well-deﬁned thermal conditions (Ref 55, 56). The phase-ﬁeld model has also been used to simulate the competitive growth between grains with different misorientations with respect to the thermal gradient (Ref 57). Further development in phaseﬁeld models includes the extension into three dimensions (Ref 58, 59) and multicomponent systems (Ref 60, 61). Usually, a regular grid composed of square elements is used in the phase-ﬁeld models (Ref 52, 53), but an unstructured mesh composed of triangular elements has also been used, which enables the phase-ﬁeld method to be applicable in a domain with complex geometry shape and also in a large scale. From a physical point of view, the phase-ﬁeld method requires knowledge of the physical nature of the liquid-solid interface. However, little is known about its true structure. Using Lennard-Jones potentials, molecular dynamics simulations of the transition in atomic positions across an interface have suggested that the interface width extends over several atomic dimensions (Ref 62). At present, it is difﬁcult to obtain usable simulations of dendritic growth with interface thickness in this range due to the limitations of computational resources. Thus, the interface width will be a parameter that affects the results of the phase-ﬁeld method. It should be realized that in the limit as the interface thickness approaches zero, the phase-ﬁeld equations converge to the sharp interface formulation (Ref 63, 64). In contrast to CA models, which adopt a pseudo-front-tracking technique, phase-ﬁeld models express the solid/liquid interface as a transitional layer that usually spreads over several cells. The diffusion equation for heat and solute can be solved without tracking the phase interface using a phase-ﬁeld variable and a corresponding governing equation to describe the state in a material as a function of position and time. This method has been used extensively to predict dendritic, eutectic, and peritectic growth in alloys, and solute trapping during rapid solidiﬁcation. (Ref 65). The interface between liquid and solid can be described by a smooth but highly localized change of a variable between ﬁxed values such as 0 and 1 to represent solid and liquid phases. The problem of applying boundary conditions at an interface whose location is an unknown can be avoided. Phase-ﬁeld models have recently become very popular for the simulation of microstructure evolution during solidiﬁcation processes (Ref

66–70). While these models address the evolution of a solid-liquid interface using only one phase-ﬁeld parameter, interaction of more than two phases or grains, and consequently the occurrence of triple junctions, needed to be included into the multiphase-ﬁeld approach (Ref 71–74). The deterministic model is capable of tracking the evolution of the macroscale or average variables, for example, average temperature and the total fraction of solid, but it cannot simulate the structure of grains. The CA models can simulate the macroscale and mesoscale

grain structures, but it has difﬁculty in resolving the microstructure. The phase-ﬁeld method can well reproduce the microstructure of dendritic grains (Fig. 11, 12). However, with the current computational power, phase-ﬁeld models can only work well on a very small scale (up to hundreds of Micrometers). The typical scale of laboratory experiments is 1 cm, and the scale of an industry problem can be up to 1 m. Both of them are beyond the capability of the phase-ﬁeld method. In the industry, for larger castings, deterministic micromodeling still is a main player.

Predicted results for two-dimensional dendritic solidiﬁcation of a binary alloy into an undercooled melt with coupled heat and solute diffusion for Le = 50 at tD/d0 2 = 3500. The upper- and lower-right quadrants show the dimensionless concentration U and temperature ﬁelds, respectively; the left quadrants both show concentration c/c0 ﬁelds, with different scales used in the upper and lower quadrants to better visualize the concentration variations in the solid and liquid, respectively. Source: Ref 74

Fig. 11

Fig. 12

Melting of dendritic structure and formation of fragments when temperature is increased from the growth temperature of 1574 K shown in (a) to 1589 K shown at later times in (b–d). Source: Ref 65

44 / Casting Design and Performance Micromodeling Applications in the Industry Here, the focus is on deterministic modeling due to its wide application in the casting industry. Thermodynamic calculations are coupled with the macroscale thermal and ﬂuid ﬂow calculations (Ref 75). Ductile iron is chosen as an example to demonstrate the capability of deterministic micromodeling because of its complex solidiﬁcation behavior. Ductile irons are still important engineering materials due to their high strength and toughness and relatively low price. In the foundries, ductile irons suffer from shrinkage porosity formation during solidiﬁcation, which is detrimental to the mechanical properties. To minimize porosity formation, large risers are normally used in the design, which reduces porosity level sometimes but leads to a low yield. Due to the complex solidiﬁcation behavior of ductile irons and their extreme sensitivity to the process, it is very difﬁcult to optimize the casting design to ensure the soundness of castings. To better understand the shrinkage behavior of ductile iron during solidiﬁcation, a micromodel was developed to simulate the microstructure formation. The density change during solidiﬁcation and the room-temperature mechanical properties can be calculated based on the microstructure. The simulation has been compared with the experimental results and found to be in good agreement. Cast iron remains the most important casting material, with over 70% of the total world tonnage now (Ref 76). Based on the shape of graphite, cast iron can be lamellar (ﬂake) or spheroidal (nodular). In the last 40 years, many papers have been published on the modeling of ductile iron solidiﬁcation. It started with the computational modeling by analytical heat transport and calculation of transformation kinetics (Ref 77–83). The computer model can calculate the cooling curve with an analytical method together with the kinetics calculation of carbon diffusion through the gamma-phase shell. In 1985, Su et al. coupled heat transfer and a solidiﬁcation kinetics model ﬁrst, using the ﬁnite-difference method (Ref 84). After that, many papers have been published on micromodeling of ductile iron solidiﬁcation (Ref 85–95). The carbon diffusioncontrolled growth through the gamma shell was modeled. In those models, the nodule count, graphite radius, and austenite shell radius were calculated. Reference 96 used the internal state variable approach to model the multiple phase changes occurring during solidiﬁcation and subsequent cooling of near-eutectic ductile cast iron. In this simulation, the effects on the microstructure evolution at various stages of the process by the alloy composition, graphite nucleation potential, and thermal progress were illuminated. The heat ﬂow, fading effect, graphite/ austenite eutectic transformation, ledeburite eutectic transformation, graphite growth in austenite regime, and the eutectoid transformation were all modeled. A comprehensive micromodel

is developed that can give accurate microstructure information as well as the mechanical properties, such as yield strength, tensile strength, and hardness. The density of austenite, ferrite, pearlite, graphite, liquid, and ledeburite are all calculated. The prediction has been compared with the experimental results and found to be in good agreement (Ref 75). Nucleation Model. Here, Oldﬁeld’s nucleation model is applied. In this mode, bulk heterogeneous nucleation occurs at foreign sites that are already present within the melt or intentionally added to the melt by inoculation: No ¼ AðT Þn

(Eq 34)

where rG ,rg , and rl are the densities of graphite, austenite, and liquid, respectively, and the calculation can be found in the next section; RG , Rg , andRl are the radii of graphite, austenite, and the ﬁnal grain, respectively; and mav is the average mass of the grain. Assuming complete mixing of solute in liquid, the overall solute balance is written as:

(Eq 32)

where A is the nucleation constant, No is the nucleation number per unit volume, T is the undercooling, and n is another constant that depends on the effectiveness of inoculation. Fading Effect. Fading is the phenomenon whereby the effectiveness of inoculation diminishes as the time between inoculation and casting increases. It is believed that the nucleation of graphite occurs on small, nonmetallic inclusions that are entrapped in the liquid after inoculation (Ref 93). The small particles will grow with time. The particle diameter can be calculated by: d ¼ ðd3o þ ktÞ1=3

4 4 4 rG R3G þ rg ðR3g R3G Þ þ rl ðR3l R3g Þ 3 3 3 ¼ mav

(Eq 33)

where d is the particle diameter with time, do is the particle diameter at the beginning of the inoculation, and k is a kinetic constant. Graphite/Austenite Eutectic Transformation. The eutectic growth process in ductile iron is a divorced growth of austenite and graphite, which do not grow concomitantly. At the beginning of the liquid/solid transformation, graphite nodules nucleate in the liquid and grow in the liquid to a small extent. The formation of graphite nodules and their limited growth in liquid depletes the carbon in the melt locally in the vicinity of the nodules. This facilitates the nucleation of austenite around the nodules, forming a shell. Further growth of these nodules is by diffusion of carbon from the melt through the austenite shell. Once the austenite shell is formed around each nodule, the diffusion equation for carbon through the austenitic shell is solved in one-dimensional spherical coordinates. The boundary conditions are known from the phase diagram because thermodynamic equilibrium is maintained locally. Conservation of mass and solute is maintained in each grain. Because of the density variation resulting from the growth of austenite and graphite, the expansion/contraction of the grain is taken into account by allowing the ﬁnal grain size to vary. Toward the end of solidiﬁcation, the grains impinge on each other. This is taken into consideration by using the Johnson-Mehl approximation. Using a spherical coordinate system, a mass balance is written as:

rG 1

4 3 R þ 3 G

R ða

rg cðr; tÞ4r2 dr Rc

4 þ rl cl ðR3l R3g Þ ¼ cav 3

ðEq 35Þ

Differentiation of the previous two equations and the use of Fick’s law in spherical coordinates lead to two equations for graphite and austenite growth rates, following some manipulation. Ledeburite Eutectic Transformation. When the temperature reaches below the metastable eutectic temperature, the metastable phase forms. The metastable cementite eutectic is also called ledeburite, in which small islands of austenite are dispersed in the carbide phase. It has both direct and indirect effects on the properties of ductile iron castings. Increasing the volume percent of the hard, brittle carbide results in an increase in the yield strength but a reduction in the tensile strength. Following the assumptions from Onsoien (Ref 96, 97) that the graphite/ austenite nodule distribution is approximated by that of a close-packed face-centered space lattice, and that the ledeburite eutectic appears in an intermediate position, the total number of ledeburite nucleation sites would be the same for graphite/austenite nodules. The grain is assumed to be spherical. Hence, the growth of the ledeburite can be calculated as: dRLE ¼ 30:0 106 ðT Þn dt

(Eq 36)

Thus, the fraction of ledeburite can be written as: 4 fLE ¼ NR3LE 3

(Eq 37)

Eutectoid Transformation. The eutectoid reaction leads to the decomposition of austenite into ferrite and graphite for the case of the stable eutectoid and to pearlite for the metastable eutectoid transformation. Usually, the metastable eutectoid temperature is lower than the stable eutectoid temperature. Slower cooling rates result in more stable eutectoid structure. Following solidiﬁcation, the solubility of carbon in austenite decreases with the drop in temperature until the stable eutectoid temperature is reached.

Modeling of Casting and Solidiﬁcation Processing / 45 decreases the tensile strength. The yield strength and hardness continuously decrease as the cooling rate decreases. The yield strength is very high on the left part because of the formation of carbide. On the other hand, the carbide decreases the tensile strength. The results are shown in Fig. 16. Experimental Validations. A series of experiments was performed to validate the micromodel (Ref 99). The three-part cast iron foundry mold containing the gating system is shown in Fig. 17. The casting is GGG60 ductile iron. The pouring temperature is 1400 C, the initial die

temperature is 165 C, and the initial sand temperature is 20 C. To establish the structure of the casting and the morphology of graphite, specimens were taken, as shown in Fig. 18. The specimens were then ground, polished, and etched for structure evaluation. It can be seen in the pictures of the microstructure that graphite was segregated in the form of spheroids. Because of the rapid cooling, a large amount of the metastable phase, ledeburite, was formed in the corner. The ledeburite phase reduces gradually as the cooling rate decreases. In the center of the casting, no ledeburite phase was

9

110 100

8

Solidification time

7

Nodule count

Solidification time, s

80 6

70 60

5

50

4

40

3

30 2

20

1

10 0

0 0

2 4 6 8 Distance from the cooling end X, cm

10

Solidiﬁcation time and nodule count at various distances from the chill

1.0

9

0.9

8

0.8

7

0.7

6 Fraction of ledelurite

0.6

5

Fraction of ferrite

0.5

Fraction of pearlite

0.4

4

Elongation

3

0.3 0.2

2

0.1

1

0.0

0 0

Fig. 14

1

2

3 4 5 6 7 8 Distance from the cooling end X, cm

Phase fractions and elongation of the casting at various distances from the chill

9

10

Elongation, %

Fig. 13

Nodule count, 107/cm3

90

Phase fraction

The rejected carbon migrates toward graphite nodules, which are the carbon sinks. This results in carbon-depleted regions in austenite around the graphite nodules. This provides favorable sites for ferrites to nucleate, which grow as a shell around the graphite nodules. If the complete transformation of austenite is not achieved when the metastable temperature is reached, pearlite forms and grows in competition with ferrite. The ultimate goal of process modeling is to predict the ﬁnal mechanical properties. The mechanical properties (hardness, tensile strength, yield strength, and elongation) of ductile iron castings are a function of composition and microstructure. The graphite shape, graphite structure, graphite amount, carbide content, and matrix structure (pearlite, ferrite) all affect the mechanical properties of ductile iron castings. As for the matrix structure, the increasing of pearlite increases the strength and hardness but reduces the elongation (Ref 98). To show the capability of this model, a simulated ductile iron casting with a simple geometry was investigated. The dimension of the casting is 10 by 10 by 200 cm. On the left face, it is cooled by contact with a constanttemperature media (15 C) at a heat-transfer coefﬁcient of 500 W/m2K. All the other faces are adiabatic. The initial melt temperature is 1400 C. According to the boundary condition, the left side cools faster than the right side. Figure 13 shows the solidiﬁcation time for different distances from the cooling end. At the very left end, the solidiﬁcation time is less than 1 s. On the other hand, the solidiﬁcation at 10 cm from the cooling end is more than 100 s. Because of the different cooling, the nodule count varies and is shown in the same ﬁgure. The metastable phase forms when the cooling is too fast. Figure 14 shows the volume fraction of different phases at room temperature. On the very left end, there is approximately 90% volume fraction of ledeburite phase. It reduces gradually from left to right until approximately 3 cm from the chill end. There is no ledeburite phase after 3 cm. At the same time, as cooling decreases, the volume fraction of ferrite increases and that of pearlite decreases. Ledeburite is a very hard, brittle phase. The pearlite phase is harder than ferrite. Hence, the ductility increases as the cooling rate decreases. From the micromodeling, the calculated grain and graphite radius at different distances from the chill are shown in Fig. 15. Faster cooling results in smaller grain and graphite sizes. The ratio of the radius of graphite and austenite increases as cooling decreases but reaches a constant value of approximately 0.44, even though the radius of graphite and austenite still continue to increase. This constant ratio is determined by the initial carbon content. It can determine the expansion level during solidiﬁcation. Based on the microstructure, the mechanical properties can be calculated. As mentioned previously, carbide increases the yield strength but

46 / Casting Design and Performance 0.03

0.50 0.48 0.46 0.44

Radius, mm

0.02

0.40 0.38 0.01

Porosity

0.36

Grain radius

0.34

Graphite radius Rg/Ra

0.32

0.00

0.30 0

Fig. 15

Rg/Ra

0.42

2 4 6 8 Distance from the cooling end X, cm

10

Grain and graphite size of the casting at various distances from the chill

650

400 380

600 360 340

500 450

Tensile strength

320

Yield strength

300

Hardness

280

Hardness, HB

Strength, MPa

550

260

400

240 350 220 300

200 0

Fig. 16

1

2 3 4 5 6 7 8 Distance from the cooling end X, cm

9

10

Mechanical properties of the casting at various distances from the chill

found. The radius of the black balls, graphite, increases as cooling decreases. The structure of the metal is formed by pearlite and ferrite. Figure 19 shows the volume fraction of metastable phase (top) and volume fraction of ferrite (bottom). It is difﬁcult to measure the yield strength of the sample at different locations because the strength could change dramatically based on the microstructure variation. On the other hand, hardness is an excellent indicator of strength and relatively easy to measure. Figure 20 shows the hardness measurement points on the sample. Table 2 shows the

the microstructure. There are many kinds of casting defects. Those defects are dependent on the chemistry of alloys, casting design, and casting processes. Defects can be related to thermodynamics, ﬂuid ﬂow, thermal, and/or stress. For most cases, all of those phenomena are correlated. It is necessary to consider every aspect to prevent a defect forming. Here, some common casting defects in foundries are discussed. They are porosity, hot tearing, and macrosegregation.

comparison between the measurement and prediction results of the hardness at different locations. It can be concluded that the prediction matches the experiments very well.

Defect Prediction Defects reduce the performance and increase the cost of castings. It is critical to understand the mechanism of defects and microstructure on the performance so that an effective tool can be developed to prevent defects and control

Porosity formed in castings leads to a decrease in the mechanical properties (Ref 100–106]. This porosity may be a combined result of solidiﬁcation shrinkage and gas evolution. They can occur simultaneously when conditions are such that both may exist in a solidifying casting. One of the most effective ways to minimize porosity defects is to design a feeding system using porosity prediction modeling. In such a way, the model can determine the location of microporosity so that the feeding system can be redesigned. This process is repeated until microporosity is minimized and not likely to appear in the critical areas of the castings. There are many models that can predict the shrinkage porosity from the pressure drop during interdendritic ﬂuid ﬂow and gas evolution (Ref 104–111). The model of Felicelli et al. can predict the pressure and redistribution of gas and the region of possible formation of porosity by solving the transport of gas solutes (Ref 112). A comprehensive model should calculate the shrinkage porosity, gas porosity, and pore size. As for many casting defects observed in solidiﬁcation processes, the mushy zone is the source of microporosity. The basic mechanism of microporosity formation is pressure drop due to shrinkage and gas segregation in the liquid (Ref 101, 102, 104). The liquid densities of many alloys are lower than that of the solid phase. Hence, solidiﬁcation shrinkage happens due to the metal contraction during the phase change. The dynamic pressure within the liquid decreases because of the contraction and sometimes cannot be compensated by the metallostatic pressure associated with the height of the liquid metal. The decrease of pressure lowers the solubility of gas dissolved in the liquid. If the liquid becomes supersaturated, then bubbles can precipitate (Ref 113). Most liquid metals can dissolve some amount of gas. The solubility of gases in the solid phase is usually much smaller than that in the liquid phase. Normally, the rejected gases during solidiﬁcation do not have enough time to escape from the mushy zone into the ambient air. Being trapped within the interdendritic liquid, the gas can supersaturate the liquid and eventually precipitate under the form of pores if nucleation conditions are met. The formation of bubbles requires overcoming the surface tension (Ref 114). Homogeneous

Modeling of Casting and Solidiﬁcation Processing / 47 Metal die

Specimen Sand mold

Metal base Gating system

Fig. 17

Experiment setup

to shaped castings was done by Kubo and Pehlke (Ref 114). A more accurate ﬂuid ﬂow model was presented by Combeau et al. (Ref 105). In their study, the interdendritic ﬂow for three-dimensional simulations of mold ﬁlling was included but without considering microporosity. Among all these models, no one calculated the diffusion and convection of gas. Felicelli et al. (Ref 112) calculated the redistribution of gas during solidiﬁcation but did not predict how much porosity forms and the sizes of the pores. To predict microporosity defects in casting processes accurately, the following factors that contribute to microporosity formation should be considered: macroscopic heat transfer, interdendritic ﬂuid ﬂow, gas redistribution by diffusion and convection, microstructure evolution, and microporosity growth. The conservation equation beside the equations mentioned early on is gas conservation. Gas conservation: @r c @ð1 fl Þr þ rui rcl ¼ ð1 kÞcl þ rðfl rrcl Þ @t @t (Eq 38)

where c ¼ fl cl þ ð1 fl Þcs ,cs ¼ kcl ;cl is the gas concentration in the liquid;cs is the gas concentration in the solid; and k is the partition coefﬁcient. In addition to the liquid and solid fraction, which are calculated from the energy equation, the dendrite spacing is needed to estimate pore curvature and permeability in the mushy zone. The pore radius or curvature is taken to be proportional to the dendrite cell spacing through the relationship: 1 r ¼ fl d 2

where d is the secondary dendrite arm spacing. Gas Porosity Evolution. Pores will form in a solidifying alloy when the equilibrium partial pressure of gas within the liquid exceeds the local pressure in the mushy zone by an amount necessary to overcome surface tension. Hence, gas porosity develops when:

Fig. 18

Microstructure of ductile iron at indicated points

nucleation is very difﬁcult. In castings, nucleation of pores can be expected to occur primarily on heterogeneous nucleation sites, such as the solid-liquid interface and inclusions (Ref 114, 115). Generally speaking, there are two ways to predict the level of microporosity in castings. One is a parametric method derived from ﬁrst principles by using a feeding resistance criteria function combined with macroscopic heat ﬂow calculations (Ref 116–119). Parametric models are easy to apply to shaped castings and have been mainly directed at the prediction of centerline shrinkage. Another approach is a direct simulation method (Ref 104–108, 112, 114).

Pg > Pa þ Pm þ Pd þ Pd

Fig. 19

Simulation results of fraction of metastable phase (top) and fraction of ferrite (bottom)

They usually derive governing equations based on a set of simplifying assumptions and solve the resulting equations numerically. By combining the cellular automata technique, some models can not only predict the percentage porosity but also the size, shape, and distribution of the pores (Ref 109–111). There has been some research that attempted to understand the physics of microporosity formation, too (Ref 103). The earliest work to directly predict microporosity distribution that was general and applicable

(Eq 39)

where Pg is the Sievert pressure, Pa is the ambient pressure,Pm is the metallostatic pressure, Pd is the pressure drop due to the friction within the interdendritic liquid, Pd ¼ 2s=r is the surface tension, s is the surface tension, and r is the pore radius. The maximum dissolved gas (hydrogen or nitrogen) in the liquid, gl , or solid, gs , and the gas pressure are related through Sievert’s law: gl ¼

Kg 1=2 P gs ¼ kgl fg g

(Eq 40)

where Kg and fg are the equilibrium constant and activity coefﬁcient, respectively, for hydrogen. The activity coefﬁcient for hydrogen in

48 / Casting Design and Performance aluminum alloy is estimated (Ref 112, 113) as, for example: ln fH ¼

N X

ajH C j

þ

j¼1

N X

bjH ðC j Þ2

(Eq 41)

j¼1

where ajH and bjH are interaction coefﬁcients, and C j is the concentration of solute element j. The detailed coefﬁcients can be found in Ref 112 and 113. The equilibrium constant is calculated as: ln KH ¼ 3:039

6198:47 T

(Eq 42)

The volume fraction of gas porosity is: cl fl þ cs ð1 fl Þ ¼ gl fl þ gs ð1 fl fv Þ þ

Pg fv T (Eq 43)

where fv is the volume fraction of gas porosity, and is the gas conversion factor. If no pore has formed yet, then: cl fl þ cs ð1 fl Þ ¼ gl fl þ gs ð1 fl Þ

(Eq 44)

Shrinkage Porosity. If the pressure drops below the cavitation pressure, it is assumed that liquid feeding ceases and that the solidiﬁcation shrinkage in that computational cell is compensated only by pore growth. In general, cavitation pressure is very small. When the liquid pressure drops below the cavitation pressure, the porosity is determined such that it compensates for the entire solidiﬁcation shrinkage within the current time step: fvnþ1 ¼ fvn þ

rnþ1 rn rnþ1 =ð1 fvn Þ

Macrosegregation Modeling and simulation of macrosegregation during solidiﬁcation has experienced explosive growth since the pioneering studies of Flemings and co-workers in the mid-1960s (Ref 120–122). Beckerman did a comprehensive review of recent macrosegregation models and their application to relevant casting industries (Ref 123). There are numerous factors that can cause macrosegregation during casting solidiﬁcation processes. Those include thermal- and soluteinduced buoyancy, forced ﬂow, solid movement, and so on. Macrosegregation models have been applied extensively in the casting industry, such as steel ingot castings, continuous and direct chill castings, nickel-base superalloy single-crystal castings (freckle simulation), and shape castings as well. Freckles have been the subject of intense research efforts for approximately 30 years due to their importance as a defect in alloy casting (Ref 124, 125). They represent a major problem in directionally solidiﬁed superalloys used in the manufacture of turbine blades (Ref 124– 128). Upward directional solidiﬁcation provides an effective means of producing a columnar

microstructure with all the grain boundaries parallel to the longitudinal direction of the casting. In conjunction with a grain selector or a preoriented seed at the bottom of the casting, directional solidiﬁcation is used to make entire castings that are dendritic single crystals. During such solidiﬁcation, the melt inside the mushy zone can become gravitationally unstable due to preferential rejection of light alloy elements (for a partition coefﬁcient less than unity) into the melt. Since the mass diffusivity of the liquid is much lower than its heat diffusivity, the segregated melt retains its composition as it ﬂows upward and causes delayed growth and localized remelting of the solid network in the mush. Ultimately, a pencil-shaped vertical channel, devoid of solid, forms in the mushy zone through which low-density, highly segregated liquid ﬂows upward as a plume or solutal ﬁnger into the superheated melt region above the mushy zone. This ﬂow is continually fed by segregated melt ﬂowing inside the mushy zone radially toward the channel. At the lateral boundaries of the channel, dendrite arms can become detached from the main trunk, and those fragments that remain in the channel are later observed as freckle chains. The complex convection phenomena occurring during freckle formation represent a formidable challenge for casting simulation

(Eq 45)

Experimental Validation. In this example, a set of castings with different initial hydrogen content using an iron chill plate was simulated and compared with the experiment for an A319 casting. The geometry and mesh are shown in Fig. 21. The casting is 132 mm in height, 220 mm in length, and of varying thickness. Wedges are cut horizontally at 35 mm from the bottom end and with a thickness of 12 mm. The initial pouring temperature is 750 C. Initial hydrogen contents are 0.108, 0.152, and 0.184. The experiments and simulation results are taken at different distances from the chill end. The comparison of the value of percentage porosity against local solidiﬁcation time and hydrogen content between simulation and experiment is shown in Fig. 22. It shows that increasing solidiﬁcation time and hydrogen content increases considerably the percentage of porosity. Numerical simulation results give excellent agreement with the measurements of percentage of porosity. The results also show that local solidiﬁcation time and initial hydrogen content are very important factors inﬂuencing the formation of porosity.

Ingate

Fig. 20

Sample for hardness measurement

Table 2 Comparison between measured and predicted hardness Location

1A B 2 3 4

Dimension x, mm

4 10 50 50 50

Dimension y, mm

4 7 4 10 48

Measurement, HB

368 313 249 236 209

Simulated, HB

371 320 255 245 203

Modeling of Casting and Solidiﬁcation Processing / 49 (Ref 123, 129–131). In 1991, Felicelli et al. simulated channel formation in directional solidiﬁcation of Lead-tin alloys in two dimensions (Ref 132). Since then, numerous studies have been performed to simulate and predict freckling in upward directional solidiﬁcation (Ref 133–142). Neilson and Incropera performed the ﬁrst three-dimensional simulations of channel formation in 1993 (Ref 136). However, the coarseness of the mesh caused a serious lack of resolution and inaccuracies. Threedimensional simulations have also been performed by Poirier, Felicelli, and co-workers for both binary and multicomponent alloys (Ref 141–143). Freckle formation can be simulated with a commercial package, such as ProCAST. Figure 23 illustrates the freckle formation for a Pb10%Sn binary alloy directionally solidiﬁed in a simple geometry. A temperature gradient is initially imposed to simulate a directional solidiﬁcation system. Cooling is achieved by lowering the temperatures of the upper and lower walls of the cavity at a constant rate, such that the overall temperature gradient is maintained over the height of the cavity. The lateral walls of the cavity are taken as adiabatic.

Hot Tearing Hot tearing is one of the most serious defects encountered in castings. Many studies have revealed that this phenomenon occurs in the late stage of solidiﬁcation when the fraction of solid is close to one. The formation and propagation of the hot tearing have been found to be directly affected by the cooling history, the chemical composition and mechanical properties of the alloy, as well as the geometry of the casting. Various theories have been proposed in the literature on the mechanisms of hot tearing formation. Detailed reviews on the theories and

experimental observations of the formation and evolution of hot tearing can be found in Ref 144 and 145 and the references therein. Most of the existing hot tearing theories are based on the development of strain, strain rate, or stress in the semisolid state of the casting. For strain-based theory, the premise is that hot tearing will occur when the accumulated strain exceeds the ductility (Ref 146–148). The strain-rate-based theories suggest that hot tearing may form when the strain rate, or strainrate-related pressure, reaches a critical limit during solidiﬁcation (Ref 149, 150). The stress-based criteria, on the other hand, assumes that hot tearing will start if the induced stress in the semisolid exceeds some critical value (Ref 151, 152). Although these theories were proposed independently as distinct theories, they can, indeed, be considered as somewhat related due to the relationship between strain, strain rate, and stress. It is such a relationship that motivates the development of a hot tearing indicator which uses the accumulated plastic strain as an indication of the susceptibility of hot tearing. This considers the evolution of strain, strain rate, and stress in the last stage of solidiﬁcation. A Gurson type of constitutive model, which describes the progressive microrupture in the ductile and porous solid, is adopted to characterize the material behavior in the semisolid state. The proposed hot tearing indicator, while veriﬁed speciﬁcally for magnesium alloys, has much wider application. To reliably predict the formation and evolution of hot tearing in casting by numerical simulations, it is critical to have accurate thermophysical and mechanical properties, especially in the mushy zone. It is also essential that the solidiﬁcation path of the alloy be accurately described. The prediction of the thermophysical and mechanical properties has recently become possible by using the

knowledge of the microstructure, phase fractions, and defects present in a metallic part (Ref 100). The solidiﬁcation path can be obtained with the help of thermodynamic calculations of phase stability at given temperatures and compositions. A comprehensive multicomponent alloy solidiﬁcation model, coupled with a Gibbs free-energy minimization engine and thermodynamic databases, has been developed to facilitate such calculations (Ref 8). With the integration of a back-diffusion model in the calculation, solidiﬁcation conditions, such as cooling rate, can also be taken into account. Hot Tearing Indicator. A constitutive model used to describe the material behavior in the semisolid state is the Gurson model (Ref 153–155), which was originally developed for studying the progressive microrupture through nucleation and growth of microvoids in the material of ductile and porous solid. When the material is considered as elasticplastic, the yield condition in the Gurson model is of the form: fðs; x; T; ep ; Gu Þ ¼ F ðsÞ Gu ðs; ep ; fv Þkðep ; T Þ ¼ 0 (Eq 46) 1=2

where F ðsÞ ¼ ð3ðs xÞ : ðs xÞ=2Þ is the Mises stress in terms of the deviatoric stress s ¼ s ðtrsÞI=3. k represents the plastic ﬂow stress due to isotropic hardening, and x denotes back stress due to kinematic hardening. The accumulated effective plastic strain is written as: ep ¼

ð t pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ð2=3Þ_ep : e_ p dt

(Eq 47)

0

with e_ p ¼ g_

@f @s

(Eq 48)

and g_ being the plastic ﬂow parameter. The Gurson coefﬁcient, Gu, is deﬁned as: trðsÞ Gu ¼ 2f q1 cosh þ f1 þ ðq1 f Þ2 g 2k

Chill

Casting

(Eq 49)

in which, q1 is a material constant and: Percentage of porosity, %

132 mm

2.5

Porosity, %

2.0 1.5 H=0.108ppm H=0.108ppm H=0.152ppm H=0.152ppm H=0.184ppm H=0.184ppm

1.0 0.5 0.0

60 mm

Fig. 21

Geometry and mesh

150 mm

0

100

70 mm

Fig. 22

200

300 400 500 600 700 Local solidification time, s

800

900

Comparison between experiment (symbols) and calculation (lines)

50 / Casting Design and Performance

Fig. 23

Macrosegregation for directional solidiﬁcation of a Pb-10%Sn binary alloy. (a) Final tin composition after solidiﬁcation. (b) Cut-off view

f_nucleation

f ¼ fv

for fv fc fu fc f ¼ fc þ ðfv fc Þ for fv > fc fF fc

(Eq 51)

with the rate of void growth deﬁned as: p f_growth ¼ ð1 f Þtrð_ e Þ 3f q1 trðsÞ _ f Þ sin h ¼ gð1 2k k

(Eq 53)

where

Here, fu = 1/q1, fc is the critical void volume fraction, and fF is the failure void volume fraction. Their values should be different for different materials. In the calculation for an indicator for hot tearing, constants are used. Here, q1 = 1.5, fc = 0.15, and fF = 0.25, as used in (Ref 154). The Gurson coefﬁcient characterizes the rapid loss of material strength due to the growth of void volume fraction, fv . When fv = fF, then f = fu = 1/ q1 , are Gu = 0, for zero stress; that is, the stresscarrying capacity of the material vanishes. The evolution of the void volume fraction is described by the nucleation of the new void and the growth of the existing void: f_v ¼ f_nucleation þ f_growth

¼ e_ ht

(Eq 50)

(Eq 52)

In this study, the nucleation of the void is assumed to be strain controlled and is written as:

eht ¼

ð t pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ð2=3Þ e_ p : e_ p dt

tc t ts

(Eq 54)

tc

is deﬁned as the hot tearing indicator. tc represents time at coherency temperature, and ts denotes time at solidus temperature. It is observed that the hot tearing indicator is in fact the accumulated plastic strain in the semisolid region, and it corresponds to the void nucleation. Therefore, it should provide a good indication for the susceptibility of hot tearing during solidiﬁcation. The value of the hot tearing indicator is determined by ﬁnite-element analysis (Ref 156). For materials described by a viscoplastic or creep model, yield condition does not exist. The function f deﬁned in Eq 46 can be used as a potential for the inelastic ﬂow, so that the inelastic part of the strain rate can still be given in the form of Eq 48. Experiment Validation. Cao et al. performed some experiments to study the hot tearing formation during solidiﬁcation of binary Mg-Al and ternary Mg-Al-Ca alloys in a steel mold (Ref 157–160), which is shown in Fig. 24. A hot cracking susceptibility was

introduced, which is a function of maximum crack width, crack length factor, and the crack location. It was found that it is easier to have a crack at the sprue end than at the ball end. It is less likely to have a crack in the middle of the rod. Also, a longer rod is easier to crack. Figure 25 shows the simulated results of a hot tearing indicator for a Mg-2%Al alloy casting. The computed hot tearing indicator agrees very well with the experiments. Figure 26 shows the experimental results of hot tearing at the sprue end of the rods for three different alloys. The calculated hot tearing indicators are shown in Fig. 27 accordingly. It can be seen that the hot tearing is less severe as the aluminum content increases from 2% to 4% and then to 8% at the same location for the same casting with the same casting conditions. Again, the simulated hot tearing indicators agree well with the observations. The susceptibility rises sharply from pure magnesium, reaches its maximum at Mg-1%Al, and decreases gradually with further increase in the aluminum content. The hot tearing indicator is calculated at the end of the sprue for the longest rod with a different alloy composition. For comparison, the hot tearing indicators as well as a crack susceptibility coefﬁcient (CSC), which is deﬁned as the temperature difference between the fraction of solid

Modeling of Casting and Solidiﬁcation Processing / 51

Fig. 24

Steel mold for constrained rod casting. Source: Ref 157

at 0.9 and at the end of solidiﬁcation, are shown in Fig. 28. Same as the experiment, the susceptibility of hot tearing rises sharply from pure Magnesium, reaches its maximum at Mg-1% Al, and decreases gradually with further increase in the aluminum content. Similarly, different ternary magnesium alloys have different hot tearing susceptibility. The calculated CSC by this model for different alloys is shown in Fig. 29. The experimental hot tearing susceptibility (Ref 158) is included in the same ﬁgure for comparison. The addition of calcium to Magnesium-aluminum alloys can reduce the temperature range between fraction of solid at 0.9 and end of solidiﬁcation, which is shown in Fig. 30; hence, susceptibility to hot tearing formation will decrease as well. It indicates that the current hot tearing indicator can predict hot tearing trends very well. The alloy chemistry, casting geometry, and cooling conditions all contribute to the formation of hot tearing, and they are included in this model directly or indirectly.

Examples of Modeling Applied in Casting Industries

Fig. 25

Hot tearing indictor for a Mg-2%Al alloy casting

Fig. 26

Close-up views of hot tearing (cracks) in the bottom rods near the sprue. (a) Mg-2%Al. (b) Mg-4%Al. (c) Mg8%Al

As mentioned earlier, radiation is very important for modeling of investment casting. In this example, radiation is discussed extensively.

High- or Low-Pressure Die Casting To demonstrate the application of modeling in the casting industry, a couple of examples are presented next. First, a die casting process is modeled, while focusing on the stress analysis of the casting and die and the gap formation between mold and part. Next, an example of modeling of an investment casting is presented.

For such casting processes, metallic molds are used, occasionally with cooling or heating channels. It is called permanent mold casting if the ﬁlling is by gravity. Otherwise, it is called die casting, for which a pressure is applied to provide rapid ﬁlling. These processes are

mainly focused on high-volume components such as automobile parts. Fluid ﬂow and stress analysis for both casting and mold are important to other casting processes. Because the molten metal is introduced into the mold by gravity, permanent mold casting ﬂow analysis is similar to that of sand casting. However, back pressure or trapped gas cannot be ignored, since a metal mold is not permeable, unlike that of a sand mold. The ﬂow analysis can be complicated because of the inclusion of the effect of back pressure in mold

52 / Casting Design and Performance 200

HTI CSC

150

0.1

100

50

0.0

0 0

Fig. 28 (CSC)

20

Crack susceptibility coefficient

105

CSC

100

HTC

15

95 10 90 5 85

80

0 Mg-4Al Mg-4Al- Mg-4Al- Mg-4Al- Mg-4Al- Mg-5Al- Mg-6Al0.5Ca 1.5Ca 2.5Ca 3.5Ca 2.5Ca 2.5Ca

Fig. 29

Comparison between hot tearing susceptibility (HTC) and crack susceptibility coefﬁcient (CSC)

1.00 Mg-4Al Mg-4Al-0.5Ca Mg-4Al-1.5Ca Mg-4Al-2.5Ca Mg-4Al-3.5Ca Mg-5Al-2.5Ca Mg-6Al-2.5Ca

Fraction of solid

0.98

0.96

0.94

0.92

0.90 420

440

460

480

500

Temperature,°C

Fig. 30

1

2 3 4 5 6 Aluminum content, wt%

7

8

Comparison between hot tearing indicator (HTI) and crack susceptibility coefﬁcient

Hot tearing indicator in the bottom rods near the sprue. (a) 2% Al. (b) 4% Al. (c) 8% Al

Hot tearing susceptibility

Fig. 27

CSC (Ts0.9-Ts (K))

Hot tearing indicator

0.2

Solidiﬁcation paths for some Magnesium alloys

520

540

560

ﬁlling. If a relatively low pressure is applied to the sealed furnace, it is called low-pressure casting process. The low pressure pushes the molten metal to ﬁll the mold cavity slowly. For low-pressure casting ﬁlling analysis, the boundary conditions are pressure, which is a function of time instead of an inﬂow velocity. This is to simulate the furnace pressure controlling ﬂow in the low-pressure casting presses. In permanent mold casting, the molds are used repeatedly. Hence, the molds develop a nonuniform temperature distribution during the initial cycles of the casting process that approaches a periodic quasi-steady-state condition. For a cyclic analysis, all casting parameters, such as the liquid pouring, dwell time, open time, and spraying conditions, must be considered for the calculation before the solidiﬁcation analysis. As the modern foundry continues to evolve in implementing new technology, process modeling must also advance to meet the next hurdles facing foundry engineers. It is critical that these new hurdles — stress and deformation in the casting and mold, micro- and gas porosity, as-cast mechanical properties — be accurately predicted and corrected. Beyond simply identifying shrink porosity and ﬁll problems, numerical tools have been developed to predict defects at a microstructural level that can be used effectively by the foundry engineer. Knowing the answers to these questions early in the manufacturing process affords signiﬁcant time and cost-savings. The main goal of the casting process is to approach and achieve net shape. Large deformation or distortion in the part requires more rework, such as hot pressing, even after a heat treatment operation. To aid in the understanding of why parts deform or develop residual stresses, simulation software can now predict thermally induced stresses, which include the effects of the strength or constraint of the mold. Simulating the strength of molding components has been proven as a necessary input in determining the amount of distortion in the casting, because the main prevention against signiﬁcant warpage is the mold itself.

Modeling of Casting and Solidiﬁcation Processing / 53

ð ð t dud s gradðduÞd b dud s ð xðuÞgðuÞn dud ¼ 0 ðEq 55Þ þ

ð

c

Here, a frictionless contact is considered for simplicity. In the previous equation, O represents the geometry of casting and all the mold parts, and G represents all the contact interfaces between all parts. The body forces and surface tensions are denoted by b and t, respectively. The augmented penalty function is given by x. Thermal contact between parts is considered by adjusting the interface heat-transfer coefﬁcient with respect to either the air gap width or the contact pressure, as computed by the mechanical contact algorithm. When the gap width is greater than zero, the adjusted heattransfer coefﬁcient has the form: heff ¼

1 h0

1 1 þ ðhair þh rad Þ

(Eq 56)

where h0 is the initial value of the heat-transfer coefﬁcient, hair is the conductivity of air divided by the gap width (if a vacuum is used, this term equates to zero), and hrad is the radiation heat-transfer coefﬁcient. If the contact pressure is greater than zero, the effective heat transfer is increased linearly with that pressure up to a maximum value.

used to make jewelry. The investment casting process does not require elaborate or expensive tooling and has the ability to produce several casting in one pour. The various stages include creating the wax pattern, growing the investment shell, dewaxing, and pouring the casting. The investment casting process uses ceramic shells as molds. The ceramic shells can be preheated to very high temperatures, up to and above the liquidus temperature. Investment castings are mostly used for aerospace and medical implant applications. Investment castings are commonly made from nickel-base superalloys, titanium alloys, aluminum alloys, cobalt-base alloys, and steels. The most important application for investment casting is in the aerospace industry. Nickel-base alloys are used for jet engine structural castings as well as turbine airfoils. Structural castings normally have an equiaxed grain structure, whereas turbine airfoils have equiaxed grain, columnar grain, and single-crystal types of structure. The unique features of the investment casting process include high-temperature ceramic molds, frequently casting in a vacuum, a furnace withdrawal process for directional solidiﬁcation, strict requirements for microstructure, and sometimes applying centrifugal force for mold ﬁlling. Hence, it is critical to apply some special treatments for modeling of the investment casting process. The mold of investment castings, called a shell, is one of the unusual features of the process. The outer contour of a shell mold generally follows the shape of the casting, unlike most other molds. Since the shape of the shell mold is determined by the procedure of repeatedly dipping the wax part into a ceramic slurry, followed by drying, it is somewhat freeform. It is thus difﬁcult to generate a computer-aided design representation of the shell geometry. Hence, it is necessary to use specialized meshing techniques to create a model of the shell. Some commercial packages can generate the shell automatically based on the investment process, such that the thickness of the shell will be different at different geometrical locations, such as convex sides, concave sides, ﬂat surfaces, or some corners.

When the casting is ejected from the die, the mechanical contact is no longer applied to the casting/die interfaces. Care must then be taken to apply an appropriate displacement constraint to prevent solid body movement. In the ﬁrst example (Fig. 31), a simple Tshaped casting of A356 in an H13 mold is simulated. The effective interface heat-transfer coefﬁcient at two different points on the casting is plotted. The top curve is from a point experiencing increasing contact pressure as the casting contracts. The middle curve is from a point where a gap is opening up between casting and mold, assuming the presence of air. The bottom curve is from that same point but assuming a vacuum. The large variation in the coefﬁcient illustrates the importance of accounting for local conditions. In addition, this example illustrates the value of the reverse coupling of the mechanical deformations with the energy solution. This effect can be seen in Fig. 31 on the right, where the heat ﬂux contours are plotted. The heat ﬂux is greatest where the contact pressure is highest. Postejection springback and further relaxation after mold removal can also be tracked, completing the necessary cycle to determine ﬁnal part shape (Fig. 32). Finally, heat treatment processes may further deﬁne the ﬁnal net shape. By appropriately simulating the thermal loading on the part during heat treatment, any additional shape change may be captured, either independent of the casting process or as a continuance of the deformation and residual-stress evolution formed from casting.

Investment Casting The investment casting or lost wax process, as it is commonly called, is one of the oldest known manufacturing processes. The ability to produce near-net shapes leading to minimized machine cost makes the investment casting process one of the most attractive casting processes, especially for making exotic castings with expensive alloys. This process can be used to make complex shapes, from aircraft jet engine components to small, intricate castings 1.40000E+03 1.20000E+03 2 Effective heatflux, W/m . K

Coupling the stress analysis with the thermal and ﬂuid calculation gives a clearer understanding of the physical phenomenon. By comparing the mushy-zone location and the evolution of stresses while cooling, problems such as hot tearing and cracking can be clearly indicated. Hot tearing and cracking can occur when the stresses in the local region go beyond the yield stress, thus requiring a tight coupling of the thermal and stress calculation. Another typical phenomenon involves the gap formed between casting and mold as the casting solidiﬁes and shrinks. When this gap develops, there can be a signiﬁcant reduction in heat transfer from the casting to the mold components where contact is lost (Ref 101). Conversely, there may be locations where, due to the shrinkage of the casting onto the core, an increasing contact pressure will increase the heat-transfer rate from casting to core. By accurately tracking the heat transfer, a better indication of surface shrink can be achieved. A multibody mechanical contact algorithm is employed to compute the contact and gap formation between the castings and die parts. Contact between different die parts is also considered. An automatic penalty number adjustment technique is implemented in the contact algorithm. Such a technique greatly enhances the stability and robustness of the contact computation algorithm. The variational form of the equilibrium equation with mechanical contact at any time, t, is written as (Ref 161):

Gap closed 1.00000E+03 8.00000E+02 6.00000E+02 4.00000E+02

Gap air 2.00000E+02

Gap vacuum

0.00000E+00 0

200

400

600

800

1000

Time (s)

Fig. 31

The T-bar example shows the heat ﬂux increasing in locations of higher contact pressure and decreasing where gaps develop.

54 / Casting Design and Performance

Fig. 32

Stresses in the component (a) before and (b) after ejection

Another unique feature for investment casting has to do with solving the heat-transfer problems. In general, the model should handle heat conduction in the core and the mold, convection and radiation across the metal and mold interface, and radiation and convection at the mold outer interface. If the alloys are cast in a vacuum environment, such as for most nickeland titanium-base alloy castings, radiation heat transfer is the only method for heat loss from the mold surface to the furnace. For a radiation calculation, proper view-factor calculations are very important. Sometimes, the view factors can change, for example, during the withdrawal process for direct chill and single-crystal castings. To change the cooling rate for some investment casting processes, some insulation materials such as kaowool can be used to wrap some speciﬁc areas of the shell. For such situations, this can be represented in the model by changing the shell surface emissivity or heattransfer coefﬁcient. The most signiﬁcant mode of heat transfer in an investment casting process is through radiation. It is critical that this be fully understood and accounted for when planning the process parameters for any casting. One must

understand the radiation effects during the metal pouring and cooling cycle, because self-radiation and external casting conditions will affect cooling rates and solidiﬁcation patterns. Additionally, prepouring soaking determines the shell temperatures before the pour, and if this is not taken into account accurately, it will adversely affect the desired outcome. This section delves into radiation and the downstream effects in the casting of certain process scenarios, such as heat loss in the mold before pouring, self-radiation affects on solidiﬁcation, cooling of the shell and pouring castings in ambient conditions versus in an enclosed chamber (i.e., a can), and the effects of kaowool and other insulating media. These scenarios are demonstrated using computer analyses with the ProCAST casting simulation software by ESI. For a simulation program to model the physics of the process, the computer program must accurately mimic the various stages, from wax injection to the actual pouring of the molten metal and the solidiﬁcation process. In addition to the mold ﬁlling and solidiﬁcation analysis, advanced computer simulation programs such as ProCAST can also evaluate the thermally

induced stress in the casting and underlying grain structure and mechanical properties of the ﬁnal cast part. Accurately modeling the trapped air and shrink porosity is critical to determine the validity of an existing gating system and feeder locations. Defects are usually observed in the last areas to ﬁll, especially if the permeability of the shell is not appropriate and if the last liquid area during solidiﬁcation is in the casting and not properly fed. In badly designed gating systems, the liquid metal can prematurely solidify, leading to cold laps and other such defects. To accurately model the physics, it is essential that the proper heat extraction from the casting to the investment shell is calculated. There are three modes of heat transfer: conduction, convection, and radiation. In the investment casting process, the radiative heat transfer is dominant over the other modes of heat transfer. Due to the elevated temperature of the mold relative to its surroundings and the nature of the system setup, radiative heat transfer is the dominant mode of heat transfer. Since each location on the mold surface sees a different view of its environment, each will have a different rate of heat loss. For example, faces around the outer perimeter of the mold are exposed mostly to the environment, while the view space of those facing the inside of the mold see mostly other hot regions of the casting. As a result, radiative view-factor calculations are required to accurately account for the heat-exchange variations across the geometry. The initial mold temperature and heat loss to the environment prior to casting dictate the thermal proﬁles that exist at the time of pour. This creates a thermal gradient from the inside to the outside of the shell, depending on the self-radiating effects of the whole assembly. Most analysis packages on the market today assume basic radiation heat-transfer calculations, making single-body assumptions that do not account for multiple bodies or shapes. Calculating view factors allows for multiple bodies and the shapes of those bodies to be involved in the radiation heat transfer to give a much more precise and accurate calculation. By not including multibody radiation, part-to-part radiation effects or self-radiation effects would not be seen. As a result, regardless of the casting orientation on the tree, each casting would heat or cool exactly the same — even if it was a casting on the end of the runner or a part that was fully surrounded by other hot parts. Planning setup design for investment casting involves visualizing the “invisible” radiation heat-transfer effects. Therefore, to further understand radiation effects, analysis tools that are able to calculate the view-factor radiation and shadowing effects are used to present various common investment casting scenarios. A simple example given subsequently shows four test bar castings, one with the effects of selfradiation and the other without. Figure 33(a), which does not have the effects of self-radiation, shows all four test bars losing the same amount

Modeling of Casting and Solidiﬁcation Processing / 55 of heat at the given time. Even though the radiative heat effects are considered, there is no reﬂective heat back from the hot surfaces; thus, all the test bars show exactly the same temperature proﬁle, regardless of their location in the model. Figure 33(b), on the other hand, shows the effects of self-radiation on each specimen. As expected, the interior test samples are hotter than those on the outside, because they receive reﬂected heat from the test bars on either side of them. The outside bars only receive heat from one side. This temperature gradient effect is also reﬂected in the solidiﬁcation rate and porosity location and size in both the test conﬁgurations. To prevent freezing of the metal or cold ﬂow during ﬁlling, it is desirable to pour the casting as soon as possible when the mold is removed from the furnace. This is especially so for thin-shell molds or very thin parts. For “small” castings, where the mold is typically handmoved from the furnace to the pouring bed, it is typical to have 10 to 20 s elapse before the metal is poured. Even this small amount of time can cause a signiﬁcant reduction in temperature on shell faces that are open to the environment, especially if a thin-shell mold or highly conductive shell material is used. Figures 34(a) and (b) show the effects of shell cooling between the time it takes the shell to be extracted from the furnace to the beginning of metal entering the mold cavity. In the casting shown, a 25 to 30 s time delay can cause certain regions of the casting to drop 300 to 400 . Note how the shell on the end has cooled much more rapidly than the interior regions. With large shells or ones that must be moved from a furnace into a vacuum chamber, the elapsed time can be much more, and thus, the difference between internal shell temperatures and external temperatures may be quite large — potentially resulting in unexpected ﬁlling issues or patterns. In many investment castings, when the casting is poured, the shell glows, indicating the large amount of heat inside. During and after mold ﬁlling, solidiﬁcation starts to take place. The energy content of the metal continually decreases by heat conduction with the cooler mold and by radiation from the mold surface. Exterior faces radiate this heat freely and allow for a relatively high rate of cooling. However, internal surfaces or locations where there is mostly part-to-part radiation are, in essence, insulating each other by radiating heat onto itself. It is quite common to have internal parts or parts involved in a high amount of radiation have a solidiﬁcation pattern that is much different than parts that may be on the end of a row or more exposed to the ambient conditions. Therefore, it may be beneﬁcial to design casting setups such that all of the parts experience the same heat transfer (that is, radiation) effects. By having the same cooling pattern, any changes to the design or rigging of the part will apply evenly to each part. Otherwise, the casting engineer may be chasing a defect that occurs in one part position that does not occur in another.

The case study shown in Fig. 35(a) and (b) demonstrates an in-line gating design. The initial setup shows nine castings stacked next to each other, with a common top-fed runner design. A coupled ﬂuid and solidiﬁcation analysis was run to understand the problems associated with molten metal ﬂow and porosity. A quick review of the computer analysis run in ProCAST gave a very good indication on the ﬂow pattern of the mold. The molten metal seems to ﬁll most of the central casting, because the sprue is located right above it, then the other castings ﬁll from the center outward. Since the central casting is already ﬁlled before the other castings, the temperature in the center casting is initially lower than the others. As the ﬁlling progresses to ﬁll the entire mold cavity, the shell starts to become heated, thus reﬂecting heat to the outside. The central casting remains

much hotter than the outer casting region due to the self-radiation effects, as explained in the earlier section. A quick look at the solidiﬁcation plots shows similar results. The outside castings solidify quicker than the inside ones, even though the ﬁlling sequence was reverse, where the hot metal ﬁlled the outside casting last. Figure 35(b) shows the porosity plot for the in-line gating design. The varying degree and size of porosity validates the fact that even though each part is exactly the same, the location on the tree greatly inﬂuences the solidiﬁcation rate and porosity. To minimize the erratic ﬂow and solidiﬁcation pattern, a circular gating design is introduced. A ring gate on the top fed the castings from a central sprue. Due to the symmetric nature of the gating design, the ﬂow pattern for the entire casting is very similar. Also, the

Fig. 33

Four test bar castings. (a) Without self-radiation. (b) With effects of self-radiation

Fig. 34

(a) Cooldown inside. (b) Cooldown outside

56 / Casting Design and Performance onto the part. With a casting chamber, the enclosure emissivity is controlled with speciﬁc chamber wall materials, and the temperature can be controlled with heaters or cooling systems placed on or in the chamber.

Conclusions

Fig. 35

Fig. 36

Modeling of casting and solidiﬁcation has been used extensively in foundries to solve routine production problems. Casting defects such as those related to ﬁlling, solidiﬁcation, stress, and microstructure can be predicted with conﬁdence, thanks to comprehensive models and the ability to compute thermophysical and mechanical properties of multicomponent alloys. As always, better understanding and accurate material properties, including mold materials, lead to improved predictive capabilities. Further development efforts should emphasize the enhancement of the accuracy of predicting and eliminating various casting defects. Coupling heat treatment simulation with casting simulation can then predict the ﬁnal mechanical properties of the part in service.

Temperature plot. (b) Porosity plot

Temperature plot with local insulation

ﬁll rate for each part on the tree is the same, so the effects seen in the in-line design, where the central casting ﬁlled much earlier than the others, is not observed. Also, due to the circular ring gate design, the radiation view factor that each part “sees” is the same for each part, leading to a similar behavior in radiative and reﬂective heat from each other. An additional beneﬁt of using a ring gate was that this design accommodated 12 castings on the tree instead of 9 in the in-line design. The ﬁlling pattern improved with the ring gates design, but the solidiﬁcation pattern was similar to the one observed in the

in-line design. The porosity magnitude was reduced by approximately 30% compared to the worst porosity observed in some of the inline design castings. The casting engineer has a few variables with which to adjust and optimize when considering the rigging design. The preceding paragraphs focused on the conﬁguration of the tree—how to place the various parts in reference to each other. The other parameters involved in radiation are also controllable by the engineer: the emissivity off of the shell faces, the temperature surrounding the shell (ambient or controlled enclosure temperature), and the emissivity of the casting environment (open-air cooling versus cooling in some chamber). Going to an extreme case, radiation can be removed completely by the depth at which the shell is buried in a sand bed. Figure 36 displays the cooling pattern on the shell for the in-line casting design when local insulation is applied on the shell. By putting a ½ or 1 in thick kaowool insulation on the critical regions that require insulation to enhance the feeding of the casting, the temperature in any area of the shell can be effectively controlled by controlling the radiation heat-transfer loss through that region. To aid a more balanced and even cooling of the exterior faces of the shell compared to the interior shell locations, adding a can or other enclosing structure can provide this effect. Technologically advanced chambers, such as those used in singlecrystal casting, are even engineered to force certain solidiﬁcation behaviors and resultant microstructures by controlling the amount of radiation to and from the casting. With a can, the temperature is not controlled; however, the can itself holds in heat by reﬂecting some of the heat back

REFERENCES 1. K.-O. Yu, Modeling for Casting and Solidiﬁcation Processing, Marcel Dekker, Inc., 2002 2. Y.A. Chang, S. Chen, F. Zhang, X. Yan, F. Xie, R. Schmid-Fetzer, and W.A. Oates, Phase Diagram Calculation: Past, Present and Future, Progress in Materials Science, Vol 49, 2004, p 313–345 3. H.L. Lukas, J. Weiss, and E.Th. Henig, Strategies for the Calculation of Phase Diagrams, CALPHAD, 1982, Vol 6 (No. 3), p 229–251 4. U.R. Kattner, The Thermodynamic Modeling of Multicomponent Phase Equilibria, JOM, 1997, Vol 49, p 14–19 5. N. Saunders and A.P. Miodownik, CALPHAD: Calculation of Phase Diagrams, A Comprehensive Guide, Elsevier Science Ltd., New York, NY, 1998, p 299–411 6. N. Saunders, Z. Guo, X. Li, A.P. Miodownik, and J.-Ph. Schille, Using JMatPro to Model Materials Properties and Behavior, JOM, 2003, Dec., p 65 7. J. Guo and M.T. Samonds, Property Prediction with Coupled Macro-Micromodeling and Computational Thermodynamics, H. Weng-Sing, Ed., Modeling of Casting & Solidiﬁcation Processes, Taiwan, 2004, p 157–164 8. J. Guo and M. Samonds, Thermophysical Property Prediction Coupled Computational Thermodynamics with Back Diffusion Consideration, J. Phase Equilibrium and Diffusion, Volume 28, Number 1/ February 2007, pp 58–63 9. J. Guo and M. Samonds, The Thermophysical and Mechanical Property Prediction

Modeling of Casting and Solidiﬁcation Processing / 57

10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

22. 23.

of Multicomponent Alloys by the Coupling of Macro-Micromodeling and Computational Thermodynamics, Modeling of Casting, Welding and Advanced Solidiﬁcation Processes XI, 2006, Opio, France H.D. Brody and M.C. Flemings, Solute Redistribution during Dendrite Solidiﬁcation, Trans. Met. Soc. AIME, 1966, Vol 236, p 615 T.W. Clyne and W. Kurz, Solute Redistribution during Solidiﬁcation with Rapid Solid State Diffusion, Metall. Trans., 1981, Vol 12A, p 965 I. Ohnaka, Mathematical Analysis of Solute Redistribution during Solidiﬁcation with Diffusion in Solid Phase, Trans. ISIJ, 1986, Vol 26, p 1045 C.Y. Wang and C. Beckermann, Uniﬁed Solute Diffusion Model for Columnar and Equiaxed Dendritic Alloy Solidiﬁcation, Mater. Sci. Eng., 1993, Vol 171, p 199 L. Nastac and D.M. Stefanescu, An Analytical Model for Solute Redistribution during Solidiﬁcation of Planar, Columnar and Equiaxed Morphology, Metall. Trans., 1993, Vol 24, p 2107 V.R. Voller and C. Beckermann, A Uniﬁed Model of Microsegregation and Coarsening, Metall. Trans., 1999, Vol 30, p 2183 V.R. Voller, On a General Back-diffusion Parameter, J. Cryst. Growth, 2001, Vol 226, p 562–568 M.C. Schneider, J.P. Gu, C. Beckermann, W.J. Boettinger, and U.R. Kattner, Modeling of Micro- and Macrosegregation and Freckle Formation in Single-Crystal Nickel-Base Superalloy Directional Solidiﬁcation, Met. Mat. Trans. A, 1997, Vol 28A, p 1517–1531 X. Yan, “Thermodynamic and Solidiﬁcation Modeling Coupled with Experimental Investigation of the Multicomponent Aluminum Alloys,” Ph.D. Thesis, University of Wisconsin-Madison, Madison, WI, 2001 C.Y. Wang and C. Beckermann, A Multiphase Solute Diffusion Model for Dendritic Alloy Solidiﬁcation, Met. Trans. A, 1993, Vol 24A, p 2787–2802 I. Ohnaka, Mathematical Analysis of Solute Redistribution during Solidiﬁcation with Diffusion in Solid Phase, Trans. ISIJ, 1986, Vol 26, p 1045–1052 X.L. Yang, P.D. Lee, R.F. Brooks, and R. Wunderlich, The Sensitivity of Investment Casting Simulations to the Accuracy of Thermophysical Property Values, Superalloys, The Minerals, Metals & Materials Society, 2004 B. Alkan, R. Karabulut, and B. Unal, Electrical Resistivity of Liquid Metals and Alloys, Acta Phys. Pol. A, 2002, Vol 102, p 385–400 P.L. Rossiter, The Electrical Resistivity of Metals and Alloys, Cambridge University Press, New York, NY, 1987, p 137–272

24. A. Rudajevova, M. Stanek, and P. Lukac, Determination of Thermal Diffusivity and Thermal Conductivity of Mg-Al Alloys, Mat. Sci. Eng., A, 2003, Vol A341, p 152–157 25. P.G. Klemens and R.K. Williams, Thermal Conductivity of Metals and Alloys, Int. Mat. Rev., 1986, Vol 31 (No. 5), p 197–215 26. D. Sichen, J. Bygden, and S. Seetharaman, A Model for Estimation of Viscosities of Complex Metallic and Ionic Melts, Met. Mat. Trans. B, 1994, Vol 25B, p 519–525, August 27. N. Saunders, X. Li, A.P. Miodownik, and J.P. Schille, Modeling of the Thermophysical and Physical Properties for Solidiﬁcation of Al-Alloys, Light Metals, 2003, p 999–1004 28. ProCAST user’s manual and technical references, ESI, 2008 29. M. Rappaz and W.J. Boettinger, On Dendritic Solidiﬁcation of Multicomponent Alloys with Unequal Liquid Diffusion Coefﬁcients, Acta Mater., Vol. 47 (No. 11), pp. 3205–3219, 1999 30. W.J. Boettinger, S.R. Coriell, A.L. Greer, A. Karma, W. Kurz, M. Rappaz, and R. Trivedi, Solidiﬁcation Microstructures: Recent Developments, Future Directions, Acta Mater., Vol 48, 2000, p 43–70 31. J. Guo and M. Samonds, Microstructures and Mechanical Properties Prediction for Ti Based Alloys, Journal of Materials Engineering and Performance, May 17, 2007 32. J. Guo and M. Samonds, Modeling of Microstructures and Mechanical Properties of a + b Ti Based Alloys, Materials Science and Technology, Volume 3, 2005, p 41–50 33. A. Kermanpur, M. Mehrara, N. Varahram, and P. Davami, Improvement of Grain Structure and Mechanical Properties of a Land Based Gas Turbine Blade Directionally Solidiﬁed with Liquid Metal Cooling Process, Materials Science and Technology, Vol 24 (No. 1), January, 2008, p 100–106 34. I. Steinbach, C. Beckermann, B. Kauerauf, Q. Li, and J. Guo, Three-Dimensional Modeling of Equiaxed Dendritic Growth on a Mesoscopic Scale, Acta Materialia, Vol. 47, 1999, pp. 971–982 35. M. Rappaz and P. H. Thevoz, Solute Diffusion Model for Equiaxed Dendritic Growth, Acta Metall., Vol 35, p 1487–1497, 1987 36. M. Rappaz and P. H. Thevoz, Solute Diffusion Model for Equiaxed Dendritic Growth: Analytical Solution, Acta Metall., Vol 35, p 2929–2933, 1987 37. P. Thevoz, J.L. Desbiolles, and M. Rappaz, Modeling of Equiaxed Microstructure Formation in Casting, Metall. Trans. A, 1989, Vol 20, p 311–322 38. M. Rappaz, Modelling of Microstructure Formation in Solidiﬁcation Processes, Int. Mater. Rev., Vol 34, 1989, p 93

39. C.Y. Wang and C. Beckermann, Equiaxed Dendritic Solidiﬁcation with Convection: Part I. Multiscale/Multiphase Modeling, Metall. Mater. Trans., Vol 27A, 1996, p 2754–2764 40. Ch.-A. Gandin and M. Rappaz, Coupled Finite Element-Cellular Automation Model for the Prediction of Dendritic Grain Structures in Solidiﬁcation Processes, Acta Metall. Mater., Vol 42, 1994, p 2233–2246 41. Ch.-A. Gandin, J.-L. Desbiolles, M. Rappaz, and Ph. Thevoz, Three-Dimensional Cellular Automaton-Finite Element Model for the Prediction of Solidiﬁcation Grain Structures, Metall. Mater. Trans., Vol 30A, 1999 p 3153–3165 42. H. Takatani, Ch.-A. Gandin, and M. Rappaz, EBSD Characterization and Modelling of Columnar Dendritic Grains Growing in the Presence of Fluid Flow, Acta Mater., Vol 48, 2000, p 675–688 43. G. Guillemot, Ch.-A. Gandin, and H. Combeau, Modeling of Macrosegregation and Solidiﬁcation Grain Structures with a Coupled Cellular Automaton-Finite Element Model, Solidiﬁcation Processes and Microstructures—A Symposium in Honor of Wilfried Kurz, T.M.S., Warrendale, PA, 2004, pp. 157–163 44. N. Ahmad, H. Combeau, J.-L. Desbiolles, T. Jalanti, G. Lesoult, J. Rappaz, M. Rappaz, and C. Stomp, Numerical Simulation of Macrosegregation: A Comparison Between Finite Volume Method and Finite Element Method Predictions and a Confrontation with Experiments, Mater. Trans., Vol 29A, 1998, p 617–630 45. J.-L. Desbiolles, Ph. Thevoz, and M. Rappaz, Micro-/Macrosegregation Modeling in Casting: A Fully Coupled 3D Model, Modeling of Casting, Welding and Advanced Solidiﬁcation Processes X, T. M.S., Warrendale, PA, 2003, pp. 245–252 46. M. Bellet, V. D. Fachinotti, S. Gouttebroze, W. Liu, and H. Combeau, “A 3DFEM Model Solving Thermomechanics and Macrosegregation in Binary Alloys Solidiﬁcation, Solidiﬁcation Processes and Microstructures—A Symposium in Honor of Wilfried Kurz, T.M.S., Warrendale, PA, 2004, pp. 41–46 47. D.J. Hebditch and J.D. Hunt, Observations of Ingot Macrosegregation on Model Systems, Metall. Trans., Vol 5, 1974, p 1557 48. W. Wang, A. Kermanpur, P.D. Lee, and M. Mclean, Simulation of Dendritic Growth in the Platform Region of Single Crystal Superalloy Turbine Blades, Journal of Materials Science, Vol 8, 2003, p 4385 – 4391 49. J.S. Langer, Existence of Needle Crystals in Local Models of Solidiﬁcation, Phys. Rev. A, Vol 33, 1986, p 435 50. D.A. Kessler H. Levine, Velocity Selection in Dendritic Growth, Phys. Rev. B, 1986, Vol 33, p 7867

58 / Casting Design and Performance 51. R.J. Braun, G.B. Mcfadden, and S.R. Coriell, Morphological Instability in Phase-Field Models of Solidiﬁcation, Phys. Rev. E, Vol 49, p 4336–4352, 1994 52. J.A. Warren and W.J. Boettinger, Prediction of Dendritic Growth and Microsegregation Patterns in a Binary Alloy Using the Phase-Field Method, Acta Mater., Vol 43, p 689–703, 1995 53. W.J. Boettinger and J.A. Warren, The Phase-Field Method: Simulation of Alloy Dendritic Solidiﬁcation during Recalescence, Metall. Mater. Trans. A, Vol 27, p 657–669, 1996 54. I. Loginova, G. Amberg, and J. Agren, Phase-Field Simulation of Nonisothermal Binary Alloy Solidiﬁcation, Acta Mater., Vol 49, p 573–581, 2001 55. W.J. Boettinger and J.A. Warren, Simulation of the Cell to Plane Front Transition during Directional Solidiﬁcation at High Velocity, J. Cryst. Growth, Vol 200, p 583–591, 1999 56. H.-J. Diepers, D. Ma, and I. Steinbach, History Effects during the Selection of Primary Dendrite Spacing. Comparison of Phase-Field Simulations with Experimental Observations, J. Cryst. Growth, Vol 237–239, p 149–153, 2002 57. J. Tiaden and U. Grafe, A Phase-Field Model for Diffusion and Curvature Controlled Phase Transformations in Steels, Proceedings of the International Conference on Solid-Solid Phase Transformations ’99, M. Koiwa, K. Otsuka, and T. Miyazaki, Eds., 1999 58. A. Karma and W.J. Rappel, Phase-Field Method for Computationally Efﬁcient Modeling of Solidiﬁcation with Arbitrary Interface Kinetics, Phys. Rev. E, Vol 53, p 3017–3020, 1996 59. I. Steinbach, B. Kauerauf, C. Beckermann, J. Guo, and Q. Li, Three Dimensional Modeling of Equiaxed Dendritic Growth on a Mesoscopic Scale, Proceedings of the Eighth International Conference on Modeling of Casting and Welding Processes, B. G. Thomas and C. Beckermann, Eds. San Diego, California, p 565–572, 1998 60. I. Steinbach, F. Pezzolla, B. Nestler, M. Seebelberg, R. Prieler, G.J. Schmitz, and J.L.L. Rezende, A Phase Field Concept for Multiphase Systems, Physica D, Vol 94, p 135–147, 1996 61. J. Tiaden, B. Nestler, H.J. Diepers, and I. Steinbach, The Multiphase-Field Model with an Integrated Concept for Modelling Solute Diffusion, Physica D, Vol 115, p 73–86, 1998 62. J.Q. Broughton, A. Bonissent, and F.F. Abraham, The Fcc (111) and (100) Crystal-Melt Interfaces: A Comparison by Molecular Dynamics Simulation, J. Chem. Phys., Vol 74, p 4029, 1981 63. A.A. Wheeler, W.J. Boettinger, and G. B. Mcfadden, Phase-Field Model for

64. 65.

66.

67.

68. 69. 70.

71.

72.

73.

74.

75.

76.

77. 78.

Isothermal Phase Transitions in Binary Alloys, Phys. Rev. E, Vol 45, p 7424– 9439, 1992 G. Caginalp and W. Xie, Phase-Field and Sharp-Interface Alloy Models, Phys. Rev. E, Vol 48, p 1897–1910, 1993 W.J. Boettinger, J.A. Warren, C. Beckermann, and A. Karma, Phase-Field Simulation of Solidiﬁcation, Annu. Rev. Mater. Res., 2002, Vol 32, p 163–194 W.J. Boettinger, A.A. Wheeler, B.T. Murray, G.B. McFadden, and R. Kobayashi, Calculation of Alloy Solidiﬁcation Morphologies Using the Phase-Field Method, Modeling of Casting, Welding and Advanced Solidiﬁcation Processes, 1993, p 79 A.A. Wheeler, W.J. Boettinger, G.B. McFadden, Phase-Field Model for Isothermal Phase Transitions in Binary Alloys, Phys. Rev. A, 1992, Vol 45, p 7424 G. Caginalp and W. Xie, Phase-Field and Sharp-Interface Alloy Models, Phys. Rev. E, 1993, Vol 48, p 1897 S.G. Kim, W.T. Kim, and T. Suzuki, Phase-Field Model for Binary Alloys, Phys. Rev. E, 1999, Vol 60, p 7186 A. Karma, Y.H. Lee, and M. Plapp, Three-Dimensional Dendrite-Tip Morphology at Low Undercooling, Phys. Rev. E, 2000, Vol 61, p 3996 I. Steinbach, F. Pezzolla, B. Nestler, M. Seelberg, R. Prieler, G.J. Schmitz, et al., Phase Field Concept for Multiphase Systems, Physica D, 1996, Vol 94, p 135–147 J. Tiaden, B. Nestler, H.J. Diepers, and I. Steinbach, Multiphase-Field Model with an Integrated Concept for Modelling Solute Diffusion, Physica D, 1998, Vol 115, p 73 I. Steinbach, and F. Pezzolla, A Generalized Field Method for Multiphase Transformations Using Interface Fields, Physica D, 1999, Vol 134, p 385 J.C. Ramirez, C. Beckermann, A. Karma, and H.-J. Diepers, Phase-Field Modeling of Binary Alloy Solidiﬁcation with Coupled Heat and Solute Diffusion, Physical Review E, Vol 69 (No. 5 2), May, 2004, p 051607-1 to 051607-16 J. Guo, “Modeling and Experimental Validation of Cast Iron Solidiﬁcation,” 5th Decennial International Conference on Solidiﬁcation Processing SP07, July 23– 25, 2007, University of Shefﬁeld, U.K D.M. Stefanescu, Solidiﬁcation and Modeling of Cast Iron — A Short History of the Deﬁning Moments, Materials Science and Engineering A, Vol 413–414, 2005, pp. 322–333 W. Oldﬁeld, A Qualitative Approach to Casting Solidiﬁcation, ASM Trans, Vol 59, 1966, p 945 D.M. Stefanescu and S. Truﬁnescu, Kinetics of Solidiﬁcation of Gray Iron, Z. Metall. Vol 9, 1974, p 610

79. H. Fredriksson and L. Svensson, Computer Simulation of Structure Formation and Segregation during the Solidiﬁcation of Cast Iron, The Physical Metallurgy of Cast Iron, H. Fredriksson and M. Hillert, Eds., Elsevier, 1985, p 273–284 80. H. Fredriksson and L. Svensson, Simulation of Grey Cast-Iron Solidiﬁcation in a Shaped Casting, Solidiﬁcation Processing of Eutectic Alloys, D.M. Stefanescu, G.J. Abbaschian, and R.J. Bayuzick, Eds., The Metallurgical Soc., Warrendale, PA, 1988, p 153–162 81. D.M. Stefanescu and C. Kanetkar, Computer Modeling of the Solidiﬁcation of Eutectic Alloys: The Case of Cast Iron, Computer Simulation of Microstructural Evolution, D.J. Srolovitz, Ed., The Metallurgical Soc., Warrendale, PA, 1985, p 171–188 82. J. Lacaze, M. Castro, C. Selig, and G. Lesoult, Solidiﬁcation of Spheroidal Graphite Cast Irons, Modeling of Casting, Welding and Advanced Solidiﬁcation Processes, V.M. Rappaz, Ed., The Metallurgical Soc., Warrendale, PA, 1991, p 473–478 83. E. Fras, W. Kapturkiewicz, and A.A. Burbielko, Micro-Macro Modeling of Casting Solidiﬁcation Controlled by Transient Diffusion and Undercooling, Modeling of Casting, Welding and Advanced Solidiﬁcation Processes VII, M. Croos and J. Campbell, Eds., The Metallurgical Soc., Warrendale, PA, 1995, p 679–686 84. K.C. Su, I. Ohnaka, I. Yamauchhi, and T. Fukusako, Modelling of Solidiﬁed Structure of Castings, The Physical Metallurgy of Cast Iron, H. Fredriksson and M. Hillert, Eds., Elsevier, 1985, p 181–189 85. S. Chang, D. Shangguan, and D.M. Stefanescu, Prediction of Microstructural Evolution in SG Cast Iron from Solidiﬁcation to Room Temperature, Metall. Trans., Vol 22A, 1991, p 915 86. D.M. Stefanescu and C.S. Kanetkar, Computer Modeling of the Solidiﬁcation of Eutectic Alloys: Comparison of Various Models for Eutectic Growth of Cast Iron, AFS Trans., 1987, pp. 139–144 87. L. Nastac and D.M. Stefanescu, Prediction of Gray-to-White Transition in Cast Iron by Solidiﬁcation Modeling, AFS Trans., Vol 103, 1995, pp. 329–337 88. K.M. Pedersen, J.H. Hattel, and N. Tiedje, Numerical Modelling of Thin-Walled Hypereutectic Ductile Cast Iron Parts, Acta Materialia, Vol 54 (No. 19), November, 2006, p 5103–5114 89. G. Lesoult, M. Castro, and J. Lacaze, Solidiﬁcation of Spheroidal Graphite Cast Irons — I. Physical Modelling, Acta Mater., Vol. 46 (No. 3), pp. 983–995, 1998 90. G. Lesoult, M. Castro, and J. Lacaze, Solidiﬁcation of Spheroidal Graphite Cast Irons — II. Numerical Simulation,

Modeling of Casting and Solidiﬁcation Processing / 59

91.

92.

93.

94.

95.

96.

97.

98. 99. 100.

101. 102.

103.

104.

Acta Mater., Vol. 46 (No. 3), pp. 997–1010, 1998 F.J. Bradley, A Stereological Formulation for the Source Term in Micromodels of Equiaxed Eutectic Solidiﬁcation, Metall. Trans. B, Vol 24B, p 539, 1993 R. Vijayaraghavan and F.J. Bradley, Micro-Model for Eutectoid Phase Transformations in As-Cast Ductile Iron, Scripta Materialia, Vol. 41 (No. 11), pp.1247–1253, 1999 D. Venugopalan, Prediction of Matrix Microstructure in Ductile Iron, Transactions of the American Foundrymen’s Society, 1990, p 465 L. Nastac and D.M. Stefanescu, Prediction of Gray-to-White Transition in Cast Iron by Solidiﬁcation Modeling, AFS Trans., 1995, Vol. 103, pp. 329–337 Q. Chen, E.W. Langer, and P.N. Hansen, Comparative Study on Kinetic Models and Their Effects on Volume Change during Eutectic Reaction of S.G. Cast Iron, Scand. J. Metall., 1995, Vol. 24, pp. 48– 62 M.I. Onsoien, O. Grong, O. Gundersen, and T. Skaland, Process Model for the Microstructure Evolution in Ductile Cast Iron: Part I. The Model, Metallurgical and Materials Transactions A, Vol. 30A, 1999, pp. 1053–1068 M.I. Onsoien, O. Grong, O. Gundersen, and T. Skaland, Process Model for the Microstructure Evolution in Ductile Cast Iron: Part II. Applications of the Model, Metallurgical and Materials Transactions A, Vol. 30A, 1999, pp. 1069–1079 J.R. Davis, ASM Specialty Handbook: Cast Irons, ASM International, Materials Park, OH V. Krutisˇ and J. Roucˇka, “Shrinkage of Graphitic Cast Irons,” ESI Group Internal Report J. Guo and M. Samonds, Microporosity Simulations in Multicomponent Alloy Castings, Modeling of Casting, Welding and Advanced Solidiﬁcation Processes X, D.M. Stefanescu et al., Ed., Sandestin, FL, 2003, pp 303–310 J. Campbell, Castings, Butterworth-Heinemann, Oxford, United Kingdom, 1991 A.V. Kuznetsov and K. Vafai, Development and Investigation of Three-Phase Model of the Mushy Zone for Analysis of Porosity Formation in Solidifying Castings, International Journal of Heat and Mass Transfer, Vol 38 (No. 14), Sept, 1995, p 2557–2567 G.K. Sigworth and C. Wang, Mechanisms of Porosity Formation during Solidiﬁcation: A Theoretical Analysis, Metallurgical Transactions B, Vol. 24B, April, 1993, pp. 349–364 D.R. Poirier, K. Yeum, and A.L. Maples, A Thermodynamic Prediction for Microporosity Formation in Aluminum-Rich

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115. 116.

117.

Al-Cu Alloys, Metallurgical Transactions A, Vol 18A, Nov. 1987, pp. 1979–1987 H. Combeau, D. Carpentier, J. Lacaze, and G. Lesoult, Modelling of Microporosity Formation in Aluminum Alloys Castings, Materials Science and Engineering, Vol A173, 1993, pp. 155–159 P.K. Sung, D.R. Poirier, S.D. Felicelli, E.J. Poirier, and A. Ahmed, Simulations of Microporosity in IN718 Equiaxed Investment Castings, Journal of Crystal Growth, Vol 226, 2001, pp. 363–377 J.D. Zhu and I. Ohnaka, Modeling of Microporosity Formation in A356 Aluminum Alloy Casting, Modeling of Casting, Welding and Advanced Solidiﬁcation Processes V, 1991, pp. 435–442 S. Sabau and S. Viswanathan, Microporosity Prediction in Aluminum Alloy Castings, Metallurgical and Materials Transactions B, Volume 33B, April 2002, pp. 243–255 D. See, R.C. Atweed, and P.D. Lee, A Comparison of Three Modeling Approaches for the Prediction of Microporosity in Aluminum-Silicon Alloys, Journal of Materials Sciene, Vol 36, 2001, pp. 3423–3435 R.C. Atwood and P.D. Lee, A ThreePhase Model of Hydrogen Pore Formation during the Equiaxed Dendritic Solidiﬁcation of Aluminum-Silicon Alloys, Metallurgical and Materials Transactions B, Vol. 33B, April, 2002, pp. 209–221 P.D. Lee, R.C. Atwood, R.J. Dashwood, and H. Nagaumi, Modeling of Porosity Formation in Direct Chill Cast Aluminum-Magnesium Alloys, Materials Science and Engineering, Vol A328, 2002, pp. 213–222 S.D. Felicelli, D.R. Poirier, and P.K. Sung, Model for Prediction of Pressure and Redistribution of Gas-Forming Elements in Multicomponent Casting Alloys, Metallurgical and Materials Transactions B, Vol 31B, Dec. 2000, pp. 1283–1292 P.N. Anyalebechi, Analysis and Thermodynamic Prediction of Hydrogen Solution in Solid and Liquid Multicomponent Aluminum Alloys, Light Metals, 1998, pp. 827–842 K. Kubo and R.D. Pehlke, Mathematical Modeling of Porosity Formation in Solidiﬁcation, Metallurgical Transactions B, Vol 16B, June 1985, pp. 359–366 B. Chalmers, Principles of Solidiﬁcation, John Wiley & Sons, Inc, 1964 E. Niyama, T. Uchida, M. Morikama, and S. Saito, Predicting Shrinkage in Large Steel Castings from Temperature Gradient Calculations, AFS Int. Cast Met. J., 1981, Vol. 6, pp. 16–22 Y.W. Lee, E. Chang, and C.F. Chieu, Modeling of Feeding Behavior of Solidifying Al-7Si-0.3Mg Alloy Plate Casting, Metall. Trans. B, 1990, Vol. 21B, pp. 715–722

118. V.K. Suri, A.J. Paul, N. El-Kaddah, and J.T. Berry, Determination of Correlation Factors for Prediction of Shrinkage in Castings — Part I: Prediction of Microporosity in Castings; A Generalized Criterion (AFS Research), AFS Transactions, Vol. 103, 1995, pp. 861–867 119. R.P. Taylor, H.T. Berry, and R.A. Overfelt, Parallel Derivation and Comparison of Feeding-Resistance Porosity Criteria Functions for Castings, HTD-Vol. 323, National Heat Transfer Conference Vol. 1, ASME, 1996, pp. 69–77 120. M.C. Flemings and G.E. Nereo, Macrosegregation: Part I, Transactions of the Metallurgical Society of AIME, 1967, Vol 239, p 1449–1461 121. M.C. Flemings, R. Mehrabian, and G.E. Nereo, Macrosegregation, Part II, Transactions of the Metallurgical Society of AIME, Vol. 242, Jan., 1968, pp. 41–49 122. M.C. Flemings and G.E. Nereo, Macrosegregation, Part III, Transactions of the Metallurgical Society of AIME, Vol. 242, Jan., 1968, pp. 50–56 123. C. Beckermann, Modeling of Macrosegregation: Application and Future Needs, International Materials Reviews, 2002, Vol. 47 (No. 5), pp 243–261 124. A.F. Giamei and B.H. Kear, On the Nature of Freckles in Nickel Base Superalloys, Metall. Trans., Vol. 1, pp. 2185– 2192, 1970 125. S.M. Copley, A.F. Giamei, S.M. Johnson, and M.F. Hornbecker, The Origin of Freckles in Unidirectionally Solidiﬁed Castings, Metall. Trans., Vol. 1, pp. 2193–2204, 1970 126. A. Hellawell, J.R. Sarazin, and R.S. Steube, Channel Convection in Partly Solidiﬁed Systems, Phil. Trans. R. Soc. Lond. A, Vol. 345, pp. 507–544, 1993 127. M.G. Worster, Convection in Mushy Layers, Ann. Rev. Fluid Mech., Vol. 29, pp. 91–122, 1997 128. T.M. Pollock and W.H. Murphy, The Breakdown of Single-Crystal Solidiﬁcation in High Refractory Nickel-Base Alloys, Metall. Mater. Trans. A, Vol. 27A, pp. 1081–1094, 1996 129. J. Guo, “Three-Dimensional Finite Element Simulation of Transport Phenomena in Alloy Solidiﬁcation,” The University of Iowa, Iowa City, Iowa, 2000 130. J. Guo, and C. Beckermann, ThreeDimensional Simulation of Freckle Formation during Binary Alloy Solidiﬁcation: Effect of Mesh Spacing, Numerical Heat Transfer, Part A, Vol. 44, pp. 559– 576, 2003 131. M.C. Schneider, J.P. Gu, C. Beckermann, W.J. Boettinger, and U.R. Kattner, Modeling of Micro- and Macrosegregation and Freckle Formation in Single-Crystal Nickel-Base Superalloy Directional Solidiﬁcation, Metallurgical and Materials

60 / Casting Design and Performance

132.

133.

134.

135.

136.

137.

138.

139.

140.

Transactions A: Physical Metallurgy and Materials Science, Vol 28A (No. 7), July 1997, p 1517–1531 S.D. Felicelli, J.C. Heinrich, and D.R. Poirier, Simulation of Freckles during Vertical Solidiﬁcation of Binary Alloys, Metall. Trans. B, Vol. 22B, pp. 847–859, 1991 S.D. Felicelli, D.R. Poirier, A.F. Giamei, and J.C. Heinrich, Simulating Convection and Macrosegregation in Superalloys, JOM, Vol 49 (No. 3), March 1997, p 21–25 P.J. Prescott and F.P. Incropera, Numerical Simulation of a Solidifying Pb-Sn Alloy: The Effects of Cooling Rate on Thermosolutal Convection and Macrosegregation, Metall. Trans. B, Vol. 22B, pp. 529–540, 1991 C.S. Magirl and F.P. Incropera, Flow and Morphological Conditions Associated with Unidirectional Solidiﬁcation of Aqueous Ammonium Chloride, ASME J. Heat Transfer, Vol. 115, pp. 1036–1043, 1993 D.G. Neilson and F.P. Incropera, Effect of Rotation on Fluid Motion and Channel Formation during Unidirectional Solidiﬁcation of a Binary Alloy, Int. J. Heat Mass Transfer, Vol. 36, pp. 489–505, 1993 D.G. Neilson and F.P. Incropera, ThreeDimensional Considerations of Unidirectional Solidiﬁcation in a Binary Liquid, Numer. Heat Transfer A, Vol. 23, pp. 1–20, 1993 H.-W. Huang, J.C. Heinrich, and D.R. Poirier, Numerical Anomalies in Simulating Directional Solidiﬁcation of Binary Alloys, Numer. Heat Transfer A, Vol. 29, pp. 639–644, 1996 M.C. Schneider, J.P. Gu, C. Beckermann, W.J. Boettinger, and U.R. Kattner, Modeling of Micro- and Macrosegregation and Freckle Formation in Single-Crystal Nickel-Base Superalloy Directional Solidiﬁcation, Metall. Mater. Trans. A, Vol. 28A, pp. 1517–1531, 1997 S.D. Felicelli, D.R. Poirier, and J.C. Heinrich, Macrosegregation Patterns in

141.

142.

143.

144. 145.

146.

147.

148.

149.

150.

Multicomponent Ni-Base Alloys, J. Crystal Growth, Vol. 177, pp. 145–161, 1997 S.D. Felicelli, J.C. Heinrich, and D.R. Poirier, Finite Element Analysis of Directional Solidiﬁcation of Multicomponent Alloys, Int. J. Numer. Meth. Fluids, Vol. 27, pp. 207–227, 1998 S.D. Felicelli, J.C. Heinrich, and D.R. Porier, Three-Dimensional Simulations of Freckles in Binary Alloys, J. Crystal Growth, Vol. 191, pp. 879–888, 1998 S.D. Felicelli, D.R. Poirier, and J.C. Heinrich, Modeling Freckle Formation in Three Dimensions during Solidiﬁcation of Multicomponent Alloys, Metall. Mater. Trans. B, Vol. 29B, pp. 847–855, 1998 J. Guo and J.Z. Zhu, Hot Tearing Prediction for Magnesium Alloy Castings, in preparation for Material Science Forum J. Guo and J.Z. Zhu, “Hot Tearing Indicator for Multi-Component Alloys Solidiﬁcation,” 5th Decennial International Conference on Solidiﬁcation Processing SP07, July 23–25, 2007, University of Shefﬁeld, U.K A. Stangeland, A. Mo, M. M’Hamdi, D. Viano, and C. Davidson, Thermal Strain in the Mushy Zone Related to Hot Tearing, Metallurgical and Materials Transactions A: Physical Metallurgy and Materials Science, Vol 37 (No. 13), March, 2006, p 705–714 D.G. Eskin, Suyitno, and L. Katgerman, Mechanical Properties in the Semi-Solid State and Hot Tearing of Aluminium Alloys, Progress in Materials Science, Vol 49, 2004, p 629–711 C. Monroe and C. Beckermann, Development of a Hot Tear Indicator for Steel Castings, Materials Science and Engineering A, Vol 413–414, Dec 15, 2005, p 30–36 B. Magnin, L. Maenner, L. Katgermann, and S. Engler, Ductility and Rheology of an Al-4.5% Cu Alloy from Room Temperature to Coherency Temperature, Mater. Sci. Forum, 1996, Vol 1209, p 217–222 J. Zhang and R.F. Singer, Hot Tearing of Nickel-Based Superalloys during

151. 152.

153.

154.

155.

156.

157. 158. 159.

160.

161.

Directional Solidiﬁcation, Acta Materialia, Vol 50 (No. 7), Apr 19, 2002, p 1869–1879 M. Rappaz, J.M. Drezet, and M. Gremaud, New Hot-Tearing Criterion, Metall. Mater. Trans A, 1999, Vol 30A, pp 449 D.G. Eskin, Suyitno, and L. Katgerman, Mechanical Properties in the Semi-Solid State and Hot Tearing of Aluminium Alloys, Progress in Materials Science, Vol 49 (No. 5), 2004, p 629–711 Z.L. Zhang, C. Thaulow, and J. Odegard, Complete Gurson Model Approach for Ductile Fracture, Engineering Fracture Mechanics, Vol 67 (No. 2), Sept 2000, p 155–168 J. Langlais and J.E. Gruzleski, Novel Approach to Assessing the Hot Tearing Susceptibility of Aluminium Alloys, Mater. Sci. Forum, 2000, Vol 167, p 331–337 A.L. Gurson, Continuum Theory of Ductile Rupture by Void Nucleation and Growth: Part 1 — Yield Criteria and Flow Rules for Porous Ductile Media, Journal of Engineering Materials and Technology, Transactions of the ASME, Vol 99, Ser H (No. 1), Jan, 1977, p 2–15 V. Tvergaard and A. Needleman, Analysis of the Cup-Cone Fracture in a Round Tensile Bar, Acta Metall., 1984, Vol 32, p 157 A. Needleman and V. Tvergaard, Analysis of Ductile Rupture Modes at a Crack Tip, J. Mech. Phys. Solids, 1987, Vol 35, p 151 O.C. Zienkiewicz, R.L. Taylor, and J.Z. Zhu, The Finite Element Method: Its Basis and Fundamentals, 2005, Elsevier G. Cao, S. Kou, and Y.A. Chang, Hot Cracking Susceptibility of Binary Mg-Al alloys, Magnesium Technology 2006, A. A. Luo, Ed., TMS, 2006, p 57–61 G. Cao and S. Kou, Hot Tearing of Ternary Mg-Al-Ca Alloy Castings, Met. and Mat. Trans. A, Vol 37A, Dec. 2006, pp 3647–3663 J.C. Simo and T.A. Laursen, An Augmented Lagrangian Treatment of Contact Problems Involving Friction, Comput. & Structures, Vol. 42, pp 97–116, 1992

Casting Design and Performance Pages 61–72

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Riser Design RISER DESIGN, or risering, deals with the development of suitable reservoirs of feed metal in addition to the desired casting shape so that undesirable shrinkage cavities in the casting are eliminated or moved to locations where they are acceptable for the intended application of the casting. When metals solidify and cool to form a casting, they generally undergo three distinct stages of volume contraction, or shrinkage. Exceptions to this shrinkage behavior of some graphitic cast irons are noted later in this article. These stages, shown schematically in Fig. 1, are: Liquid shrinkage: The liquid metal loses

volume as it gives up superheat and cools to its solidiﬁcation temperature. Solidiﬁcation shrinkage: The metal freezes, changing from a liquid to a higher-density solid. For pure metals, this contraction will occur at a single temperature, but for noneutectic alloys, it will take place over some temperature range or freezing interval. Solid shrinkage: The solid casting cools from its solidiﬁcation temperature to room temperature.

Visible signs of shrinkage-induced casting defects include internal shrinkage voids, surface deformation or dishing, and surface puncture. These defects will vary with different alloys; for example, internal shrinkage may be more dispersed, or alloys with strong skinforming behavior may not exhibit surface deformation. To eliminate these defects in the casting, a riser is added to accommodate the liquid shrinkage and to supply liquid feed metal to compensate for the solidiﬁcation shrinkage within the casting (Fig. 3). Therefore, the shrinkage in the riser/casting system is concentrated in the riser, which will then be removed from the ﬁnished casting. As illustrated in Fig. 3, the riser may be larger than the casting it feeds, because it must supply feed metal for as long as the casting is solidifying. Various methods are used to reduce the size of the required riser, including chilling the casting, that is, reducing its solidiﬁcation

time or insulating the riser, that is, extending its solidiﬁcation time.

Optimum Riser Design The role of the methods engineer in designing risers can be stated simply as making sure that risers will provide the feed metal: In the right amount At the right place At the right time

To this list can be added several other considerations: The riser/casting junction should be designed

to minimize riser removal costs.

The number and size of risers should be

minimized to increase mold yield and to reduce production costs.

Solid shrinkage, also called patternmaker’s shrinkage, is accommodated by making the pattern and, therefore; the mold cavity, somewhat larger than the desired dimensions of the ﬁnal casting. Liquid shrinkage and solidiﬁcation shrinkage are the concern of risering practice. In the absence of risers, a casting would otherwise solidify as shown in Fig. 2.

Schematic of the shrinkage of low-carbon steel. The contribution of each one of the three distinct stages of volume contraction is shown: liquid shrinkage, solidiﬁcation shrinkage, and solid contraction. Source: Ref 1

Fig. 1

Fig. 2

Schematic of sequence of solidiﬁcation shrinkage in an iron cube. (a) Initial liquid metal. (b) Solid skin and formation of shrinkage void. (c) Internal shrinkage. (d) Internal shrinkage plus dishing. (e) Surface puncture

62 / Casting Design and Performance Riser placement must be chosen so as not to

exaggerate potential problems in a particular casting design; for example, tendencies toward hot tearing or distortion. In practice, these considerations are often in conﬂict, and the ﬁnal riser design and pattern layout represent a compromise.

Fig. 3

Feed Metal Volume In addition to the rise volume required to satisfy liquid and solidiﬁcation shrinkage in the casting, the riser itself will be solidifying, so the total shrinkage requirement will be for the riser/casting combination. The total feeding requirement will depend on the speciﬁc alloy,

Methods of controlling shrinkage in an iron cube to reduce riser size. (a) Open-top riser. (b) Open-top riser plus chill. (c) Small open-top riser plus chill. (d) Insulated riser. (e) Insulated riser plus chill

the amount of superheat, the casting geometry, and the molding medium. Liquid shrinkage depends on the alloy and the amount of superheat. Liquid shrinkage for carbon steels is generally in the range of 1.6 to 1.8%/100 C (0.9 to 1.0%/100 F) superheat. For graphitic cast irons, liquid shrinkage has been reported in the range of 0.68 to 1.8%/ 100 C (0.38 to 1.0%/100 F) (Ref 2–5). Solidiﬁcation Shrinkage. Table 1 indicates that solidiﬁcation shrinkage varies considerably according to the alloy melted and that, within the graphitic cast irons, expansion may occur. This phenomenon is often ascribed to the precipitation of the less dense graphite phase overcoming the contraction associated with the solidiﬁcation of austenite. Theoretical calculations indicate that such density differences cannot account for the higher reported expansion percentages (Ref 2, 5). Practice shows that, with proper control of metallurgical and mold conditions, expansion phenomena can be used to reduce greatly or eliminate risers; with the liquid shrinkage accommodated by the gating system instead of the riser system (Ref 5). Mold Dilation. Mold wall movement after a mold cavity has been ﬁlled with liquid metal can enlarge the casting and thus increase the feed metal requirements. Such mold dilation is a function of the molding medium, the mold ﬁlling temperature, and the alloy. With gray and ductile irons, mold dilation may result partially from expansion pressures within the solidifying casting generated by the precipitation of graphite. In soft green sand molds, such mold dilation may produce an additional 15% feed metal requirement above that needed to satisfy the calculated liquid and solidiﬁcation shrinkages (Ref 7). In copper-base alloys, it has been suggested that an additional 1% volumetric shrinkage should be expected as a result of mold cavity expansion in green sand molds (Ref 8). Casting Geometry. The shape of a casting will affect the size of the riser needed to meet its feed requirements for the obvious reason that the longer the casting takes to solidify, the longer the riser must maintain a reservoir of liquid metal. For rangy, thin-section castings,

Table 1 Solidiﬁcation contraction for various cast metals Metal

Carbon steel 1% carbon steel White iron Gray iron Ductile iron Copper Cu-30Zn Cu-10Al Aluminum Al-4.5Cu Al-12Si Magnesium Zinc Source: Ref 6

Percentage of volumetric solidiﬁcation contraction

2.5–3 4 4–5.5 Varies from 1.6 contraction to 2.5 expansion Varies from 2.7 contraction to 4.5 expansion 4.9 4.5 4 6.6 6.3 3.8 4.2 6.5

Table 2

Minimum volume requirements for risers on steel castings Minimum riser volume/casting volume (Vr/Vc), % Insulated risers

Type of casting

Very chunky (cubes, etc.): dimensions in ratio 1:1.33:2 Chunky: dimensions in ratio 1:24 Average: dimensions in ratio 1:3:9 Fairly rangy: dimensions in ratio 1:10:10 Rangy: dimensions in ratio 1:15:30 Very rangy: dimensions in ratio 1:>15:>30 Source: Ref 9

Sand risers

H/D = 1:1

H/D = 2:1

H/D = 1:1

H/D = 2:1

32

40

140

198

26

32

106

140

19

22

58

75

14

16

30

38

9

10

13

15

8

8

11

13

Riser Design / 63 where solidiﬁcation will be rapid, feed metal requirements may be smaller than what would ordinarily be calculated. This is because a portion of the liquid and solidiﬁcation shrinkages will be fed by liquid metal entering the mold from the gating system. Table 2 indicates the effect of differences in casting geometry on minimum riser volume requirements for steel castings.

Riser Location To determine the correct riser locations, the designer should make use of the concept of directional solidiﬁcation. If shrinkage cavities in the casting are to be avoided, solidiﬁcation should proceed directionally from those parts of the casting farthest from the riser, through

Fig. 4

Directional and progressive solidiﬁcation in a casting equipped with a riser. Source: Ref 10

Schematic of mode of freezing in pure metals. Crystallization begins at the mold wall and advances into the casting interior on a plane solidiﬁcation front. Source: Adapted from Ref 11

the intermediate portions of the casting, and ﬁnally into the riser itself, where the ﬁnal solidiﬁcation will occur. Shrinkage at each step of solidiﬁcation is thus fed by liquid feed metal being drawn out of the riser. The ability to achieve such directional solidiﬁcation will depend on: The alloy and its mode of solidiﬁcation The mold medium The casting design

Two distinct types of castings must be considered: castings with uniform wall thickness and castings with wall sections of varying thickness. Progressive and Directional Solidiﬁcation. Figure 4 illustrates the interplay of progressive and directional solidiﬁcation in a casting. With the mold cavity ﬁlled, solidiﬁcation will generally proceed from the mold wall, where a skin of solid metal will form. As heat is lost to the mold, that skin will grow progressively inward. Two conditions serve to change the rate of this growth. At the casting edge, where the greater surface area allows more rapid transfer of heat to the mold, the solidiﬁcation rate will be faster. At the riser, where the mass of the riser provides more heat, and where heat transfer to the mold is reduced at the internal angle of the riser/casting junction, the rate of skin formation will be reduced. This combination of edge effect, or end effect, and riser effect provides directional solidiﬁcation.

If the wedge-shaped pattern of the solidiﬁcation front begun at the casting edge can be maintained, a channel of liquid feed metal should be available throughout its progress toward the riser. If, however, the parallel advancing walls progressively solidifying in the intermediate zone begin to meet, movement of liquid feed metal will be restricted, and centerline shrinkage will result. Solidiﬁcation Mode. The ability to promote and sustain directional solidiﬁcation will depend greatly on the manner in which an alloy solidiﬁes. Alloys can be classiﬁed into three types based on their freezing ranges: C (<90 F) Intermediate: interval of 50 to 110 C (90 to 200 F) Long: interval >110 C (>200 F)

Short: liquidus-to-solidus interval <50

This classiﬁcation is not precise, but the general solidiﬁcation mode of each type is illustrated in Fig. 5 to 8. For pure metals (Fig. 5), in which the freezing range approaches zero, the solidifying casting walls progress inward as a plane front. Short-freezing-range alloys (Fig. 6) will show a strong tendency toward skin formation, and the fronts of the crystals solidifying inward (start of freeze) will not advance much faster than their bases (end of freeze). Such relative, short crystalline growth helps keep liquid

Fig. 5

Fig. 6

Diagram of mode of freezing in alloys having a short freezing range

Fig. 7

Schematic of mode of freezing in alloys having a long freezing range

64 / Casting Design and Performance feed metal in contact with all the solidifying surfaces. Such strong progressive solidiﬁcation in these short-freezing-range alloys promotes the development of directional solidiﬁcation along any temperature gradients in the solidifying casting. For example, in carbon steel, gradients of only 0.022 to 0.045 C/mm (1 to 2 F/ in.) in plates and 0.135 to 0.269 C/mm (6 to 12 F/in.) in bars are sufﬁcient to produce a shrinkage-free casting section through directional solidiﬁcation. For long-freezing-range alloys (Fig. 7), the development of directional solidiﬁcation is difﬁcult. Although a thin skin may initially form on the mold walls, solidiﬁcation does not proceed progressively inward. Rather, it develops throughout the solidifying casting at scattered locations, forming equiaxed islands. This mushy or pasty mode of solidiﬁcation results in the development of numerous small channels of liquid metal late in solidiﬁcation. Feeding through these channels is restricted, and dispersed shrinkage porosity occurs throughout the casting. Such solidiﬁcation is typical of many commercial copper-base alloys, in which the difﬁculty in feeding caused by shrinkage porosity is aggravated, especially in thick sections, by the high thermal conductivity of the alloys, which helps maintain a nearly uniform temperature throughout the solidifying casting. To promote directional solidiﬁcation in such alloys may require temperature gradients as high as 1.46 C/mm (65 F/in.), which can usually be achieved only by severely

chilling one portion of the solidifying casting. Generally, the goal in risering such alloys is not to eliminate shrinkage but to ensure that it is ﬁnely dispersed (microporosity). For alloys with an intermediate freezing range (Fig. 8), the mode of solidiﬁcation will combine elements of both the skin-forming and mushy solidiﬁcation modes. Short-freezing-range alloys may shift to a more intermediate mode of solidiﬁcation in heavy casting sections, in which heat loss from the casting surface will be slowed as the molding medium heats up. As temperature gradients from the center of the solidifying section to the casting edge are reduced, crystal growth will change from the columnar pattern growing in from the mold walls to an equiaxial pattern dispersed throughout the still-liquid center. The various solidiﬁcation modes result in very different typical shrinkage conﬁgurations in the casting and riser (Fig. 9 and 10) and present the methods engineer with distinctly different problems to overcome in riser and casting design. Selection of appropriate methods will depend largely on the possibility of promoting directional solidiﬁcation. Figure 11 illustrates the effects of several mold and metal variables on the development of progressive and, therefore, directional, solidiﬁcation. Castings of Uniform Wall Thickness. The speciﬁc alloy and the section conﬁguration will combine to impose a limiting feeding distance (FD) over which a casting can solidify free of centerline shrinkage. As shown in Fig. 12, total FD in a section of steel plate (thickness, T; width, >3T) with a cooling edge is the sum of the riser effect and edge effect. Several key points are shown: The contribution from the edge effect is

generally greater than that from riser effect.

In the absence of cooling edges, FD between

risers is dramatically reduced.

If the maximum FD in a section is exceeded,

the edge effect will give a sound edge to its usual length, but centerline shrinkage may extend for some variable distance into the area that would ordinarily be expected to be sound because of riser effect.

Fig. 8

Schematic of intermediate mode of freezing in alloys having a moderate freezing range

Fig. 9

Forms of shrinkage porosity in the sand castings of alloys that freeze in a pasty manner

Figure 13 illustrates the same relationships in steel bars (width equals T). When compared with Fig. 12, Fig. 13 also highlights the fact that bar-shaped sections will have shorter FDs than platelike sections of the same thickness. Figure 14 shows the use of chills to extend FD. When applied to the edge of a casting section, the chill will withdraw heat rapidly, enhancing the development of directional solidiﬁcation away from the edge. This will add to the length of the zone that will be sound due to end effect. In addition, if a chill is placed between risers in a casting section where there is no natural cooling edge, it can be used to establish an artiﬁcial end effect. In this way, the distance between risers can be dramatically increased, thus reducing the number of risers needed to ensure a sound casting. Such a use of chills is illustrated for a steel ﬂange (Fig. 15). The ﬁrst attempt (Fig. 15a) at subdividing this casting into feedable sections with riser placement based on the absence of end effect (except at the periphery of the ﬂange) requires eight risers (two on the hub and six on the ﬂange). The overlapping feeding zones of the riser (based on riser effect) cover the feeding requirements of most of the ﬂange, but there still remain unfed regions (in which centerline shrinkage would occur) that would probably require the addition of at least one more riser on the ﬂange. The second subdivision of the casting (Fig. 15b), using chills to establish artiﬁcial end effects, reduces the number of risers needed to only ﬁve (one on the hub and four on the ﬂange). Such an application of chills, in addition to ensuring a sound casting, can provide economic advantages by simplifying molding and patternmaking procedures, increasing casting yield, and reducing riser removal costs. It should be reemphasized that FDs in sections of uniform thickness depend on alloy characteristics and section conﬁguration (that is, whether barlike or platelike) to determine how far directional solidiﬁcation can be sustained. If the walls of the progressively solidifying intermediate sections begin to come together, disrupting and constricting the feeding

Fig. 10

Shrinkage cavities produced by skin formation in alloys

Riser Design / 65

Feeding distance relationships in steel plates (section width greater than 3T, where T – thickness). Source: Ref 13

Fig. 12

Schematic of the effect of mold and metal variables on progressive solidiﬁcation. (a) Effect of mold conductivity on solidifying metal. (b) Effect of liquidus-to- solidus range of solidifying metal. (c) Effect of conductivity on solidifying metal. (d) Effect of temperature level on solidiﬁcation. Source: Ref 12

Fig. 11

channels through which feed metal can move, centerline shrinkage will occur. Once the FD of a section has been exceeded, the development of centerline shrinkage will not be overcome by increasing the size of the riser. Feeding Distance. It should also be noted that most of the data on FD is for steel. A variety

of nomograms and tables have been widely used for decades (Ref 13–16). These served the industry well, but in the 1990s it was felt these FDs were overly conservative. After researching low-alloy plate casting throughout North America, analyzing foundry results, and testing simulation software, the Steel Founders’ Society of

Feeding distance relationships in steel bars (section width equal to thickness, T ). Source: Ref 13

Fig. 13

66 / Casting Design and Performance America (SFSA) presented FD rules (Ref 17). They address the section conﬁguration (bar or plate) by expressing riser and end zone effects as a function of width to thickness, W/T. Multiplier’s are provided for differences in alloy composition, mold materials, pouring superheats, and level of casting soundness desired. For most other metals, such precise data are not available, so their FDs are often characterized by their similarity (or lack of similarity) to carbon steel. One method of making this comparison is by the calculation of centerline feeding resistance (Ref 6). This measurement indicates that some alloys (for example, Monel) will have FDs very much like those of carbon steel, and the methods engineer can use FD tables and nomograms established for the latter. Some alloys, such as 18-8 steel, 12% Cr steel, 99.8% Cu, and 60-40 brass, will have greater FDs in similar casting sections, so a multiplier can be applied to steel-base rules. Multipliers for high-alloy steel castings have been developed for steel alloys CF-8M, CA-15, HH, HK, and HP (Ref 18). In other alloys, such as 88-10-2 bronze, 85-55-5 bronze, Al- 8Mg, and Al-4.5Cu, the centerline feeding resistance is so great that FD is virtually nonexistent unless severe chilling is used. Graphitic Cast Irons. In these materials, the crystallization of the low-density graphite as ﬂakes or nodules should promote self-feeding behavior in the solidifying casting and allow inﬁnite FDs as long as the early liquid feed

demand of the casting is satisﬁed by the gating system or by a riser. Very large gray and ductile iron castings are often produced to meet radiographic or ultrasonic soundness standards with minimal risering. The key to such practice is a rigid molding medium to minimize mold wall movement.

Castings with Sections of Varying Thicknesses. Most commercially produced castings consist of sections of varying thickness and conﬁguration. Thicker, more slowly solidifying sections are separated from each other by thinner, rapidly solidifying sections. The heavier sections will then act as risers, supplying the feed metal demands of the lighter sections. The selection of a risering method changes from a problem of riser spacing to one of riser placement so that each of the late- solidifying sections has its feed requirements met. Therefore, the methods engineer must divide a casting into sections requiring risers by determining the feeding paths by which solidiﬁcation will move directionally from early- to late- solidifying sections (Ref 19). These feed paths can often be manipulated by proper application of chilling or insulating materials to minimize risering needs. Several methods of risering isolated heavy sections are shown in Fig. 16, using a casting with two heavy sections joined by a thinner connection. In Fig. 16(a), with no risers, shrinkage develops in the two heavy sections. When an adequate riser is applied to one side (Fig. 16b), shrinkage remains unfed in the other heavy section, or hot spot, because the connecting section freezes ﬁrst. A simple solution is to use risers on both sides (Fig. 16c). Two feed paths are thus established, running from the center section outward to the two risers. Two alternative methods are shown by which a single feed path can be generated. In Fig. 16(d), a chill is applied to the isolated section to reduce its solidiﬁcation time below that of the connection. In Fig. 16(e), the solidiﬁcation time of the connection is extended by applying an insulating or exothermic pad to the casting walls.

Duration of Liquid Feed Metal Availability

Use of chills to reduce the number of risers (T) on a steel ﬂange casting. (a) Side and top view of the casting illustrating locations of the eight risers used when the workpiece is divided into feeding areas without considering end effects. (b) Top view of identical casting showing locations of ﬁve risers used when the workpiece is divided into feeding areas in which riser effect and end effect considerations are accounted for through the use of chills. Source: Ref 14

Fig. 15

Fig. 14

Effect of chills on feeding distance relationships in steel bars. Source: Ref 13

A variety of methods have been devised to calculate the riser size needed to ensure that liquid feed metal will be available for as long as the solidifying casting requires. Several of the most commonly used methods are discussed brieﬂy. Shape Factor Method. Drawing on the theoretical work of Caine (Ref 20), researchers at the U.S. Naval Research Laboratory (NRL) devised a method to determine riser size by calculating a shape factor by adding the length and width of a casting section and dividing this sum by the section thickness (Ref 21). The NRL method is illustrated in Fig. 17 with the example of a 508 mm (20 in.) square plate 50 mm (2 in.) thick. According to the earlier discussion of FD, this casting (at least in steel) should be able to be made free of centerline shrinkage with a single top riser in the center of the casting. Figure 12 shows that the FD of this 50 mm (2 in.) plate, with end effect, will be 229 mm (9 in.) from the edge of the riser. If the proposed central riser is at least 50 mm (2 in.) in diameter, the FD requirements should be met.

Riser Design / 67

Procedure for determining a minimum riser size using the shape factor method. (a) The shape factor, derived from the dimensions of the casting, is used to determine the minimum ratio of riser volume to casting volume. (b) The same dimensions used to obtain the shape factor provide the data for casting volume, Vc, from which riser height and riser diameter can be determined. Dimensions given in inches. Source: Ref 21

Fig. 17

Risering of isolated heavy sections joined by a thinner section to minimize shrinkage and number of risers. (a) Workpiece with no risers. (b) Riser added to one side. (c) Risers located on both ends. (d) Chill applied to one end and riser to other end. (e) Riser used on one end and insulator or exothermic pad on opposite end

Fig. 16

Figure 17 shows the calculation of the shape factor (SF) and volume of the casting. In Fig. 17(a), one reads up from the SF (20) on the x-axis to intersect the NRL-developed curve. From this point, one reads across to the y-axis to ﬁnd a minimum riser volume, Vr, needed to feed the casting volume, Vc. Figure 17(a) shows that, for this example, the required Vr/Vc ratio is 0.25; that is, the riser must be designed to contain a minimum of 25% as much metal as is required by the casting itself. The SFSA guideline (Ref 17) makes use of the data (Ref 21) to express the volume ratio SF relation similar to that shown in Fig. 17(a).

The actual curve is slightly different, being plotted from the equation (Ref 22):

in a riser as it simultaneously feeds metal into the casting and solidiﬁes inward from its walls (Fig. 18) and simpliﬁes it to the conﬁguration of a cylindrical pipe (Ref 23). The size of this pipe depends on the weight of the casting section and the shrinkage percentage (both liquid and solidiﬁcation shrinkages) of the casting alloy. Simple nomograms and tables are available to determine pipe sizes for a given alloy and casting weight and for a variety of ratios for the height of the pipe, Hp, divided by the diameter of the pipe, Dp. As shown in Fig. 19, the actual riser size is determined by surrounding the pipe by a riser wall, W, that is expected to solidify at the same rate as the walls of the casting. The calculation of the required wall thickness will thus vary with casting conﬁguration:

Vr =Vc ¼ 2:51SF0:74

For platelike castings, W = casting thickness. For cube-shaped castings, W = 35% of the

Formation of (a) the conical type of shrinkage cavity due to (b) the accumulation of solidiﬁed layers on the outer walls of the riser

Fig. 18

(Eq 1)

The volume ratio can be calculated directly from this equation. To ﬁnd the appropriate riser size, simple nomograms like the one shown in Fig. 17(b) are used; risers of various height-to-diameter ratios can be found to satisfy the calculated volume requirement. As shown in Fig. 17(b), most such nomograms assume a riser H/D ratio of approximately 1 to 1. Geometric Method. Originally devised for determining side risers for malleable iron castings, the geometric method takes the typical conical shape of the shrinkage cavity formed

edge length of the cube.

For round or square bars, W = twice the bar

diameter or thickness. The ﬁnal riser diameter, Dr, thus equals Dp + 2W. The overall riser height, Hr, is determined by adding together: The pipe height, Hp A middle section, Hm, above the riser con-

nection with the casting (neck) to provide a safety margin A base section, Hb, including the depth of the riser neck

68 / Casting Design and Performance This concept was developed by Wlodawer for practical riser calculations by eliminating the need to calculate actual solidiﬁcation times in favor of simply determining the relative solidiﬁcation times of casting sections and risers (Ref 14). Chvorinov’s rule is thus simpliﬁed to: t

Vc Ac

(Eq 3)

The volume-to-area ratio of the casting is termed the casting modulus, Mc: Mc ¼

Fig. 19

Fig. 20

Cross section of a piping side riser designed by the geometric method

Example of a design system for tapered tall risers. Source: Ref 24

Casting yield is usually maximized by choosing a ratio for Hp/Dp of 2.5 to 1, generally resulting in a riser with an H/D ratio (above the neck) of approximately 1 to 1. Design rules for the geometric method have been further developed to design tapered risers, as shown in Fig. 20. This design not only improves casting yield but also promotes the development of deﬁnite piping behavior in the riser, ensuring that metal will feed out from riser to casting. The modulus method is based on the concept that the freezing time of a casting or a casting section can be approximated by using Chvorinov’s rule: t ¼ k2

Vc Ac

2 ¼ k2 Mc2

(Eq 2)

where t is the freezing time of the casting, Vc is the volume of the casting, Ac is the surface area of the casting, and k is the constant governed by metal and mold properties.

Vc Ac

(Eq 4)

The freezing times of risers and castings are proportional to their respective moduli, and if the modulus of the riser, Mr, is sufﬁciently greater than the modulus of the casting, Mc, good feeding will be obtained. In steel, if Mr = 1.2 Mc, feeding will be satisfactory. For other skin-forming alloys, including many aluminum- and copper-base alloys, the Mr/Mc ratio of 1.2 to 1 is appropriate. With gray and ductile irons, depending on the carbon equivalent, the required Mr/Mc ratio can range from 0.8:1 to 1.2:1 because the riser may not be required to supply feed metal throughout the entire solidiﬁcation time of the casting. Wlodawer simpliﬁed the modulus method by showing that many casting sections can be reduced to simple geometric shapes for which Mc can be readily found without elaborate calculations of actual surface areas and volumes (for example, for a plate section, Mc = half the plate thickness). Wlodawer further simpliﬁed the method by providing simple conversion charts, such as that shown in Fig. 21. This chart allows easy calculation of risers of various shapes if the necessary modulus is known. One note of caution must be kept in mind when using the modulus approach to riser design. This method can recommend risers that are too small on very rangy, thin-section castings. In such cases, the thermal requirements of the casting would indicate very small risers, but the demand for feed metal volume can still be substantial. For such volumecontrolled castings, the riser size arrived at through modulus calculations must be veriﬁed against riser volume charts such as that given in Table 2. Computerized Methods. Over the past 15 years, a virtual revolution has taken place regarding computerized modeling methods. Programs can be grouped into two very general categories. The ﬁrst category includes empirical programs to assist with riser sizing (Ref 25–29). These programs are largely based on practical experience and run nearly instantaneously on the most basic computer platforms. The basis of the calculations is generally one or more of the manual calculation methods discussed previously coupled with measured performance data from foundry trials. These methods

usually contain algorithms to assist in calculating casting weights, section modulus, and FD. In addition, riser calculations generally require simple inputs of such factors as section weight and shape, shrinkage percentage, section modulus, mold medium, ingate location, and desired riser shape for the program to provide a variety of riser alternatives. Quite often, the riser recommendation from these types of programs represents the starting point for a more detailed analysis using the second category of programs. The second category comprises programs that simulate the entire casting process using ﬁrstprinciples physics equations. Casting ﬁlling, solidiﬁcation, heat transfer, residual stress/ strain, heat treatment, and other important casting processes are modeled using complicated, detailed algorithms based on the underlying physics equations. As an example, the casting ﬁlling process for programs in this category would use the Navier-Stokes equations to predict the ﬁlling characteristics of the gating system and mold cavity. Computational power for simple personal computers has increased substantially over the past few years, which has allowed the use of these types of programs to become routine. In many foundries today (2008), every new part is modeled using casting process simulation, and the gating and risering are optimized before the casting is ever produced on the shop ﬂoor. The reference on feeding and risering (Ref 18) provides an extensive set of simulations for high-alloy steel casting.

Feeding Systems Used in Riser Design Feeding systems are widely used by the foundry industry to increase casting soundness and to reduce the cost of casting manufacture. Feeding systems reduce the rate of heat transfer from the riser to the molding medium and to the atmosphere. In riser design, three types of feeding systems are commonly used: Riser sleeves or panels are used to insulate

the riser sidewall or riser top from the mold.

Topping compounds are used to insulate the

top of open risers from the atmosphere.

Breaker cores are used between the riser and

the casting to facilitate the removal of the riser from the casting; the use of these materials is illustrated in Fig. 22. Metal is transferred from the riser by three mechanisms: gravity, atmospheric pressure (or pressure applied by other means, as in die casting), and capillary action. For most casting processes, atmospheric pressure is the most important.

Feeding System Advantages Feeding systems delay solidiﬁcation and the formation of a skin of solid metal in the riser.

Riser Design / 69

Schematic showing the application of feeding systems for (a) a closed-top riser and (b) an open-top riser

Fig. 22

Feeding systems, by reducing the rate of heat transfer between the riser and the mold, increase the effective modulus of a riser when compared with the geometric modulus of that riser. For a given riser modulus, a riser incorporating feeding systems will be smaller than an uninsulated riser. Less liquid metal is needed to produce a given casting, and the cost of producing that casting is reduced (Ref 30). Reducing riser size may allow a given casting to be produced in a smaller mold or may allow the production of more castings in a given mold, both resulting in lower cost. The value of feeding systems in delaying riser solidiﬁcation is indicated in Table 3, in which the solidiﬁcation time of a 102 mm (4 in.) diameter, 102 mm (4 in.) high riser is calculated in three alloys with various combinations of insulation on the riser sidewall and top. Also given is the percentage of total heat lost from the solidifying riser through radiation from the riser top. This highlights the fact that hot topping to reduce radiative heat loss is signiﬁcantly more important in steel casting than in aluminum casting.

Fig. 21

Riser conﬁgurations and their characteristic values (Mr, Vr, D, H, and so on). Source: Ref 14

Thermal Properties of Materials This promotes atmospheric puncture of the riser, assisting in the transfer of metal from the riser to the casting. Alloys, speciﬁcally those that have a wide freezing range, beneﬁt the most from feeding systems.

Feeding systems increase the temperature gradient between riser and casting. As a result, directional solidiﬁcation from casting to riser is promoted; that is, a stronger feeding path is established.

Riser sleeves and topping compounds are classiﬁed by thermal properties as exothermic, insulating, and exothermic-insulating. The generalized thermal properties of each of these classiﬁcations are shown in Fig. 23.

70 / Casting Design and Performance Exothermic feeding systems are based on the oxidation of aluminum to produce heat. These feeding systems tend to be of relatively high density, and the matrix has thermal properties similar to those of a sand mold after the exothermic reaction has subsided. They display an initial chilling effect on the molten metal and rely on a strong exothermic reaction to remelt any metal that may have solidiﬁed. Exothermic materials tend to be used on small and intermediate-sized risers and are not generally recommended for large risers having a long solidiﬁcation time. Exothermic materials also have the greatest likelihood of metal contamination. Metal-producing topping compounds are a special class of exothermic material, and they have a composition based on the thermite reaction. These materials are sometimes used on large risers. Contamination must be carefully monitored; the metal produced by the thermite reaction should not be allowed to feed into the casting. Insulating feeding systems are formulated from refractory materials to have a low density. They exhibit a very low initial chill of the molten metal and rely on their low density to minimize heat loss from the riser. Insulating feeding systems tend to be used on small to intermediate-sized risers and alloys having lower pouring temperatures. They are not

generally recommended for large risers, because these low-density materials thermally degrade when exposed to high pressure and high temperature for extended periods of time. Insulating feeding systems are least likely to cause problems with metal contamination. Exothermic-insulating feeding systems consist of exothermic materials surrounded by a highly refractory, insulating matrix. These materials are the most versatile, having a low initial chill, extended exothermic reaction, and good insulation after the exothermic reaction has subsided. Exothermic-insulating materials are used for the entire range of riser sizes. Their propensity for metal contamination lies between insulating and exothermic materials.

Riser insulation in the form of a sleeve and hot topping may be regarded as effectively decreasing the surface area of the riser. The effect of the sidewall (sleeve) insulation may be represented by a factor x and that of a hot topping by a factor y, both relative to sand (where x = 1 for sand). The factors x and y have been termed apparent surface alteration factors (ASAF). The effective modulus Mr of a cylindrical riser incorporating feeding systems on both the sidewall and top is given by:

Feeding systems are formulated to minimize the possibility of metal contamination. Depending on the formulation of the feeding system and its method of application, certain elements may be absorbed by the metal in the riser. Carbon, silicon, aluminum, oxygen, nitrogen, and sulfur may be absorbed. The widespread use of feeding systems in riser design is an indication that contamination is not normally a problem, but the possibility should not be overlooked.

Mr ¼

Heat loss from a riser as a function of time for the major classiﬁcations of feeding systems, with sand included as the basis for comparison

Table 3 Alloy cast

Steel Copper Aluminum Source: Ref 31

(Eq 7)

This should be compared with the modulus of a similar 1 to 1 sand-lined riser, which is 0.2D. The ASAF values for feeding systems are generally available from manufacturers of these products. Although the size of a riser incorporating feeding systems can be readily calculated using the approach described previously, computer programs are now widely available to speed the process of designing risers—both conventional sand-lined risers and risers incorporating feeding systems.

The thermal properties of feeding systems can be readily incorporated into the modulus method of riser calculation discussed previously (Ref 32). Cylindrical risers are the most common because this shape has a high modulus for a given volume of metal and because this shape is easy to mold. The modulus of a cylindrical riser, Mr, is given by:

Fig. 23

D2 D2 ¼ 4Dð0:65Þ þ Dð0:7Þ 3:3D

¼ 0:303D

Factors in Riser Size

Riser Necks and Breaker Cores

2

Vr R H RH ¼ ¼ Ar 2RH þ R2 2H þ R DH ¼ 4H þ D

(Eq 6)

The ASAF values of insulating and exothermic feeding system materials vary and generally range from 0.50 to 0.90. The smaller the ASAF value, the more efﬁcient the insulation. For example, if x = 0.65 and y = 0.7, then for a cylindrical riser with a height-to-diameter ratio of 1 to 1 (where H = D):

Metal Contamination

Mr ¼

DH 4Hx þ Dy

Mr ¼

To promote directional solidiﬁcation from the casting into the riser, the modulus of the riser neck, Mn, should be intermediate to the moduli of the casting and riser, Mc and Mr (Ref 14). The general rule given previously for the required riser modulus was Mr = 1.2 Mc.

(Eq 5)

where R is the riser radius, D is the riser diameter, and H is the riser height.

Effect of risering systems on the solidiﬁcation times of various alloys Solidiﬁcation time, min

Radiation loss through top, %

Sand riser/open end

Sleeved riser/open top

Sand riser/insulated top

Sleeved riser/insulated top

42 26 8

5 8.2 12.3

7.5 15.1 31.1

13.4 14.0 14.3

43.0 45.0 45.6

Riser Design / 71

Fig. 24

General design rules for riser necks used in iron casting applications (side view and top view, respectively). (a) General type of side riser. (b) Side riser for plate casting. (c) Top round riser. Source: Ref 33

The general rule for riser neck design—at least for skin-forming alloys—is Mn = 1.1 Mc. Once again, the graphitic cast irons are an exception because the graphite expansion phase makes it unnecessary for the riser neck to stay open for feed metal transfer throughout the entire solidiﬁcation time of the casting. Depending on the chemistry of the casting, riser necks for gray and ductile irons can have moduli in the range of 0.67 to 1.1 times the casting modulus (Ref 5). General design rules for riser necks for iron castings are widely available (Fig. 24). For the economical removal of risers, breaker cores can be used between the riser and the casting, as shown in Fig. 22. Breaker cores are typically made of bonded sand or fused ceramic materials. For risers lined with sleeves, the breaker core thickness is generally 10% of the riser diameter, and the breaker core opening is generally in the range of 40 to 50% of the riser diameter. By keeping the mass of the breaker core small, the breaker core quickly reaches the temperature of the surrounding metal and does not appreciably affect the solidiﬁcation of the riser.

Optimum Riser and Neck Conﬁgurations Applying the required riser to a speciﬁc casting may generate problems. For example, the necessary riser or neck size may not easily ﬁt the casting conﬁguration. An important problem may be created simply by attaching the riser, because the new riser/casting conﬁguration will have its own solidiﬁcation pattern, sometimes with unintended results. This is illustrated in Fig. 25, in which the attachment of a side riser at the center of the rim of a gear blank casting generates a hot spot inside the casting, where shrinkage would be expected to occur after the neck freezes. This problem can be avoided either by moving the attachment of the side riser or by using a top riser on the rim. The top riser has the added beneﬁt of improving casting yield. ACKNOWLEDGMENT We thank Richard Williams and Tony Midea, Foseco, Inc., for their assistance in

Effect of risers and riser contacts on solidiﬁcation wave fronts. Last points to solidify indicated by +. (a) Gear blank casting. (b) The side riser attached directly at rim center generates a hot spot inside the casting itself. (c) and (d) The addition of a thin section between the workpiece and the riser overcomes the problem of hot spots inside the casting. Dimensions in inches

Fig. 25

reviewing this article and revising the section “Computerized Methods.” The remainder of the article, with the exception of material related to new references, is substantially as written by L.A. Plutshack and A.L. Suschil in “Riser Design,” Casting, Vol 15, ASM Handbook, 1988. REFERENCES 1. H.F. Taylor, M.C. Flemings, and J. Wulff, Foundry Engineering, John Wiley & Sons, 1959 2. C.E. Bates and B. Patterson, Volumetric Changes Occurring During Freezing of Hypereutectic Ductile Irons, Trans. AFS, Vol 87, 1979, p 323–334

72 / Casting Design and Performance 3. B.P. Winter, T.R. Ostrom, D.H. Hartman, P.K. Trojan, and R.D. Pehlke, Mold Dilation and Volumetric Shrinkage of White, Gray, and Ductile Cast Irons, Trans. AFS, Vol 92, 1984, p 551–560 4. R.W. Heine, The Fe-C-Si Solidiﬁcation Diagram for Cast Irons, Trans. AFS, Vol 94, 1986 5. S.I. Karsay, Ductile Iron III: Gating and Risering, QIT—Fer Et Titane Inc., 1981 6. R.A. Flinn, Fundamentals of Metal Casting, Addison-Wesley, 1963 7. J. Briggs, Risering Gray and Ductile Iron, 90% Yield Seminar: Gating and Risering of Iron Castings, Foseco, Inc., 1983 8. B.P. Winter, R.D. Pehlke, and P.K. Trojan, Volumetric Shrinkage and Gap Formation During Solidiﬁcation of Copper-Base Alloys, Trans. AFS, Vol 91, 1983, p 81–88 9. R.W. Ruddle, Inﬂuence of Feeding Practice on Energy Conservation and Production Economics in Steel Foundries, Br. Foundryman, Sept 1978, p 197–222 10. J.L. Francis and P.G.A. Pardoe, The Feeding of Iron Castings, Applied Science in the Casting of Metals, K. Strauss, Ed., Pergamon Press, 1970 11. R.W. Ruddle, Solidiﬁcation of Copper Alloys, Trans. AFS, Vol 68, 1960, p 685–690 12. R.W. Heine, C.R. Loper, Jr., and P.C. Rosenthal, Principles of Metal Casting, McGraw-Hill, 1967 13. W.S. Pellini, Factors Which Determine Riser Adequacy and Feeding Range, Trans. AFS, Vol 61, 1953, p 61–80

14. R. Wlodawer, Directional Solidiﬁcation of Steel Castings, Pergamon Press, 1966 15. H.F. Bishop and W.S. Pellini, The Contribution of Riser and Chill-Edge Effects to Soundness of Cast Steel Plates, Trans. AFS, Vol 58, 1950, p 185–197 16. E.T. Myskowski, H.F. Bishop, and W.S. Pellini, Feeding Range of Joined Sections, Trans. AFS, Vol 61, 1953, p 302–308 17. Feeding and Risering Guidelines for Steel Castings, Steel Founders’ Society of America, 2001 18. O.U. Shouzhu, K.D. Carlson, and C. Beckermann, Feeding and Risering of HighAlloy Steel Castings, Metall. Trans. B, Vol 36 (No. 1), Feb 2005, p 97–116 19. R.W. Heine, Feeding Paths for Risering Castings, Trans. AFS, Vol 76, 1968, p 463–469 20. J.B. Caine, A Theoretical Approach to the Problem of Dimensioning Risers, Trans. AFS, Vol 56, 1948, p 492–501 21. E.T. Myskowski, H.F. Bishop, and W.S. Pellini, A Simpliﬁed Method of Determining Riser Dimensions, Trans. AFS, Vol 63, 1955, p 271–281 22. W.D. Spiegelberg, “Computation of Solidiﬁcation Gradients in Cast Steel Sections,” Ph.D. thesis, Department of Metallurgy and Materials Science, Case Western Reserve University, 1970 23. R.W. Heine, Riser Design for Mold Dilation, Mod. Cast., Feb 1965 24. R.W. Heine, Design Method for Tapered Riser Feeding of Ductile Iron Castings in

25.

26. 27.

28.

29.

30. 31. 32. 33.

Green Sand, Trans. AFS, Vol 90, 1982, p 147–158 R.W. Ruddle and A.L. Suschil, Riser Sizing by Microcomputer, Modeling of Casting and Welding Processes II, 1983 Engineering Foundation Conference, Conference Proceedings, The Metallurgical Society, 1983, p 403–419 C.F. Corbett, Methoding of Castings Using a Microcomputer, Br. Foundryman, Aug 1983, p 67–78 G.L. Moffat, FEEDERCALC—The Development of a Micro-Computer Program to Determine Optimum Feeder Sizes for Ductile Iron Castings, Foundry Pract., No. 210, Jan 1985, p 3–8 J.E. Pickin, The Development of a Computer Assisted Feeding System, Proceedings of 1981 Annual Conference, Steel Castings Research and Trade Association, 1981 W.T. Adams and K.W. Murphy, Optimum Full Contact Top Risers to Avoid Severs Under Riser Chemical Segregation in Steel Casting, Trans. AFS, Vol 88, 1980, p 389–404 R.W. Ruddle, Proﬁt Through Steel Risering Technology, Trans. AFS, Vol 87, 1979, p 423–432 C.M. Adams, Jr., and H.F. Taylor, Fundamentals of Riser Behavior, Trans. AFS, Vol 61, 1953, p 686 R.W. Ruddle, “Risering of Steel Castings,” Foseco, Inc., 1979 J.F. Wallace and E.B. Evans, Risering of Gray Iron Castings, Trans. AFS, Vol 66, 1958, p 49

Casting Design and Performance Pages 73–80

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Gating Design A GATING SYSTEM is the conduit network through which liquid metal enters a mold and ﬂows to ﬁll the mold cavity, where the metal can then solidify to form the desired casting shape. The basic components of a simple gating system for a horizontally parted mold are shown in Fig. 1. A pouring cup or a pouring basin provides an opening for the introduction of metal from a pouring device. A sprue carries the liquid metal down to join one or more runners, which distribute the metal throughout the mold until it can enter the casting cavity through ingates.

Design Variables Methods used to promote any of the desirable design considerations discussed subsequently often conﬂict with another desired effect. For example, attempts to ﬁll a mold rapidly can result in metal velocities that promote mold erosion. As a result, any gating system will generally be a compromise among conﬂicting design considerations, with the relative importance of the consideration being determined by the speciﬁc casting and its molding and pouring conditions. Rapid mold ﬁlling can be important for several reasons. Especially with thin-section

Fig. 1

Basic components of a simple gating system for a horizontally parted mold

castings, heat loss from the liquid metal during mold ﬁlling may result in premature freezing, producing surface defects (for example, cold laps) or incompletely ﬁlled sections (misruns). Superheating of the molten metal will increase ﬂuidity and retard freezing, but excessive superheat can increase problems of gas pickup by the molten metal and exaggerate the thermal degradation of the mold medium as well as increase costs. In addition, the mold ﬁlling time should be kept shorter than the mold producing time of the molding equipment to maximize productivity. Minimizing Turbulence. Turbulent ﬁlling and ﬂow in the gating system and mold cavity can increase mechanical and thermal attack on the mold. More important, turbulence may produce casting defects by promoting the entrainment of gases into the ﬂowing metal. These gases may by themselves become defects (for example, bubbles), or they may produce dross or inclusions by reacting with the liquid metal. Turbulent ﬂow increases the surface area of the liquid metal exposed to air within the gating system. The susceptibility of different casting alloys to oxidation varies considerably. For those alloys that are highly sensitive to oxidation, such as aluminum alloys; magnesium alloys; and silicon, aluminum, and manganese bronzes, turbulence can generate extensive oxide ﬁlms that will be churned into the ﬂowing metal, often causing unacceptable defects. Avoiding Mold and Core Erosion. High ﬂow velocity or improperly directed ﬂow against a mold or core surface may produce defective castings by eroding the mold surface (thus enlarging the mold cavity) and by entraining the dislodged particles of the mold to produce inclusions in the casting. Removing Slag, Dross, and Inclusions. This factor includes materials that may be introduced from outside the mold (for example, furnace slags and ladle refractories) and those that may be generated inside the system. Methods can be incorporated into the gating system to trap such particles (for example, ﬁlters) or to allow them time to ﬂoat out of the metal stream before entering the mold cavity. Promoting Favorable Thermal Gradients. Because the last metal to enter the mold cavity will generally be the hottest, it is usually desirable to introduce metal in those parts of the casting that would already be expected to be

the last to solidify. One obvious method of accomplishing this is to direct the metal ﬂow from the gating system into a riser, from which it then enters the mold cavity. Because the riser is designed to be the last part of the riser/casting system to solidify, such a gating arrangement will help promote directional solidiﬁcation from the casting to the riser. If the gating system cannot be designed to promote desirable thermal gradients, it should at least be designed so that it will not produce unfavorable gradients. This will often involve introducing metal into the mold cavity through multiple ingates so that no one location becomes a hot spot. Maximizing Yield. A variety of unrecoverable costs reside with the metal that will ﬁll the gating system and risers. These components must then be removed from the casting and generally returned for remelt, where their value is downgraded to that of scrap. Production costs can be signiﬁcantly reduced by minimizing the amount of metal contained in the gating system. The production capacity of a foundry can also be enhanced by increasing the percentage of salable castings that can be produced from a given volume of melted metal. Economical Gating Removal. Costs associated with the cleaning and ﬁnishing of castings can be reduced if the number and size of ingate connections with the casting can be minimized. Again, it may be advantageous to introduce metal into the mold cavity through a riser, because the riser neck can also serve as an ingate. Avoiding casting distortion is especially important with rangy, thin-wall castings, in which uneven distribution of heat as the mold cavity is ﬁlled may produce undesirable solidiﬁcation patterns that cause the casting to warp. In addition, the contraction of the gating system as it solidiﬁes can pull on sections of the solidifying casting, resulting in hot tearing or distortion. Compatibility with Existing Molding/ Pouring Methods. Modern high-production molding machines and automated pouring systems often severely limit the ﬂexibility allowed in locating and shaping the pouring cup and sprue for introducing metal into the mold. They also generally place deﬁnite limits on the rate at which metal can be poured. Controlled Flow Conditions. A steady ﬂow rate of metal in the gating system should be

74 / Casting Design and Performance established as soon as possible during mold ﬁlling, and the conditions of ﬂow should be predictably consistent from one mold to the next.

Principles of Fluid Flow Proper design of an optimized gating system will be made easier by the application of several fundamental principles of ﬂuid ﬂow. Numerous software packages employ these design principles. Chief among these ﬂuid ﬂow principles are Bernoulli’s theorem, the law of continuity, and the effect of momentum.

Bernoulli’s Theorem This basic law of hydraulics relates the pressure, velocity, and elevation along a line of ﬂow in a way that can be applied to gating systems. The theorem states that, at any point in a full system, the sum of the potential energy, kinetic energy, pressure energy, and frictional energy of a ﬂowing liquid is equal to a constant. The theorem can be expressed as: wZ þ wP v þ

wV 2 þ wF ¼ K 2g

(Eq 1)

where w is the total weight of the ﬂowing liquid, Z is the height of the liquid, P is the static pressure in liquid, v is the speciﬁc volume of liquid, g is the acceleration due to gravity (9.807 m/s2, or 386.4 in./s2), V is the velocity, F is the friction loss per unit weight, and K is a constant. If Eq 1 is divided by w, all the terms reduce to dimensions of length and will represent:

Potential head Z Pressure head, Pv Velocity head V2/2g Frictional loss of head F

Equation 1 allows prediction of the effect of the several variables at different points in the gating system, although several conditions inherent in foundry gating systems complicate and modify its strict application. For example: Equation 1 is for full systems, and at least at

the start of pouring, a gating system is empty. This indicates that a gating system should be designed to establish as quickly as possible the ﬂow conditions of a full system. Equation 1 assumes an impermeable wall around the ﬂowing metal. In sand foundry practice, the permeability of the mold medium can introduce problems, for example, air aspiration in the ﬂowing liquid. Additional energy losses due to turbulence or to friction (for example, because of changes in the direction of ﬂow) must be accounted for.

Fig. 2

Schematic illustrating the application of Bernoulli’s theorem to a gating system. Source: Ref 1

Heat loss from the liquid metal is not consid-

ered, which will set a limit on the time over which ﬂow can be maintained. Also, solidifying metal on the walls of the gating system components will alter their design while ﬂow continues. Bernoulli’s theorem (Eq 1) is illustrated in Fig. 2, and several practical interpretations can be derived. The potential energy is obviously at a maximum at the highest point in the system, that is, the top of the pouring basin. As metal ﬂows from the basin down the sprue, potential energy changes to kinetic energy as the stream increases in velocity because of gravity. As the sprue ﬁlls, a pressure head is developed. Once ﬂow in a ﬁlled system is established, the potential and frictional heads become virtually constant, so conditions within the gating system are determined by the interplay of the remaining factors. The velocity is high where the pressure is low, and vice versa.

The Law of Continuity This law states that, for a system with impermeable walls and ﬁlled with an incompressible ﬂuid, the rate of ﬂow will be the same at all points in the system. This can be expressed as: Q ¼ A1 v1 ¼ A2 v2

(Eq 2)

where Q is the rate of ﬂow, A is the cross-sectional area of the stream, v is the velocity of the stream, and the subscripts 1 and 2 designate two different locations in the system. Again, the permeability of sand molds can complicate the strict application of this law, introducing potential problems into the casting process.

One practical implication of the law of continuity is shown in Fig. 3, which illustrates the ﬂow of metal from a pouring basin. Due to gravitational acceleration, velocity increases as the stream falls, so the cross-sectional area of the stream must decrease proportionately to maintain the balance of the ﬂow rate. The result is the tapered shape typical of a free-falling stream shown in Fig. 3(a). If the same ﬂow is directed down a straightsided sprue (Fig. 3b), the falling stream will create a low-pressure area as it pulls away from the sprue walls and will probably aspirate air. In addition, the ﬂow will tend to be uneven and turbulent, especially when the stream reaches the base of the sprue. The tapered sprue shown in Fig. 3(c) is designed to conform to the natural form of the ﬂowing stream and therefore reduces turbulence and the possibility of air aspiration. It also tends to ﬁll quickly, establishing the pressure head characteristic of the full-ﬂow conditions required by Eq 1. Many types of high-production molding units do not readily accommodate tapered sprues, so the gating system designer will try to approximate the effect of a tapered sprue by placing a restriction, or choke, at or near the base of the sprue to force the falling stream to back up into the sprue (Fig. 4).

Momentum Effects Newton’s ﬁrst law states that a body in motion will continue to move in a given direction until some force is exerted on it to change its direction. Reynold’s Numbers and Types of Flow. The ﬂow of liquids can be characterized by the Reynold’s number:

Gating Design / 75

Schematic showing the advantages of a tapered sprue over a straight-sided sprue. (a) Natural ﬂow of a free-falling liquid. (b) Air aspiration induced by liquid ﬂow in a straight-sided sprue. (c) Liquid ﬂow in a tapered sprue

Fig. 3

Schematic showing the formation of lowpressure areas due to abrupt changes in the cross section of a ﬂow channel. (a) Sudden enlargement of the channel. (b) Sudden reduction of the channel

Fig. 6

Choke mechanisms incorporated into straightsided sprues to approximate liquid ﬂow in tapered sprues. (a) Choke core. (b) Runner choke

Fig. 4

NR ¼

vdr m

(Eq 3)

where NR is the Reynold’s number, v is the velocity of the liquid, d is the diameter of the liquid channel, r is the density of the liquid, and m is the viscosity of the liquid. As shown in Fig. 5, if the Reynold’s number is less than 2000, the ﬂow is characterized as laminar, with the molecules of the liquid tending to move in straight lines without turbulence (Fig. 5a). For NR between 2000 and 20,000, some mixing and turbulence will occur (Fig. 5b), but a relatively undisturbed boundary layer will be maintained on the surface of the stream. This type of turbulent ﬂow, common in most foundry gating systems, can be considered relatively harmless so long as the surface is not

Reynold’s number, NR, and its relationship to ﬂow characterization. (a) NR < 2000, laminar ﬂow. (b) 2000 NR < 20,000, turbulent ﬂow. (c) NR 20,000, severe turbulent ﬂow

Fig. 5

ruptured, thus avoiding air entrainment in the ﬂowing stream. With NR of 20,000 or greater, ﬂow will be severely turbulent (Fig. 5c). This will lead to rupturing of the stream surface, with the strong likelihood of air entrainment and dross formation as the ﬂowing metal reacts with gases. Abrupt Changes in Flow Channel Cross Section. As shown in Fig. 6, low-pressure zones—with a resulting tendency toward air entrainment—can be created as the metal stream pulls away from the mold wall. With a sudden enlargement of the channel (Fig. 6a), momentum effects will carry the stream forward and create low-pressure zones at the enlargement. With a sudden reduction in the channel (Fig. 6b), the law of continuity shows that the stream velocity must increase rapidly. This spurting ﬂow will create a low-pressure

zone directly after the constriction. The problems illustrated in Fig. 6 can be minimized by making gradual changes in the ﬂow channel cross section; abrupt changes should be avoided. Abrupt Changes in Flow Direction. As shown in Fig. 7, sudden changes in the direction of ﬂow can produce low-pressure zones, as described previously. Problems of air entrainment can be minimized by making the change in direction more gradual. Abrupt changes in ﬂow direction, in addition to increasing the chances of metal damage, will increase the frictional losses during ﬂow. As shown in Fig. 8, a system with high frictional losses will require a greater pressure head to maintain a given ﬂow velocity. Using a Runner Extension. Use of a runner extension beyond the last ingate is illustrated in Fig. 1. The ﬁrst metal entering the gating system will generally be the most heavily damaged by contact with the mold medium and with air as it ﬂows. To avoid having this metal enter the casting cavity, momentum effects can be used to carry it past the ingates and into the runner extension. The ingates will then ﬁll with the cleaner, less damaged metal that follows the initial molten metal stream. Equalizing ﬂow through ingates by decreasing the runner cross section after the ingate is illustrated in Fig. 9; this is done for systems with multiple ingates. As noted earlier, in the ﬁlled system shown, potential and frictional energies become constants, so they can be dropped from consideration in Eq 1 to show the interaction of pressure and velocity effects. At the ﬁrst ingate, velocity is high as momentum effects carry the ﬂowing stream past the gate. At the second ingate, velocity decreases in the runner as it reaches the end,

76 / Casting Design and Performance

Schematic illustrating ﬂuid ﬂow around right-angle and curved bends in a gating system. (a) Turbulence resulting from a sharp corner. (b) Metal damage resulting from a sharp corner. (c) Streamlined corner that minimizes turbulence and metal damage

Fig. 7

Effect of pressure head and change in gate design on the velocity of metal ﬂow. A, 90 bend; B, r/d = 1; C, r/d = 6; D, multiple 90 bends. The variables r and d are the radius of curvature and the diameter of the runner, respectively. Source: Ref 2

Fig. 8

causing higher pressure and resulting in higher ﬂow through the gate. By stepping down the runner after the ﬁrst ingate, metal velocities and pressures at the two ingates can be equalized. This effect can be achieved by gradually tapering the runner to a smaller cross section along its length, but patternmaking limitations usually make it simpler to incorporate actual steps in the runner.

Design Considerations In applying ﬂuid ﬂow principles to the design of a speciﬁc gating system, several design decisions must be made before the actual sizes of the various components can be calculated.

Runner and Ingate. Figures 1 and 2 show gating systems with ingates coming off the top of the runner and then into the casting. This arrangement of cope ingates and drag runners is common and has the advantages that the runner will be full before metal enters the ingates. This establishes the full-ﬂow conditions discussed earlier in this article. A full runner will reduce turbulence and will help to allow any low-density inclusions in the ﬂowing stream to ﬂoat out and attach themselves to the mold wall. A system of cope runners and drag ingates (or ingates coming off the base of a cope runner) is also common and has strong proponents (Ref 3). The basis of this design is that momentum effects will carry the ﬁrst metal past the ingates, and if the runner can be quickly ﬁlled

(at least above the level of the ingates), clean metal will ﬂow from the bottom of the runner, while inclusions carried along in the metal stream will ﬂoat above the ingates. A common element of this system is that the total cross-sectional area of the ingates should be smaller than the cross-sectional area of the runner. Such a pressurized system is intended to force the metal to back up at the ingates and rapidly ﬁll the runner, although complete ﬁlling of a cope runner will often depend on at least partial ﬁlling of the casting. During this period of incomplete ﬁlling, turbulence and the potential for air entrainment and dross generation are increased. Techniques for prepriming and countergravity ﬂow are discussed in the article “Filling and Feeding System Concepts” in Casting, Volume 15 of the ASM Handbook. Pressurized versus Unpressurized Systems. The difference between these two systems is in the choice of the location of the ﬂow-controlling constriction, or choke, that will determine the ultimate ﬂow rate for the gating system. This decision involves the determination of a desired gating ratio, that is, the relative cross-sectional areas of the sprue, runner, and gates. This ratio, numerically expressed in the order sprue:runner:gate, deﬁnes whether a gating system is increasing in area (unpressurized) or constricting (pressurized). Common unpressurized gating ratios are 1:2:2, 1:2:4, and 1:4:4. A typical pressurized gating ratio is 4:8:3. Both types of systems are widely used. The unpressurized system has the advantage of reducing metal velocity in the gating system as it approaches and enters the casting. Lower velocities help encourage laminar (or less turbulent) ﬂow, so unpressurized systems are recommended for alloys that are highly sensitive to oxide and dross formation. Pressurized systems generally have the advantage of reduced size and weight for a given casting, thus increasing mold yield. The single greatest disadvantage of pressurized systems is that, by design, stream velocities are highest at the gates just as the metal enters the casting. This increases the potential for mold or core erosion and places a premium on proper location of ingates to minimize such damage. Vertical versus Horizontal Gating Systems. This decision may be imposed by the orientation of the mold parting line. Figure 1 shows conventional, horizontally parted molds with the gating system most conveniently arranged along the parting line. Vertically parted molds impose a vertical placement of gating components, but the design considerations for these components are often the same as those for horizontal systems, for example, the desirability of tapered sprues and runner extensions. In addition, the need for stepped-down sprues to equalize ﬂow through multiple ingates is, if anything, more critical than in a horizontal system. This element of a properly designed vertical gating system is illustrated in Fig. 10.

Gating Design / 77

Properly designed bottom gate that ensures smooth ﬁlling of the casting while producing minimal turbulence

Fig. 11

Applying Bernoulli’s theorem to ﬂow from a runner at two ingates for a ﬁlled system and comparing velocity and pressure at the ingates for two runner conﬁgurations. (a) Same runner cross section at both ingates. (b) Stepped runner providing two different runner cross sections at each ingate. Source: Ref 1

Fig. 9

system necessary to deliver the metal with the minimum ﬂow rate required, which in turn determines the cross-sectional area of the choke. Once that is calculated, the rest of the components are easily calculated, moving downstream from the choke in an unpressurized system or upstream from the choke (that is, the gates) in a pressurized system.

Ceramic Filters in Gating Design Ceramic ﬁlters are extensively used in the foundry industry to improve casting cleanliness and to reduce the cost of casting manufacture. Incorporated into the gating system, ceramic ﬁlters remove slag, dross, and other nonmetallic particles from the metal stream before the metal enters the mold cavity. Most casting alloys are subject to the presence of particles that can deleteriously affect the physical properties and appearance of the casting. These particles commonly include: Oxides formed during melting, metal trans-

fer, and pouring

Fig. 10

Comparison of ﬂow patterns in two vertical gating systems. (a) Poorly designed system. (b) Properly designed system using a tapered runner that equalizes ﬂow through the ingates

Refractory particles from the furnace and

ladle

Refractory particles present in the gating

One form of vertical gating common in both vertically and horizontally parted molds is called bottom gating and countergravity ﬁlling. Elements of a properly designed bottom gating system are shown in Fig. 11. This method has the particular advantage of introducing metal into the casting cavity at its lowest point, thus helping to ensure smooth ﬁlling of the casting with minimal turbulence. Numerous examples of calculations for gating system design are

available in the References cited at the end of this article and are not covered in detail here. Also see the article “Filling and Feeding System Concepts” in Casting, Volume 15 of the ASM Handbook. Optimum mold ﬁlling times are determined by such factors as metal type, casting weight, and typical casting section thickness. Once the optimum time is established, principles of ﬂuid ﬂow are used to determine the size of the

system or dislodged from the mold or cores during pouring Reaction products from metallurgical operations Undissolved metallic or nonmetallic particles made as additions to the molten metal for metallurgical modiﬁcations These particles, or inclusions, act as discontinuities in the metal matrix of a casting and can have a variety of adverse effects:

78 / Casting Design and Performance Large inclusions can reduce mechanical prop

erties such as tensile strength and elongation. Fatigue life can be reduced (Ref 4–6). Machining can become more difﬁcult, and the rate of tool wear can increase. Surface ﬁnish can deteriorate in appearance. Lack of pressure tightness can occur. Subsequent surface treatments such as anodizing or ceramic coating can be adversely affected.

The conventional approach taken to remove inclusions from the metal stream is through gating system design. Using this approach, gating systems are designed to encourage the separation of particles from the metal due to the density difference between the metal and the inclusion (Table 1). Both pressurized and unpressurized gating systems are used for this purpose. To be effective, the gating system must have sufﬁcient length to allow the lower-density particles sufﬁcient time to ﬂoat and adhere to the mold surface before they can enter the casting cavity. In practice, this approach does not always provide castings of adequate quality, and casting yield is often reduced, especially with unpressurized systems. Filter Advantages. Ceramic ﬁlters, when correctly applied, can be relied on to trap particles before they can enter the casting cavity. This feature can result in the following advantages:

Machine stock allowances can be reduced. Casting physical properties can be increased. Casting surface ﬁnish can be improved. Reliability of the casting process can be increased.

designed to incorporate a ceramic ﬁlter are shown in Fig. 13. By relying on a ﬁlter to remove particles from the metal stream, the gating system can be designed with the following features and advantages:

The use of a ceramic ﬁlter in a gating system of conventional design can be effective in reducing inclusion-related defects, but a gating system designed speciﬁcally to incorporate ceramic ﬁlters will be more effective. Typical gating systems for horizontally parted molds

The system can be unpressurized, resulting

Two crankshafts produced using a ceramic foam ﬁlter positioned vertically in the drag just downstream of the sprue. Casting yield is 91%.

Fig. 12

in reduced metal velocity as metal enters the casting cavity. Runners can be reduced in size—both length and cross-sectional area—thus increasing casting yield. Runners are in the drag, resulting in more rapid, complete ﬁlling of the runner and reduced opportunity for metal oxidation. Filter Types. Ceramic ﬁlters are available in a wide variety of materials and in many different forms. Mullite, alumina, cordierite, silica, zirconia, and silicon carbide are all commonly used. The most common forms are open-weave cloth, reticulated foam, extruded cellular shapes, pressed shapes, and perforated sheets (Fig. 14). Ceramic foams and cellular ceramics have been shown to be the most effective for inclusion removal and are the most widely used. Strainer cores were widely used prior to the advent of ceramic ﬁlters, and it is important to note the distinction between these two devices and their applications.

Inclusion-related scrap and inclusion-related

rectiﬁcation can be reduced.

Gating system size can be reduced and cast-

ing yield increased (Fig. 12).

Casting machinability can be improved. Machine tool life can be increased.

Table 1 Densities of metals and metal oxides commonly used in casting alloys Casting alloy

Density, g/cm3

Aluminum alloys Al Al2O3 3Al2O3 2SiO2

2.41 3.96 3.15

Magnesium alloys Mg MgO

1.57 3.58

Copper alloys Cu CuO ZnO SnO BeO

8.00 6.00 5.61 6.45 3.01

Iron alloys Cast iron Low-carbon steel 2% C steel FeO Fe2O3 Fe3O4 FeSiO4 MnO Cr2O3 SiO2

6.97 7.81 6.93 5.70 5.24 5.18 4.34 5.45 5.21 2.65

Fig. 13

Gating system designs for optimizing the effectiveness of ceramic ﬁlters in horizontally parted molds having sprue:ﬁlter:runner:ingate cross-sectional area ratios of (a) 1:3–6:1.1:1.2 and (b) 1.2:3–6:1.0:1.1

Gating Design / 79

Several common ﬁltration and ﬂow modiﬁcation devices (from left to right): strainer core, extruded ceramic ﬁlter, ceramic foam ﬁlter, mica screen, and woven fabric screen. The two types of ceramic ﬁlters are by far the most widely used.

Fig. 14

Cross section through a runner bar containing a ceramic foam ﬁlter. Particle capture occurs at the front face (top) and inside the ﬁlter. The alloy is aluminum-base LM25.

Fig. 15

Filter type must be correct for the application. Gating system design must provide mini-

mum metal turbulence downstream from the ﬁlter and in the casting cavity. Gating system size must be kept to a minimum.

Common methods of ﬁlter placement in horizontally parted molds. (a) Parallel to parting line. (b) Between 0 and 90 to parting line. (c) 90 to parting line. Arrows indicate the direction of metal ﬂow.

Fig. 16

Strainer (choke) cores are designed to function as a restriction or choke in the gating system. The open frontal area—the ratio of the area available for metal passage to total part cross-sectional area—is in the range of 20 to 45%. Strainer cores are designed into a gating system to control the rate of mold ﬁlling and can aid rapid ﬁlling of the gating system. This action can assist in the ﬂotation and separation of large particles from the molten metal stream, but strainer cores are not intended to remove particles. Their large, widely spaced holes split the metal ﬂow into several separate metal streams, often resulting in aspiration and subsequent oxide formation downstream. Ceramic ﬁlters, on the other hand, are designed to remove inclusions from molten metal. Filtering occurs by two mechanisms: physical screening (or sieving) and chemical

Common methods of ﬁlter placement in vertically parted molds. (a) Filter located in pouring basin. (b) Filter located inside the mold. Arrows indicate the direction of metal ﬂow.

Fig. 17

attraction (Fig. 15). When properly designed into a gating system, ﬁlters do not act as a signiﬁcant restriction to metal ﬂow. The open frontal area of most ceramic ﬁlters is in the range of 60 to 85%, and the ﬂow downstream of the ﬁlter is much less turbulent than with strainer cores. Use of Ceramic Filters. Design of the optimum gating system for ceramic ﬁlter use incorporates the following principles: Filter placement must be easily accomplished. Mold ﬁlling time must be constant and not

affected by the presence of the ﬁlter.

Filter Placement. The location and position of a ceramic ﬁlter are inﬂuenced by the molding method, pattern layout, and any metallurgical processes performed inside the mold cavity, such as the nodularizing and inoculating additions made in certain iron castings. The molding method is particularly important in determining ﬁlter position. In molding processes that use an expendable pattern, ﬁlters are most easily incorporated into the pouring basin. In horizontally parted molds, ﬁlters are commonly positioned as shown in Fig. 16. Filters should not be placed at the base of the sprue, because this increases the possibility of ﬁlter breakage and reduces ﬁlter effectiveness. In vertically parted molds, ﬁlters are commonly positioned as shown in Fig. 17. Although the pouring basin location is often used, ﬁlters are more effective when located farther down the gating system. When metallurgical additions are made in the mold cavity, ﬁlters must be located downstream, as shown in Fig. 18. Filter Area/Choke Area Ratio. The size and number of ceramic ﬁlters required are determined by the rate of mold ﬁlling required and the volume of metal to be ﬁltered. As ﬁltration occurs, individual cells in a ﬁlter become

80 / Casting Design and Performance

Plot of ﬂow rate versus time showing the behavior of a ceramic ﬁlter during the course of pouring until complete blockage occurs

Fig. 19

Fig. 18

Filter placement when a metallurgical operation (that is, magnesium treatment or inoculation of iron) occurs in the mold

blocked, and the rate at which the ﬁlter can pass metal is reduced. This phenomenon is illustrated in Fig. 19. Filter size is determined such that the ﬁlter operates in the range termed normal ﬂow in Fig. 19. In general, the necessary ratio of ﬁlter area to choke area is in the range of 2:1 to 4:1 at the start of pouring. If the ﬁlter is greatly undersized, complete ﬁlter blockage can occur, and incomplete mold ﬁlling will result.

REFERENCES 1. J.F. Wallace and E.B. Evans, Principles of Gating, Foundry, Vol 87, Oct 1959 2. J.G. Sylvia, Cast Metals Technology, Addison-Wesley, 1972 3. S.I. Karsay, Ductile Iron III: Gating and Risering, QIT—Fer et Titane, Inc., 1981 4. P.R. Khan, W.M. Su, H.S. Kim, J.W. Kang, and J.F. Wallace, Flow of Ductile Iron Through Ceramic Filters and the Effects on the Dross and Fatigue Properties, Trans. AFS, Vol 95, 1987, p 106–116

5. W. Simmons, The Filtering of Molten Metal to Improve Productivity, Yield, Quality and Properties, Proceedings of the 52nd International Foundry Congress, 1985 6. W. Simmons and H.A. Bowes, Efﬁcient Filtration Improves Casting Quality, Foundry Pract., Vol 209, 1984 SELECTED REFERENCES R.A. Flinn, Fundamentals of Metal Casting,

Addison-Wesley, 1963

L.F. Porter and P.C. Rosenthal, Fluidity Test-

ing of Gray Cast Irons, Trans. AFS, Vol 60, 1952 J.M. Svoboda, Basic Principles of Gating and Risering, American Foundrymen’s Society, 1973 H.F. Taylor, M.C. Flemings, and J. Wulff, Foundry Engineering, John Wiley & Sons, 1959

Casting Design and Performance Pages 81–87

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design for Economical Sand Molding BASIC PRINCIPLES of sand molding are shown in Fig. 1. The cylindrical shape of this casting provides natural clearance or draft along the length of the casting, as seen in Fig. 1. The tapered ends of the pattern permit it to be removed from the sand without restriction. The sand core is positioned in the core print cavities provided in the molds by the ends of the pattern (the core prints). For processes other than sand casting, these molding principles are equally valid. In die casting and investment casting, metal dies and cores are used. In permanent mold casting, metal molds are used with either expendable or metal cores; with expendable cores the process is referred to as semi-permanent molding. When four external ribs are added to the tubular shape (Fig. 2), molding practice remains unchanged. However, when six ribs are added (Fig. 3), either cores or loose pieces must be used to avoid locking the pattern in the mold. This principle is applicable to all shapes, regardless of their complexity.

General Design Factors Parting Lines. The location of a parting line may be dictated either by the shape of the casting or by special requirements such as those illustrated in Fig. 4. Parting the casting as shown in Fig. 4(a) to (e) makes it possible to cast faces that are ﬂat and parallel. The parting lines shown in Fig. 4(f) and (g) necessitate that the faces be tapered in order to provide draft. Other special requirements are illustrated in Fig. 4(a) to (f). For example, the method of parting in Fig. 4(a) provides normal draft in the hole and on the sides of the casting; the hole and the sides are concentric. Any ﬂash formed at the parting line can be easily removed. Figure 4(b) provides parallel sides and normal draft in the hole and on the sides of the casting. It is more difﬁcult, however, to maintain concentricity between the hole and the sides, because of the possible misalignment of the mold halves after the pattern has been withdrawn.

In Fig. 4(c), hole taper is reduced by 50%. Only a minimum of material need be machined from the hole to provide perfect straightness, a possible advantage when difﬁcult-to-machine materials are involved. As in Fig. 4(b), however misalignment of the mold halves after the pattern has been withdrawn is a potential disadvantage. Figure 4(d) is similar to 4(c), except that taper remains the same in the hole and is reduced by 50% on the outer walls. In Fig. 4(e), taper is reduced by 50% both in the hole and on the outer walls. Metal requirements in casting and machining are minimized to a greater degree than in any of the other designs. Figure 4(f) requires taper on two sides, and 4(g) provides sides that are parallel. These seven examples illustrate the adaptability of parting line location to casting design requirements. For the bell-shaped casting of Fig. 5 locating the parting line at the base of the bell, as in Fig. 5(a), would eliminate parting line reﬂection from the body of the casting. However, because the core cannot be vented at the top, trapped gases might cause defects in the casting

Fig. 1

A simple casting that illustrates the principle of molding a pattern and withdrawing it from a sand mold

Fig. 2

Adding four ribs to a tubular casting introduces no problems in removal of pattern from mold. Molding practice is the same as for the cylindrical shape of Fig. 1.

82 / Casting Design and Performance

Fig. 3

Adding six ribs to a tubular casting necessitates core prints on the pattern, to permit withdrawal of pattern from mold. Cores are necessary to complete the mold.

Fig. 4

Seven different parting lines are possible in producing this simple casting. Each parting line has a different inﬂuence on the casting. (See text for discussion.)

Parting the casting as in (a) eliminates any parting line reﬂection from the sand mold on the casting, and circularity would be good. A riser at the isolated boss could eliminate porosity if it were a problem. Gas trapped by the metal could cause defects. Parting the casting as shown in (b) would eliminate any gas problems but would adversely inﬂuence stability of diametral dimensions. The parting line would appear as a seam along the length of the casting.

Fig. 5

metal. If this shape were to be cast in some metals, a second riser might be required at the top of the casting, in order to assure sound metal. Placing the casting on its side so that the parting line is at right angles to the base, as in Fig. 5(b), would permit adequate venting of the core, provide an improved means of gating, and eliminate the need for a second riser; a parting line seam is unavoidable, but it can be easily removed. Figure 6 provides an example of how the shape of a cross section of a casting can dictate location of the parting line. Figure 6(a) shows a

Fig. 6

Redesign to eliminate the need for a large core and to permit relocation of the parting line

casting as originally designed, with a cross section having an I-shape. This shape can be produced only by the use of a core, and the parting line must be located as shown in Fig. 6(a). When the casting is redesigned to have an X-shaped (or a T-shaped) cross section, the need for a core is eliminated and the parting line can be relocated, as shown in Fig. 6(b). This revision appreciably reduces production cost. A stepped parting line results in the formation of ﬁns that may be difﬁcult to remove. It should be avoided in casting design whenever possible. Figure 7 shows a bearing housing in the original design of which, Fig. 7(a), a

stepped parting line was used to allow for simple coring of the hole. To provide a straight parting line, this casting could be produced without the cored hole, Fig. 7(b). The hole could be drilled later, increasing the cost of machining but saving the cost of coring. The relative economy of coring and not coring can be determined only by analysis of the total cost of ﬁnished parts produced by each method. Several examples are given in the next chapter. Location of Radii. The casting sections shown in Fig. 8 illustrate how a minor design concession serves to avoid possible mismatch and simpliﬁes removal of ﬁns at the parting line. The original design, Fig. 8(a), was modiﬁed to eliminate radii and thus enable the parting line to be located at the top surface of the casting, as shown in Fig. 8(b). A similar concession applied to coring is shown in Fig. 9. Here the possibility of core shift may be a problem, but it can be avoided by eliminating the radius at the end of the core. If such a radius is required, it can be provided easily by machining. A more complicated design involving radii is shown in Fig. 10. Figure 10(a) represents the more difﬁcult approach to the problem of coring the ﬂanged perimeter. Here the core forms both the slot and the radii on the inner faces of the ﬂange. The major disadvantage of this design is increased core cost. Figure 10(b) shows how the same ﬂange can be designed without radii, at a lower cost. A simple modiﬁcation, providing one radius only, is shown in Fig. 10(c). Bosses and Undercuts. Frequently, it is necessary to locate a boss some distance from the parting line. The section shown in Fig. 11(a) illustrates the positioning of a boss well below a ﬂange whose upper surface serves as a parting

Design for Economical Sand Molding / 83

Fig. 7

Eliminating the cored hole in this sand casting permitted a simple ﬂat parting plane. Previously, a stepped parting plane was required, to permit removal of the core print from the mold.

A radius where the ﬂat face joins the edges of the casting would require parting line as shown in (a). Seams or mismatch may result. By eliminating the radius, the parting line can be located as in (b).

Fig. 8

line. In this design, a core is required, to permit removal of the pattern from the mold. In producing the casting as shown, accurate positioning of the core is difﬁcult, and any shifting of the core results in surface irregularities. A less complicated design, shown in Fig. 11(b), extends the boss to the ﬂange, eliminating the undercut and the need for a core. A design in which the boss has been extended downward to the parting line to satisfy molding requirements is shown in Fig. 11 (c) and (d). The boss can be extended beyond the parting line and provided with a radius to improve appearance, as shown in Fig. 11(d). (Another method would be to form the boss by means of a “loose piece” on the pattern, the loose piece being drawn sideways from the mold after the pattern has been withdrawn vertically.) Figures 12 and 13 deal with the elimination of undercuts in achieving simpler and less expensive designs. As originally designed, the undercuts shown in Fig. 12(a) were needed to provide clearance for other components in the assembly. An improved design without undercuts is illustrated in Fig. 12(b); however, metal is added and the weight of the casting is increased. The design of a wheel hub, Fig. 13(a), required a ring-shaped core to form the exterior of the hub, because of the undercut adjacent to the ﬂange. Eight ribs were needed to strengthen the part and assure its satisfactory performance. In the improved design of Fig. 13(b) the hub section is tapered, eliminating the undercut and the need for a ring core. The strength and rigidity of the casting are improved enough so that the supporting ribs are no longer required. Rib Locations. The housing shown in Fig. 14 can be molded in two different ways at an appreciable difference in cost, depending principally on a design simpliﬁcation involving

A radius at the junction of a cored hole and a sand casting face requires a shaped core, as in (a). Mismatch could result. Elimination of radius, as in (b), simpliﬁes the core and removes the possibility of mismatch as a result of core shift.

Fig. 9

the number and location of ribs. The original design, Fig. 14(a), speciﬁed three ribs (one of which is the tubular shape) projecting radially from the center of the housing and spaced 120 apart. With this arrangement, it was most economical to locate the parting line at the face of the small ﬂange, Fig. 14(b), in order to guarantee stability of the two internal cores. The method of coring is shown in Fig. 14(c). Core box details for the entire casting are presented in Fig. 14(d) to (i). A study of these details indicates the complexity of the coring arrangement. In a redesign, Fig. 14(j), the parting line was located at the centerline of the housing, and the boss was placed in the cope half of the mold at right angles to the parting line. Four ribs (of which one is the tubular shape) spaced 90 apart replaced the three-rib arrangement of the original design. The twelve bolting bosses were revised to eliminate an undercut and permit straight withdrawal from the mold. Core box details for the redesigned casting are shown in Fig. 14(k) and (l). The core box for core Z, Fig. 14(j), is not shown but is similar to core box X, Fig. 14(k). The pattern and rigging for the revised casting are shown in Fig. 14(m). In the original design, the heavy sections of the casting were not accessible for risering. In the new design, risers were easily located where they were needed. Also, the revised design permits the use of an alternate sprue.

Molding and Coring Details The sand casting shown in Fig. 15 was designed solely to demonstrate alternative design solutions to molding and coring problems. Although there are other ways to produce this casting, the methods described here are intended to indicate the important inﬂuence of design on the cost of a casting and to

(a) A radius on the inside edges of the two ﬂanges would increase the cost of the core, because of the requirement for a loose piece in the core box. (b) Eliminating the radius simpliﬁes the core. (c) One radius can be incorporated without complicating the core of this hypothetical casting. (See Chapter “Designing for Economical Coring” for an explanation of the term “drier”.)

Fig. 10

emphasize the relation of design to molding and coring practice. The pattern for this casting, as originally designed, is shown in Fig. 15(a). The pattern is made in ﬁve sections and requires a special four-part ﬂask. Although the end sections of the ﬂask (for mold sections 1 and 4) may be made to any height that provides enough sand to retain the metal within the mold cavity, the height of the center sections must correspond to the height of the sections of the pattern involved. The molding sequence is: 1. Pattern sections 2 and 3, Fig. 15(a), are assembled, ﬂask section 2 is set in place, and mold section 2 is produced. 2. Pattern sections 4 and 5 are placed in position on pattern section 3. Flask section 3 is set in place, and mold section 3 is produced. 3. An end section of the ﬂask is set in place on ﬂask section 3, and mold section 4 is produced. 4. The three sections of the mold that have been produced are turned over as a unit, exposing the mold surface at parting line A and pattern section 2. 5. Pattern section 1 is set in place on pattern section 2, and the remaining end ﬂask section is set in position. Mold section 1 is produced. At this point, the four ﬂask sections are in place. The mold is complete, but the patterns have not been removed. The steps that follow concern removal of the pattern.

84 / Casting Design and Performance

Fig. 11

An undercut created by an isolated boss on the side of a sand casting requires either a core as in (a) or continuation to a ﬂange as in (b). Absence of a ﬂange as in (c) requires continuation of the boss to the parting line as in (d).

Fig. 12

(a) Undercuts required cores as shown, to permit withdrawal of the pattern from the sand mold. (b) Redesign eliminated the undercuts and the need for cores.

Fig. 13

(a) A reduced diameter adjacent to the ﬂange of this sand cast wheel hub necessitated a core. In addition, eight ribs were required, to provide the desired strength. (b) Redesign as shown eliminated the ring core and the ribs, and a stronger casting was produced.

6. Mold section 1 is removed and turned over to expose pattern section 1. Pattern section 1 is then removed from the mold. 7. Pattern section 2 is removed from mold section 2. Mold section 2 is removed and turned over. Pattern section 3 is withdrawn from the mold. 8. Pattern section 4 is withdrawn from the mold. Mold section 3 is then removed and turned over. Pattern section 5 is withdrawn from the mold. In the foregoing description of the molding sequence, details of gating and risering have been omitted, in the interests of simplicity. Because two of the outside diameters of the casting are smaller than three of the inside diameters, a one-piece core cannot be placed in the mold. A three-piece core, Fig. 15(b), is required. The preparation of each section of the threepiece core entails as much work as would the making a single large core for the entire casting (if such were possible). To understand the details of preparing a core section, the reader

may begin by considering the cross section of the core box used for core 2, which is shown in Fig. 15(c). To produce the required core shape, both a loose ring and a loose center plug are needed. The sand used to ﬁll the core box must be introduced in the limited space separating the ring from the plug. Thus, packing the sand in the core is difﬁcult and timeconsuming. After packing the sand and removing the ring and plug, it is not possible to set the core down on its upper face without sagging of the core. As shown in Fig. 15(c), this upper face or contact area consists of a relatively narrow ring beyond which a portion of the core extends. If only a few castings were to be produced, it would be feasible to prevent core sag by ﬁlling the cavities formed by the ring and plug with bedding sand. For production quantities, however, a core drier, Fig. 15(d), is required. The core drier provides for greater stability when the core is removed from the box. An individual drier is required for each core placed in the baking oven.

An alternate method for making core 2 is shown in Fig. 15(e). Here, the core is produced in two halves, each having a large ﬂat face that makes the use of core driers unnecessary. The large core box opening also simpliﬁes packing of the sand. After baking, the two halves are pasted together to form the complete core. The possible disadvantages of this method are the chance of misaligning the halves in assembly and the effect of parting on the accuracy of dimensions measured across the parting line. Cores 1 and 3 are prepared in a manner similar to that used for core 2. Core 1 is placed in mold section 1, and the remaining cores and mold sections are assembled alternately as shown in Fig. 15(b). Although sectional cores and molds provide a high degree of ﬂexibility in forming complicated shapes, they have certain inherent disadvantages. Among these is the possibility of movement, which would cause dimensional variation in the casting. Also, at every junction of the loose pieces in the pattern or core box, and at joints between mold and core sections, there is a possibility

Design for Economical Sand Molding / 85

Housing producible (with permissible modiﬁcation) by either of two molding methods with different parting line locations. For the casting as originally designed (a), parting line location is shown in (b), and method of coring in (c); core box details are shown in (d) to (i). Redesign of casting is shown in (j); note relocation of parting line. Core box details for redesign are shown in (k) and (l); pattern and rigging, in (m).

Fig. 14

86 / Casting Design and Performance

A hypothetical casting illustrating core and mold problems if placement of cores is not considered in the design of a casting. The required patterns and the necessary mold sections are shown in (a). The core and related mold sections are shown in (b). Views (c) and (d) show core box and drier to produce core 2 in one piece. View (e) shows core 2 produced in two pieces. Revision as in (f) would simplify mold and cores. Revision as in (g) would simplify still more the core necessary to produce the outside shape. (See text for discussion.)

Fig. 15

of forming seams or ﬁns, which would have to be removed at extra cost. The ﬁrst of two revisions of the original design of this casting is shown in Fig. 15(f). This redesign provides a maximum core diameter smaller than the minimum mold diameter through which the core must pass. The contour of the outer midsection of the casting requires the use of loose pieces in the core box forming core 1. This design eliminates sectional cores and permits the use of a conventional copeand-drag mold. A further design improvement, Fig. 15(g), simpliﬁes the core box producing core 1, and results in a cleaner casting with less dimensional variation.

Conventional molding of a circular sand casting is shown in (a). A speciﬁcation requiring the large face to be smooth necessitated placing this face in the drag half of the mold, as shown in (b), to obtain maximum smoothness of surface.

Fig. 16

Design for Economical Sand Molding / 87 Draft refers to the amount of taper given to the sides of a pattern to enable it to be withdrawn easily from the mold. For this reason, draft requirements are intimately related to casting design. A simple example demonstrating the inﬂuence of a casting requirement on the direction of draft is given in Fig. 16.

Conventionally, this design would be molded as shown in Fig. 16(a). However, when the largest plane surface of the casting is required to be straight and smooth, as cast, the casting is molded as shown in Fig. 16(b), to avoid risering into the plane surface and the possibility of inclusions ﬂoating to this surface during

pouring. This reverses the draft direction on the sides of the casting (or moves the casting into the cope). In this example, the importance of draft direction increases with an increase in the over-all dimensions of the casting. To some extent, this problem parallels that of parting line location discussed with Fig. 4.

Casting Design and Performance Pages 89–99

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design for Economical Coring CORES are separate shapes, of sand, metal or plaster, that are placed in the mold to provide castings with contours, cavities and passages not otherwise practical or physically obtainable by the mold. In general, the principles that apply to molding a pattern apply also to molding a sand core. The core must be freed from the core box by moving the box away from the core in a direction generally perpendicular to the face of the core box. Loose pieces that have been placed in the core box to assist in producing the shape of the core will remain with the core sand when the box is removed from the core, and must be removed from the core by a movement in the appropriate direction.

General Principles General principles of coremaking are illustrated in Fig. 1 by four basic examples that depict, in simpliﬁed form, most of the problems entailed in molding a core. At the top of each illustration, the core box is shown in section after ram-up; in the lower part, the core box has been inverted onto a plate upon which the core rests preparatory to being baked. Figure 1(a) shows a core that is semicylindrical — one of the most common of core shapes. No problem is presented in removing this core

Fig. 1

Basic principles of core molding.

from the core box; natural draft is provided by the shape of the core, and no obstructions exist. Figure 1(b) shows two bosses formed by the core box. Boss A is simple to produce and adds nothing to the cost of producing the core, although it does add a small amount to the cost of the core box. Boss B, however, creates an undercut that requires a loose piece, as shown, to permit the core to be freed without destroying the desired shape. This adds to the cost of the core box. It also adds to the cost of producing the core, and may increase the dimensional variations in the location of the boss. Figure 1(c) illustrates a core box into which two more conditions are incorporated that require loose pieces. Loose piece A permits the side of the core to be produced without draft. Bosses or ribs could be appended to this face without adding to molding problems, provided they could be drafted to permit easy withdrawal of the loose piece. Loose piece B forms a cavity in the core that assists in producing a rib on the casting. The overhang of the core box shape would lock the core in the box unless a loose piece were provided as shown. Figure 1(d) illustrates the additional problem that arises when the loose piece at the face of the core box projects excessively into the core. When the core box is inverted and the loose piece removed, excessive overhang of the core could cause the sand to sag and lose its

dimensional accuracy. To prevent this, support must be provided for the core. In the core box shown, the loose piece is removed and replaced with bedding sand before the core box is inverted. (Bedding sand is a special sand that does not adhere readily to the core and so can be removed cleanly after the core has been baked.) Core Driers. In the core box shown in Fig. 1 (d), a core drier could be the loose piece. (A core drier is a contoured metal form or “cradle” that remains with and supports an irregularly shaped core while it is being baked.) However, since a drier is required for each core being baked, the expense of producing all the necessary driers adds substantially to core cost. Although supporting an overhanging or irregular section of a core with metal core driers provides greater accuracy and production speed, the use of bedding sand is usually more economical. The tolerance speciﬁed will inﬂuence the choice. Advantages. Although cores increase cost and tolerance requirements, they enable the foundryman to cast intricate internal shapes not producible by any other process. Examples are curved passages of various kinds and large interior cavities connected to the outside surface by relatively small openings. The interior passages need not have uniform shape or cross section. The internal cavities can be of almost

90 / Casting Design and Performance any shape dictated by the requirements of design. Cores also allow the casting designer to eliminate superﬂuous metal, thus reducing the weight of a casting and the cost of subsequent machining operations. Cores may be used to produce draft-free surfaces, to simplify parting lines, or to eliminate the use of loose pieces in the pattern. Cores may also provide an additional beneﬁt, in that patterns that would be too fragile to withstand normal handling will, if made with cores, be braced by the core prints, thus permitting shapes and cross sections not otherwise practicable. The part sketched in Fig. 2 was easily producible as a magnesium alloy sand casting. The twist at its center section, which was readily achievable by the use of cores, would pose a problem in the production of this part by any process other than casting. The intricately cored shell mold casting shown in Fig. 3, a manual control valve body

A skewed wall readily producible as a casting but not economical by any other manufacturing process

Fig. 2

An intricately cored value body cast in a shell mold. This shape was economically producible only by the casting process. Cored passages between individual chambers are not visible in the above views.

Fig. 3

Fig. 4

for an automatic transmission, is another part that would be economically impractical to produce by any process except casting. Core Size Limitations. Figure 4 presents recommended relations between core length and core diameter. These maximums were established by one foundry and, although other foundries may be more or less restrictive, they can serve as guides in the design of cored holes. Recommended maximum lengths for horizontal sand cores supported at both ends are given in Fig. 4(a), for steel, aluminum, and magnesium castings. Such cores often bend upward because of the buoyant force of the molten metal. Steel causes more bending than does aluminum because, being heavier, it exerts a greater upward pressure on the core. Also, the higher temperature at which steel is cast is more likely to break down the binder that holds the core sand together. One foundry producing a casting with a horizontally supported core reported an upward bending of 1/32 in. in a tubular casting with a 1½-in. OD, 1-in. ID, and an over-all length of 10 in. Occasionally, one of these cores bent 1/16 in. Cantilever-supported cores, Fig. 4(b), are even more likely to bend, because they are supported at only one end and are less able to resist the buoyant force of the molten metal. A larger tolerance is needed on dimensions at the unsupported end of the core, because of the necessity for a small amount of side clearance between the core and the mold at the opposite end. This clearance permits a displacement of the core when the molten metal enters the mold. The displacement is ampliﬁed as the core extends into the casting, and has a pronounced inﬂuence on dimensional discrepancies. A blind hole whose length is approximately equal to its diameter usually offers no cleaning problems, because excessive heat concentration is unlikely in the core. As the length of the blind hole increases in relation to its diameter, the possibility of burnt-in sand increases, making cleaning progressively more difﬁcult. It is for these reasons that blind cored holes should be kept to minimum length; as indicated

in Fig. 4, the maximum length recommended is less than for cores supported at both ends. For cantilever-supported cores, the absolute minimum that must be allowed for coreprint lengths is 1/3 L, as shown in Fig. 4(b). This minimum length is suggested for use only in dry sand molds and in shell molds, where the ﬁt is accurate enough to hold the core ﬁrmly in place. In green sand molds, enough mass must be provided in the core print section to balance the core so that it stays in place while the mold is being closed. The designer must keep in mind the requirements for core print lengths and mass when it is necessary to locate a part of the casting close to the opening of a cored hole. Vertical cylindrical cores supported at both ends are not distorted by the buoyant action of the metal. As long as they are free to expand, they remain straighter than horizontal cores. With such vertical cores, however, it may be difﬁcult to close the mold without either disturbing the core location or shaving sand from the mold. This sand may fall down into the mold and appear as a casting defect. Figure 4(c) shows maximum recommended lengths. Although vertical cores supported only at the bottom are not subject to distortion by the buoyant effect of the casting metal, they can ﬂoat out of position if they are not anchored ﬁrmly. Furthermore, such cores vent less efﬁciently than horizontal cores or cores with a core print into the cope. Another method of anchoring such a core in place is illustrated here in Fig. 5(a). The printed section of the core extends past the cavity and is held ﬁrmly in place by the cope half of the mold. This method requires the use of more core sand than a ram-up core and is more expensive, but it is also more accurate. The insertion and removal of ram-up cores abrade the pattern and permit variation in core location from casting to casting. Because of the difﬁculties usually present in molding and coring vertical blind holes, the maximum length of the cores should be reduced to about half the values given in Fig. 4(c).

A comparison of maximum recommended core length versus core diameter for baked sand cores positioned in the mold (a) at the parting line and ﬁrmly anchored at both ends, (b) at the parting line but anchored at one end only, and (c) vertical to the parting line and anchored at both ends

Design for Economical Coring / 91 Vertical cores supported from the top, as in Fig. 5(b), are not subject to distortion from the buoyant effect of the molten metal, and they are prevented from ﬂoating by the core print, which is held in place by the cope half of the mold. Such cores are easily vented, because core gas ordinarily travels upward to escape. For vertical cores supported at the top only, recommended maximum lengths with reference to diameters are the same as those given in Fig. 4(b) for cantilever-supported cores. Green Sand Versus Dry Sand Cores. Some cavities can be cored by the use of “green sand cores” (cores formed from the molding sand and generally an integral part of the pattern and mold). Because these cores are formed when the mold is made around the pattern, they usually add no expense to the production of a casting, and save the cost of producing and placing in the mold equivalent dry sand cores. Although green sand cores are less expensive than dry sand cores, they are at a disadvantage in several respects. Because they are less rigid than dry sand cores, they are less resistant to movement or deformation under the static pressure of the molten casting metal. This is an important consideration when wall thicknesses must be held to close tolerances. Dry sand cores can provide smoother casting surfaces, and so would be preferable when surface ﬁnish is a consideration. Furthermore, green sand cores usually require more draft than would be normal for the same length of draw on an outside surface of the pattern. This is a precaution to

prevent the green sand core from breaking away from the rest of the mold when the pattern is withdrawn. Figure 6 shows a thrust bearing housing, sand cast of malleable iron, that was successfully redesigned to employ green sand cores to form the entire inside diameter, and thereby to lower production costs substantially. As originally designed, Fig. 6(a), this casting required a baked sand core and one of green sand. It had a deep circular lightener pocket on one face and a circular oil passage opening to its outside diameter. The lightener pocket was too deep to be made in green sand with normal processing, so a ram-up core was utilized. This core also formed a portion of the inside diameter of the casting, as shown. (A “ram-up” core is one which is placed in the pattern before the ﬂask is ﬁlled with sand, so that as the mold is rammed-up the core becomes an integral part of the mold.) A ring core was required also, to form the circular oil passage. Since the pocket was merely a lightener and the function of the oil passage was noncritical, no precise dimensions or tolerances were involved. A redesign, shown in Fig. 6(b), combined the oil passage and the lightener, was no heavier, and required only one baked sand core. The inside diameter of the casting is now completely formed using green sand cores, half in the drag and half in the cope. They meet at one of the steps and are produced as the pattern is molded.

(a) Core ﬂoat was prevented by a large coreprinted section clamped between the mold halves. Venting of the core would be difﬁcult. (b) By inverting the pattern, better venting of core gas was possible. A large core print was still necessary to position the core securely and to prevent shifting.

Fig. 5

Fig. 6

A redesign combined two cores into one, resulting in substantial cost savings.

Designing for the Use of Sand Cores Cored holes should be designed as simply as the intended function of the casting permits. An effort should be made to eliminate projecting and overhanging parts, pockets and recesses, and to use simple and smoothly contoured core shapes. The casting design should take into consideration the requirements for draft necessary for successful removal of the core from the core box. Although the incorporation of loose pieces permits cores with straight sides, overhangs and undercuts, these add to dimensional discrepancies and increase the cost of producing the core. Bosses should be extended to the parting line if this can be done without adding impermissibly to the weight of the part. When this is a questionable procedure, the undesirability of additional weight should be balanced against the desirability of savings in the casting cost through the design simpliﬁcation. Of course, if bosses can be located on the parting line, both cost and weight are held to a minimum. Core Fragility. It is important that the smallest dimensions of the core be large enough to facilitate production of the core. Thin or small-diameter cores are extremely fragile and are a cause of both core scrap and casting scrap. The armament stores pylon casting shown in Fig. 7, which was produced from 4330 steel, presented a problem because of core fragility. View (a) shows the as-cast conﬁguration of the part. View (b) is the machined casting cut in half. In this casting as originally designed, holes A and B, formed by cores 1 and 2, respectively, were all ½ in. in diameter. Additional support for core 1 was provided by the core print for the 1-in.-diameter core at its left end Core 2 was supported solely by the four (two on each side) ½-in.-diameter core prints. This design was unsatisfactory because, with normal foundry handing, the ½-in.-diameter core sections were too readily broken off the main bodies of the cores. In addition, it was difﬁcult to remove the core sand through the small holes in the casting. The casting was redesigned. Holes A were enlarged to be 3/4 in. in diameter. These four holes (two on each side of the casting), plus the 1-in.-diameter core at the left end, now provided adequate support for core 1. In core 2, holes B were enlarged to 5/8 in. in diameter. Hole C, 3/4 in. in diameter, was added for additional support. Increasing the size of holes A and B not only eliminated the problem of core breakage but, also increased the dimensional accuracy provided by the cores, through more positive placement and anchorage during the pouring and solidiﬁcation of the metal. Venting of the cores was also improved, and removal of core sand was made easier. Thin Core Sections. The separation of heavy masses of metal with thin core sections should

92 / Casting Design and Performance

Fig. 7

Small appended core sections were broken off the main core in handling. Increasing the size of the cored holes reduced breakage, improved core venting, and facilitated removal of core sand after casting this steel part.

Fig. 8

By eliminating a thin core between two heavy sections of this AZ63 magnesium alloy casting, a problem of heat retention was solved. The casting was slotted by a subsequent machining operation: (See next page for discussion.)

be avoided. If too much metal is adjacent to a core or surrounds it, the core will be unable to dissipate the heat from the molten metal. This will result in hot spots in the casting during cooling and solidiﬁcation. Such hot spots inevitably give rise to shrinkage or porosity, which can cause rejection of the castings. An additional problem is the inability of these thin cores to vent core gas adequately, which can result in blow-holes. Also, metal penetration may cause difﬁculty in core removal. The designer should attempt to forestall these problems by increasing the cored area or reducing the metal walls that surround it, or both. Often, the thickness of metal wall sections can be decreased, without sacriﬁcing strength, by the use of ribs. The presence of a thin core section between two heavier masses of metal presented a problem in production of the turbine door ﬁtting shown in Fig. 8. Made from magnesium alloy AZ63, this sand casting, as originally designed, had two isolated walls, each 0.38 in. thick, separated from each other by a core section 0.23 in. thick. When the casting was poured, this core section was too thin to dissipate the heat from the surrounding molten metal. A hot spot developed, which resulted in a heavy shrinkage defect on each inside surface of the bosses. Rejection rate was 90%.

This aluminum plaster mold casting is an example where thin cores became excessively hot when surrounded by molten metal, causing shrinkage porosity.

Fig. 9

The casting was redesigned to eliminate the core section between the bosses, so that the metal section was cast solid in this region. Although this now required a milling operation to remove the added metal, rejections dropped to 3%. Another casting with production problems caused by cored areas being-too thin with respect to surrounding metal masses is shown in Fig. 9. When this casting (an aluminum alloy impeller, cast in a plaster mold) was poured, the inability of the thin cores to dissipate heat adequately from the molten metal caused hot spots that produced shrinkage porosity.

Well-proportioned cored passages. Sand core was able to withstand normal foundry handling, provided stable positioning in the mold, and was easily vented.

Fig. 10

Figure 10. shows a sand cast ﬁtting with cored passages exemplifying good design for any process that uses expendable cores, because:

Design for Economical Coring / 93 1. The cores are large enough in cross section to enable them to support themselves without sagging and also to withstand normal foundry handling without breaking. 2. The cores have sufﬁcient mass, in relation to that of the surrounding metal, to dissipate heat rapidly enough to prevent hot spots that can cause shrinkage defects. 3. The cores can be easily and adequately vented, to eliminate any problems arising from generation of core gas. 4. The presence of a core print at each of the three areas where the core emerges from the casting assures that the core can be accurately positioned and that it can be ﬁrmly anchored to remain in place while the molten metal is ﬁlling the cavity. 5. The straight-through shapes of the cored sections permit easy removal of core sand after the casting has cooled. Thin-Wall Casting Sections Formed by Cores. The minimum thickness of walls formed entirely by cores is governed by the same considerations that control the minimum thickness of walls formed by the mold, as detailed in Chapter 3. Most important is that thin sections be given access to an adequate supply of molten metal, to prevent cold shuts or misruns. This molten metal can be supplied either from adjacent heavier sections or from risers. The thin-wall magnesium sand casting illustrated in Fig. 11 had a center web formed by two adjacent cores. The necessity for core prints to anchor the two cores made it difﬁcult to gate and riser this casting properly. Consequently, misruns and cold shuts, particularly in the thin center web, occurred frequently when castings were made in production. In producing the wave guide power divider shown in Fig. 12, an aluminum alloy plaster mold casting, the 0.040-in. vertical sections,

Six additional cored holes (three on each side) were incorporated to provide the core support necessary to prevent core distortion when molten metal ﬁlled the cavity. These added holes were later tapped and plugged.

formed by cores, were difﬁcult to ﬁll completely with casting metal. The metal cooled so quickly when poured into the mold at normal casting temperature that it lost the ﬂuidity necessary to ﬁll the thin sections. Increasing the temperature of the metal caused shrinkage of the heavy sections and excessive formation of gas. Whenever possible, extremes in section size, as in this example, should be avoided, preferably by increasing the thin sections.

Support for Sand Cores The fact that certain complicated shapes can be produced economically only by the casting process may require the design of cored sections difﬁcult to support. It is not always possible to provide as many or as large core prints as necessary to support the core properly. And, although a core may appear to be rigid, if supports for it are too far apart it can warp, sag or ﬂoat when surrounded by the molten casting metal. This is particularly true with a long core of relatively small cross section. Inadequate support was a problem with the core necessary to form the inside conﬁguration of the oil slinger shown in Fig. 13, sand cast in aluminum alloy. The original design had six cores, three on each side, spaced as shown. After casting, the 5/16-in. wall thickness varied (in some instances, to as little as 3/32 in.), and

rejection rate was high. It was necessary to add six more holes, three on each side, as shown, to provide adequate support for the core. These additional holes were later tapped and plugged, to permit the casting to function as originally intended. When cores are produced in halves, baked and pasted together, any distortion will prevent the faces from meeting ﬁrmly. The gap between the cores will cause ﬁns to be present in the cored passage unless special processing is given the cores. In the pilot’s steering wheel shown in Fig. 14 (a magnesium sand casting), a large ﬁn on the inside passage wall was a hazard to the electrical wires that ran through this passage to

A wave guide power divider casting difﬁcult to produce because 0.040-in. sections, formed by cores, froze prematurely. Increasing the temperature of the metal caused excessive gas to form in areas where the cores were in contact with the heavier sections. Increasing the thin sections would be a practical solution in this aluminum plaster mold casting.

Fig. 12 Fig. 11 (See text.)

Thin center section, formed by two cores, was not readily accessible to gating and risering.

Fig. 13

A heavy ﬁn occurred on the inside of this pilot’s wheel (a magnesium sand casting) as a result of a gap between the two core halves that formed the inner shape. Special processing of the core was necessary to eliminate this condition and prevent damage to electrical wires that were installed in this cored passage.

Fig. 14

94 / Casting Design and Performance the control switches at the tips of the wheels. The ﬁns are shown in the cutaway sections in the illustration. It was necessary to specify maximum allowable ﬁns, and to add appropriate inspection procedures. Chaplets. It is sometimes permissible to provide the necessary support for cores by the use of chaplets, which are metal spacers placed in the mold to support a core or to space it a ﬁxed distance from the mold wall. Chaplets (illustrated in Fig. 15) are usually made of the same metal as the casting in which they are to be used. In theory, the chaplet fuses with the casting, to form a continuous wall. In practice, this does not always happen, and poor fusion may create a weak spot in the wall that will leak if subjected to pressure. Fusion between aluminum or magnesium chaplets and castings is even more uncertain than in steel, because of the lower heat capacity of aluminum and magnesium and the fact that an oxide ﬁlm is usually present on the surface of aluminum and magnesium chaplets. A long, slender core supported at one end only, such as the core shown in Fig. 16(a), requires support to prevent it from sagging or ﬂoating and thus causing distortion. Chaplets would provide this support and would also help the molder to position the core properly. Two cored openings, in the same positions as the chaplets, would provide support for the core (from their core prints) and would also allow for better venting, Fig. 16(b). The cored holes would be tapped and plugged later. The enlarged end of the core print, Fig. 16(a), positions the core axially. The addition of the small core prints, Fig. 16(b), eliminates the need for this enlarged section.

During the past several years, however, considerable progress has been made in the development of techniques employing metal tubing for these cores. These give the designer greater freedom in the inclusion and arrangement of passages, and permit highly complex design. Three different techniques are currently available: 1. The older method of drilled passages plus complex, bulky external tubing. When feasible, this may still be the most economical method. 2. Lined passages produced by pouring aluminum or magnesium around tubing which remains in the casting. 3. The most recent method, which involves the introduction of unlined passages into sand and permanent mold castings. The unlined passages (3, above) require cores that can be dissolved or otherwise removed from the casting. These cores are usually made of steel, copper, glass or some siliceous refractory material. For practical purposes, the unsupported length of such cores depends not only on the diameter but also on the design of the core. Large-diameter cores support themselves over a longer span than do cores of smaller diameter, because increasing the diameter increases the rigidity of the core. Best design practice allows for at least a ¼-in. wall of cast metal around a passageway. In some castings a 3/16-in. wall may be adequate, but using a wall of less than 3 /16 in. invites trouble.

Cores in Permanent Mold Castings Coring of Tortuous Passages For reasons of function, many castings are required to have tortuous internal passages of small diameter. Sand cores for these passages are difﬁcult to produce. Ramming of the sand in the core box, in conformity with the outlines of the intended core, is often impossible to do properly. Because of the shape of the cores, they must be supported by core driers during baking. Also, such cores are easily broken in normal foundry handling. The designer of castings has been greatly restricted as to the passages that could be case economically, using sand cores.

Cores in the permanent mold process may be of gray iron, steel, sand or plaster. Metal cores may be either movable or stationary. Stationary cores must be perpendicular to the parting line to permit removal of the casting from the mold, and must be shaped so that the casting is readily

As originally designed, chaplets were required for core support. An improved design eliminated the need for chaplets by incorporating two holes which provided adequate support by the core itself and permitted better venting of the core gas. These two holes were later tapped and plugged.

Fig. 16

Chaplets may be used to position cores not readily supported by core-printed sections. As illustrated, chaplets prevent core sag before pouring and core ﬂoat after pouring.

Fig. 15

freed. Metal cores not perpendicular to the parting line must be movable, so they can be withdrawn from the casting before it is removed from the mold. When baked sand or plaster cores are used in a permanent mold, the process is referred to as semipermanent molding. With sand or plaster cores, which are expendable, more complex shapes can be produced than with metal cores. The two cored passageways shown in Fig. 17 illustrate the limitations of cores in permanent mold castings. The cored passageway shown in Fig. 17(a) could be produced with metal cores. Sand or plaster cores would be needed to provide the inside radius of the passageway shown in Fig. 17(b), which would also be more expensive to produce. Cored holes in permanent mold castings can usually be held to closer tolerances, of both size and location, than in sand castings. Both movable cores and stationary cores can be machined to close dimensions and accurately located. Although sand cores in permanent molds are no more accurate dimensionally than in sand molds, in the metal molds they may be positioned and retained in place with more accuracy than in sand molds. In general, the problems of expendable cores in permanent mold castings and in sand castings are similar. The problems with metal cores are similar to those in die castings.

Investment Castings In sand casting, wood or metal patterns are used to make the necessary impression in the molding material. Although the pattern may be re-used, the mold is expendable. In investment casting, a metal die is used to produce wax or plastic patterns. These patterns, in turn,

Coring limitations of permanent mold castings are illustrated by this simple casting. (a) The corner radius of the cored passage must be sacriﬁced if metal cores are to be used to achieve production at the most economical level. (b) Plaster or sand cores can provide the inside radius but casting cost will be higher than for (a).

Fig. 17

Design for Economical Coring / 95 are used to produce ceramic molds, either solid or shell. Both the patterns and the ceramic molds are expendable. The crux of the molding problem involves the withdrawal of the wax or plastic pattern from the metal die and the withdrawal of cores, when used, from the pattern (Fig. 18). In other respects, the principles applicable to sand molding apply to investment molding. The wax pattern for investment molding shown in Fig. 19 incorporates an inside radius at the junction of the two cores. Metal cores A and B are required to be withdrawn from the wax pattern in the directions indicated in Fig. 19(a). Such withdrawal cannot be achieved without damaging the wax pattern. In the redesigned casting, shown in Fig. 19(b), the inside radius has been eliminated in order to circumvent this problem. To retain the original radius would have required a costly change from metal cores to “soluble cores” made of a material capable of being etched out or dissolved out of the wax pattern without damaging it. Figure 20 illustrates a curved core passage. Again, a metal core cannot be used in the design shown in Fig. 20(a). A two-section core,

Fig. 18

Metal tooling for making the wax or plastic pattern for tubular investment custing shown at left. Note draft to permit withdrawal of the core from the pattern.

(a) In making the wax pattern for investment molding, an inside radius is impossible using two cores; core A cannot be withdrawn without damaging the pattern. (b) Sharp corner permits easy withdrawal of cores.

Fig. 19

Fig. 22

eliminating the inside radius, is shown in the design of Fig. 20(b). Where grooves are required for tool clearance in subsequent machining, the correct procedure for incorporating these grooves into the design of a casting is that shown in Fig. 21(b). The method of Fig. 21(a) does not permit withdrawal of a metal core. In the original design of a four-blade propeller, Fig. 22(a), the hub was designed with a pronounced taper. This necessitated the use of four loose pieces, located at the hub and equally spaced between blades. Further details of this design, which more fully explain the location of the loose pieces, are given in Fig. 22(b). A simpliﬁcation of the original parting line location, Fig. 22(c), is shown in Fig. 22(d). Note that the radii at the leading and trailing edges of the blade have been greatly modiﬁed in the revised design. When the hub taper was eliminated, use of a simple two-part mold was possible. The original design of a blade for a gas turbine called for notches in the cast airfoil section, as shown in Fig. 23(a). To provide such notches required the attachment of gates to a

(a) In making the wax pattern for investment molding, special coring procedure would be required, to produce radius Y. (b) By permitting the inside corner to be sharp, standard coring will sufﬁce.

Fig. 20

contoured surface. The gate contacts were removed by an expensive grinding operation. A less expensive approach is shown in Fig. 23(b). The notch was eliminated from one end of the blade, permitting attachment of an adequate gate at this end. The large gate encouraged directional solidiﬁcation and permitted feeding the casting as it froze. This gate was easily removed, and the required notch was subsequently machined in place. Figure 24 shows another example of simpliﬁed gate removal, in connection with the design of a two-blade propeller. Originally, Fig. 24(a), a hole was to be cored through the hub of the propeller. This required double gating to a contoured surface. By eliminating the cored hole and providing a simple pad at one end of the hub, the use of a single gate was made possible. Removal of this gate was inexpensive and there was no longer a risk of damaging the airfoil section. With some designs of this type, cast in some metals, gating on the hub as in Fig. 24(b) could be used with a core in position. Holes in Investment Castings. Holes of various shapes can usually be formed with little or no difﬁculty by the investment casting process. Limiting factors are the size of the opening, the ratio of the size of opening to the length of core, and the complexity of the shape as it affects metal ﬂuidity and wall thicknesses. Generally, in investment casting, through holes of less than 3/32-in. diameter and blind holes of less than 3/16-in. diameter require special coring techniques that add signiﬁcantly to the cost of the part. Blind holes that have a length-to-diameter ratio of more than 2:1 are more difﬁcult to cast, because of breaking or shifting of the unsupported end of the core. General rules for designing cored holes in investment castings are embodied in the

(a) Tool relief undercut is incorrectly located here, because it prevents withdrawal of a metal core. (b) Relocation of the relief notch as shown permits easy withdrawal of core.

Fig. 21

Tapered hub on this investment cast propeller (a) would require special core pulls to free the wax pattern from undercuts (b). A simpliﬁed parting line would be possible if full radii were not required at blade edges (c) and (d).

96 / Casting Design and Performance dimensional recommendations of Table 1. Because other factors of the design can inﬂuence these values, and because some foundry methods have greater capabilities than others, these ﬁgures are useful primarily as a general guide. Seven typical examples of cored holes and cavities easily incorporated in investment castings are illustrated in Fig. 25. None of these coring situations presented any difﬁculty. Small Cavities. The investment casting process permits coring of cavities that would be too small to be cored in sand castings. Typical of such cavities is the internal oil slot shown in Fig. 26. When the part shown was produced as a sand casting, this slot had to be milled in, and the hole leading to it drilled. Producing the part as an investment casting permitted the slot to be cored, and no subsequent machining on it was required. The change in method resulted in total savings of 41¢ per casting. Abutting Cores. In investment mold castings, as in permanent mold castings (Fig. 17),

(a) In the original design of this investment cast turbine blade, careful removal of the gates located on the contoured surface was required, to maintain the airfoil shape. (b) Simpliﬁed gating and gate removal were made possible by eliminating one notch and gating into a ﬂat surface. The notch was machined later.

Fig. 23

where holes connect internally to form a sharp corner, the use of special cores is not required. In an investment casting, a special core would be required for the curved hole in Fig. 17(b), whereas that in Fig. 17(a) could be produced with regular mold equipment. In die casting, the design shown in Fig. 17(a) would be possible to manufacture, but the design shown in Fig. 17(b) would be impossible. Where sharp corners will reduce product performance (as by turbulence in fuel valves, fraying of insulation in electrical wire housings, or signal error in radar components), ﬁllet radii should be incorporated. This requires expendable cores, which are more expensive than steel core pins. Figure 27(a) shows an example of an investment casting that required expendable cores. As designed, this casting could not be produced with steel core pins in the die used to produce the wax or plastic patterns. To permit the use of steel core pins, the casting would have to be redesigned as shown in Fig. 27(b). Six core pins would then be necessary. They would be located and pulled from the pattern as shown. An added operation would be required to plug the hole at core 1 after casting. This redesign has ﬁve core-pin junctions. It would be virtually impossible to produce these castings without prominent ﬂash at these junctions, because of the difﬁculty of maintaining ﬁrm butt-type joints. Removing the ﬂash would be an added expense. Unequal Expansion of Mold and Core. In investment castings, unequal heating and expansion of the mold and core can cause nonuniformity of wall thickness. This is particularly likely in castings that have tortuous passages formed by cores completely surrounded by metal. Typical of such castings is the one shown in Fig. 28. When this casting was poured, the mold was able to dissipate heat at a greater rate than the core. Consequently, the core became hotter than the mold and expanded more. The rigid mold resisted the expansion of the core, causing the core to bend. This bending moved part of the core closer to one wall and farther away from the other, thus producing a thinner wall on one side of the casting than on the other. Recommended minimum tolerances for the thickness of walls enclosing tortuous cored passages in investment castings are as follows:

(a) The cored hole in this investment cast propeller hub necessitates a relatively expensive gating arrangement. Because the gates were attached to the blades, contour machining was necessary for complete removal of the gates. (b) By eliminating the cored hole, a simple and less costly gate is possible. (The center hole was subsequently drilled.)

Fig. 24

Length of hole, in.

Under 2 2 to 4 Over 4

Wall thickness tolerance, in.

þ0.005 þ0.010 þ0.012

Designing to Eliminate Cores Useful as they are, cores add to the cost of producing castings; hence the designer should be alert to possibilities for avoiding them. A decision often depends on cost analysis. An example is shown in Fig. 29. In the original design of this casting, Fig. 29(a) a core is required to permit molding of the “hook” shape. The possible redesign shown in Fig. 29(b) would permit easy removal of the pattern from the sand, eliminate the need for a core, and effect a saving in molding cost. Another example is provided in Fig. 6 of the preceding article. Figure 30 shows a sand cast malleable iron wheel hub for which redesign eliminated a ring core and at the same time provided a stronger casting. As originally designed, Fig. 30(a), the eight ribs and eight small bosses prevented this casting from being molded with the parting line parallel to the axis of the hole. Furthermore, adjacent to the ﬂange, the casting had a cross section smaller than either the ﬂange or the extreme end of the casting. The undercut section that was thus formed prevented the pattern from being withdrawn from the mold in a direction perpendicular to the mounting ﬂange. A ring core, as shown, was necessary to produce the shape. By revising the casting as shown in Fig. 30(b), the need for the ring core was eliminated and the shape could be withdrawn easily from the mold. By broadening the base of the tubular section the eight ribs were also eliminated. In the original design, the small diameter of the tubular section at the junction with the ﬂange section was unable to withstand the forces of service. Eight strengthening ribs were required, to assure satisfactory performance of the casting in application. As redesigned, the broader base of the tubular section provided sufﬁcient strength to permit elimination of the ribs. Figure 31 shows a sand cast launching frame, produced from 4330 steel, that was redesigned to eliminate cores. The original design of this casting called for the coring of two pockets approximately 100 mm (4 in.) deep, to reduce weight. In production, however, it was difﬁcult to attain sound metal in this cored area. At the expense of adding weight, the cored pockets (and the cores) were eliminated in a redesign as shown. This not only corrected the defects but also facilitated production. The solid center was easily risered and encouraged a beneﬁcial freezing pattern in the entire casting. Simpliﬁed Cores. Not all designs or redesigns can eliminate cores, but often cores can be simpliﬁed. This was accomplished for the fuel cell interconnect, an aluminum casting, shown in Fig. 32. A semi-permanent mold

Design for Economical Coring / 97 Table 1

General dimensional relations for cored cavities in investment castings

Length X, in.

Diameter Y, in.

Min wall, in.

For a typical ferrous casting 0.250 0.125 0.500 0.250 0.750 0.375 1.000 0.500 1.250 0.750 1.500 1.000 2.000 1.500 2.500 1.750

0.030 0.040 0.050 0.060 0.060 0.060 0.060 0.060

Inside diameter A, In.

As-east roundness (TIR), in.

0.092 to 0.250 0.251 to 0.500 0.501 to 1.000 over 1.000, add 0.020 in. per in. A diameter, in.

0.250 0.500 0.750 1.000

Coring Versus Drilling

0.012 0.016 0.020

B diameter, in.

As-cast concentricity (TIR), in.

0.750 1.000 1.500 2.000

0.008 0.010 0.016 0.020

Minimum diameter Y Depth W, in.

Ferrous, in.

Nonferrous, in.

0.188 0.250 0.500 0.625 0.750 0.750 1.000 1.000

0.188 0.250 0.312 0.375 0.438 0.500 0.562 0.625

0.250 0.500 0.750 1.000 1.250 1.500 2.000 2.500

Min draft

0 0 0 0 0 0 0 0

15´ 15´ 15´ 15´

H, in.

D, in.

Z, in.

Parallelism(a), in.

0.250 0.500 0.750 1.000 1.250 1.500 2.000 2.500

0.125 0.250 0.375 0.500 0.625 0.750 0.875 1.000

0.500 0.750 1.000 1.250 1.500 1.750 2.000 2.250

þ0.005 þ0.005 þ0.005 þ0.010 þ0.012 þ0.013 þ0.015 þ0.015

(a) Parallelism may be regarded as the tolerance for dimension Z. T, in.

X, in.

a´, in.

Y, in.

b´, in.

0.125 0.250 0.125 0.250 0.500 1.000 2.000

0.250 0.500 0.750 1.000 0.500 0.500 2.500

þ0.004 þ0.005 þ0.005 þ0.005 þ0.008 þ0.010 þ0.015

0.125 0.250 0.500 0.750 0.250 0.250 1.500

þ0.003 þ0.004 þ0.005 þ0.005 þ0.005 þ0.008 þ0.010

Y, in.

X

a´ min

0.250 0.500 1.000

15 30 45

þ1 þ1 þ1

X diameter, in.

Y, in.

Z

0.250 0.500 0.250 0.500 0.750 and up

0.500 0.750 0.500 0.750 1.500

15 30 60 90 120

R, in

/64 /32 1 /16 1/16 1 /16 1 /16 1 /8 1 1

a´

þ0 þ0 þ0 þ0 þ0

casting as originally designed, it required a contoured sand core to form the inside conﬁguration shown. A revision to eliminate defects encouraged a further revision to simplify the core. By moving the wall to coincide with the inside contour of the ﬂanges, a simple straight-through metal core sufﬁced. The redesigned castings (now permanent mold, because of the metal core) were more accurate and less costly.

30´ 30´ 30´ 45´ 60´

A, in.

B, in.

C, in.

D, in.

E, in.

n´, in.

0.500 0.750 1.000 1.250

1.000 1.500 2.000 2.500

0.250 0.500 0.750 1.000

0.750 1.000 1.500 2.000

0.375 0.500 0.750 1.000

þ0.004 þ0.006 þ0.008 þ0.010

Although for many castings coring is essential to successful production, for others it is advisable to omit cores and to remove excess metal by other means. The choice may be based on considerations of soundness, dimensional accuracy, economy, or producibility. For example, if a casting is to have one or more round holes, these may be produced with greater accuracy or economy by subsequent boring or drilling, rather than by cores. As a general rule, it is often cheaper to core larger holes if an appreciable amount of metal can be saved, or if machining operations can be speeded up or eliminated. It is usually more expensive to core smaller holes unless a large amount of machining expense can be saved. Eliminating cored cavities in favor of subsequent drilling may also permit more efﬁcient risering in some designs, thus preventing defects. In the casting illustrated in Fig. 33, The core adjacent to the gate impeded ﬂow of metal into the mold. This caused cold shuts at the end of the casting opposite the gate. Dispensing with this core permitted metal to ﬂow freely to all parts of the cavity, so that the cold shuts were eliminated and sound castings produced. The previously cored hole was later drilled. In the sand casting shown in Fig. 34, it was cheaper to core and ﬁnish-drill the 0.84-in.diameter hole than to cast solid and produce the hole entirely by drilling. The cost of metal saved was greater than the cost of coring. Actual savings in this instance amounted to 1 ½¢ for each casting. If the cored hole had been much smaller, however, this would not have been true, because the core cost would have remained relatively constant whereas the value of the metal saved would have been reduced in proportion to the volume of the hole.

98 / Casting Design and Performance

Fig. 25

Examples of cored holes and cavities easily produced in investment castings

Oil slot was easily cored when part was produced as an investment casting, but had to be milled in when part was sand cast. Total saving of 41¢ per casting was realized by changing from sand to investment casting.

Fig. 26

Because it was completely surrounded by molten metal, the core that formed the passage through this investment casting expanded more in length than the mold when the metal was poured. This caused bending of the core, which inﬂuenced the casting wall thickness.

Special coring practice involving additional cost would be required to produce this investment casting as shown in (a). Normal practice using metal core pins in the pattern die is possible for the casting as shown in (b). Design (b) is less expensive than (a).

Fig. 27

Fig. 28

Fig. 30

Fig. 29

A redesign that eliminated a core and reduced the cost of the casting

An improved design that eliminated one core and eight ribs from a sand casting. This resulted in a stronger, more economical part.

Design for Economical Coring / 99

A cored hole 0.84 in. in diameter was incorporated in this casting to reduce machining cost. It was less expensive to core and drill than to drill through solid metal.

Fig. 34

Fig. 31

By eliminating two cored pockets from this 4330 steel sand casting, a better solidiﬁcation pattern was established and defects were eliminated.

A sand core was required to produce the center conﬁguration of this semipermanent mold aluminum casting. By increasing the wall thickness and placing the inner face of the wall in line with the inner edge of the ﬂanges, defects were eliminated and a simple metal core produced the center cavity.

Fig. 32

Fig. 33

Better ﬂow of molten metal was obtained by eliminating the 0.406-in.-diameter core adjacent to the gate of this investment casting. This enabled the metal streams at the extreme end of the cavity to unite properly.

Casting Design and Performance Pages 101–119

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Casting Design and Geometry Michael Gwyn, Advanced Technology Institute and American Foundry Society Technical Department

CASTINGS possess many inherent advantages that have long been accepted by the design engineer and metal parts user. In terms of component design, casting offers a great amount of ﬂexibility. The casting process permits the formation of streamlined, intricate, integral parts of strength and rigidity that are not obtainable by other methods of fabrication. The shape and size of the part are also primary considerations in the design and application of cast parts. In this category, the possibilities of metal castings are unsurpassed. The ﬂexibility of cast metal design offers a wide scope in converting ideas into an engineered part. The freedom of design offered through the metalcasting process allows the designer to accomplish several tasks simultaneously. These include the following: Freedom of design to optimize functionality

and manufacturability

Net or near-net shape design Intricate components can be produced as a

single cast part.

Few restrictions on part weight or size Almost all metals and alloys can be cast. Optimal appearance

Structural design engineers who work successfully with castings commonly design in a narrow group of casting types poured from familiar alloys (such as the family of irons or the 300 series of aluminum) and molded from familiar metalcasting processes (such as green sand or nobake). Rules of thumb and lists of design considerations are available in many design handbooks and references as an aid for common design situations. However, close examination of these recommended rules often reveal conﬂicting guidance in many cases. For example, the use of gusseting instead of mass for stiffness may be labeled “recommended” in one set of design rules and “poor” in another. Further, when a design engineer goes from a familiar casting design realm to an unfamiliar one, unexpected results may occur. For example, when substituting aluminum bronze for ductile iron (for a given metalcasting mold process), there is likely to be trouble with the usual ductile ironstyle geometry. Good aluminum-bronze geometry is different from typical ductile iron geometry, and

the molding process may need to supplement the different geometry with heat-transfer techniques. Not suspecting this, the design may suffer from no-quotes, higher-than-expected prices, or metalcaster requests for design changes. How are design engineers to know the differences in a successful casting geometry between that of an aluminum bronze and a ductile iron? Moreover, if the design engineer did know that information, what would be the proper course of design action? The answer lies in a better understanding of the relationship among geometry, metallurgy, and physics.

Casting Design Parameters The key to consistent and good casting design is how to choose a suitable geometry. There are six parameters (based on physics) that underlie cost-effective casting design. Four are based on casting properties:

Liquid metal ﬂuid life Solidiﬁcation shrinkage type Slag/dross formation tendency Pouring temperature

Two other design parameters are based on structural properties: Section modulus Modulus of elasticity

These six parameters, as a system, can drive the geometry of casting design from a process standpoint. Geometry is not only the result of product function design, but it also has a powerful inﬂuence on the castability of cost-effective castable designs that are economically produced, machined, and assembled into a ﬁnal product.

Fluid Life Fluid life more accurately deﬁnes the alloy liquid characteristics than does the traditional term ﬂuidity. Molten metal ﬂuidity is a dynamic property that depends on several factors that include: Mold materials and surface characteristics Alloy composition (and alloying effects on

freezing range)

Surface tension and surface ﬁlms Gas content and suspended inclusions

(which may alter ﬂuidity directly or indirectly by effects of surface tension) Degree of superheat Rate of pouring Rate of heat transfer to the surroundings Heat of fusion and solidiﬁcation Viscosity

Viscosity is a factor, but ﬂuidity should not be confused with viscosity. In fact, molten metals have a relatively low kinematic viscosity (viscosity divided by density). Fluid life is affected more by other factors that change as the alloy is delivered from a pouring ladle, diecasting chamber, and so on into a gating system and ﬁnally into the mold or die cavity. Heat transfer reduces the metal temperature, and oxide ﬁlms form on the metal front as this occurs. Fluidity decreases most rapidly with temperature loss, and it can decrease signiﬁcantly from the surface tension of oxide ﬁlms. Temperature also is not the sole basis of ﬂuidity at a given moment. For example, some aluminum alloys at 650 to 750 C (1200 to 1400 F) have excellent ﬂuid life. However, some molten steels at 1650 C (3000 F) have much shorter ﬂuid life. In other words, a molten alloy ﬂuid life also depends on chemical, metallurgical, and surface tension factors. Fluid life can also be related to freezing range. In alloys with a wide freezing range, casting is more difﬁcult, because there is a stronger tendency for nonuniform solidiﬁcation. Extra effort may be needed with the use of chills to create a steeper thermal gradient in portions of the mold when casting alloys with a wide freezing range. Fluid life thus affects castability and the possible design conﬁguration of a casting. The design limitation may be the minimum section thickness that can be cast reliably, the maximum length of a thin section, the ﬁneness of cosmetic detail (such as lettering and logos), or the accuracy with which the alloy ﬁlls the mold extremities. However, it is important to understand that moderate or even poor ﬂuid life does not limit the cost-effectiveness of design. Knowing that an alloy has limited ﬂuid life, the designer will make sure the part features:

102 / Casting Design and Performance Softer shapes and larger lettering Finer detail in the bottom portion of the

mold, where metal arrives ﬁrst, fastest, and generally hottest Coarser detail in the upper portions of the mold, where the metal is slower to arrive and more affected by oxide ﬁlms and solidiﬁcation “skin” formation. Even an alloy with good ﬂuidity, when overexposed to oxygen, may form a high-surface-tension oxide ﬁlm that makes the ﬂuidity die, rounding off the leading metal front as it ﬂows. More taper toward thin sections Some alloys, such as 356 aluminum, have been speciﬁcally designed metallurgically to enhance ﬂuid life. In the case of 356, the high heat capacity of silicon atoms revives aluminum atoms as their ﬂuid life begins to wane.

Solidiﬁcation Shrinkage The three distinct stages of shrinkage as molten metals solidify are liquid shrinkage, liquidto-solid shrinkage, and patternmaker’s contraction. Liquid shrinkage is the contraction of the liquid before solidiﬁcation begins. It is not an important design consideration, because the liquid metal will feed itself with additional metal Table 1

while in the mold. Feeding refers to liquid metal ﬂowing into a mold to replace the volume of metal lost due to shrinkage. Feeding is more important for liquid-to-solid shrinkage. Liquid-to-solid shrinkage is the shrinkage of the metal mass as it transforms from the disconnected atoms and molecules of the liquid into the structured building blocks of solid metal. The amount of solidiﬁcation shrinkage varies greatly from alloy to alloy. Table 1 provides a guide to the liquid-to-solid shrinkage of common alloys. As shown, shrinkage can vary from nearly no shrinkage to high shrinkage volumes. Alloys are further classiﬁed based on their solidiﬁcation type: directional, eutectic-type, and equiaxed (Table 1). The type of solidiﬁcation shrinkage in a casting is just as important as the amount of shrinkage. Speciﬁc types of geometry can be chosen to control internal integrity when solidiﬁcation amount or types are a problem. Solidiﬁcation on and perpendicular to the casting surfaces is known as progressive solidiﬁcation. At the same time, solidiﬁcation moves at a faster rate from the ends of the section(s) toward the source of feed metal (risers)—this is known as directional solidiﬁcation. Directional solidiﬁcation moves faster from the ends of the sections because of the greater amount of surface area through which the solidifying

metal can lose its heat. The objective is for directional solidiﬁcation to beat out progressive solidiﬁcation before it can block the source of the feed metal. Directionally solidifying alloys require extensive risering and tapering, but they also have the capability for excellent internal soundness when solidiﬁcation patterns are designed properly. The eutectic-type alloy is the most forgiving of the three. Such alloys typically have less solidiﬁcation shrinkage volume. Risers are much smaller and, in special cases, can be eliminated by strategically placed gates. The key feature with these alloys is the extended time that the metal feed path stays open. The plate solidiﬁes more uniformly all over and all at once, similar to eutectic solidiﬁcation. Eutectic-type alloys are less sensitive to shrinkage problems from abrupt geometry changes. Alloys that exhibit equiaxed solidiﬁcation respond the most dramatically to differences in geometry. Shrinkage in these alloys tends to be widely distributed as micropores, typically along the center plane of a casting section. The reason is that solidiﬁcation occurs not only progressively from casting surfaces inward and directionally from high-surface-area extremities toward lower-surface-area sections but also equiaxially via solidiﬁed islands in the middle of the liquid metal. These islands of solidiﬁcation interrupt the liquid pathway of directional

Guide to the liquid-to-solid shrinkage of common alloys

Alloys are further classiﬁed based on their solidiﬁcation type: directional, eutectic-type, and equiaxed. Solidiﬁcation shrinkage Alloy group

Irons Gray iron Compacted graphite Ductile iron Austempered ductile iron I White iron Malleable iron High-alloy iron High-silicon iron Carbon and low-alloy steels Carbon Low alloy High-alloy steels Martensitic Partly austenitic Fully austenitic High manganese Superalloys Nickel base Cobalt base Nonferrous Aluminum 356, 357 Aluminum 201, 206

Fluid life

Type

Amount

Pour temperature

Gas microporosity

Slag/dross

Excellent Excellent/good Good Good Excellent Excellent Fair/good Good

Eutectic-type Eutectic-type Eutectic-type/directional Eutectic-type/directional Directional Directional Equiaxed Equiaxed

Very small Small Small/moderate Small/moderate Moderate/large Moderate/large Moderate Moderate

2500–2600 2500–2600 2500–2600 2550–2650 2500–2600 2500–2600 2600–2700 2550–2650

Poor Poor

Directional Directional

Large Large

2850–3000 F 2850–3000 F

Poor/fair Fair Fair/good Fair

Directional Eutectic-type Equiaxed Equiaxed

Moderate/large Moderate Moderate Moderate/large

2850–2950 2800–2900 2775–2875 2725–2850

Poor Excellent

Directional Directional

Moderate Little

2600–2800 F 2600–2800 F

Moderate(+) tendency Moderate(+) tendency

Moderate Little/moderate

Excellent Good/fair

Eutectic-type Equiaxed

Little Moderate

1300–1400 F 1300–1400 F

High tendency High tendency

Moderate Moderate/large

F F F F

Higher Higher Higher Higher

tendency tendency tendency tendency

Little/moderate Little/moderate Little Little/moderate

Moderate (+) tendency Moderate () tendency Low tendency Moderate tendency High(++) tendency High(++) tendency

Large Large Little Moderate/large Very large Very large

Aluminum die-cast alloys are dominated by the diecasting process, and castability characteristics are less signiﬁcant in this context. Magnesium ZE43 Excellent Directional Moderate 1300–1400 Magnesium ZK51A Excellent Directional Moderate 1300–1400 Magnesium AM60A Excellent Eutectic-type Little 1300–1400 Magnesium AZ91C Excellent Eutectic-type Little 1300–1400 Magnesium die-cast alloys are dominated by the diecasting process, and castablity characteristics are less signiﬁcant in this context. Aluminum bronze Fair Equiaxed Moderate/large 2000–2150 Silicon bronze Fair Eutectic-type Little 1900–2050 Red brass 85-5-5-5 Fair (+) Equiaxed Moderate 1950–2100 Yellow brass Fair () Eutectic-type Moderate 1800–1950 Titanium Very good Eutectic-type Little 3300–3400 Zirconium Fair (+) Eutectic-type Little 3300–3400

F F F F F F F F

F F F F

F F F F F F

Low()tendency Low()tendency Low tendency Low(+)tendency Low()tendency Low()tendency Low(+)tendency Low(+)tendency

Little Little Little/moderate Little/moderate Little Little Little/moderate Moderate

Low tendency Low tendency

Moderate Moderate

Moderate Moderate Moderate Moderate

Moderate Moderate Little/moderate Little/moderate

tendency tendency tendency tendency

Casting Design and Geometry / 103 solidiﬁcation. Gradually, the pathways freeze off, leaving micropores of shrinkage around and behind the islands that grew in the middle of the pathway. Larger risers, thicker sections, and tapering are counterproductive, causing micropores to coalesce into larger pores across more of the casting cross section. Microporosity is kept small and conﬁned to a narrow midplane in the casting section by more thermally neutral geometry with smaller, further-spaced risers. A signiﬁcant bilateral and reciprocal relationship exists between solidiﬁcation shrinkage and geometry. Most simply, eutectic-type solidiﬁcation is tolerant of a wide variety of geometries; the least reciprocity is required. Most complexly, equiaxed solidiﬁcation requires the most engineering foresight in the choice of geometry and may require supplemental heat-transfer techniques in the mold process. In the middle lies directional solidiﬁcation. While capable of the worst shrinkage cavities, it is the most capable of very high internal integrity when the geometry is properly designed. Well-planned geometry in a directionally solidifying alloy can eliminate not only shrinkage but the need for any supplemental heat-transfer techniques in the mold. In fact, the real mechanism behind the bilateral and reciprocal relationship between solidiﬁcation shrinkage and geometry is heat transfer. All three modes of heat transfer—radiation, conduction, and convection—are involved in solidiﬁcation of castings, and all three depend on geometry for transfer efﬁciency. Convection and conduction are very important in casting solidiﬁcation, and transfer rates are highly affected by geometry. Thermal expansion/contraction, or patternmaker’s contraction, is the contraction that occurs after the metal has completely solidiﬁed and is cooling to ambient temperature. This contraction changes the dimensions of the casting from those of liquid in the mold to those dictated by the alloy rate of contraction. So, as the solid casting shrinks away from the mold walls, it assumes ﬁnal dimensions that must be predicted by the pattern- or diemaker. This variability of contraction is another important casting design consideration, and it is critical to dimensional accuracy. Tooling design and construction must compensate for it. Achieving dimensions that are “just like the blueprint” requires the metalcaster’s patternand/or diemaker to be included. In some cases involving metal molds, such as diecasting and permanent mold casting, the die material is stronger than the casting while it cools. Geometric features may actually pin a casting in the die. If this occurs, the casting is deformed or stretched during cooling, making the prediction of the ﬁnal geometric dimension impossible. The unpredictable nature of thermal expansion/contraction makes tooling adjustments inevitable. For example, a highly recommended

practice for critical dimensions and tolerances is to build the patterns/dies/coreboxes with extra material on critical surfaces, so that the dimensions can be ﬁne-tuned by removing small amounts of tooling stock after capability castings have been made and measured.

Structural Properties As described, the properties of castability inﬂuence geometry, and a well-chosen geometry by designers can offset the metallurgical nature of the more difﬁcult-to-cast alloys. Likewise, casting engineers can beneﬁt from an understanding of the structural design requirements of a section. Geometry is the key to consistency, and both the designer and metalcaster must understand the interplay of geometry on structural design and castability. In terms of structural geometry, castings can easily apply shape to structural requirements. Most casting designs are used to statically or dynamically control forces. When designing a component structurally, a design engineer is generally interested in safely controlling forces through choice of allowable stress and deﬂection. Although choice of material affects allowable stress and deﬂection, a very signiﬁcant factor is geometry. Geometry has a direct inﬂuence on the stiffness and stresses in a structure. Casting has wide capability in producing various shapes, and computer-generated solid models and rapid prototypes provide ﬂexibility in the design stage. However, optimal geometry for allowable, uniform stress may not be acceptable geometry for castability. When a foundry engineer quotes a design that considered structural geometry only, requests for geometry changes are likely. At this point, the geometry adjustments for castability may be more substantial than the solid model software can “tweak.” The result can be no-quotes, higher-than-expected casting prices, or starting over with a new solid model. A practical solution to this problem is to concurrently engineer product geometry while considering structural, metalcasting, and downstream manufacturing needs. The result can be a casting geometry optimized for both functionality and manufacturability. The most efﬁcient technique is to make engineering sketches or marked sections and/or views on blueprints. The idea is to explore overall geometry before locking into a solid model too quickly. Engineering sketches or markups are easy and quick to change—even dramatically—in the concurrent brainstorming process; solid models are not. Solid models can provide important analysis and reﬁnement of design, but two key structural properties that should be emphasized for castable design are section modulus and modulus of elasticity. The modulus of elasticity deﬁnes the inherent stiffness or rigidity of a material,

while the section modulus (or section stiffness) depends on the geometric (section proﬁle) conﬁguration of the part. These two structural factors are critical in knowing how to choose a suitable geometry for a castable design. Modulus of elasticity is an important parameter in structural design, because it is directly related to geometric deﬂection under load. A larger modulus of elasticity means less deﬂection. Modulus of elasticity varies widely among materials, and it varies signiﬁcantly among metals; that is, some metals are considerably stiffer than others. A given alloy family (say steel or aluminum) also tends to have the same modulus value; for example, the entire family of steels (carbon, low alloy, and high alloy) has the same elastic moduli value of approximately 30 106 psi. In contrast, all aluminum alloys have elastic moduli that are three times smaller ( 10 106 psi). Thus, a steel casting is three times stiffer than an aluminum casting of identical geometry, simply because steel is stiffer than aluminum. Section Modulus. Evaluation of section modulus depends on the geometric conﬁguration of the part. As the design engineer well knows, the classic formulas for bending stress, torsional stress, and deﬂection are relatively simple. Each, however, contains the same parameter, section modulus, which is a function of shape and difﬁcult to compute. However, a quick, simple way to compute or estimate section modulus can be helpful in evaluating options for a more castable geometry. The key is to estimate the area moment of inertia, which is the basic measure of section modulus (see the section “Quick Method for Estimating Area Moment of Inertia from Sketches” in this chapter).

Applying the System This six-faceted system (based on four casting properties and two structural properties) is capable of optimizing: Geometry for castability, structure, and

downstream processing (machining and assembly) Process geometry (risering, gating, venting, and heat-transfer patterns) in the mold The process geometry forms the casting geometry. Quickly sorting through possible casting and process geometries by marking up blueprints or by making engineering sketches is an effective way to rapidly improve system geometry. An elegant result of a good brainstorming markup session can be a solid model of the casting and its process geometry, the basis of rapid prototyping and/or computerized testing. The objective is to explore geometry possibilities, looking for an ideal shape that is both castable in the chosen alloy and allowable in stress and deﬂection for that alloy. As noted, there is

104 / Casting Design and Performance great variety in the four metallurgical characteristics that govern alloy castability. Similarly, great variety exists among metals in their allowable stress and deﬂection. Therefore, an ideal casting shape for all six of the casting design factors is not necessarily a trivial exercise. For alloys that have good castability, choosing geometry for allowable stress and deﬂection is the best place to start. For alloys with less-than-the-best castability, it is better to ﬁrst ﬁnd geometry that assists castability and then modify it for allowable stress and deﬂection. Not all alloys are like ductile iron, which is highly castable, relatively resistant to stress, and moderately resilient against deﬂection. For ductile iron, many geometries may be equally acceptable. Martensitic high-alloy steel has fair-to-poor castability but can have amazing resistance to stress and can tolerate very large deﬂections without structural harm. Therefore, structural geometry is easy to develop, but a coincidental castable shape is more difﬁcult to design. Premium A356 aluminum has good castability but rather weak resistance to stress and low tolerance for deﬂection. Carefully chosen structural geometry, however, combined with solidiﬁcation enhancements in the molding process, has resulted in extremely weight-effective A356 structural components for aircraft, cars, and trucks. Optimizing casting geometry using the sixparameter system is not difﬁcult. The six casting and structural characteristics inﬂuence important variables in designing, producing, and using metal castings. These variables include:

Casting method Design of casting sections Design of junctions between casting sections Surface integrity Internal integrity Dimensional capability Cosmetic appearance

Both the designer and metalcaster possess a vital ally to streamline any casting design: casting geometry. Casting geometry is the most powerful tool available to improve castability of the alloy and mechanical stiffness of the casting.

Design of Junctions A junction is a region in which different section shapes come together within an overall casting geometry. Simply stated, junctions are the intersection of two or more casting sections. The four junction types are L-, T-, X-, and Ydesigns. Designing junctions is the ﬁrst step to ﬁnding castable geometry via the six-faceted system for casting design. There are major differences in allowable junction geometry, depending on alloy shrinkage amount and pouring temperature. One alloy may allow abrupt section changes and tight geometry, while another alloy requires considerable adjustment of junction geometry, such as radiusing, spacing, dimpling, and feeding.

Considerations of Secondary Operations in Design System-wide thinking also must include the secondary operations, such as machining, welding and joining, heat treating, painting, and plating. One aspect that affects geometry is the use of ﬁxturing to hold the casting during machining. Frequently, the engineers who design machining ﬁxtures for castings are not consulted by either the design engineer or the foundry engineer as a new casting geometry is being developed. Failure to do so can be a signiﬁcant oversight that adds machining costs. If the casting geometry has been based on the four casting characteristics of the alloy, then the designer knows the likely surfaces for riser contacts and may have some idea of likely parting lines and core match lines. These surfaces and lines will be irregularities on the casting geometry and will cause problems if they contact ﬁxturing targets. It is best to deﬁne the casting dimensional datums as the signiﬁcant installation surfaces, in order of function priority, based on how the casting is actually used. Targets for machining ﬁxtures should be consistent with these datum principles. Nothing is more signiﬁcant in successful computer numerical control and transfer line machining of castings than the religious application of these datum ﬁxture and targeting principles.

Drawings and Dimensions The tool that has had the most dramatic positive impact on the manufacture of parts that reliably ﬁt together is geometric dimensioning and tolerancing (GD&T), as deﬁned by ANSI Y14.5M—1994. When compared to traditional (coordinate) methods, GD&T: Considers tolerances, feature-by-feature Minimizes the use of the title block toler-

ances and maximizes the application of tolerances speciﬁc to the requirement of the feature and its function Is a contract for inspection (tells the person reading the drawing how to measure the part), rather than a recipe for manufacture In other words, GD&T speciﬁes the tolerances required feature-by-feature in a way that does not specify or suggest how the feature should be manufactured. This allows casting processes to be applied more creatively, often reducing costs compared to other modes of manufacture, as well as ﬁnish machining costs. GD&T encourages the manufacturer to be creative in complying with the dimensional speciﬁcations of the drawing, because the issue is compliance with tolerance, not necessarily compliance with a manufacturing method. By forcing the designer to consider tolerances

feature-by-feature, GD&T often results in broader tolerances in some features, which opens up consideration of lower-cost manufacturing methods, such as castings.

Factors That Control Casting Tolerances How a cast feature is formed in a mold has a signiﬁcant effect on the tolerance capability of the feature. The following six parameters (in order of preference) control the tolerance capability of castings. Molding Process. The type of molding process has the greatest single inﬂuence on tolerance capability. How a given molding process is mechanized and the sophistication of its pattern or die equipment can reﬁne or coarsen its base tolerance capability. Casting Weight and Longest Dimension. Logically, heavier castings with longer overall dimensions require more tolerance. These two parameters have been deﬁned statistically in tolerance tables for some alloy families. Mold Degrees of Freedom. This parameter is least understood. Just as some molding processes have more mold components (mold halves, cores, loose pieces, chills, etc.) than others, some casting designs require more mold components. Each mold component has its own tolerances, and tolerances are stacked as the mold is assembled. More mold components mean more degrees of freedom, hence more tolerance. Good design for tolerance capability minimizes degrees of freedom in the mold for features with critical dimensions. Design engineers must be aware of this and consider what really needs to be controlled when tolerancing a design. For example, is a position tolerance really needed relative to a datum or another speciﬁc feature? Interdependent features with tight position tolerances often can be located in a single-piece portion of the molding tool, reducing the number of degrees of freedom and improving dimensional accuracy. Draft. It is common for casting designs to ignore the certainty of draft, including mold draft, draft on wax and/or Styrofoam patterns made from dies, and core draft. Since 1 of draft angle generates 0.017 in. of offset per inch of draw (approximately 0.5 mm/30 mm), draft can quickly use up all of a tolerance zone and more. Patternmaker’s Contraction. The uncertainty of patternmaker’s contraction is why metalcasters normally recommend producing ﬁrst-article and production-process veriﬁcation castings (sometimes called sample or capability castings) to establish what the dimensions really will be before going into production. A common consequence of patternmaker’s contraction uncertainty is a casting dimension that is out of tolerance, not because it varies too much but because its average value is too far from nominal. In other words, the dimension contracted more or less than what was expected.

Casting Design and Geometry / 105 Cleaning and Heat Treating. Many casting dimensions are affected by downstream processing. At the least, most castings are touched by abrasive cutting wheels and grinding—even precision castings. Many castings are heat treated, which can affect straightness and ﬂatness. When considering the breadth and depth of the importance of geometry in casting design, from its inﬂuence on castability, the geometry of gating/risering, structural form, cosmetic appearances, and downstream ﬁxturing, extensive brainstorming of geometry is highly recommended. The standard for optimal casting geometry is high, but the possibilities for geometry are limitless. Find ways of exploring geometry quickly, such as engineering sketching, before committing to a print or solid model.

Quick Method for Estimating Area Moment of Inertia from Sketches As noted, estimating the area moment of inertia is the basic measure of section modulus. The difﬁculty in computing area moment of inertia for casting shapes is one of the hidden reasons for the design and use of fabrications. Fabrications are made from building blocks of wrought shapes, such as I-beams, rectangular bars, angles, channels, and tubes. These shapes, which are simple and constant over their length, have area moments of inertia that are easy to calculate or are available in handbooks. Consequently, stress and deﬂection calculations are relatively easy. Fabricated designs, however, are heavy and nonuniform in stress compared to a well-designed casting for the same purpose. Section modulus (or, more speciﬁcally, the area moment of inertia) is a function of shape and difﬁcult to compute. Therefore, a quick, simple way to compute or estimate the area moment of inertia can be useful in brainstorming for a casting geometry with both structural ﬁtness and castability. Solid modeling is used for both structural analysis by design engineers and for mold ﬁlling and solidiﬁcation analysis for the casting engineer. However, it can be helpful to markup and brainstorm engineering drawings before building a solid model. This can be an efﬁcient structural evaluation of geometry in casting design. To take full advantage of engineering sketching/print marking as a way to brainstorm geometry, one must be able to quickly evaluate stress and deﬂection at important cross sections in the sketches Although there are ﬁve kinds of stress (tension, compression, shear, bending, and torsion), the interesting ones for complex structures are bending and torsion, and their equations are shown in Table 2. (If more than one type of stress is involved in the same section, the principle of superposition allows the individual stress types to be analyzed separately and then added together; once again, the larger of the stresses to be combined is usually from bending or torsion.)

Equations for deﬂection are more complexlooking and are different for each type of loading geometry. However, a simpliﬁed relationship for bending deﬂection can be used to estimate whether deﬂection will increase or decrease with a given geometry. The relationship for deﬂection (d) from bending around axis x-x is: dxx /

ðBending momentÞðSection lengthÞ2 E Ixx

where E is the modulus of elasticity, and Ixx is the moment of inertia. This is not an exact analytical relation but rather a proportional (/) relation to estimate how changes in geometry affect deﬂection. If the area moment of inertia can be quickly estimated, engineering isometric sketches can be roughly evaluated in estimating the effects of stresses on deﬂection. Maximum tensile stress in bending is often most critical in structural design. Section modulus is deﬁned as the area moment of inertia divided by the maximum distance from the center of bending (centroid) to the outermost edge of the casting cross section. Section modulus is similar to a stiffness index, because it considers not only magnitude of area moment of inertia but also maximum section depth. If maximum section depth increases faster than area moment of inertia, a geometry change can actually increase maximum tensile stress rather than reduce it. This index, termed section modulus, accounts for that potential problem. The estimation method recommended is based on three principles. One is intuitive, and the other two are from the mathematics of engineering mechanics. The Design Engineer’s Sense of Load Magnitudes and Component Size/Shape. Engineers routinely use this sense to sketch sized shapes that are in the ballpark of the ﬁnal design. Casting engineers can learn this sense, and when they do, they become effective concurrent engineering partners in their customers’ casting designs.

The Equation for Area Moment of Inertia (Fig. 1). Although the calculus for an interesting casting cross section can be very difﬁcult, the relationship expressed between depth of section (Y) and change in cross-sectional area (dA) is very simple. The position and shape of the two rectangles in Fig. 1 (top) demonstrate this simple yet powerful relationship. The change in shape of the inside of the tube (at bottom of Fig. 1) is an even more dramatic illustration. Calculations were not made in either case, but the qualitative impact of Y2dA on stiffness and stress is unmistakable. Area Moment of Inertia. Once the engineering sense of structural size and Y2dA has been applied qualitatively to a sketched cross section, the parallel axis theorem can be applied to simple building blocks in the cross section to estimate area moment of inertia quantitatively. A numerical value for area moment of inertia is required to calculate the stress level in the sketched cross section. The parallel axis theorem is illustrated in Fig. 2.

Speciﬁc Design Tips for Castings The optimal geometry for casting and its structural properties offers many opportunities for improving quality while decreasing costs

Table 2 General relationships between stress and geometry sxx

taverage

Bending around axis x-x tension/ compression, max Tension/ compression, max Average compression Shear

ttorsional

Shear, max

saxial sbearing

thorizontal Shear from bending at slice n-n

max Þ ¼ ðBending momentÞðY Ixx

force ¼ CrossAxial -sectional area

Bearing force ¼ Projected contact area force ¼ CrossShear -sectional area max Þ ¼ ðTorqueÞðRadius J

forceÞðStatic moment;partial areaÞ ¼ ðShear ðlxx ÞðWidth of partial area slice n-nÞ

Geometric aspects in calculating the area moment of inertia in rectangular coordinates (Ixx) or polar coordinates (J). The essence of area moment of inertia can be quickly gleaned from its base formula. Moment of inertia is an important geometric factor because stresses in bending and torsion are inversely proportional (Table 2).

Fig. 1

106 / Casting Design and Performance Area moment of inertia Parallel axis theorem

Area moment of inertia Axial - vs.- polar

+Y

Y

Rule of pythagoras 2

2

x

2

r =y + x 2

2

dA

2

r dA = y dA + x dA 2

2

2

òr dA = òy dA + òx dA –X

Ix-x

=

ò or

y

r

XMAX

YMAX Y2dA YMIN

Iy-y R

J

=

ò0

J

=

Ix-x

=

ò

+X

2

X dA YMIN

Find centroid of overall cross-section by weighted average of component shape

X

positions:

YG

R1 Y1

YG1

X

YG = Sum of AR1 Y1+ AR2 Y2

Axis 0

+ .... + ASHAPEn Yn

2

r dA + I y-y

Y YG12AR1

Divided by

IR1X-X = IR1 +

Sum of all areas

IR1X-X = b1h1 /12 + YG1 (b1h1)

3

2

Area moment of inertia Centroid position Rectangles

AXls 2

Area = bh

Y

Centroid h X

X b Y

X-X = h/2

Y-Y = b/2 Axls 1 Moment of inertia Ix-x = bh3/12 Iy-y = b3h/12

Parallel axis theorem and general equations in calculating area moment of inertia. Good casting designs typically have complex cross sections, and although the general equations for moment of inertia are not trivial, the parallel axis theorem provides an estimating technique.

Fig. 2

to add value. The objective of good casting design is to explore geometry possibilities, looking for an ideal shape that is both castable in the chosen metalcasting alloy and allowable in stress and deﬂection. For alloys that have good castability, choosing geometry for stress and deﬂection is the best place to start. For alloys with less-than-the-best castability, it is better to ﬁrst ﬁnd geometry that assists castability and then modify it for allowable stress and deﬂection.

Design for Functional Packaging Product engineers and designers are faced with the challenge of adding more functions at reduced costs, often with additional constraints on physical space. No matter how well an individual component or subsystem is designed, the product is useless if it does not ﬁt into a limited envelope. This is especially true in transportation vehicles. Transportation system design is becoming more complex because the leading requirement has become functional packaging. Components must ﬁt into a conﬁned environment and survive interactions with mating and nearby components. Although functional packaging often is the foremost requirement in vehicle design, few product engineers and designers understand all aspects of functional packaging. Following are ten key points to consider when

developing a functional package in the case of components for transportation vehicles. Hard Points. Each vehicle has hard points used in its manufacture that typically remain ﬁxed for many years across model changes and sometimes across completely different models. These points are used for locating and handling the vehicle as well as major assemblies during manufacture. The most obvious are the lifting points on vehicle frames used to carry them down the assembly line before suspensions and wheels have been mounted. Diesel engines for the truck industry are prime examples in which hard points for mounting the engine have remained constant across engine models for years. Begin every design with an understanding of the customer’s hard points. When developing a three-dimensional (3-D) computer-aided design (CAD) model, start by creating the customer’s hard points, then package the design around these constraints. If possible, model the hardware associated with the hard points as well. Mating Parts. The most obvious design consideration when developing a functional package is mating parts. Solid models of mating parts are routinely imported into CAD systems prior to developing a design. Although most designers and product engineers consider mating parts when developing a design, the allowable variation associated with them often is

overlooked. Dimensional variation due to tolerancing can cause numerous failure modes, including interference conditions and part mismatch. Using today’s design tools, 3-D models can be modiﬁed to examine these phenomena. Imported electronic models of mating parts can be modiﬁed to investigate the allowable variation of a given component. Nearby Parts. As with mating parts, other nearby components must be considered when developing a design package. The same issue of allowable variation often is neglected. As with mating parts, 3-D models can be modiﬁed to examine the maximum and minimum material conditions of nearby parts as well as positional variation allowed in tolerancing. Close Mechanical Members. When designing a product, nearby moving parts, such as linkages and control arms, must be understood. Often, random vibrations caused by road conditions may excite nearby components, causing movement. Many knocks and other noiserelated issues stem from such interactions. When developing a CAD design, model the 3-D envelope required by the movement of a nearby member as a solid space. The design must package around this space. Thermal Source. Transportation components and assemblies must be capable of surviving a wide range of temperatures. For example, the temperature in an engine compartment may be as low as 40 C (40 F) when starting a vehicle that has been parked outside in the cold winters of Manitoba, Canada. During operation a few minutes later, exhaust components, such as the catalytic converter and exhaust gas recirculation tube (EMF), may reach temperatures in excess of 600 C (1110 F). Consider the operating temperatures of mating and nearby components. When packaging a design, sufﬁcient space must be allowed between a hightemperature member to avoid interactions. In some cases, heat shielding may be required, making packaging more difﬁcult. EMF Sources. With the increased use of electronics in transportation products, the electromechanical output of components has become a consideration when packaging a design. If necessary, space must be maintained between nearby members to avoid EMF interactions. In some cases, EMF shielding may be required. Run Downs. Assembly of vehicle components and systems typically involves the use of threaded fasteners. When packaging a design, do not encroach on the space needed by tooling to “run down” a fastener. Consider the run downs needed on mating parts, nearby parts, and the component being designed. When developing a CAD design, model the 3-D cylinder required by a tool to run down a fastener. The design must package around this space. In some cases, the envelope needed for a rundown tool can be minimized by using specialty fasteners with internal drives, such as Allenheaded screws.

Casting Design and Geometry / 107 Future Parts. Often, components that are designed ﬁrst occupy maximum space, making subsequently developed components more difﬁcult to design and more costly to produce. When designing a component, consider what parts and assemblies will be added to complete the vehicle. The most neglected components have historically been wire harnesses and hoses of all types. Even fuel lines often are afterthoughts on many vehicles, designed after all other systems. Gaging Points. In many cases, transportation vehicle components and assemblies must be aligned or modiﬁed during manufacture. Although such cases are undesirable, they do exist. At least three points in space are required to locate a component. When developing a package, include and specify these points and allow clearance so that these points may be reached by the gage hardware. When developing a CAD design, the gage hardware should be modeled as a solid. This is yet another consideration that consumes functional design space. Service Window. Service requirements for transportation vehicles range from changing lubricants to replacing perishable components such as batteries and oil ﬁlters. Clearance for service tools must be computer-modeled as a solid space. A functional package must be developed around these pathways. Access to perishable components also must be maintained. Modeling extraction paths for such components is difﬁcult. At the very least, a minimum clearance must be maintained to allow for the removal of such parts.

Add Strength, Not Mass Often, when designers and product engineers are faced with making a component stronger, the simplest solution at hand is to add thickness. Adding more material can be a quick ﬁx. Using basic physical theory, doubling the material thickness in a select location should double the strength. However, such a basic, straightforward theory makes many assumptions that are rarely true. The theory that increasing thickness adds strength assumes that the added material is perfect. When using net shape manufacturing processes, this is rarely the case. Material integrity can change drastically as the wall thickness increases. This is particularly true with casting processes. Of the casting processes, diecasting is among the most sensitive to increases in material thickness. An atomized spray of metal is injected into the die, ﬁlling the cavity from the outside inward. The metal, which contacts the die, freezes quickly and creates a thin, high-integrity skin. The remaining molten metal freezes at a slower rate. As heat is removed at the die surface, the casting solidiﬁes inward, leaving trace amounts of porosity in the center. This porosity is often

referred to as centerline porosity. As the wall thickness of a die-cast component increases, the amount of centerline porosity increases. Although actions can be taken in tool design and process settings to minimize the effects of centerline porosity, increases in wall thickness often create casting defects in the areas that need to be of the highest integrity. Slow-cooling casting processes, which use risers and other methods to minimize porosity, often exhibit extreme changes in material properties as the thickness is increased. This is primarily due to variations in the solidiﬁed microstructure, which differ as the distance from the mold surface increases. Near the surface is a thin chill zone in which the casting is of the highest integrity with a ﬁne, equiaxed structure. Immediately following this zone is a columnar region that forms as heat ﬂows from the liquid through the chill zone and into the mold. Note that columnar structures associated with different surfaces in the mold can run into each other. In the center of the casting is a large, grained, equiaxed region, which is the last portion to solidify. The material properties in each of these regions are very different. Adding quality material is the key to improving strength. Adding more of the high-integrity skin (chill zone) is the best way to add strength. To add more of the high-integrity skin, the surface area of the component must be increased. This can be done efﬁciently by using ribs of a uniform wall thickness. Ribs serve to improve component stiffness while reducing weight by removing material from the casting. Correct rib depth and spacing during component design depends on the overall component engineering. Ribs are recommended for reinforcing or stiffening most webbed sections and are required for thin webs. When designing castings with webs, ribs will provide the following beneﬁts: Reduce web thickness and save weight Reduce web load deﬂection by strengthening

and stiffening

Prevent warping of unsupported web areas Prevent hot tearing or cracking of the casting

during solidiﬁcation by acting as a chill An easy method for computing proper web section thickness does not exist. However, for both the casting design engineer and metalcaster, webs must be neither too thick nor too thin for the speciﬁc design. Web thickness depends on various factors, including web area, alloy used, structural requirements, and the number and geometry of joining webs and stiffeners. For good aluminum casting design, rib junctions should be staggered to avoid a large metal mass that will be present at the junction of multiple sections. This mass can cause uneven cooling of the casting (preventing directional solidiﬁcation) and defects. Staggering the ribs reduces the strength of the section, but both the loss in strength and the amount of metal

can be held to the minimum at the intersections by staggering the rib junctions a distance less than or equal to the thickness of the rib. Another factor to consider in designing junctions in castings is the use of ﬁllets at all sharp corners. By rounding junctions to the proper radii, ﬁllets prevent cracks, tears, and shrinkage at re-entry angles. They also make corners more moldable for manufacturing and reduce stress concentrations in the casting when in service. Fillets must be large enough to meet engineering stress requirements and reduce stress concentrations but not so large that they cause shrinkage during casting solidiﬁcation. When possible, ribs should be designed appreciably deeper than their widths. The recommended width for ribs in aluminum castings is 1.5 times the thickness of intersecting walls. The minimum rib width is equal to the wall thickness. Ribs in compression while in service offer a greater factor of safety than ribs in tension. Because thin ribs are subject to failure in service, they generally are regarded as poor design practice. Figure 3 provides four degrees of rib design for aluminum cast components. The rib atop the section in Fig. 3 is termed a bead and is rated highly as a design for load-carrying capability. However, it can be difﬁcult to cast. To aid in manufacturing a component with a beaded rib, design the feature so it is located on the parting plane of the mold. Cross ribbing or ribbing on both sides of a casting (Fig. 4) is undesirable from a casting manufacturing perspective. Cross ribbing increases the rate of casting defects and the total cost of the component. This design also creates hot spots in the casting during cooling and solidiﬁcation, creating feeding difﬁculties and potential shrinkage-related defects. If design requirements dictate the placement of cross ribbing on a casting, the designer should adhere to the recommended practice of staggering the location of the ribs, as shown in Fig. 4 (right). This will alleviate the problems encountered in casting cross ribs. Cored rectangular holes in ribs and webs (Fig. 5) promote difﬁculties during casting manufacturing. In ribs and/or webs, round or oval cored holes are preferred. However, if other shaped holes are required, all corners should be rounded to enhance casting solidiﬁcation. In highly stressed webs, ribs, and walls, beading cored holes (Fig. 6) is critical for increasing the component service life. If a component has been designed with a heavy section that has been cored, ribs may be added for strength and to achieve a uniform wall thickness. During design stages, however, remember that if a component is designed with deep ribs, these ribs require adequate draft to ensure removal or ejection of the mold from the pattern during the casting manufacturing process. The goal in all designs should be to create a component without ribs, if possible. If the casting wall alone has adequate strength and

108 / Casting Design and Performance t

R= t 4

t

(a) Poor

(b) Acceptable R=t

t

T 4

T

3t

R= t 4

Cross ribbing (left) should be avoided in cast component design due to the hot spots and potential shrinkage defects it creates in casting solidiﬁcation. Staggered ribs (right) provide for better manufacturing.

Fig. 4

t

R=t T+t 2

t R = T+t t 2

R=t

t

3t - 3T

(c) Good

(d) Excellent

Fig. 3

Four degrees of rib design for aluminum cast components. The beaded rib in is rated highly for load-carrying ability. This type of rib feature is cast most easily when located on the parting plane of the mold.

Fig. 5

When designing a cored hole between ribs or in a web design, design the hole to be round or oval with rounded corners (right). Rectangular holes (left) create manufacturing difﬁculties.

T

T

1.5T 1.5T

T

T

–2T

T

Fig. 6

Although the design on the left is an acceptable design for a cored hole in a highly stressed rib design, the design on the right is preferred to improve the service life of the component.

stiffness, the elimination of ribs results in lower stresses and improved stress distribution in the component.

Designing Castings with Uniform Wall Thickness In many castings, functional requirements dictate that walls be uniform or nearly uniform in thickness. However, if these castings are not properly designed, the molten metal will freeze prematurely during casting before all parts of the mold cavity are ﬁlled. Even if the mold cavity is properly ﬁlled, portions of the casting may not be fed before the onset of metal solidiﬁcation, causing centerline shrinkage or porosity.

To alleviate the problem of incomplete mold ﬁll with uniform section thickness castings, walls may be tapered with extra padding during the design, and/or ribs and webs may be added to the casting to provide feed paths for the molten metal. When pouring a ﬂat plate with uniform wall thickness, the metal enters the mold cavity and attempts to ﬂow in all directions. If the distance from the gate to the extremities of the mold cavity is too great, the metal freezes prematurely, resulting in misruns. When ribs are added to the plate, the metal ﬂows readily to the extremities, and successful cavities are produced. The same success with the plate is attained when the casting gate is enlarged and uniformly tapered to the extremities. The metal ﬂows into

the mold cavity at the heaviest section, thus preventing an initial chilling of the metal by the mold, as occurs with a small volume of metal. Since the molten metal in heavy sections retains more of its heat as it ﬂows through the mold, it more readily can reach the extremities of the mold cavity. The goal of casting solidiﬁcation design is that the volume of metal lost because of shrinkage can be replaced by the riser. Although measures are taken with uniform-thickness castings to ensure proper freezing, there is no assurance that freezing will start at points furthest from the gate and progress in an orderly manner to the gate. The result could have metal nearer to the gate freezing ﬁrst, isolating pockets of molten metal, which results in voids or porosity at these pockets. One solution is the use of two risers to feed metal from opposite ends of the casting. However, risers add to the cost of producing a casting. Another option is designing feed paths to allow molten metal to ﬂow easily to all sections of the mold cavity. If the enlarged sections are gradually tapered from the gate to the extremities, solidiﬁcation follows an orderly progression through the casting, terminating at the riser. Compensation for normal shrinkage is provided throughout the casting, and sound metal is assured. The same beneﬁcial freezing pattern is obtained by tapering the plate, provided the angle of taper is sufﬁcient and the source of the feed metal is at the heaviest section. Under these conditions, a favorable freezing pattern is assured because the metal that has ﬂowed the farthest (and is therefore the coldest) is deposited in the thin section. In contrast to a heavy section, the thin section of molten metal has less heat to transmit to the mold before freezing can start. Thus, the direction of solidiﬁcation can be predicted in tapered sections, provided the metal enters at the heavy end. If the need for feed paths is anticipated in the early stages of casting design, padding can be made a functional part of the casting, or it can be located so that its removal adds little to no cost.

Designing Castings with Unequal Sections Casting processes provide design freedom by allowing engineers to place metal where needed to meet desired criteria. Typically, functional

Casting Design and Geometry / 109 demands of the part as well as induced stress distribution result in a nonuniform mass distribution in a designed structure. This means that a casting shape can assume nonuniform distribution of interconnected thick and thin sections. For example, if a heavy section is surrounded with thin sections, strategically located risers may be needed to feed isolated heavy sections. If each heavy section requires a separate riser, the result is lower yield, increased labor for casting cleaning, and increased ﬁnal part cost. However, there are certain instances that can lend themselves to redesign to make the shape more readily castable without additional risers. In those circumstances, a designer’s knowledge of the metalcasting process or close collaboration between the designer and casting engineer can play a signiﬁcant role in modifying the shape to suit the process. For example, one design option may be a feeding pad, which is essentially additional metal added to the casting that allows directional solidiﬁcation to progress toward the side riser. The addition of feeding pads can result in a solid casting without a top riser, thus reducing yield. This type of modiﬁcation can be made in many instances without affecting the part functionality and only minimally increasing its weight. In application, when the part is rotating around a central axis, the additional pad can be added on the opposite side to maintain dynamic balance. Another geometric modiﬁcation that can be considered to improve component castability with isolated masses is to reallocate the existing geometric features. For example, ribs are used to provide additional structural support to isolated bosses. The size of the ribs is an important factor in casting, since it determines how far the molten metal can ﬂow and how it will solidify. If the ribs are too thin, this will promote rapid cooling and premature solidiﬁcation and will result in either a cold shut during pouring or inadequate feeding of the boss, causing shrinkage defects. Alternately, large ribs will add to the weight of the casting. Hence, ribs should be dimensioned properly to satisfy structural integrity while allowing castability.

Designing Thin Sections The lightweight structure of thin-wall castings allows for increased payload and reduced energy consumption in transportation applications. However, thin-wall castings also can pose manufacturability problems associated with mold ﬁlling. Rapid cooling of thin-wall sections of the casting reduces the ﬂuidity of the molten metal, which could cause the molten metal to prematurely freeze before it can completely ﬁll the mold cavity. The mold-ﬁlling capability in thin-wall casting depends on a number of different process- and design-related variables, including the type of metal, pouring temperature, casting shape, and metalcasting process.

The practical minimum wall thickness for a particular casting conﬁguration can be determined based on experiments using inexpensive and accurately made loose patterns. Often, a metalcasting engineer will recommend a value to start at, based on prior experience. The wall thickness then should be adjusted, depending on if the defect rate is above or below the acceptable range. Typically, mold-ﬁlling problems are rarely an issue when the wall thickness exceeds 9.5 mm (0.375 in.). Alternately, when the wall thickness is reduced below 3.2 mm (0.125 in.), adequate mold ﬁlling is virtually impossible except for short distances. The problem is premature freezing of the molten metal. In such instances, a designer can allow for certain geometric features on the casting that can ensure ﬂuidity of the molten metal for the entire length of the mold cavity. The chapter “Design Problems Involving Thin Sections” in this book describes various examples. In some cases, the selection of a different alloy can improve the castability of thinwalled castings. For example, aluminum can achieve thinner walls more successfully than steel. However, a drastic change in metal is not an option in many cases. If a casting is not structural or load bearing, a designer can take this fact into consideration to achieve thinner walls. Even in a similar group of materials, small deviations in alloy composition can have a signiﬁcant impact on material ﬂuidity. For example, 300-series stainless steels have better ﬂuidity than 300-series stainless steels with the addition of molybdenum. Hence, whenever possible, the designer should allow a change in the material if that could lead to the desired casting conﬁguration at a reduced cost.

Designing for Economical Coring One of the major advantages of the metalcasting process is its ability to produce components with complex internal cavities (such as undercuts and skewed internal walls). The shapes used to form internal cavities within the casting are called cores, which can be made from different materials (sand, ceramic, and metal) depending on the application. The same principles that govern moldmaking also are valid for making cores. The cores are made in coreboxes and must be removed from the corebox after curing, so they cannot have undercuts that would prevent them from being removed from the corebox. In addition, deep pockets in the corebox must have a sufﬁcient draft angle to prevent high friction between the core and corebox during removal, which could cause a core to break. For deep pockets, additional loose pieces are used in the corebox to assist the core removal. Although cores increase the cost of castings, they provide a number of advantages. The most important is the shaping of internal passages,

which can be of almost any arbitrary shape and with nonuniform cross sections. Another advantage is shaping the geometry of the part to optimize wall thickness. Cores can provide casting geometry that can make stress distribution uniform throughout the part, thereby reducing its weight. This capability has not been extensively exploited in the past due to the complexity of stress analysis in highly complex shapes, but with the advancements in ﬁnite-element analysis tools, this design advantage of metalcastings becomes a signiﬁcant advantage. Cores also can form zero draft angles on castings and can be used to allow for undercuts on the vertical walls of the casting to reduce or eliminate secondary machining. The process used for making the core also has signiﬁcant inﬂuence on the core strength and its ability to hold under its own weight. For example, shell cores typically are stronger than oil-based cores and can allow more freedom with respect to core location within the mold. In addition, some geometries and coremaking processes allow the cores to be hollow, reducing the core weight and stresses. Cores can be placed within the mold cavity in a number of different ways, each offering certain advantages and also posing certain limitations. For example, a core for a bell-shaped casting could be in the drag section of the mold, with a portion protruding into the cope section. Alternatively, the core placed in the cope section may hang into the drag section. The ﬁrst case allows for almost any core size, while the second limits the core size because of the stress in a core induced by its own weight. Sand cores are fairly weak and can break if not supported properly. Thus, it is important to support cores within the mold cavity in such a way as to prevent their fracture and/or excessive deﬂection. Due to limited ﬂask size and gating system issues when the casting is very long, supporting the core may require horizontal positioning of the mold cavity and core within the mold. In this case, orientation of the core presents additional problems. The stresses caused by the weight of the core can be higher than the core tensile strength, thus causing the core to break. Even if the core does not break, the body force on the core will cause it to deﬂect downward before the metal is poured, while buoyancy forces will cause it to bend upward after the metal has been poured. To prevent or minimize these problems, designers redesign the casting to allow for additional support. The resulting stresses and deﬂection can be reduced by an order of magnitude with this simple modiﬁcation. If necessary, holes on the casting can be plugged after casting and cleaning. Another method to support the core in the mold cavity to lower stress and minimize deﬂection is to use chaplets. Chaplets are metal shapes made of the same material or compatible material to the one used for casting. Since they are placed within the mold cavity, chaplets become an integral part of the casting. Hence, the casting engineer must ensure tight bonding is achieved between the casting and chaplet.

110 / Casting Design and Performance The stresses and deﬂection of cores caused by the body and buoyancy forces are of great importance in manufacturing good castings. These two phenomena depend on the core shape, core material, nature of core support within the mold, and metal being poured. To determine the stress distribution and deﬂection of the core before and after the metal has been poured, one must apply the basic principles of statics, strength of materials, and ﬂuid dynamics to the particular core conﬁguration. The most important factor for maximum core length is the type of core support within the mold cavity. Vertical placement of the core within the cavity allows for longer cores than horizontal orientation (due to the elimination of a buoyancy force acting on the core). However, this orientation of the core within the mold cavity can cause damage to the mold during mold assembly. In the case of horizontally positioned cores, the core supported on both ends can be longer than the one supported only on one end. In the case of the cantilevered core, the minimum length of the core print should be 33%. In addition, the core print section should be of sufﬁcient mass to allow the core to remain in place for mold closure. The maximum allowable core length also increases as the core diameter increases (as strength of the core increases due to the increased cross section). The casting wall thickness also inﬂuences the core length (casting wall thickness decreases as the maximum recommended core length increases). This is caused by the decreased amount of metal in the mold cavity and thus reduced buoyancy force. The type of metal also has a certain level of inﬂuence on the recommended maximum length of the core. Steel has a higher density than most metals and results in a higher buoyancy force than aluminum and magnesium. As a result, the maximum recommended core length is smaller for steel than for aluminum and magnesium.

Core Design Principles Cored holes should be designed as simply as possible while satisfying functional requirements. Several basic principles exist that should be observed when designing cored holes: Assure the strength of the core during core-

making, handling, and pouring.

Provide sufﬁcient mass to the core sections

to allow proper cooling during solidiﬁcation.

Avoid thin casting sections formed by cores. Avoid or minimize the need for complex

core handling. This will result in a scrapped core or a defective casting if the core breaks during pouring, due to the high pressure exerted by the metal stream. In some cases, thin core sections can cause signiﬁcant rework on the casting and/or a scrap rate of more than 50%. When designing narrow holes in a casting, designers must allow for the possibility that the particular feature should be machined instead of cast to allow for additional elements to strengthen the core. Thin core sections also pose a problem for gating and feeding the casting. The ingate should be placed in such a way that the incoming stream does not impinge directly onto the thin core section; otherwise, the core could fracture. Having too many thin core features will limit the possible location for ingates and thus proper gating of the casting. In addition, too many thin core sections protruding to the surface of the casting will affect the location of a riser (if required to feed the casting during solidiﬁcation). Thin core sections surrounded by heavy metal sections should be avoided, due to the excessive amount of heat to which they are exposed without the ability to dissipate it properly. In this design, hot tears (cracks) can be expected at these locations. These cracks represent signiﬁcant stress raisers in the casting, reducing its fatigue strength. Gas-related defects also are common in these conﬁgurations, due to the inability to properly vent thin cores. The gas formed by a sand binder will not be able to dissipate through the mold material and could end up in the mold cavity and the metal itself. As a result, gas porosity could be expected adjacent to thin and wide core sections. Another factor is metal penetration, which can be a common problem if there is premature decomposition of the core binder. Last, cleaning of narrow cored sections can be a very difﬁcult, labor-intensive job. Designers must allow for an increase in the cored sections according to the speciﬁcation provided by the foundry engineer and according to the casting process and type of alloy. There is a minimum casting wall thickness that can be physically or economically achieved. As noted, the minimum casting wall thickness is governed by a number of different variables, including casting process, type of alloy, and pouring temperature. Potential problems in the production of thin-wall castings are misruns and cold shuts. The cores that form thin-wall sections on the casting should be avoided or redesigned to provide uniform ﬁll of the entire mold cavity.

and/or stacked cores.

Minimize the number of required cores.

Cores are typically made from sand and can be fragile, depending on the cross-sectional area of the particular feature on the core and the core manufacturing method. Long, thin sections on the core can easily be broken during

Design Conversions to Casting Welded assemblies, fabrications, or forgings may be done more effectively by casting. How does one spot a conversion opportunity for casting? This chapter describes how one can

identify components worth considering as castings for possible savings or improved performance. On the ﬂoor of manufacturing and assembly operations, countless components are made up of several stamped, wrought, or machined metal parts. Engineers everywhere stipulate fabrications and forgings without weighing all of their options. Could those parts be redesigned to a single cast metal component for improved performance? Is there a potential performance gain or cost-savings that would make the redesign viable? The choice of whether a component is best manufactured as welded, assembled, fabricated, forged, machined, or cast is based on the component geometry, production costs, and requirements in application. This section looks at these issues and provides a framework for analyzing all manners of manufacturing as possible conversion candidates.

Weldments and Assemblies Time to market can be a major focus, and components often are designed to be manufactured by the process(es) that will deliver them with the shortest lead time. This leads to original equipment manufacturers (OEMs) designing and manufacturing weldments and assemblies of weldments in-house, using their available capacity and labor and eliminating the demands of sourcing a component (such as design changes and pattern costs). Design engineers generally feel more comfortable designing building blocks of simple wrought shapes. This is the method commonly taught in structural design, and manufacturing engineering and purchasing colleagues are comfortable with bringing weldments and assemblies of weldments into production. Another quick-to-market option is to hog-out (machine) a shape from stock to produce a required component, simplifying the sourcing process to a one-step machining operation. Although both approaches deliver the product to the customer in a time-efﬁcient manner, they may not deliver it in the most cost-effective manner, particularly when volume increases. Weldments and assemblies of weldments create a signiﬁcant volume of part numbers to produce (or buy), schedule, and track. They require high levels of labor to ﬁxture (weld, bolt, align, etc.) and typically are not as dimensionally or structurally consistent due to the inherent variances in manufacturing. On the other hand, hog-outs have the advantage of known wrought stock mechanical properties (making them a popular choice for critical aerospace applications) but suffer from high costs, because a large percentage of the bar stock they are borne from ends up on the machining room ﬂoor. Because mechanical properties of casting alloys are not as well deﬁned nor as widely available in properties literature (casting mechanical properties depend

Casting Design and Geometry / 111 on the process and involve liquid ﬂow and solidiﬁcation gradients), some structural designers feel more comfortable with wrought metal properties. As a result, however, the opportunity to learn more about the casting structural properties is often missed, along with the prospects of lighter, stronger, and more cost-effective parts. In prototype situations when time is of the essence, weldments, assemblies, and/or machining are sometimes selected over casting to bring the product to market quickly. The key is to convert these fabricated components to the most efﬁcient manufacturing method when volume production is ready to begin. The ﬁrst thing to look for in a possible conversion to casting is a component with a complex geometry. This could be a single component that was machined or forged, but more often than not, the most impressive cost/ weight-savings is with a series of stampings and/or other wrought shapes that are welded and/or bolted together. Components often are designed with the building-block mindset during the prototype stages due to design familiarity. However, the building blocks usually can be streamlined to one-piece cast components. The type of complex geometry components that make good conversion candidates to casting often have a high surface-area-to-volume ratio. In addition, they have a high number of inches of weld, which leads to high fabrication costs due to the weld time, material, and complicated ﬁxtures. The complex geometry components often also have high aspect ratios, which mean that they are long with respect to their width (rangy components). These factors typically indicate where the most cost-savings potential is in a casting conversion. When a fabricated component has a high number of inches of weld in critical stress areas, a greater chance exists that problems in the weld itself or from microstructural effects in the neighboring heat-affected zones will cause failure in the ﬁeld. High-stress welded joints are generally less capable in overload instances or cyclic fatigue than junctions formed in the metalcasting process. Another factor to consider with weldments or assemblies of weldments is the number of separate parts necessary to form one component. Beyond the inconsistency in mechanical and dimensional properties developed throughout the multipiece fabrication, the OEM must inventory all the part numbers and absorb the labor costs to manufacture the component. In addition, the in-house manufacturing facility is responsible for ensuring dimensional consistency from part to part as well as maintaining a low scrap rate during production. Many of these costs are found in burden rates or are activity-based and go unrecognized in a cursory value analysis of component cost. The last factor to consider is ﬁnal component performance. The well-known mechanical properties of wrought metals are directional

(stronger in the wrought direction, weaker in the transverse direction) and may be compromised signiﬁcantly by welding. A properly produced, high-integrity casting with isotropic mechanical properties (equal in all directions) enjoys uniformity of properties in continuous sections as well as junctions. In addition, visual appearance must be evaluated. While fabrications offer smooth surfaces, their welded junctions are not as pleasing to the eye as the continuity of a complex cast shape. If a component or series of components have been identiﬁed for possible conversion to casting, the next step is to bring in a metalcaster to determine the of the component castability and possible cost/weight reductions. Typically, redesigns to casting aim for a 40% cost reduction from the weldment, assembly, or hog-out to overcome the labor and lead-time beneﬁts associated with the other manufacturing methods. The ﬁrst step in identifying a casting partner is to determine the material in which the redesigned component could be made. Weldments and assemblies of weldments often are fabricated from wrought steel shapes because carbon and low-alloy steel is the most easily weldable of metals and provides high toughness, strength, ductility, yield stress, and stiffness. The problem is that steel often is chosen solely on its weldability. If the component is going to be a casting instead of a weldment, then weldability may not be as important. Therefore, other materials should be considered. Potential questions to ask to determine the correct material include: What material is required for the application from a physical and mechanical property standpoint? Will the component be used in a severe service application with high fatigue and impact resistance, or is it a cosmetic part? Does it need to adsorb vibration? Does it need to control sparks? Does it need to resist corrosion? Does it need to be lightweight? Steel is the material of choice when high impact resistance and fatigue life are required for severe service applications. In addition, if the cast component is going to be welded to another component, then steel is a logical choice. Generally, however, the opportunity exists to replace a steel weldment with a casting from another alloy family, such as iron or aluminum. If the component is going to be used in a fatigue application but is not a severe service application, then iron may be the logical choice because of its castability (ease of casting). Ductile iron, in particular, has made a name for itself in the conversion of steel fabrications to castings, because it combines excellent castability with reasonable toughness. When weight is an issue without severe service application, aluminum often is the casting material choice. With new technology in alloying and molding, aluminum is increasingly becoming a material of choice in important fatigue applications, such as replacing stamped steel fabrications with squeeze-cast aluminum in automotive suspension components.

Regardless of choice of alloy family, the most signiﬁcant factor in the success of castings in structural applications is the use of geometry to control stress and stiffness. Only the casting process offers so much variety in shape at low cost. Even though some alloys are stiffer than others (for example, steel is three times stiffer than aluminum), it is stiffness from shape that makes castings a breakthrough metal product form. Once a material is selected, design engineers must then go to their purchasing colleagues to begin the search for a casting partner to assist in the redesign of the component to casting. Ask purchasing colleagues to ﬁnd casting sources who appreciate the importance of structural geometry in casting design. Such a casting producer will offer the most overall capability in cost reduction and component improvement when converting stamped steel fabrications, weldments, and weldment assemblies to metal castings. When examining a weldment or assembly for a possible conversion to casting, the following are seven considerations to determine the feasibility of the redesign: Geometry:

Is the component complex enough to warrant a redesign to casting? Length of welding required: The more length of weld needed to make up a component from several smaller components, the more likely it is a candidate for conversion to casting. Number of parts: Is the component made up of several smaller parts that have been attached together? If it is, would a single casting reduce the cost associated with part number inventory as well as free up in-house operations for other, more important manufacturing tasks? Dimensional consistency: How tight are the tolerances for the component? Is warping ever an issue? The inherent dimensional inconsistency with fabrications is eliminated with a single cast component. Scrap rates: Are scrap rates high for the component during in-house fabrication? If so, the component may be better manufactured out-of-house via a different production method. Appearance: Is the component visible? Does it require a clean, streamlined appearance with a ﬁne surface ﬁnish? Field problems: Does the component fail in use? Are the stresses exerted on it too great? Castings provide consistent properties throughout each component, eliminating much of the variability found in assemblies and weldments.

Fabrications Geometry can help identify successful casting design from former fabrications. Even for castings that do not serve a meaningful

112 / Casting Design and Performance structural purpose, geometry is critical. Geometry for casting design takes into account alloy castability, deﬂection, stress concentration, risering/gating/venting, and efﬁcient machining and assembly. As a result, well-selected casting geometries are capable of surpassing other methods, particularly fabrications, in making metal shapes that are limited by their process. Flexibility in geometry choice is more restricted among fabricating, as well as stamping, forging, or machining from stock. Since stress is a function of loads applied and geometry of structure (geometry alone controls the amount of stress for a given system of forces on a structure; material choice controls how much stress is allowable), castings make efﬁcient, cost-effective structures and allow the most freedom of geometry. In redesigning fabrications to castings, opportunities exist to use casting geometry to simplify design and reduce weight and cost by designing for more uniform stress throughout the part. Since fabrications are made from building blocks of uniform cross section, they tend to be overdesigned in sections away from the critical stress areas. They also tend to suffer from heat-affected properties and stress-concentration factors in the weld joint areas. By using engineering sketching with basic stress analysis techniques, design engineers can formulate an estimated casting design geometry ready for validation via ﬁnite-element stress analysis and casting process analysis using a solid model. Since casting design geometry selection requires brainstorming, sketching is the best way to ﬁnd a shape that is structurally efﬁcient, castable, and manufacturable downstream. A dimensioned, well-proportioned sketch also is an effective platform for building a solid model. Alternative design methods based on engineering sketching are viable for the following reasons. First, dimension-engineered sketches are probably the most efﬁcient starting point for construction of solid models. Second, sketching allows the integration of the four disciplines of good casting geometry:

Geometry of castability Geometry of structure Process geometry Geometry to minimize/reduce machining and assembly

cost

of

Third, a metalcasting geometry can be quickly sketched using the designer’s sense of scale, proportion, and load-carrying capability of shape and material. Last, reality checks of sections where stress and/or deﬂection are a concern can be quickly calculated using classic stress analysis; methods of estimating þ20% make this step practical. Computer solid modeling can validate or reﬁne the estimates. The method of sections, vector algebra, the principle of superposition, and methods of transforming stress and checking allowable transformed stress are all highly graphical

techniques. Nothing in stress analysis is more valuable than a simple sketch of the loads and reactions expressed in a free-body diagram. This methodology as a precursor, coupled with today’s model building and simulation technology, is the most effective pathway to efﬁcient structural metalcasting designs. Remember, this method uses engineering sense of scale and load-carrying capability combined with computer modeling to verify intuition. The goal is to obtain reasonable estimates of geometry that will be relatively uniformly stressed and will respond well to the alloy to be cast. The goal is not to compete with the computer; it is to help the computer ﬁnish the job quicker. When performing a component redesign from fabrication to casting, do not start from the fabricated form. Start with a clean sheet of paper and think only about what the functional areas and surfaces of the component design must be. The free-body diagrams at important slices in the overall structure come together to form efﬁcient casting design. A principle in stress analysis called superposition allows loads and reactions to be analyzed separately and then added up. Some of them are positive and negative, one partially canceling another out. It is very much like algebra—factoring the equation in clever ways makes the analysis easier. In a broad sense, the stress analysis of metal castings can be described as “the unique capability to easily and cost-effectively change area moment of inertia over section length.” This capability—plus the ready ability to soften surface shapes where stresses concentrate—underlies the power of castings to control stress and deﬂection in structural applications. Some important elements and steps of the principles of engineering mechanics for linearly elastic solids at rest are summarized as follows. Stick-Figure Sketches. Draw a picture of the loads on the structure and the reactions to the loads using stick ﬁgures. Recognize that these loads and reactions are vectors with magnitude and direction. They also work like algebra because þ signs are important. Using Method of Sections. This is the basis of free-body diagrams. Any loaded static structure can be sliced at any location in any plane, and reactions can be deﬁned at the slice to hold the remaining part of the structure in equilibrium. The reactions at the slice are the basis of classical stress analysis. Stress Correlates to Geometry. Stress is 100% dependent on geometry. Choice of material determines how much stress can be tolerated, but the amount of stress always is strictly dependent on shape. This simple fact is why metal castings are signiﬁcant as structural components. The choice of low-stress geometry is helpful in crack-sensitive structures because geometry reduces the crack propagation driving force. A misconception about using geometry to lower and improve the orientation of stress is that “lowering stress in one region of a structure

simply raises it in another region.” The fact is that increasing area moment of inertia lowers the stress at that section without necessarily affecting any other section negatively. Stress Correlation to Area. Areas are the key parameter in geometry to control shape for stress and deﬂection. The essence of area moment of inertia can be gleaned quickly from its base formula. Considering the axial formula, as cross-sectional area becomes taller in the Y-direction, the area moment of inertia (Ixx) becomes larger quickly because the distance (Y) is squared. At the same time Y2 becomes signiﬁcantly larger (if the cross section starts becoming wider), then Ixx increases even faster (dA is the change in cross-sectional area with each small increase in Y; when Y starts to become a large distance, a big change in area width becomes important). This formula explains why I-beams, with their skinny center web but wide ﬂanges separated substantially from the center, are good at controlling stress. Considering the polar form for the area moment of inertia (J), as the radius of a cross section increases, J becomes larger quickly because the radius is squared. For resisting torque and reducing torsional stresses, intuitively the area in the center of the shape is relatively meaningless. This explains why cylindrical tubes such as drive shafts and torque tubes are effective in controlling torsional stress. Deﬂection. Controlling deﬂection involves both stiffness from area moment of inertia and choice of material. Stiffness of metalcasting alloys (or any material) is deﬁned by Young’s modulus, also called modulus of elasticity. Formulae for deﬂections, even for relatively simple beams, are complex-looking equations. However, the relationships for limiting or allowing deﬂection are as simple as the formulas for bending stress or torsional stress. An interesting example of these relationships in deﬂection is the design of a cast aluminum front suspension component for an automobile. The modulus of elasticity for aluminum is relatively low, 10 106 lbf/in.2 (68,900 MPa). Therefore, the objective should be to keep deﬂection low. To do so, the formula calls for the casting geometry to be “stubby,” that is, short in section length relative to stiffness in area moment of inertia. On the other hand, steel castings often are used as stress nodes in complex welded steel fabrications. They are designed to absorb stress and allow deﬂection in a way that protects the welds, particularly from fatigue stress. In this case, the casting design can take advantage of the high modulus of elasticity of steel (30 106 lbf/in.,2 or 206,800 MPa), and the shape can be rangy, with longer sections relative to the section area moment of inertia, allowing ﬂex. Simplifying Stresses with Mohr’s Circle. Using Mohr’s circle, complicated stress combinations can be transformed into the maximum

Casting Design and Geometry / 113 shear stress and principal stresses (maximum tensile or compressive stress) to ensure that stresses are kept below safe limits for the material involved. If a stress is found to be too high for static overload, fatigue life, or propagation of known or potential cracks, simply change the cross-sectional area of the casting (even a dramatic change in cross section) so the maximum shear stress and/or principal stresses are low enough to be safe. Mohr’s circle and Tresca’s hexagon (which checks maximum shear and principal stresses against allowable stress) are graphical methods. They ﬁt the overall technique of using sketches to get in the “good casting design ballpark” and let the computer ﬁnish the job. Steel castings and fabrications can seem interchangeable because they share many qualities; cast steels and wrought steels have such similar mechanical properties that the American Society of Mechanical Engineers Code does not differentiate between steels on the basis of their manufacturing process but by their chemical composition. However, noticeable differences do exist between the two that can affect the design and cost-effectiveness of a component. Wrought products (rolled or forged) exhibit a characteristic known as directionality. This characteristic, also known as anisotropy, means that a component has strength and ductility in the working direction but has lower transverse properties. Cast steel products do not exhibit directionality; rather, they can be described as isotropic. Steel castings can be stressed in any direction without concerns over the lower strength, ductility, and toughness that are exhibited in the transverse direction of wrought products. Designers of fabrications must be aware of the directional properties and incorporate them into the component design, or it could become overstressed when a load is applied in the transverse direction. When compared to their wrought equivalents, the mechanical properties of cast steels are approximately the average value of the longitudinal and transverse directions in the wrought product. Transverse properties always are lower than the properties obtained in cast steels, which means that casting offers more design ﬂexibility than fabrications. Properties for wrought products are determined by performing mechanical tests in the longitudinal direction. Tests in the transverse direction usually are performed only when specially requested. Steel castings are tough and ductile, contrary to the common belief that they are brittle and subject to abrupt failure. Brittleness is a function of metallurgy, not of process, and steels are not brittle alloys. Because steel castings are isotropic, uniformly heat treated, and more stress relieved than a fabrication, longer fatigue lives and more deformation without sudden failure are common in castings. One factor that must be considered in any analysis of the design of fabrications using

wrought products such as bar plate or tube is that the welds usually are placed in the highest stressed location in the component. Unless the fabricator follows a design code, most welds are placed at high-stressed section changes or design features such as corners, which limit the load-bearing strength of the component. To combat this, welded fabrications often can be redesigned to one-piece castings. Then, the casting, which welds as well as or better than the fabrication, can allow the designer to locate the welds away from the highly stressed areas. Often, the weldability of the materials used in fabrications is taken as a given. They must be weldable, otherwise they would not be used. In addition, there also is the misconception that wrought steels are easier to weld than cast steels. Cast steels have a lower susceptibility to underbead cracking—the cracking that occurs from the introduction of hydrogen into the liquid metal. The isotropic qualities of the casting provide a good (not too hard) welding surface. Welds made in the production of castings are almost always stress relieved. Castings tend to undergo heat treatment (a strengthening process) after welding occurs, keeping the component strong even in the weld areas. Fabrications usually are not stress relieved after welding, which can lead to a weaker area around the weld. Field welds can be difﬁcult to stress relieve due to the location or size of the part, but if the component is a casting, it can be designed to place the unstress-relieved weld in a lower stress location, which improves the component life. It has been shown that when castings are used at junctions such as a node on an oil production platform, the weight of the connecting area can be reduced by as much as 50%, stemming from the elimination of joints and welds to attain the required strength. In addition, the position of the welds also is moved out of the high-stress area, and the welds become simpler (circumferential instead of complex and irregular in form). Placing a weld in a position that makes it easier and simpler to perform also can minimize the nondestructive testing required and the number of discontinuities or stress raisers associated with welding in difﬁcult positions and inaccessible areas. Designers sometimes say that it is easier to design fabrications from plate and bar because it is easier to visualize a component made as a series of right-angled connections. Visualizing a component where there is total freedom of form can be more difﬁcult—it requires the designer to think in three dimensions. Yet, this freedom permits designers to design only what the component needs—no extra material, edges, or welds. Conventional fabrication generally is a compromise between material availability, fabrication capability, design codes and engineering requirements. However, creative design of

castings often will enable steel sections to be tailored to meet speciﬁc engineering loads, thus improving engineering efﬁciency. This often leads to substantial engineering beneﬁts and weight reduction elsewhere in the surrounding steelwork through the elimination of offset work points and their associated bending moments. Alignment or dimensional problems due to production are likely to be greater with fabrications than castings, because of the distortion that may occur during manufacture. Straightening operations carried out on components can have a more detrimental effect on fabrications because they will be plastically deformed in the high-stress areas. These high-stress areas often are associated with the welded joint. Although castings also will be plastically deformed during straightening, the location of the deformation is unlikely to be in a notched area such as a weld bead. Because steel castings typically are welded into a larger fabricated steel structure, it is important to consider deﬂection. Allowing steel castings to ﬂex protects the weld joints in the base structure from fatigue failure. The combinations of section length, depth, and cross section in geometry permissible with the high stiffness and high yield stress capabilities of steel can resist fatigue and protect mating sections from early fatigue failure. Allowing a steel casting to ﬂex can reduce the stress concentration in the casting weld connections to a base structure. Identifying an appropriate cast shape opens the door for designers to the opportunity for novel designs and shapes that challenge traditional fabricated concepts. Casting design should simplify the component shape as far as possible, satisfying the basic engineering requirement while reducing the overall size of the component where possible. This can be achieved by eliminating or adjusting offset work points, local stiffening, and deepened sections in plate, box girders, and tubes—all of which are necessary in fabrications to achieve acceptable designs. When designing, it is important to establish geometry that will be workable during secondary operations. This includes casting designs for welding into a larger fabricated assembly, considering the design of weld-joint geometries (compatible mating of casting and plate thicknesses and stress distribution of weld geometries) and thinking about assembly features. For example, with fabrications the capability to position and hold several pieces of plate in a weld ﬁxture is more difﬁcult than forming a mold cavity for casting. Castings allow designers to buy shape cheaply. If the desired component is mostly steel and the costs involved are mostly material, a fabrication should be used. However, the more components, the more linear inches of welding required and the more machining required per individual component, the more attractive casting becomes. Designs or

114 / Casting Design and Performance fabrications that comprise the most pieces and the most welds are ideal candidates for a casting conversion. In general, castings provide tighter tolerances and better mechanical performances. They allow the designer to shape the component exclusively for the project at hand, with no extra pieces or sections. Complex components and assemblies that can be consolidated into fewer parts become cost-effective as castings. Limiting assembly reduces cost. Castings often weigh less because the geometry can be tailored to the actual component requirements instead of being restricted by the capabilities of bars and sheets. The cost bases for fabrications and castings are different. An increase in steel thickness, shape complexity, and stiffening in a fabrication pushes up costs because of the amount of welding and nondestructive testing necessary. This also can be affected by the increased risk associated with the stress relief of highly restrained, heavy sections. Conversely for castings, castability is enhanced with increased section size, and with optimal design, the cost/ton will decrease with increased weight.

Dimensional Control Dimensional accuracy is a critical factor to be considered through the casting design, tool making, pouring, and postprocessing of the casting part. To control these sources of error, the mechanism of each error source must be understood.

Metal Shrinkage Shrinkage is a leading error source in casting. Different materials may have different contraction behaviors and thus need different shrinkage compensation factors. Shrinkage or contraction also occurs in the following stages: Contraction in volume of the liquid metal as

it cools to solidiﬁcation temperature

Shrinkage or decrease in volume, as the liq-

uid metal changes state and becomes a solid

Contraction of the solid metal as it cools fur-

ther, to room temperature

Metallurgical phase changes in the solid

state that may be accomplished by volume changes, for example, in the case of ferrous castings, the austenite-to-pearlite transformation Volume and dimensional changes that may occur if the casting is heat treated, to modify its properties To improve the accuracy of shrinkage compensation, it is necessary to determine the real sources of contraction and theoretically understand the shrinkage behaviors. Volume Contraction of Liquid Metal. This is of no practical signiﬁcance, because castings are made with risers and gating systems that contain a reservoir of molten metal that can

enter the mold cavity to replace volume changes, due to liquid contraction. Solidiﬁcation Shrinkage. The change (decrease) in volume in passing from the liquid condition to the solid condition cannot be avoided. It varies, according to the type of metal, and in some gray cast iron (hypereutectic) is almost completely offset by an accompanying expansion, resulting from graphite being deposited from solution. In the case of metals such as steel, white irons, or aluminum, if may be as high as 6% by volume, and this calls for an adequate supply of feed metal, if shrinkage cavities or porous areas are to be avoided. Crystallization proceeds from the cooler outer surface of a casting, and crystal dendrites grow in a pool of liquid, which supplies the intrametal to overcome the volume change. The problem occurs in the latter stages of solidiﬁcation, when insufﬁcient liquid metal is available, and the viscosity of the liquid is so high that it cannot adequately ﬂow into the area between growing crystal dendrites. The result is a shrinkage cavity. The metalcaster may control the solidiﬁcation rate, so that the last area to solidify is located in the riser or feedhead, which is subsequently cropped from the casting through several techniques: use of chillers, adjusting the speed of gating, and directional solidiﬁcation to avoid trapping the shrinkage cavity in a critical area. Directional solidiﬁcation occurs when the molten metal in a casting solidiﬁes in such a manner that liquid feed metal is always available for that portion that is just solidifying. When using a riser or feedhead to provide feed metal, the operator always makes this riser considerably heavier than the section of the casting that has to be fed. Doing this ensures earlier solidiﬁcation in the casting and later solidiﬁcation of the riser, which can then provide the liquid feed metal. The engineer must realize that if the casting is too complex and that if the foundry operator has to resort to too many extra steps to provide a sound casting, then the casting may cost a little more than it would if it were less complex and if it were designed allowing for these metallurgical phenomena. Sound, accurate castings of high integrity begin with a proper design, which can be properly cast in the ﬁrst place. Solid Metal Contraction. This contraction is what the foundry allows for when a shrinkage rule is used for the proper dimensions of pattern equipment. The shrinkage of different metals varies and may be as high as ¼ in./ft to as low as 1/10 in./ft. The appropriate shrinkage rule will be used in dimensioning the pattern; that is, in a 12 in. long casting, for example, to be cast in a metal with a 1/8 in./ft shrinkage, the pattern will actually measure 121/8 in. Unfortunately, shrinkage of any metal is uniform in all directions; thus, a long, thin casting may tend to shrink more in terms of length than it will in terms of width or thickness. Only experience and general foundry practices dictate exactly how and where the shrinkage will occur. The only practical way to take care of

this problem is to provide generous machine allowances or be a little more realistic in the tolerances speciﬁed in a casting. Where it is imperative that high accuracy is required in the ﬁnal casting, patterns must be built and prototypes must be cast, so that the shrinkage can be accurately determined and allowed for in that speciﬁc casting. One of the biggest problems with solid contraction lies in the development of residual stresses and strains in the casting. Obviously, it is impossible for all areas of a casting, particularly a complex one, to solidify uniformly at the same time. The position is created where thin ribs, for example, may solidify and contract before adjacent, thicker sections have experienced much of their contraction. The net result is the introduction of differential shrinkage and residual stress into the castings. This is the chief reason why complex castings requiring a high degree of accuracy and precision should be stress relieved. Stress relieving is the procedure of subsequently heating the casting into the plastic range, whereby it can deform and accommodate the stresses that have been built up during initial solidiﬁcation. Where the design is particularly bad and where contraction is hindered, the ever-present problem of cracking and actual breakage can manifest itself. Metallurgical Solid Changes. Many metals, particularly the ferrous ones, go through phase changes in the solid condition that involve dimensional changes; for example, cast iron and steel will transform from austenite (stable phase above 1330 F) to pearlite or ferrite during cooling, with an accompanying change in volume. From the foundry standpoint, these changes produce and severely aggravate casting stresses. Here again, all the foundry operator can do is persuade the casting designer to avoid abrupt section changes or to use stress-relieving heat treatments in order to remove residual stresses. In many cases, these contraction and solid-state changes are predictable and will result in a uniform degree of warping and distortion in a casting. If this is uniform, then it can be allowed for by suitable compensating variations in the dimensions of the original pattern equipment. Volume and Dimensional Changes in Heat Treatment. When a solid casting is heated to anneal or otherwise heat treat it, it is always, except perhaps in the case of stress relief, heated to a temperature above that where the metallurgical solid-state changes occur. This means it is going to go through a temperature range where volume changes will occur and where stresses will be reintroduced into the casting. Here again, the only resort is to heat and cool slowly and carefully, to avoid abrupt changes in temperature. The designer must understand and allow for the fact that the casting will be heat treated, and therefore, the design must avoid complicated and abrupt changes in casting sections. In some heat treatments (for example, where a white cast iron or a partially white cast iron, such as malleable iron, is annealed to remove carbides and replace them with graphite), the dimensional changes that occur will be permanent. Usually,

Casting Design and Geometry / 115 graphitization is accompanied by a growth in dimension, which, again, must be allowed for in the original pattern design. Metals that are white and as-cast, such as malleable iron, may often be made from a pattern made to a standard rule, because the contraction during cooling in the mold may be completely offset by the graphitization during subsequent heat treatment.

Dimensional Changes of Casting Molds During the casting process, it stands to reason that some dimensional changes of the mold can occur when a mold cavity is heated by molten metal. Fortunately, these are quite small and are insigniﬁcant in the scheme of things. Volume changes can result from the pressure of molten metal on the mold cavity coupled with a decrease in strength of the mold from heat. This strength decrease causes what is commonly called mold wall movement. Most sand molds are made from silica, which exhibits an expansion as it is heated. Unfortunately, at a temperature of approximately 1000 F, it undergoes a phase change that involves a slight contraction. Therefore, it is possible for certain areas of the mold to be expanded while other areas are contracted. This sets up severe strains in the molds, and the foundry operator must allow for the strain by the degree of mold ramming, the selection of the grain size of the silica sand, and by the judicious use of binders that may contract and offset the silica contraction. In sand casting, many individual processing steps are needed to produce a sand mold. Such steps include sand preparation, coremaking, moldmaking, mold and core assembly, and ﬁnishing. Each of these steps, as well as other factors such as pattern wear, will contribute to the overall dimensional variability of casting features. When discussing dimensional errors or variability, it is important to distinguish between dimensional accuracy and dimensional variability. Dimensional accuracy—an indication of how close the casting dimension is to the actual target value— is often referred as a system error. The main causes for poor dimensional accuracy are pattern equipment errors, pattern wear, and casting contraction uncertainty, much of which can be corrected before a production run. Dimensional consistency—the variation of individual casting dimensions about the mean casting dimension—is often referred to as random error. Many metalcasting process variables contribute to the dimensional inconsistency of castings. It is therefore important that signiﬁcant sources of dimensional variability are identiﬁed and that appropriate process controls are established to minimize them.

Design Case Studies Following are three different metalcasting design projects and how they went from idea to design to production.

Example 1: Design Conversion to Ductile Iron Casting for a Pour Chute Pivot Frame Integrating a multipiece part into a casting is often done to reduce assembly, machining, and management costs. However, the decision to convert a part to a casting is only the ﬁrst step. During the conversion process, design engineers must make several key decisions in order to achieve the most optimal design. The best material and casting process must be chosen, design iterations should be performed with the casting process capabilities in mind, and time should be spent on discovering whether more advantages can be obtained through metalcasting—either by enhancing the part performance or reducing part cost further. A manufacturer of concrete mixer trucks for the heavy construction industry faced these choices when it converted a portion of its pour chute to a one-piece casting after a proposed redesign showed it would reduce costs and improve the part. The casting design team consisted of representatives of the track manufacturer, metalcasting facility, pattern shop, and machine shop and focused on three imperatives: design for performance, design for production, and design for cost. Three casting design issues played a major role in meeting these design imperatives:

Meet or exceed the strength and durability of

the original weldment

Choose an alloy grade for sufﬁcient tough

ness and ductility for fatigue and impact resistance Design the casting with the selected alloy grade to keep stresses below the maximum design stress Minimize machining steps and costs Include the critical dimensional features— machined center hole in the hub, machined mounting holes in the triangular frame, and as-cast locking holes on the locking plate Achieve general tolerances of 0.75 mm (0.03 in.), as-cast hole tolerances of 0.5 mm (0.02 in.) and machined tolerances of 0.13 mm (0.005 in.) Achieve a clean and smooth surface ﬁnish with no visible surface defects or grinding marks

The resulting one-piece ductile iron casting weighed 18.14 kg (40 lb) and consisted of a central hub, a ﬂat locking plate, a square mounting frame, and a triangular support frame (Fig. 8). The envelope dimensions of the pivot frame casting were 48.3 by 35.6 by 15.2 cm

Select a ductile iron grade that meets the

mechanical requirements

Optimize the design for stress reduction and

weight savings

Develop a casting tool design that produces

ﬂaw-free frames at the best cost Knowing the Application. The cement truck components must be rugged and durable to withstand mixing and pouring stresses as well as road shock and vibration over the life of the truck. The metal pour chute on the rear of the truck carries the concrete mix from the drum down to the concrete forms. The chute is designed to swing across a 160 arc to facilitate easy pouring into the forms without repositioning the truck. An integral part of the pour chute is the pivot plate frame. Originally, the pivot plate frame was a steel assembly consisting of nine fabricated and machined parts that were welded together (Fig. 7). The welded assembly performed to mechanical speciﬁcation but was less than ideal for cost, assembly, variation in form and ﬁt, and inventory management. A near-net shape one-piece casting design was proposed that would cost less, improve form and ﬁt, and reduce stresses for longer life and improved durability. Part Requirements. The pivot plate frame supports the pouring chute of a concrete mixer truck, serves as a pivot point, and locks in position during the pour. The critical and production requirements for the cast plate frame were:

The original welded pivot frame assembly, shown here with the pivot shaft attached, consisted of nine pieces welded together.

Fig. 7

Fig. 8

The one-piece ductile iron casting resulted in a 50% production cost-savings.

116 / Casting Design and Performance (19 by 14 by 6 in.), the center hub had a diameter of 75 mm (3 in.) and a height of 90 mm (3.5 in.), and the minimum wall thickness was 9.5 mm (0.375 in.). The center hub of the plate frame secured the pivot shaft on which the plate and chute rotate. The locking plate had seven lock holes for securing the chute at different swing angles, and the two bushings on the top of the triangular frame secured the pouring chute to the pivot frame. Choosing the Material. The original assembly was welded from steel plates and bar stock, but ductile iron was chosen for the casting because it met the performance requirements while featuring a lower casting production cost compared to a steel alloy. Ductile iron exhibits a linear stress-strain relation, a considerable range of yield strengths, and ductility. It was clear that ductile iron was the metal of choice, but the casting design team had to choose among three grades of the ductile iron speciﬁed in ASTM A536 (Table 3). The ASTM A536 ductile iron casting speciﬁcation is based on demonstrated mechanical properties that depend on the proper cast iron microstructure. High-ductility grades have a ferritic microstructure, intermediate grades have a mixed ferrite and pearlite microstructure, and high-strength grades have a primarily pearlite microstructure. The performance and production requirements of the pivot frame called for an alloy grade with sufﬁcient toughness and ductility for fatigue and impact resistance. The microstructure of A536 grade 2 is primarily ferritic, so it has more than enough ductility/elongation and just misses the hardness requirement. However, it does not meet the minimum requirements for ultimate tensile strength and yield strength. Both A536 grades 3 and 4 feature mixed ferrite/pearlite microstructures, and both meet and exceed the requirements for tensile strength, yield strength, and hardness. However, grade 4 falls short on ductility and elongation. Because A536 grade 3 fulﬁlls all the requirements, including ductility, it was selected as the bestchoice alloy for this casting. Optimizing Design with Finite-Element Analysis. One of the design advantages of metalcasting is the freedom to optimize cross sections and shapes beyond the limits of welded

Table 3

plates and bar stock. The design team used this advantage to design the triangular support frame of the pivot plate frame for the pouring chute. The new design needed to meet two objectives: Keep the tensile stresses in critical sections

below the calculated stresses in the original welded design Minimize the weight of the overall casting to keep production costs down and save weight on the truck Three-dimensional computer-aided design is the key to rapidly optimizing the design for mechanical performance and weight reduction. This also reduces the ﬁrst-part time. The design team developed three designs and used ﬁniteelement analysis to evaluate and optimize the stresses in the triangular support frame (where the stresses were highest). The designs included a plate with perimeter ribs, a frame with cross ribs, and a plate with ribs and cutouts (Fig. 9). Design A, which used a ﬂat plate in the support section with perimeter ribs for stiffening and stress control, met target stress requirements at the joint where the triangular frame met the straight bars, but the weight of the design was higher than desired. Design B, which used a heavy frame in the support section with cross ribs for stiffening and stress control, markedly reduced the weight, but the stresses were excessive at the joint where the triangular frame met the straight bars.

Example 2: Design Conversion to a Die-Cast Aluminum Fan Housing

Grades of ductile iron alloy A536 A536 grade 2

Properties

Ultimate tensile strength, ksi Yield strength, ksl Elongation % Hardness, BHN

A536 grade 3

A536 grade 4

Performance requirement

65-45-12

80-55-06

100-70-03

>80

65 min

80 min

100 min

>55

45 min

55 min

70 min

>6 >195

12 min 156–216

6 min 187–255

3 min 241–302

Design C used a ﬂat plate in the support section, with perimeter ribs and two long ribs in the back for stiffening. In addition, three cutouts in the center panel reduced the overall weight without increasing stresses in the critical sections. This design was selected as the ﬁnal design because it kept the stresses within limits and met the weight target. Rapid prototyping using a high-speed 3-D printing system was used to check form-ﬁt function on the ﬁnal design. Saving in the Final Steps. Green sand molding was chosen as the best-value mold method to cast the pivot frame chute. Metalcasting engineers used orientation and risers to produce the desired directional ﬂow in the mold. After casting, the component was shotblasted to remove residual sand on the surface, ﬂash lines and a riser stub were ground off, and the casting was painted prior to machining. Three features on the pivot frame casting required separate machining steps. The inner diameter (6.35 cm, or 2.5 in.) of the center bore was rough drilled and ﬁnish drilled to speciﬁcation. The inner diameters (2.26 cm, or 0.89 in.) of the two bearing holes were ﬁnish drilled to speciﬁcation. The new casting, which converted a ninepiece weldment into a one-piece component, produced a 50% cost-savings on each part, compared to the original assembly, welding, and machining costs. The tool payback was four months of production. The one-piece design reduced part inventory, technical data management costs, and production-to-delivery time. Quality-management principles were applied at each stage of the casting process. Emission spectrographic analysis of the furnace and ladle chemistry assured precise control of the alloy composition, dimensional checks were performed on all features, visual examinations tested the surface appearance, and soundness checks were performed with sectioning on prototype castings. The converted component featured better form and ﬁt with less dimensional variation, lower stresses, and longer life, compared to the weldment.

Three designs were developed for the triangular support frame of the pivot frame (clockwise from top): Design A featured a plate with perimeter ribs; design B featured a frame with cross ribs; and design C featured a plate with ribs and cutouts.

Fig. 9

In cold climates, tents, sheds, and buildings without central heating use individual space heaters to provide warmth. However, the hot air from the heaters rapidly rises to the ceiling without actually warming the living space of these uninsulated structures. One solution is a thermoelectric fan (TEF), which directly circulates warm air throughout the shelters. This is an efﬁcient way to heat the whole area without being a drain on the power source. The TEF design hinged on the production of a die-cast fan housing, which provided structural support and protection for the fan, while also acting as a heat sink and thermal heat

Casting Design and Geometry / 117 conduit. Product designers from the TEF manufacturer worked closely with the diecasting supplier in each step of the casting design to achieve a ﬁnal concept that was functional and castable. Knowing the Requirements. One of the most important elements of the TEF was its ability to conduct heat efﬁciently, so thermal properties of the diecasting were of top concern. The TEF sits on a space heater and circulates the hot air from the heater downward toward the ﬂoor before it rises to the ceiling, reducing thermal variations in the shelter and providing warmer conditions for the occupants. With the TEF, the space heater is able to run at a lower fuel rate while achieving the proper temperature levels throughout the enclosure. The bottom surface of the TEF is heated by direct contact with the hot stove, absorbing a small amount of heat that is converted by a solid-state thermoelectric generator into electricity, which drives an axial motor. The fan is directly attached to the upper shaft of the motor. No other external power source is needed. The fan housing is a 35.6 cm (14 in.) diameter, open-top cylinder with a complex slotted and ﬁnned base (Fig. 10). The cylinder is 25.4 cm (10 in.) tall and weighs 3.4 kg (7.5 lb). The wall thicknesses of the housing range from 2.24 cm (0.88 in.) at the base to 0.31 cm (0.124 in.) on the sidewalls. Slots in the base provide paths for the forced air to exit close to the hot surface of the stove, picking up heat and moving down away from the fan. In addition to high thermal conductivity to maximize heat transfer, the casting needed to have controlled wall thickness and uniform metal ﬁll for structural integrity and rigidity. The machined features of the housing had to be free of porosity for good thermal conductivity and to prevent off-tolerance drilling and tapping. Dimensional tolerances for precisionmachined features needed to be þ0.127 mm (þ 0.005 in.) to meet form-and-ﬁt requirements. The diecasting process was chosen to produce the fan housing because of its ability to

produce complex shapes with close tolerances, limited need for machining, integrated fastening elements such as bosses and studs, and its ability to produce castings at high rates with low cost. Annual production for the housing was 1500 units. The diecasting engineers had three design focuses: performance, castability, and cost, which hinged on a few design issues, including: Choosing a metal alloy that best met the per-

formance and castability requirements

Optimizing the diecasting process to reduce

porosity

Tailoring the thermal management process

for quality and reproducibility Choosing an Alloy. The efﬁciency of the thermoelectric generator depended on establishing the temperature difference between its top and bottom. So, the aluminum housing had to act as a heat sink, pulling the maximum amount of heat away from the top of the generator and transferring it to the fan air. This required an alloy with high thermal conductivity for the housing. A range of different metal alloys is available to the diecasting engineer, and the performance requirements and manufacturability issues must be weighed together to ﬁnd the right ﬁt. In this case, the designers needed an alloy with high thermal conductivity, but for mechanical performance, they also needed it to be both strong and stiff. Because the housing had a machined mating surface and numerous screwholes to be drilled and tapped, an alloy with easy machinability was sought. From the diecasting perspective, the thin walls of the ﬁns and ribs in the base required an alloy that would ﬂow and ﬁll the die cavity rapidly and easily but not solidify too quickly. Early solidiﬁcation in the die can produce casting defects, such as lack of ﬁll and cold shuts. The design engineers determined that three types of metal alloys could be considered for this application: zinc ZA8, aluminum A380, and the proprietary aluminum Thermalcast 130

of the diecasting facility (Table 4). All the alloys met the mechanical property requirements of strength, ductility, and modulus. ZA8, however, had a thermal conductivity 4% short of the requirement and weighed 150% more than the aluminum alloys. A380, a common diecasting alloy, featured an ultimate tensile strength of 47 ksi, a yield strength of 23 ksi, and a density of 2.63 g/cm3, (0.095 lb/ in.3), but the thermal conductivity was 20% lower than the design target. The Thermalcast 130 aluminum alloy had a high thermal conductivity of 130 W/mK, with the mechanical properties of conventional aluminum diecasting alloys. As a result, it was chosen as the metal to use in casting the fan housing. Controlling Metal Flow. After choosing an alloy, the diecasting engineers determined the best ﬂow control and gate design for the diecast component. The diecasting facility uses a vacuum-assisted process in which vacuum evacuation of the die cavity reduces gas entrapment during metal injection and decreases porosity in the casting. The result is a higherquality casting; however, the vacuum system is not a substitute for good diecasting design in the engineering of the die cavity, runners, gates, and overﬂows. Control of metal ﬂow in the die is a key factor in producing sound castings. Metal must ﬂow rapidly and uniformly into the die, minimizing turbulence and entrapped air. A key feature in die design is the positioning of the runners and gates—the passages that feed molten metal into the die cavity. Well-designed gates are positioned to permit rapid ﬂow into the die cavity, minimizing turbulence and long ﬂow paths for the molten metal. For the fan housing, the diecasting engineers opted to position the gate at the center of the cavity, directly into the center hub of the casting (Fig. 11). With the centershot, the metal ﬂows uniformly and rapidly from the center hub into all the thin-wall ribs and outer rim, producing complete ﬁll. Producing the Final Part. After a ﬁnal gating design was achieved, the housing was cast in a three-plate die (Fig. 12) with a cold chamber

Table 4 Properties of zinc ZA8, aluminum A380, and proprietary aluminum Thermalcast 130 Properties

The fan housing diecasting features wall thicknesses that range from 0.88 to 0.124 in. The interior view of the fan housing shows the slots of the base that provide paths for the forced air to exit close to the hot surface of the stove, picking up heat and moving down away from the fan.

Fig. 10

Density, lb/in.3 Thermal conductivity, W/mK Ultimate tensile strength, ksl Tensile yield strength, ksl Ductility, % elongation Modulus, Msl

Zinc Target ZA8

Al Thermalcast 130

Al A380

0.095 130

0.096 96

<0.1 >120

0.24 115

>40

54

46

47

>20

42

24

23

>3

6–10

4

4

>10

10.2

10.3

10.3

118 / Casting Design and Performance gating system. Once the housing was cast and trimmed, it was checked for dimensional tolerances and surface condition. Then, it was bead blasted and machined. The base hub of the housing was ﬁnish machined to a 32 root mean square ﬁnish for good thermal conduction. Finally, the necessary mounting and assembly holes were drilled and tapped, and the housing was ﬁnished with a protective black powder coat for corrosion protection and appearance.

Example 3: Coil Rotor for Traction Motor of Railroad Wheel Assembly This example is a step-by-step look at the decision process for casting of a coil rotor for a diesel locomotive engine. Several decisions are needed on how the part will be cast that may affect the ﬁnal product, even if the design may specify casting as the production method. The casting design engineers for this component focused on three imperatives:

The diecasting designers opted for a centershot gate for molten metal ﬂow. This allowed metal to ﬂow uniformly and rapidly into all the thin-wall ribs and outer rim.

Fig. 11

Fig. 12

Design for performance Design for production/castability Design for cost

To meet the design imperatives, engineers had to select a steel composition that met the performance and casting requirements, choose a molding system that achieved tolerance and cost targets, plan an orientation in the mold that optimized the precision features in the casting, and design a gating and riser system that ensured soundness in the casting. The armature coil rotor is an integral part of each traction motor used in a diesel locomotive. A diesel locomotive is equipped with a turbocharged diesel engine that delivers the shaft horsepower to the main electrical generator to convert into electrical power to drive the locomotive. The electrical power from the generator then is distributed to the traction motors, each of which converts the electrical energy into torque on the drive wheels through a reduction gear case. The armature coil rotor has a complex shape disc on one end of the rotor shaft, which serves as an end plate containing the windings and acting as a structural feature. The coil rotor disc is the most complex casting in the rotor assembly. The 38.1 cm (15 in.) diameter disc weighs 54.4 kg (120 lbs) and has a slotted and ribbed main body and a 14.5 cm (5.7 in.) high, 17.8 cm (7 in.) diameter center hub with a 11.9 cm (4.7 in.) inside diameter center hole. The outer rim is 7.6 cm (3 in.) high and 1.27 cm (0.5 in.) thick. The slots in the disc aid in cooling the windings, and the disc has a range of section thicknesses from 4.83 cm (1.9 in.) to as thin as 0.48 cm (0.19 in.) in the ribs. The nominal performance requirements for the wheel are: Ultimate tensile and yield strengths of 60

and 30 ksi

Elongation of 22% and reduction of 30%

Hardness of 120 to 163 BHN Surface ﬁnish of 250 to 500 root mean

square on as-cast surfaces

Dimensional tolerances of þ0.08 mm (+/

0.003 in.) on machined surfaces high-magnetic-permeability steel for magnetic performance

Low-carbon,

Alloy Selection. Due to the range of section thicknesses in this casting, an alloy with excellent ﬂuidity was required. Also needed was an alloy with low manganese, phosphorus, and sulfur content to minimize the magnetic hysteresis losses. The original equipment manufactures speciﬁcations originally called for a low-carbon steel (1015) based on the mechanical requirements. However, that speciﬁc carbon level in the alloy has insufﬁcient ﬂuidity at the pour temperature for smooth and rapid metal ﬂow into this complex mold. The metalcasting facility suggested an alternate alloy (1020) that contained a higher carbon level while possessing the same levels of manganese, phosphorus, and sulfur without the loss of mechanical properties. The higher carbon level reduces the melt temperature and gives improved ﬂuidity in the mold at the pour temperature. Choosing a Molding Method. This casting could be produced by two types of molding methods: green sand or shell mold. Each of the two molding methods has relative capabilities, advantages, and costs, as shown in Table 5. After comparing the two molding methods according to the requirements for surface ﬁnish, section thickness, and heat-transfer efﬁciency for the support rotor, it was determined shell molding was the best choice. Green sand molding can meet the baseline requirements for the rotor as-cast tolerance, surface ﬁnish, detail level, and tool cost, but it does not meet the requirement for minimum wall thickness or heat-transfer efﬁciency. Shell molding, on the

The fan housing was cast in a three-plate die with a cold chamber gating system. Shown here is the cover half (left) and ejector half (right) of the die.

Casting Design and Geometry / 119 Table 5 Relative capabilities, advantages, and costs of green sand and shell molding Target

As-cast dimonsional tolerance across 1 in. Nominal surface ﬁnish Minimum section thickness Heat-transfer efﬁciency Intricacy of detail Tool/pattern cost

Green Sand

Shell-Mold

þ0.030 (0.08 cm)

þ0.030 (0.08 cm)

þ0.008 (0.02 cm)

500 0.19 in. (0.48 cm) High Fair Medium

500–900 0.25 in. (0.64 cm) Low (sand) Fair Low(wood)

300–500 0.15 in. (0.38 cm) High (metal ﬁll) Very good Medium (metal)

Note: The performance numbers for green sand and shell mold are typical. Actual values for a given metalcasting facility could be higher or lower depending on the equipment and expertise at the facility.

Fig. 13

Four gates on the rim

other hand, meets all the functional requirements and offers a thin-wall capability and heat-transfer efﬁciency that cannot be achieved with green sand. The higher cost of the metal tool for the shell mold was acceptable, given the production run and beneﬁts. Planning the Component Orientation. The orientation of the part in the mold is an important factor in producing a sound casting. Casting defects, when they occur, tend to rise and segregate in the cope of gravity castings, so machined surfaces and critical features should be molded in the drag. In the coil rotor disc, the thin walls next to the hub and the ribs on the base of the rotor are critical features that

Two-gate approach. The two-gate approach was chosen for this component in order to facilitate proper ﬁlling.

Fig. 14

carry a signiﬁcant load and must be ﬂaw-free for strength and machinability. The rotor is oriented in the horizontal plane, so the parting line is perpendicular to the rotor centerline. Designing the ribs in the cope section would allow the possibility for inclusions or porosity to occur in the ribs of the casting, putting the ribs and machined surfaces at risk. By orienting the molding with the component ribs down in the drag section, inclusions or porosity that may form would segregate in the top of the casting in the center hub, where they will have minimal effect on mechanical

properties. The rotor ribs will have no defects with this orientation. Designing the Gating System. Proper design of the gating system is critical for uniform, controlled metal ﬂow. Nonuniform, long path, and/or slow metal ﬂow may produce unﬁlled sections or solidiﬁcation shrinkage in the casting. Too many gates and runners also will require additional molten metal and decrease the casting yield. Casting engineers were faced with two choices for a gating system. One placed four gates on the outside rim of the casting (Fig. 13), and another placed two gates in the center hub of the casting (Fig. 14). In the ﬁrst approach, the four gates on the rim gave a very long metal ﬂow path into all the ribs. With the four-rim gates, a test casting showed poor ﬁll into the ribs, with resulting rib misruns and defects. The test casting with the two-gate approach, however, revealed excellent metal ﬁll. The two gates on the center hub gave fast, direct metal ﬂow into all the ribs. This approach was chosen. Because the rotor rim has signiﬁcant thermal mass and will solidify slowly, a continuous source of molten metal must feed into the rim during solidiﬁcation to prevent shrinkage porosity. Perimeter risers feed molten metal into the rim. Smaller risers would normally solidify more quickly than the rim, which would shut off metal feed into the rim before it is fully solidiﬁed. To overcome this problem and to provide constant molten metal to the rim, the risers are surrounded with a 13 mm (0.5 in.) thick exothermic sleeve. The sleeve is a source of heat for the riser and keeps the riser molten and feeding metal into the rim as the rim solidiﬁes. After solidiﬁcation is complete, the rotor casting is ﬁnished and machined before it is inspected for quality, dimensions, and ﬁnish.

Casting Design and Performance Pages 121–132

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Thin Sections THIN SECTIONS save weight and may thus contribute to a more favorable strength-toweight ratio. By requiring a smaller volume of metal, thin walls (within practicable limits) may also lower casting costs, particularly when an expensive alloy is being poured. Realization of the advantages of thin walls in a casting depends largely on certain physical limitations of metal in the casting process. These limitations are concerned with the ﬂuidity of molten metal as it affects mold ﬁlling, with the processing required for metal soundness, with differences in alloy and in casting process, with distortion and heat treating problems (in some designs), and with the cost of foundry engineering and development. Guidelines on the minimum thicknesses to which a cast section can be designed depends on various factors. For example, Table 1 lists guidelines intended to prevent an increase of casting cost because of thinness of section. Other guidelines may be based on the actual limits of the process, rather than cost, but often fail to deﬁne the applicable processing conditions. Because so many aspects of design (such as the area of the thin-wall section, its proximity to a gate and the availability of heavy sections to conduct molten metal to it) have an inﬂuence on the actual minimum wall obtainable for any particular casting, recommendations such as those in Table 1 have limited usefulness when applied to a speciﬁc part. They do, however, provide the casting designer with a starting point, by familiarizing him with minimum thicknesses that normally can be produced efﬁciently. The mutability of what was presumed to be a practical minimum wall thickness is demonstrated by the experience derived in producing the large cover casting shown in Fig. 1. Cast in 17-4 PH stainless steel by a ceramic mold process, this part was approximately 21 in. long by 6 in. wide and weighed about 15 lb. Minimum weight was desirable, and it was assumed initially that a wall thickness of 0.150 in. was the safe and practical minimum. Machined aluminum patterns were mounted on aluminum cope and drag plates, and castings produced were sound and free of cold shuts. Subsequently, one ﬂat mounting face of the pattern was machined to reduce the casting wall thickness to 0.080 in., a reduction of about 47%. Even with this large reduction, only a small percentage of castings showed signs of

cold shuts. (However, it was apparent that a further reduction in wall thickness would probably lead to difﬁculty in ﬁlling the mold.) Another general example is the following minimum wall thicknesses proposed for areas approximately 1 in. square in investment castings, depending on the metal being cast: Aluminum alloy Beryllium copper Low-alloy steel

0.050 in. 0.040 in. 0.060 in.

At least one producer of investment castings considers wall thicknesses about 0.010 in. less than those indicated to be feasible where smaller areas are to be fed. Sections with an area larger than 1 sq in. may require heavier walls than those indicated. Three examples of investment castings that were produced successfully with wall Table 1

thicknesses less than are usually recommended are shown in Fig. 2. Figure 2(a) illustrates an aluminum investment casting (alloy 356) that contained, in several locations, walls 0.031 in. thick; Fig. 2(b), a beryllium copper casting incorporating in three locations tubular sections having walls 0.031 in. thick; and Fig. 2(c), a 4140 steel investment casting with three walls 0.050 þ 0.010 in. thick. All three castings were successfully produced in quantity, incorporating wall thicknesses less than are usually considered minimum for economical production with these alloys.

Determination of Thinnest Wall Prior to the use computer-assisted simulation, practical estimates of minimum wall thickness of a casting could sometimes be established

Typical minimum section relations for three metals in ﬁve casting processes (a) Minimum section thickness, in.

Casting method

Sand Permanent mold Investment Die casting Plaster mold

Aluminum

0.125 þ 0.031 0.093 þ 0.015 0.062 þ 0.010 0.062 þ 0.010 0.080 þ 0.015

Magnesiu

Steel

0.156 þ 0.031 0.125 þ 0.015 0.062 þ 0.010 0.093 þ 0.010 ...

0.187 þ 0.031 ... 0.093 þ 0.010 ... ...

(a) For compatibility with economical production. Not the thinnest producible by the processes.

Fig. 1

A 17-4 PH stainless steel ceramic mold casting, the wall thickness of which was reduced, from the presumed practicable minimum of 0.150 to 0.080 in., without appreciably affecting the soundness of castings produced

122 / Casting Design and Performance

Fig. 2

Three investment castings that incorporate walls of less than recommended minimum thickness

reliably by producing sample castings from inexpensive loose patterns accurately produced. On these experimental patterns, wall thicknesses would progressively increased or diminished, or revisions may be made to compensate for errors in calculating contraction, at a fraction of the cost entailed in revising production patterns. This method was used for the casting shown in Fig. 3. For this ﬂanged elbow casting, the problem was to determine the minimum practical wall thickness, using aluminum alloy 356 and the plaster mold process. Wall thickness was reduced progressively by modifying the pattern equipment, and castings were produced with three different wall thicknesses. Signiﬁcant variables were carefully controlled, to eliminate any extraneous inﬂuence on the tests. The results of the tests were evaluated in terms of the percentage of rejected castings, as follows: Wall, 0.040 in. 0.060 0.080

Fig. 3 0.080-in.

An aluminum elbow (alloy 356) cast by the plaster mold process to three different thicknesses to determine the effect of wall thickness. Rejections were 80% with 0.040-in. wall, 35% with 0.060-in., and 10% with

Rejections, 80% 35 10

Because the percentage of rejected castings is normally reﬂected in the selling price of acceptable castings, the price of the thinnerwall castings with higher rejection rates could be expected to increase proportionately. Also, the problem of meeting production schedules would become increasingly critical as wall thickness decreased. These are disadvantages that the designer must evaluate in relation to the known advantages of a wall thickness approaching the absolute minimum. Stainless Steel. Figure 4 shows a thin-wall austenitic stainless steel sand casting with a number of design features that helped to simplify casting problems and permit production at minimum cost. The casting is of relatively uniform section thickness, except for the heavy boss at the top. This boss was ideally placed to facilitate ﬁlling the sand mold and to avoid

A thin-wall sand casting produced from austenitic stainless steel. One section of the casting required two revisions in wall thickness to bring rejection rate to an acceptable level. Rejections were 50% with 0.060-in. wall, 25% with 0.075-in., and 5% with 0.090-in.

Fig. 4

distortion as the steel solidiﬁed and cooled. The casting has good stability and rigidity, and its slightly curved surfaces minimize distortion. As originally designed, the walls in the lower part of the casting were 0.060 in. thick. However, because of difﬁculty in eliminating porosity in the walls, caused by the back pressure of mold gas, rejection rate was approximately 50 per cent. Increasing the thickness of these

walls to 0.075 in. reduced the rejection rate to about 25 per cent. When the wall thickness was increased to 0.090 in., the rejection rate was further decreased to 5 per cent. Even the thickest wall (0.090 in.) is well below the minimum usually suggested in tables like Table 1, which are based on economy rather than absolute process capability. This casting illustrates what can be done when the design of a casting is favorable for producibility.

Design Problems Involving Thin Sections / 123

Thin-Wall Steel Sand Castings In all casting processes, castings that are designed so that rapid and orderly ﬂow of metal to all parts of the mold is promoted are easiest to produce with a minimum of ﬂaws or defects. However, the speciﬁc casting process can inﬂuence metal ﬂow, and, because of inherent differences, each casting process exerts its particular inﬂuence on the production of thinwall castings. The relative unsuitability of green sand molds to the casting of thin-wall sections is considerably magniﬁed in the casting of either alloy steels or superalloys. For this reason, most aircraft castings made of these materials are produced in shell, ceramic or baked sand molds. An analytical approach to the design of thinwall steel castings to be produced in green sand molds necessitates an understanding of four major production problems involved: mold ﬁlling, distortion, soundness, and the inﬂuence of design on heat treating problems. Mold-ﬁlling problems become increasingly acute as wall thicknesses are decreased to less than 8/16 in. Above this thickness, mold ﬁlling is seldom critical. When wall thickness is decreased to less than 1/16 in., adequate mold ﬁlling becomes almost impossible, except for extremely short distances, and misruns result because of the freezing of metal before the mold is ﬁlled. To achieve minimum weight, the designer will often incorporate minimum section thicknesses within a casting, in all locations where high strength is not a requirement. For instance, in the turbine bearing housing shown in Fig. 5 (a green sand casting made of alloy steel), eight posts, each incorporating a stiffening rib, connect the top and bottom portions. It was calculated that a rib thickness of 0.100 in. would be adequate to meet strength requirements. However, ribs of this thickness could not be ﬁlled completely; all castings produced were rejected for misruns. No practical improvement was possible in foundry processing. Consequently,

Eight stiffening ribs in this alloy steel green sand casting were originally 0.100 in. thick; misruns made all castings produced unacceptable. Increasing thickness of these ribs to 0.125 in. almost completely eliminated rejections for misruns.

the patterns were revised to increase rib thickness to 0.125 in.; rejections for misruns were almost entirely eliminated. Alloy Selection. Important in the problem of mold ﬁlling are the ﬂuidity and ability to feed of the metal speciﬁed. These properties of castability are low in steels and superalloys, as compared with light-metal alloys. And among steels, the castability of one alloy differs from that of another. As a result, a casting that can be produced successfully in one alloy may not be castable in another. Among foundries, the comparative moldﬁlling abilities of metals are controversial, and relative ratings of metals in regard to this characteristic are to some extent empirical. Existing disagreement reﬂects the many factors that may affect the ability of metal to ﬂow in a mold. From its experience with many alloys, one foundry lists alloys in order of decreasing ability to ﬁll thin cavities as follows: Cobalt-base alloys (except precipitation-

hardening alloys)

Nickel-base alloys (except precipitation

hardening alloys) 11 to 13% Cr steels Series 4300 alloy steels Series 300 stainless steels Series 300 stainless steels molybdenum

containing

Rigging and processing procedures vary from foundry to foundry, and foundries generally specialize in respect to the range of castings and alloys with which they prefer to work. This should be recognized by the engineer when he designs a casting with rigid requirements that may tax the capabilities of the process best suited to produce the casting. The aft-fuselage speed-brake hinge shown in Fig. 6 was designed to be sand cast from 4340 steel and heat treated to a tensile strength of 180,000 to 200,000 psi. It was to be produced to close tolerances and rigid standards of inspection. The foundry accepted the job with the understanding that the material could be changed from 4340 steel to type 431 stainless, and that certain of the more stringent tolerances could be modiﬁed. The change in material was requested because of the foundry’s extensive experience with stainless steel. Although type 431 stainless costs considerably more than 4340, this increase was more than offset by

the higher percentage of acceptable castings produced in stainless steel. Friction between molten metal and mold may seriously hinder mold ﬁlling. It becomes increasingly troublesome as the ratio of mold surface to metal volume increases, a condition that is characteristic of thin-wall castings. A high ratio promotes premature freezing in thin sections. Mold gases may create back pressures that retard the ﬂow of metal. These gases usually originate from the burning of organic moldbonding material and from the heated air in the mold cavity. Gases are expelled through the mold wall in proportion to the static head of metal in the sprue and risers, depending also on the permeability of the mold material. Designs that must be oriented in the mold so as to have dome-like thin sections are especially likely to create problems in venting of mold gases, which can generate back pressure and cause the metal to freeze before the mold is ﬁlled. The most obvious solution in the foundry is to use a higher pressure head, in order to force metal into the cavity more rapidly, but this may cause undesirable turbulence of the metal at or near its point of entry into the mold. Mold-Filling Problems. An unusual moldﬁlling problem that, because of the design of the casting, had no satisfactory solution was encountered in the production of the alloy steel part shown in Fig. 7. This casting was molded in the position shown, because, had the mold been inverted, the heavy section at the junction of the shaft and plate sections could not have been fed through the long shaft. Because of its diameter, the entrance to the saucer shape is small in proportion to the volume of metal that must pass through it. Thus, the metal broke into separate streams and trickled down to various points on the periphery of the plate section. Such separate streams oxidized rapidly and resulted in misruns, cold shuts, and “BB shot”. (The BB-shot defect refers to drops of molten metal that separate from the main body of ﬂuid metal, freeze prematurely, and fail to fuse with the main body of metal on recontacting it.) A possible solution might have been provided by the addition of a reservoir (indicated by the broken line, B, in Fig. 7) surrounding the periphery of the plate. Such a ring would have been removed subsequently by machining. Although the wall thickness of 0.060 in. would probably

Fig. 5

This close-tolerance sand casting, although designed to be cast from 4340 steel, was produced in type 431 stainless because of the foundry’s greater experience with stainless steel. Increase in material cost was more than offset by higher percentage of acceptable castings.

Fig. 6

124 / Casting Design and Performance adding to the thickness of the thin sections and by tapering the walls to the center of the volute (E). Illustrating another mold-ﬁlling problem, the steel casting shown in Fig. 9 indicates how obstructions to the ﬂow of metal as it ﬁlls the mold can result in casting defects. In this casting, metal entering the mold cavity through the heavy section (B) ﬂowed into the thin plate area and ﬂanged sections. In pouring, the mold was tilted, with the heavy section down, in order to prevent a trickling of metal into the plate and ﬂanged areas. To assure metal ﬂow to all parts of the mold, this casting was poured rapidly. However, as the wall of metal advanced to each of the three sand cores C, an eddy was formed behind each core, as indicated by the shaded areas, resulting in defects. The defects became increasingly critical as the thickness of the plate was decreased. The possible revision of core shape shown in the lower portion of Fig. 9 was not permissible. Consequently, in order to produce a sound casting, it was necessary to eliminate the coring of holes and to drill these holes. Another problem in mold ﬁlling results from the cascading of molten metal down one side of the mold for a casting designed with a ladder shape. For the original design illustrated in Fig. 10, metal must enter the mold cavity at points A and rise evenly into the two outer walls. In

walls less than about 3/16 in. thick, this is difﬁcult to assure. If the outer walls ﬁll unevenly, metal will ﬂow across a cross member B, and cascade down the opposite wall. As descending metal makes contact with metal rising in the wall, bubbles are likely to be trapped, and cold shuts and other defects may appear. The weight of metal in a thin-wall casting is usually insufﬁcient to create enough pressure to force entrapped air out through the mold walls. The result is voids and porosity. Thus, this 8-in.-long casting as originally designed, with a 0.060-in. wall thickness and ﬁve cross members, is incapable of being properly ﬁlled. The problem can be solved in at least three ways. First, the outer wall thickness can be increased to 0.187 in. or more, as in Fig. 10 (a). Second, the inner cross members may be slanted, parallel to each other, as in Fig. 10(b), eliminating most of the danger of cascading. Third, the inner cross members may be slanted in opposite directions to form a truss, as in Fig. 10(c). Distortion is a major problem in the production of thin-wall steel castings. During solidiﬁcation and in cooling to room temperature, distortion may be caused by variations in the cooling rate and in the contraction of sections of different thickness in the casting. In fragile castings, distortion may be caused by mold restraint. It may also occur during heat treating.

An alloy steel sand casting that was impractical to produce because its inverted saucer shape contributed to cold shuts and misruns.

Thin-wall volute castings such as this present problems of mold ﬁlling because molten metal must travel a long distance to ﬁll the thin walls.

have caused the metal to freeze before it could ﬂow into the reservoir, increasing the plate thickness to 0.120 in. should have permitted the ﬂow of metal into the reservoir. Unfortunately, this approach would have been impractical; the additional metal would have had to be removed by machining to a curved surface. The cast volute shown in Fig. 8 is a common conﬁguration that illustrates another mold-ﬁlling problem. There are three ways to gate such a volute. The ﬁrst and most common method is to gate at the inside of ﬂange A. The second is to gate at ﬂange B. The third method is to gate at both ﬂanges A and B. If the mold is ﬁlled at ﬂange A, the metal will take a relatively long time to ﬁll the thinwall section before ﬁlling ﬂange B. Thus, if the wall of the volute is less than about 3/16 in. thick, the metal is likely to freeze before ﬁlling the outside ﬂange of the casting. If the mold is ﬁlled at ﬂange B, the same problem exists, with the added danger that metal will cascade at angle C, producing cold shuts and other defects. Finally, if the mold is ﬁlled at ﬂanges A and B, entrapped air and mold gases in the thin-wall section may cause bubbles and porosity. Also, if the distance from C to D exceeds about 6 in., this thin section will be very difﬁcult to feed well enough to make it sound. The broken lines in Fig. 8 indicate a method for improving the castability of such a part by

Fig. 7

This thin-wall steel casting required rapid pouring to ﬁll the mold completely. Three cores obstructed the free ﬂow of metal, creating eddies that resulted in defects. Redesign of cores as shown, had it been otherwise acceptable, would have solved the problem of metal ﬂow.

Fig. 8

Fig. 9

Fig. 10

Uneven ﬁlling of the sides of the “ladder” conﬁguration of this casting as originally designed caused metal to ﬂow through the cross members and cascade down either side. Three revised designs are shown.

Design Problems Involving Thin Sections / 125 Soundness. Because of the small cross-sectional area of a thin wall, it is difﬁcult for a riser to replace the metal lost in shrinkage. In very thin sections, feeding may be less of a problem, because the likelihood of centerline shrinkage is slight. Examples of Good Design. The chine ﬁtting (a steel casting) shown in Fig. 11 is an excellent example of good casting design. The heaviest section is at the center, where it can be fed readily. The legs of the casting are tapered in steps. All wall sections of uniform thickness are less than 3 in. long and can be fed adequately from the center, with metal soundness assured. Changes in section thickness are not abrupt, and all inside corners are provided with generous ﬁllets to resist tearing. Radii on outside corners serve to maintain constant wall thickness. A stout tie bar crosses the center, where the danger of distortion is greatest. The walls of the casting are thick enough to resist distortion. Finally, the radii at points C are large enough so there is no cracking or tearing

in this area. This casting was producible without difﬁculty. Another well-designed steel casting is shown in Fig. 12. This part can be fed with a riser at each end, and at each side at points A; thus, the casting can be ﬁlled from the center or either end. Wall thicknesses are approximately 0.12 in., and the T-sections are easy to feed, because they are not more than 4 in. long nor more than 2 in. from the nearest riser. As originally designed, the casting was weak at points B. However, this deﬁciency was readily corrected by the addition of outer ribs. The revised design has good rigidity, with all areas of stress well braced and with ample corner radii and ﬁllets. The casting did not distort in production. The wing spar casting shown in Fig. 13, although large and rather complex, is another example of good design. Basically, the design is that of a box section with a ﬂat side approximately ¼ in. thick. One end of this steel casting is closed and varies in thickness from 0.178 to 1.50 in. Except for the round bearing support at the other end, the remainder of the casting varies in thickness from 0.187 to 0.250 in.

An example of good design in terms of foundry producibility. The heaviest section of this cast steel chine ﬁtting is at its center and is easily risered. From the center, each section tapers in steps to the extreme ends.

Fig. 11

Fig. 12 risers

A box-shaped wing spar casting in which heavy sections can be easily risered from the top, the two sides, and the two ends. This steel casting demonstrates the rule that areas of increased mass should be restricted to ﬁve sides of a basically cubic conﬁguration.

A well-designed steel casting in which all sections are accessible to properly located

Most of the heavy boss sections are located on the inside of the ﬂat plate area. In casting, the part was inverted so that the bosses could be fed through the plate area, thus making this the cope surface. All massive areas could be reached from the top, the two ends, or the sides. The design of this casting exempliﬁes the principle that areas of increased mass should be restricted to ﬁve sides of a basically cubic conﬁguration. All corners of the casting have generous ﬁllets and radii. The box-like shape resists distortion in cooling from the liquid and during heat treatment. The numerous blanks attached to the horizontal ribs are intended to provide specimens for test purposes, and can be removed easily without damaging the casting. For thin-wall castings, such specimens provide more reliable test results than do specimens that are separately cast.

Thin-Wall Aluminum and Magnesium Castings Light loads that can be adequately supported by aluminum or magnesium castings incorporating thin sections. An example of an aircraft part is shown in Fig. 14, sand cast in aluminum alloy 356, this part was designed as an access door for the fuselage of an airplane. On the basis of experience with similar shapes, the practical minimum wall thickness for this casting was established at 0.08 þ 0.01 in. To assist in conducting the molten metal to all parts of the casting, however, the ribs were designed to a thickness of 0.12 þ 0.01 in. A substantial number of these casting were produced, with rejections from misruns or cold shuts held to acceptable limits. This degree of success indicated that the wall thicknesses chosen for the conﬁguration did not, in fact, exceed process capabilities. In thin-walled castings of certain designs, the light weight of aluminum fails to provide sufﬁcient metallostatic pressure to force the metal into the thin sections. In such castings, it may be necessary to choose a process that provides for the application of additional pressure to the metal.

Fig. 13

Fig. 14

An aluminum sand casting (alloy 356) with a minimum wall thickness for its shape. Heavier ribs assisted in providing adequate metal ﬂow and kept rejections at an acceptable low level.

126 / Casting Design and Performance The design for the ﬁn-shoe slide casting shown in Fig. 15 called for maintenance of thin-wall sections (0.100 in., tapering to 0.040 in.), a surface roughness not exceeding 250 micro-in., and the use of an aluminum alloy capable of providing a minimum tensile strength of 32,000 psi, a minimum yield strength of 20,000 psi, and a minimum elongation of 5 per cent. These mechanical property requirements were to be checked on the basis of separately cast test bars taken from the same heat from which the casting was poured. To accommodate the minimum wall thickness desired (0.040 in.), the investment process with gravity pouring was selected. Although its mold-ﬁlling ability is inferior to that of aluminum alloy 356, alloy 195 was ﬁrst chosen, primarily on the basis of mechanical properties. This combination of alloy and process proved unsuccessful; all castings were rejected because of cold shuts and misruns. Next, a close-tolerance dry sand mold containing three equally spaced cavities around a sprue was tried. By spinning the mold at approximately 800 rpm, enough pressure was exerted on the molten aluminum to produce sound castings. The resulting mechanical properties and surface conditions were satisfactory also. Analyzing the production record of this part, it was concluded that even a 0.095-in. section tapering to 0.035 in. might have been attempted without incurring an excessive number of rejections. Core Movement. In many castings whose designs incorporate cored holes, the shifting of cores as the castings are poured can reduce wall thickness to the point where complete ﬁlling of the mold becomes difﬁcult. To avoid a high rate of rejections, such designs should be modiﬁed to improve core stability. Even those castings in which a wide variation in wall thickness is permissible may encounter a high rejection rate if core movement results in the production of walls that are too thin to be cast satisfactorily. For example, because of core movement in the aluminum sand casting shown in Fig. 16, the 5/16-in. wall

was reduced to 3/32-in. in some areas. To maintain the 5/16-in. wall thickness, additional support was provided for the cores. Effect of Area. In aluminum and magnesium castings, as in all other castings, the total casting area has an important inﬂuence on the successful production of thin walls. This is evidenced by a comparison of the castings shown in Fig. 17 and 18. The casting of Fig. 17 is 3 by 2 by 1 in., with uniform walls 0.080 in. thick. This AZ63A magnesium casting was easily produced in shell molds (the rejection rate was less than 5 per cent from all causes). Compare this with the AZ91 magnesium sand casting illustrated in Fig. 18. The wall thickness of the body of the latter casting was established at 0.080 þ 0.010 in.—the same as for that in Fig. 17. However, because of the greater size of the casting in Fig. 18, ribs 0.100 þ 0.010 in. thick were required, to assist in conducting metal to all areas of the mold. In addition, it was necessary to produce this casting under closely controlled conditions. It is probable that, even with the ribs, the 0.080-in. wall would have proved too thin for ordinary commercial production and a revision to thicken the wall would have been necessary. Strength-to-Weight Ratio. Because they are rigid and have a high strength-to-weight ratio, truss-type castings are desirable for many instrument applications. However, production problems arise when the metal does not ﬂow readily through thin walls in such castings to all parts of the mold. The pilot’s scope and sight support shown in Fig. 19 was produced from magnesium alloy AZ91 in a close-tolerance dry sand mold. The ﬁrst castings produced were rejected for misruns, cold shuts and hot tears. It was obvious that the heavy ends, acting as manifolds, should be made still heavier, so that the metal could retain most of its heat as it ﬂowed from the heavy sections to the thin sections of the mold cavity. It was also evident that the frame sections of the casting would have to be made heavier to

Insufﬁcient pressure from a static head of molten aluminum caused cold shuts and misruns that resulted in rejection of all investment castings of the above design. By centrifuge casting in a close-tolerance dry sand mold, rejections were reduced to an acceptable level.

Fig. 15

improve metal ﬂow; consequently, ribs were added to these thin sections. Because the junctions, and thus the increased mass (and accompanying hot areas), extended the length of the thin truss sections, the metal was able to ﬂow to all parts of the casting. No solidiﬁcation problems resulted, because freezing progressed along the length of the ribs and feed metal was available through the junction area. Heat in the ribs kept the frame sections molten for a longer time. Thus, the junction areas of the frame sections were given a longer time to solidify and build up strength to resist the contractive forces that were causing hot tears. To prevent hot spots at those points where the newly added ribs joined with other components of the casting, the ribs were cut short and were not permitted to join with the abutting walls. However, they were brought close enough to these walls so the metal could utilize the ribs as ﬂow channels, and only a normal percentage of rejections occurred from all causes. Feeding Through Thin Sections. When metal must travel through a thin section of a

Core movement in this aluminum sand casting caused reduction of a 5/16-in. wall to 3/32 in. Additional core support was required, to hold desired wall thickness.

Fig. 16

A thin-wall casting that was successfully produced in a shell mold from magnesium alloy AZ63A. Rejection rate was less than 5%.

Fig. 17

Design Problems Involving Thin Sections / 127

Sand casting of AZ91 magnesium alloy, in common with the one shown in Fig. 17, had an 0.080-in. wall. Because of the considerably greater size of this casting, however, thicker ribs were required, to assist in conducting the metal to all parts of the mold.

Fig. 18

In this sand casting, the 3/32-in. wall froze with microporosity and shrinkage. By increasing the thickness to 5/32 in., defects were eliminated. Aluminum alloy 355

Fig. 20

Complicated thin-wall truss-type casting was produced economically by adding heavy ends to act as manifolds, and small ribs to act as guides for the molten metal. By not junctioning the ribs with the abutting walls, hot spots were avoided. Alloy was AZ91 magnesium; process was close-tolerance dry sand molding.

Fig. 19

casting to ﬁll an adjoining heavy section, temperature gradients and solidiﬁcation rates may be difﬁcult to control. In the aluminum sand casting of Fig. 20, produced for an aircraft structural application, the bottom part of the mold cavity was ﬁlled through the 3/32-in. wall. The chilled bottom ﬂange section ﬁlled satisfactorily, but all castings were rejected because of microporosity and shrinkage in the thin section. Establishing a sharper temperature gradient would have accentuated the defects. Side risering was impractical, because of the trimming and ﬁnishing problems involved. A feasible solution was to increase the 3/32-in. wall to 5/32 in. This increase in thickness, aided by the chills at the bottom face of the ﬂange, created a freezing pattern that permitted adequate

feeding of all sections as the casting froze progressively from the bottom to the top ﬂange.

Thin-Wall Permanent Mold Castings In the permanent mold process, the degree of as-cast surface smoothness required has a direct effect on the section thickness that can be achieved. Surfaces of the metal mold must be covered with an insulating coating so that they are protected against the damaging effects of thermal shock that occur when the molten metal is poured. Thin sections are coated with materials of various compositions and thicknesses that have certain heat-transfer characteristics. Heavy sections are coated with other types of materials, with different characteristics. These mold coatings

inﬂuence the directional solidiﬁcation of the casting. Expendable Cores. One advantage of the permanent mold casting process is that sand or plaster cores can be used instead of metal cores. This helps to overcome many design limitations. For example, in a wave-guide casting requiring a smooth inside surface, the use of permeable plaster cores achieves a surface ﬁnish better than that which could be obtained with coated metal cores. Plaster also has excellent insulating qualities and does not prematurely chill the metal in thin sections. Sand cores permit the casting of tortuous passages, or of chambers or passages that are larger in section than the opening through the wall of the casting. Because the sand cores are expendable, removing them presents no major problem. Nonuniform walls in a permanent mold casting are usually difﬁcult to produce efﬁciently, especially when a heavy section must be fed through a thin section. In the aluminum casting illustrated in Fig. 21, gating and risering the outer heavy section was routine, but the inner heavy section of the casting had to be ﬁlled and fed through the ﬁve thin ribs. To ﬁll this inner section properly, the metal had to be poured quickly. This caused sinks to appear on the outer surfaces of the casting. Rapid pouring also created excessive turbulence. Increasing the thickness of the rib opposite the gate would have permitted the molten metal to ﬂow with normal streamline velocity through the rib to all parts of the mold, and would also have allowed the inner sections of the casting to be fed more readily during the cooling cycle. Nonuniform wall thickness necessitated by molding and machining requirements caused

128 / Casting Design and Performance shrinkage defects in the permanent mold casting shown in Fig. 22. This aluminum casting (alloy 355) was part of an aircraft fuel system and had to be leakproof at a pressure of 35 psi. Minimum casting weight was mandatory, because the assembly was at its maximum acceptable weight. The shrinkage defects occurred at the junction of the hose-attachment tube A with the body of the casting. Stock required for machining increased the thickness of the wall of this tube. Furthermore, the internal core of this tube had to be tapered, to facilitate its withdrawal. This coring requirement also necessitated that a sharp corner remain at the junction of the large core with the small core that was withdrawn from the tubular section; this, too, increased the mass of metal at the junction. Compare the junction of tube A, with all the attendant difﬁculties it presented, with the

junction of the larger tube section B. Here, the direction of withdrawal of the casting from the mold permitted a blend of the tube section with the casting wall, thus avoiding the defects encountered with tube A. Inﬂuence of Large Areas. Again with reference to the casting shown in Fig. 22, the interrelated requirements of pressure-tightness and smoothness would have made production in a permanent mold difﬁcult had the casting been made larger but with the wall thickness left unchanged. Thin walls encompassing large areas cannot be easily fed; this may give rise to centerline shrinkage or microporosity. To prevent premature freezing and assist the ﬂow and feeding of metal to these thin sections, the mold coating in the thin-wall area must be heavier. However, as has been noted previously, heavier coatings cause rougher surfaces;

thus is demonstrated the basic incompatibility between thin sections and surface smoothness. Minimum Section Thickness. The casting shown in Fig. 23 taxed the capabilities of the permanent mold process. This casting, 3.703 in. in diameter by 30 in. long, with a uniform wall thickness of 0.148 in., was designed for a bazooka assembly. For portability, the unit had to be light in weight; the aluminum alloy speciﬁed (7% Mg, 0.1% Ti, 0.1% Mn, 0.05% Be) was sluggish and hot short. The various bosses adjoining the large thin-wall section caused cracking unless all processing conditions were optimum. Castings of this conﬁguration and size are difﬁcult to produce, and careful attention must be given to mold coatings. Fortunately, smoothness was not a requirement for this casting, although general good appearance was necessary. Producibility could have been improved by an increase in the wall section—speciﬁcally, by adding metal to the inside of the wall, from which it could be bored out later with negligible contribution to cost, as a machining operation was already required on the inside surface. Good Design. The aluminum casting shown in Fig. 24 illustrates good design for the permanent mold process. The 5.6 mm (7/32-in.) center web is heavy enough to allow complete ﬁlling of the mold and feeding of the center section. It also provides sufﬁcient wall around the inserts to prevent shrinkage cracks. Changes in section thickness are gradual, radii are large, and areas for risering are adequate. If weight were a problem, the heavy sections could be made thinner and certain cored areas

Fig. 21

In this aluminum permanent mold casting, thin connecting ribs made it difﬁcult to feed the heavy center section. Increasing the size of the rib opposite the riser would have improved feeding.

Fig. 22

Molding and machining requirements necessitated nonuniform walls at the junction of small tube A with the body of this casting. Shrinkage defects resulted at this junction. Large tube B and body of casting were blended smoothly. No defects occurred at this spot. Permanent mold casting of aluminum alloy 355

Design Problems Involving Thin Sections / 129 could be increased in size. However, making the web section thinner would be undesirable. Another example of good design is the front canopy bow casting shown in Fig. 25. This structural casting, produced in aluminum, was designed so that optimum location of gates was possible. Metal could ﬂow to all parts of the permanent mold cavity without the need for excessive mold coatings to prevent premature freezing. By taking advantage of the maximum chilling effect of the mold, the necessary strength in the metal was developed. In addition, all surfaces were produced to a ﬁnish of 125 micro-in. or better.

Thin-Wall Investment Castings The investment casting process is capable of producing intricately shaped castings of almost

This aluminum casting is representative of the maximum capabilities of the permanent mold process for the wall thickness indicated.

Fig. 23

Fig. 24

any metal, and can assist the designer in minimizing casting weight, obtaining maximum strength-to-weight ratio, and producing small parts. In designing for minimum wall thickness, the ﬂuidity of the alloy and its ability to ﬂow in the mold are important considerations. Of almost equal importance is the solidiﬁcation range of the alloy in relation to proper feeding of a section. The surface area of exposed molten metal and the feeding distance are also important in determining the castability of thin sections. Table 2 shows typical minimum-wall-thickness relations for a tube 40 mm (1½) in. long produced from various metals by investment casting. These recommendations reﬂect the experience of the industry as a whole. Although these minimum thicknesses may be subject to modiﬁcation, the relative mold-ﬁlling capabilities of the different metals will remain as indicated in the table. In any casting in which heavy sections are separated by thin sections, it is likely that there will be problems of ﬁlling the cavity and of feeding the metal during solidiﬁcation. Figure 26 illustrates how a simple revision in the design of an investment casting overcame these problems. Produced from 8630 alloy steel, this casting had thin sections connecting the two heavy end sections. This made it extremely difﬁcult to ﬁll the mold cavity completely; misruns and shrinkage occurred frequently. Because this casting was gated on the three end bosses, widening of two of the three bosses

This aluminum casting illustrates good design for the permanent mold process. There is adequate metal around the cast-in-place inserts, and the web is heavy enough to permit feeding of the center section.

shown in Fig. 26 placed a larger area of the boss directly over posts A. This created a direct path for metal ﬂow to the opposite heavy section and eliminated cold shuts. Increasing the size of these bosses added enough additional heat to the mold to keep the thin sections liquid for a longer period. In turn, the gated areas could feed the previously isolated heavy end for a longer period, thus eliminating shrinkage defects. The third boss was increased in area for uniformity of appearance only. Chilling Effect of the Mold. Even though the molten metal is poured into a heated mold, investment casting presents the same possibilities for premature freezing of thin walls as other casting processes. A high ratio of area to thickness in a section encourages quick cooling and freezing of the molten metal. However, the distance the metal must travel to ﬁll the section must also be considered. A heavier section would feed into a thin section 13mm (½ in.) long by 50 mm (2 in.) wide with considerably less difﬁculty than into a thin section 2 in. long by ½ in. wide, even though the surface areas (1 sq in.) and the thicknesses were the same. The top chart in Fig. 27 shows recommended and minimum section thicknesses in relation to the longest dimension of a casting; the lower chart shows minimum diameter of cylindrical cores as a function of maximum length of core, for both blind holes and through holes. Since many factors inﬂuence the maximum and minimum limits of any particular design detail, this information should be considered approximate, for use only as a guide. Mold and Metal Temperature. The pouring temperature of an alloy is selected partly on the basis of the thickness of the section to be poured. Metal is usually poured at the lowest possible temperature, to reduce the likelihood of casting defects from gas, dross, metal-andmold reaction, and other deleterious effects that depend on temperature and time at temperature. However, in pouring thin sections, it is often necessary to increase pouring temperature, in order to prolong the time interval between pouring and solidiﬁcation. Usually, a cooler mold results in a better surface ﬁnish, because of the faster freezing of the metal skin and the elimination—or at least

Table 2 Typical minimum section relations for a tube 40mm (1½ in.) long produced from various metals by investment casting

Good casting design, permitting ideal location of gates, made for efﬁcient production of this aluminum permanent mold casting. Although this casting had a wall thickness of 0.093 in., the recommended minimum for the material and process (Table 1), all surfaces were produced to a ﬁnish of at least 125 micro-in.

Fig. 25

Metal

mm

Min wall, in.

Carbon steel Series 300 stainless steel Series 400 stainless steel Aluminum alloy Magnesium alloy Aluminum bronze (10% Al) Berylllum copper Cobalt-chromium alloy

1.52 1.27 1.65

0.060 0.050 0.065 0.050 0.050 0.060 0.040 0.050

1.27

From “How to Design and Buy Investment Castings,” Investment Casting Institute, 1960, p 132

130 / Casting Design and Performance

Fig. 26

Providing clear channels for metal ﬂow in the mold simpliﬁed production of this casting and permitted gating at one end only. Investment mold cast; 8630 steel

reduction—of chemical reactions that may take place at the interface of the metal and the investment. However, this same cooling effect reduces the chance of ﬁlling extremely thin sections. Consequently, it may be necessary to relax the surface ﬁnish requirements in order to obtain the thin wall desired. Mold permeability partially provides for the release of entrapped gases and plays an important part in the determination of wall thickness. Two principal types of molding material are used by manufacturers of investment castings: a plaster-base investment, for the lower-melting nonferrous alloys; and a silica-base investment, for steel and other alloys poured at higher temperatures. The silica-base investment in solid mold work is more permeable than the plaster-base investment. Because of this, plaster-base molds for nonferrous castings are often spun centrifugally, to impart a higher force to the molten metal entering the mold. This is done particularly with the light alloys, where the molten metal in the gating system provides insufﬁcient

pressure to force the casting metal through thin sections. Prevention of Defects. In the investment process, as in other processes, misruns and cold shuts may be encountered in the production of castings with thin walls. In the steel investment casting illustrated in Fig. 28, cold shuts occurred in the thin section, which was farthest from the gate. The metal was required to ﬂow around the core into the cavity, and down into this 0.13-in. section. When the metal reached this point, it was too cold for the two streams to unite satisfactorily. As a result, all castings produced were rejected. Eliminating the large cored hole, which restricted metal ﬂow, solved this problem. Hot tearing and cold cracking also may be encountered in thin sections of investment castings. Hot tearing may occur when a section of the casting develops insufﬁcient strength to resist the contractive forces exerted as the metal cools in the mold. Cold cracking and distortion may occur if a highly stressed area is too thin to withstand the forces that develop during

cooling. Uneven section thickness gives rise to stress from unequal cooling rates, which can cause distortion or cracking of the casting. Although few castings can be designed to have completely uniform sections, uniformity should be an objective of most designs, and abrupt changes in section thickness should be avoided. Cooperative Designs. Difﬁculties in attaining soundness in thin walls may be minimized through design. One simple solution is to redesign other sections of a casting so that minimum wall thicknesses are not called for. For example, in a casting that has several heavier sections but employs thin sections to reduce mass, it is sometimes possible to reduce the heavier sections and add to the thinner sections. This may result in a total weight even less than that of the original part. Designing for Economy. Uniform thin sections present problems that often increase the cost of the casting. Tapering thin sections as little as 2 or 3 per side will increase producibility. (The increasing thickness should, of course, be in the direction of the heavier section from which metal is being fed.) This taper will encourage directional solidiﬁcation of the metal from the thinnest to the heaviest section. An alternate method is to provide ﬂow paths in the form of ribs. If these ribs can remain, they contribute little to cost; if they must be removed, additional expense is incurred. As an example, producing the 4140 steel investment casting shown in Fig. 29 presented a problem. Even though the most favorable gating practice was used, it was impossible to ﬁll the 0.09-in. leg sections, and misruns were common. Strictly in terms of producing a sound casting, two solutions were apparent. In the ﬁrst, and more desirable, solution, the legs would be tapered by increasing their thickness progressively as they approached the junction with the body. Normally, such additional metal could be added to the inside, the outside, or both sides of each leg. In this casting, however, adding metal was not permissible. The mating components of the assembly of which this casting was to be a part were in production, and it was mandatory that the speciﬁed dimensions be observed. The second solution would be to add ribs that would act as channels to permit the molten metal to ﬂow quickly to the extremities of the legs. These actually were added, and are shown in phantom in Fig. 29. It was practical to add these ribs only to the outside faces, because their removal to a close tolerance was necessary before shipment. This increased the unit cost of the casting. However, adding the ribs solved the problem of ﬁlling the legs, and many castings were produced successfully. Later, a similar casting, shown in Fig. 30, was designed with tapered legs such as were suggested in connection with the casting shown in Fig. 29. Both castings were produced in solid investment molds from 4140 steel, and both performed the same function in application. However, through improved design, the casting

Design Problems Involving Thin Sections / 131

Fig. 27

(a) Recommended and minimum section thicknesses with respect to the longest dimension of a casting. (b) Minimum diameter of cylindrical cores with respect to maximum length of core.

Fig. 28

In this steel investment casting, elimination of the round cored hole did away with cold shut defects in the 0.13-in. section, which were caused by the meeting of two metal streams that were too cold to unite

properly.

in Fig. 30 cost approximately 27% less than the other, and delivery time was shorter. Feeding Ribs. In the investment process, heavy sections separated by thin sections can be ﬁlled by adding gates and runners to channel metal directly to these heavy sections. However, this increases the cost, because it involves adding more gates and runners to the wax patterns and entails removing all traces of the gates from the castings. It was shown in Fig. 29 that a thin section can be ﬁlled more efﬁciently by adding ribs in strategic locations. Ribs work equally well in conducting metal across a thin section to feed a heavy section that would otherwise be isolated. Although the ribs may add a small amount to the cost of the pattern die, they reduce over-all production cost by eliminating

gates and runners that would otherwise be necessary. The addition of ribs permits the adjacent thin sections to be made even thinner, because the thin sections are no longer needed as primary feed paths. The ribs also add rigidity to the thin sections and can be considered as part of the working structure of the casting. However, their main advantage is cost reduction. For example, the inner gimbal investment casting shown in Fig. 31 was produced from 8620 steel with relatively heavy sections at both ends and in the center areas. These heavy sections were connected by a thin (0.093-in.) wall that incorporated a small rib. The ﬁrst production castings revealed that the cross-sectional area of the thin-wall body was not large enough, even with the small ribs, to permit the

entire casting to be fed from one end or from the center hubs. As shown in Fig. 31, six gates were necessary to produce sound castings. In subsequent production, it was established that sound castings could be produced by increasing the size of the ribs and using only two gates. These gates were located at one end, each opposite a rib, as shown. Castings thus produced cost 28% less than castings produced to the original design. Alloy selection is governed not only by the mechanical properties desired, but also by the ability of the metal to ﬁll a thin section. The two alloys capable of producing the thinnest sections are probably beryllium copper and silicon brass. These alloys have been poured successfully in wall thicknesses of 0.010 in. for a distance of 0.040 to 0.050 in. A thin-wall beryllium copper casting is illustrated in Fig. 2(b). Aluminum alloy 356 also is excellent for casting thin sections, although less so than beryllium copper or silicon brass. A ferrous alloy excellent for thin sections is SAE D2, a high-carbon high-chromium airhardening tool steel. This alloy is extremely ﬂuid. The series 300 stainless steels and many of the superalloys for high-temperature applications have good castability and are capable of ﬁlling sections of minimum thickness under speciﬁc conditions. Among the series 400 stainless steels, type 440 is best, followed by 420 and 430. Type 410, often speciﬁed as AMS 5350, is only fair from the standpoint of castability, and will not produce minimum wall thicknesses sometimes desired. Despite its limitations for thin walls, type 410 is often used for aircraft castings because of its corrosion resistance and strength. The low-alloy and carbon steels have only fair ability to produce thin sections, and their ﬂuidity and castability decrease with decreasing carbon content. As carbon content is increased to about 0.50%, there is a corresponding increase in ﬂuidity and ability to ﬁll thin sections, but above 0.50% C there is a tendency for gassing, which increases with the carbon content. Thin-wall castings have been produced successfully using “RH” Monel, particularly when their design is compatible with optimum production conditions. Such a casting is shown in Fig. 32. Although this casting has a wall only 0.035 in. thick, its design permits ideal gating at its center, and several foundries have produced it successfully. Unlike thin sections in permanent mold castings, thin sections in investment castings will usually have the best surface attainable. Because the cooling effect of the mold at the thin sections is usually greatest, the metal in contact with the mold freezes almost instantly. This restricts the amount of time for a possible chemical reaction between the mold and the metal, thus limiting surface imperfections. Rapid freezing reduces the depth that the molten metal can penetrate into any cracks or other defects in the mold precoat. However,

132 / Casting Design and Performance

To encourage metal ﬂow in the thin sections of this 4140 steel investment casting, the legs were tapered as shown. Compare with similar casting shown in Fig. 29

Fig. 30

By increasing the size of the casting ribs, four gates were eliminated from this investment casting (8620 steel), at a cost saving of 28%.

Fig. 32

The ideal gating area at the center of this casting permitted walls thinner than usually recommended in “RH” Monel investment castings.

Fig. 31

To encourage metal ﬂow in the 0.09-in. legs of this 4140 steel investment casting, ribs were added as shown. These were removed before shipment.

Fig. 29

because a more ﬂuid metal will penetrate mold defects more than a sluggish metal under the same conditions, the choice of a ﬂuid alloy to ﬁll thin sections may result in a rough ﬁnish on the casting, particularly in heavy sections that freeze slowly.

Casting Design and Performance Pages 133–138

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Uniform Sections IN MANY CASTINGS, functional requirements dictate that walls be uniform or nearly uniform in thickness. Many problems in producing castings having walls of uniform thickness are associated with the premature freezing of molten metal before all parts of the mold cavity have been ﬁlled. In other castings, it may be possible to ﬁll the mold cavity completely, but centerline shrinkage or areas of porosity may be encountered because portions of the casting cannot be fed before the onset of freezing. In such castings, certain walls may be tapered by the addition of padding, which may be incorporated into the design, or ribs or webs may be added to provide feed paths. The function of ribs or of sections of walls that have been enlarged in providing ﬂow and feed paths for the molten metal as it ﬁlls the mold and cools is shown schematically in Fig. 1. In pouring a ﬂat plate, Fig. 1(a), the metal enters the cavity and attempts to ﬂow in all directions. If the distance from the gate to the extremities of the mold cavity is too great, the metal freezes prematurely, and misruns result. When ribs are added to the plate, Fig. 1(b), the metal ﬂows readily to the extremities of the mold cavity, and successful castings are produced. The same success is attained when the plate is enlarged at the gate and uniformly tapered to the extremities, as shown in Fig. 1(c). The metal ﬂows into the mold cavity at the heaviest section, thus preventing an initial chilling of the metal by the mold, such as occurs with a small

Fig. 1

volume of metal. Because the molten metal in heavy sections retains more of its heat as it ﬂows through the mold, it can more readily reach the extremities of the mold cavity. A related problem is involved in the mechanics of cooling and freezing of molten metal and the attendant volumetric shrinkage. Referring to Fig. 1(a), let us assume that the metal is poured and ﬁlls the mold. For some castings where chilling of metal by the mold is possible, there is no assurance that freezing will start at points farthest from the gate and progress in an orderly manner to the gate so that the volume of metal lost because of shrinkage can be replaced from the reservoir (riser) intended for that purpose. In fact, a casting such as this may be expected to contain sections that are isolated from the feed source by metal that is closer to the gate. The metal nearer the gate freezes ﬁrst and thus surrounds and isolates these molten pockets. Shrinkage of the molten metal in these isolated pockets creates voids or porosity. With two risers feeding metal from opposite ends of the casting, more sound metal would be obtained in the plate, and with a riser feeding from each edge of the plate, a still greater percentage of sound metal could be obtained. However, risers add to the cost of producing a casting, and a method that minimizes the number of risers is more economical. This can sometimes be accomplished by providing feed paths within the casting conﬁguration, as shown in Fig. 1(b). These paths permit the molten

metal to ﬂow easily to all sections of the cavity. If the enlarged sections are gradually tapered from the gate to the extremities, solidiﬁcation follows an orderly progression through the casting and termigates in the riser. Compensation is provided for normal shrinkage throughout the casting, and sound metal is assured. The same beneﬁcial freezing pattern is obtained by tapering the plate, Fig. 1(c), provided the angle of taper is sufﬁcient and the source of feed metal is at the heaviest section. Under these conditions, a favorable freezing pattern is assured because the metal that has ﬂowed the farthest (and is therefore the coldest) is deposited in the thin section. In contrast to a heavy section, the thin section of molten metal has less heat to transmit to the mold before freezing starts. Thus, the direction of solidiﬁcation can be predicted in tapered sections, provided the metal enters at the heavy end. If the need for feed paths is anticipated in the early stages of the design of a casting, padding may be made a functional part of the casting, or it may be located so that its removal adds little or no cost. Consider the two investment castings shown in Fig. 29 and 30 of Chapter 9 (“Design Problems Involving Thin Sections”). For successful production of the casting shown in Fig. 29, ribs were required on each of the legs. These provided feed paths that encouraged complete ﬁlling of the cavity and resulted in

Flat plates (a) are difﬁcult to produce completely ﬁlled and with sound metal. By adding ribs (b), or tapering the plate (c), the problems are eliminated or substantially ameliorated.

134 / Casting Design and Performance sound castings. The casting shown in Fig. 30 of Chapter 9 was designed with each leg tapered. This permitted easy ﬁlling and feeding of the legs and induced a satisfactory freezing pattern. Although the addition of ribs to the ﬁrst casting was satisfactory from a foundry point of view, removal of the ribs was necessary for the proper functioning of the part. In the second casting, the taper was satisfactory in terms of both production and function, because the casting was designed with this in mind. In the magnesium casting shown in Fig. 19 of Chapter 9, the problem was to distribute metal efﬁciently to all ﬂow channels and all parts of the mold before freezing, and to accomplish this objective with a minimum amount of metal. (This was an aircraft application, and no excess weight was permitted.) The solution entailed greatly enlarging the two ends into which the metal was gated, thus allowing them to function as manifolds. Increasing their size kept the metal hotter and permitted unrestricted ﬂow to each connecting member of the casting. As shown, ribs were added to each thin strut section, to provide a larger ﬂow channel. Because of the larger mass, the rib and strut junctions developed hot spots in the mold and helped to prevent premature freezing. The heavy manifold ends were later machined off, but the small ribs remained intact.

Sand Castings Produced in a dry sand mold, the steel elbow casting shown in Fig. 2 was originally designed with a uniform wall 1½ in. thick. This casting, intended for use at high temperature and high pressure, was produced with a rejection rate of 16%, because of shrinkage porosity caused by insufﬁcient feeding of molten metal to all parts of the casting during solidiﬁcation. ASTM class 1 radiographic soundness was speciﬁed for all areas except the unﬂanged end, where no defects whatever were permitted because of a subsequent welding operation for connecting the elbow to an adjacent member of the assembly. To overcome this feeding problem, the casting was revised by adding ½ in. to the wall thickness at the ﬂanged end and tapering the wall uniformly to its original thickness at the opposite end. This change lowered the rejection rate from 16% to 6%. The gate valve casting of Fig. 3 suggests another method for improving the ﬂow of metal and feeding all sections adequately. These stainless steel castings were produced in sand molds and were intended for use in nuclear reactors, necessitating maximum attainable soundness. All castings were inspected radiographically, and any discernible defect was cause for rejection. Although this casting was redesigned primarily to reduce the cost of producing it, addition

As originally designed, with a constant wall thickness, this sand cast steel elbow ﬁtting was producible only at a 16% rejection rate, because of shrinkage porosity due to insufﬁcient feeding. Redesign, in which walls were increased at the ﬂanged end and tapered, permitted more successful production; the rejection rate dropped to 6%. Dimensions in inches

Fig. 2

Fig. 4

of the 75 mm (3/4-in.) webs also reduced the distance the metal had to travel as it cooled, and minimized shrinkage porosity caused by impeded feeding. The webs functioned efﬁciently as manifolds to distribute molten metal to all parts of the mold cavity. Because they abutted the tubular sections, the webs reduced the distance the molten metal had to ﬂow to reach the central portion of the casting. The webs also conducted metal to the ﬂanges, which were used as ﬂow paths to feed those parts of the casting nearest the ﬂanges. These heavier sections then fed molten metal to the thinner sections. Although the webs were approximately the same thickness as the ﬂanges, the webs were the last parts of the casting to freeze. The greater volume of metal that passed through the web-forming section of the mold heated it to a higher temperature than was reached in the ﬂange-forming sections. A related casting is shown in Fig. 4. These sand cast stainless steel ﬁttings for a nuclear reactor also were revised to increase efﬁciency of production. The addition of the web reduced by one the number of risers required to obtain

The application of this sand cast stainless steel valve body required completely sound metal. The uniform walls were fed readily from the risers through the ﬂanges and the webs. Webs were added to the original design.

Fig. 3

Adding the webs and tapering the tapping beads in the elbow ﬁtting (a) and the tee ﬁtting (b) assisted in distributing the molten metal and created a favorable freezing pattern that permitted adequate feeding of all sections of these stainless steel sand castings during solidiﬁcation.

Design Problems Involving Uniform Sections / 135 sound metal. However, in these castings there were no ﬂanges to assist the webs in the distribution of metal. Instead, internal padding of the thread beads provided the necessary ﬂow and feed paths. As shown, the padding was tapered, with the heaviest part adjacent to the web and the thinnest part diametrically opposite—that is, at the point farthest from the webs. The elbow castings, Fig. 2 and 4(a), and tee castings, Fig. 3 and 4(b), beneﬁted from a more orderly solidiﬁcation of molten metal as a result of redesign. Directional solidiﬁcation was assured by the uniform temperature gradients in the molds, created by both the size of the casting sections and the ﬂow path of the metal. Freezing of the metal started at the point farthest from the risers and progressed uniformly to the risers. All sections of the castings were adequately fed, and sound metal was obtained. The padding on the elbow and tee castings in Fig. 4(a) and (b) was machined away when the thread beads were bored to the desired diameter before being tapped. The AZ91 magnesium casting of Fig. 5 illustrates a straightforward approach to padding. This casting was produced in dry sand molds. The minimum castable section thickness, arbitrarily established at 0.130 in., was heavy enough to provide the necessary strength. However, all castings produced were defective because of cold shuts and shrinkage. The draft angle on the pattern was then increased from one degree to three degrees, and 0.125 in. was added to all wall thicknesses. With these revisions, all castings produced were sound. Although such a freehanded use of padding is not generally recommended, especially for aircraft castings, circumstances may occasionally warrant it. (The unwanted weight that was added to the casting shown in Fig. 5 was justiﬁable only because delivery of these castings was of utmost importance.) Problems presented by walls of uniform thickness are not always attributable to the distance the metal must ﬂow or be fed, but sometimes result from the way in which the walls are shaped. Figure 6(a), a section of a large aluminum sand casting for aircraft use, illustrates a tortuous conﬁguration that caused turbulence as the metal ﬂowed through the mold, resulting in the entrapment of oxides. Micro-shrinkage also occurred, and the rejection rate of this expensive casting was excessively high. By relieving the sharpness of the angles of the casting walls, as shown in Fig. 6(b), free ﬂow of metal was obtained. This eliminated both oxide inclusions and microporosity and resulted in the production of sound, acceptable castings.

Shell Mold Castings As illustrated by previous examples, casting problems presented by uniform wall thicknesses are not always attributable to the distance metal must ﬂow. Shape, too, exerts a

strong inﬂuence. The casting shown in Fig. 7, a hollow spherical casting with a heavy ﬂange, is an apparent exception to rules relating to feeding distance. Because the distance from the edge of the heavy ﬂange to the bottom of the sphere is 4 in., it might be assumed that the casting could be fed satisfactorily from the

heavy ﬂange. However, as originally designed (Fig. 7a), it could not be. The total area of the thin section greatly exceeded that of the ﬂange section from which it was to be fed, and feeding was inadequate; the part was unsound below points A, or about two thirds of the way down the wall of the sphere.

In this sand casting as originally designed, excessively thin walls resulted in cold shuts and shrinkage. Overcorrecting by nearly doubling all wall thicknesses and by increasing the draft angle eliminated defects but added unnecessarily to the weight. Alloy was AZ91 magnesium

Fig. 5

A partial section of a large, intricately shaped aluminum casting in whose original design (a) sharp bends retarded metal ﬂow and created turbulence. The resulting inclusions and porosity caused a high rejection rate. Revised design (b), in which bends were eliminated or were made less acute, was producible without defects.

Fig. 6

In this ﬂanged spherical casting as originally designed (a), feeding was restricted a short distance from the riser, resulting in unsoundness below points A. Tapering the walls, as shown in (b), provided adequate feeding. This alloy steel casting was produced in a shell mold.

Fig. 7

136 / Casting Design and Performance

Walls of uniform thickness did not deter successful production of this casting, because it was designed so as not to exceed the capabilities of the metal (magnesium alloy AZ63A) or the process (shell molding) that was used in producing it. Rejection rate was less than 5%

Fig. 8

Fig. 9

To assure soundness, this wall had to be tapered from the top to a point not more than 1½ in. from the bottom of the sphere, as indicated in Fig. 7(b). Such a taper should be no less than 1½ ; a taper of 3 would virtually guarantee soundness. When a large sphere is cast it is usually necessary to provide risers at various points around the sphere in order to feed from lower wall sections as well as from the top. Not all castings require padding or the addition of ribs to obtain sound metal. When the normal capabilities of metal and process are not exceeded, uniform walls are practical and do not defy successful production. A good example of a successfully produced casting with walls of uniform thickness is shown in Fig. 8. This AZ63A magnesium part, 3 in. long,

In this permanent mold casting, functioning of heavier ribs with thin walls induced hot tears and shrinkage at the junctions. Uniformity of wall thickness would have eliminated these defects.

2 in. wide, and 1 in. high, with 0.080-in.-thick walls, was produced in shell molds with a rejection rate of less than 5% for all causes.

Permanent Mold Castings Castings produced in metal molds have an advantage, in one respect, over those produced in sand or ceramic molds. Water lines can be incorporated in the mold to create a planned chilling effect on speciﬁc sections of the casting. By adjusting the rate of water ﬂow, the degree of chilling can be controlled accurately. This permits some leeway in the variation of wall thicknesses that can be incorporated in a single casting. This possibility of chilling the casting should not be used as a substitute for good design. Regardless of the casting process, a well designed casting can be produced more easily and more economically than one of unfavorable design, even though the unfavorable design does not tax the capabilities of the process. The permanent mold casting shown in Fig. 9 was structural in use and required radiographic quality. The inner wall was designed to have minimum thickness, to save weight. Because of the taper required for molding, ribs were heavier at the junction with the inner wall than was functionally necessary. Eliminating shrinkage areas and cracks at these junctions was a problem. Two of the ribs at the parting line were still heavier, because of their being used for the attachment of gates and risers. These ribs were insulated with extra mold coating, because smoothness of the as-cast surface was not mandatory. By water-cooling the inner steel core to obtain some chilling of the metal, the casting was made to good quality. The design of this casting could have been improved by increasing the thickness of the inner wall to equal that of the heavier adjoining ribs. This uniformity of wall thickness would have eliminated the shrinkage defects. Size can determine whether a casting design incorporating uniform walls is practical or impractical. The effect of large areas of uniform wall thickness on producibility can be illustrated with reference to the casting shown in Fig. 10. This casting was relatively easy to produce to the dimensions indicated. However, if it had been several times larger, with no change in wall thickness, progressive solidiﬁcation would have had to be induced by varying the thickness of the mold coating, even if a wall thickness of 4.06 mm (0.160 in.) were otherwise adequate. Such practice is difﬁcult, time-consuming and expensive. An alternate method would be to incorporate a taper in the walls to help produce the necessary solidiﬁcation pattern.

Investment Castings This casting was easy to produce in a permanent mold to the dimensions indicated. However, if all dimensions other than wall thickness were increased proportionally, this casting would be difﬁcult to produce. The enlarged casting would require tapered walls or controlled variation in the thickness of the mold coating, in order to permit efﬁcient production.

Fig. 10

Uniform walls in an investment casting are readily produced if the casting is not too large, in relationship to wall thickness, and if gates

Design Problems Involving Uniform Sections / 137

This 356 aluminum alloy casting, produced in a centrifuged plaster-base investment mold, represents minimum wall thickness for the shape, alloy and process.

Fig. 12

Fig. 11

A well-designed aluminum investment casting, incorporating uniform walls and - efﬁcient gating areas

A critical application necessitated class 1A quality for this type 347 stainless steel investment casting. Hot tearing at the junctions and porosity in the body, which resulted from too-rapid solidiﬁcation within the uniform wall of the tubular section, were eliminated by tapering the body toward each ﬂange and increasing the ﬁllet radii at the junctions.

Fig. 13

Fig. 14

can be placed at desirable locations. The body of the aluminum investment casting illustrated in Fig. 11 incorporates a uniform wall thickness of 0.09 in. throughout, even at the point where the wall changes direction. The irregularly shaped ﬂange also is maintained at this same thickness. The rectangular ﬂange (3.8 mm, OR 0.15 in. thick) is just enough heavier to permit efﬁcient gating and transfer of metal to all sections. This part was easy to produce, and virtually no defects were encountered. The 356 aluminum alloy part illustrated in Fig. 12 was produced in a centrifuged mold and heat treated to the T6 condition. This casting is part of the leading-edge assembly of an aileron; all walls are 0.090 in. thick. The foundry considered the 0.090-in. wall to be of minimum practical thickness for this conﬁguration. Decreasing thickness below 0.090 in. would result in rejected castings, and the rejection rate would increase with decreasing wall thickness. By centrifuging the mold, pressure is exerted on the molten metal, forcing it rapidly into all parts of the mold cavity. Centrifugal action is often used to produce castings with uniform or thin walls that cannot be produced with static pressure alone. Lack of favorable directionality of freezing was a problem with the type 347 stainless steel investment casting illustrated in Fig. 13. This inlet by-pass valve adapter was intended to carry fuel close to a hot part of an airplane engine. Failure of the casting could cause a ﬁre or explosion, with disastrous loss of the plane

The ﬂat surfaces of this investment casting as originally designed (a) created problems of thickness tolerance and surface defects from cracking of the mold precoat. A redesign (b) eliminated the ﬂat surfaces and the attendant defects.

138 / Casting Design and Performance and its occupants. Consequently, this part had to be produced to class 1A quality. As originally designed, the casting did not lend itself to efﬁcient production. The problem was not in ﬁlling the mold but in controlling solidiﬁcation. With a uniform thickness of 0.100 in., the body of the casting solidiﬁed too fast for the ﬁllet sections at the junctions of the body and the ﬂanges to develop enough strength to withstand the contractive forces of the solidiﬁed body. There were hot tears at the junctions and porosity in the 0.100-in. walls. To correct these defects, the casting wall was maintained at its original thickness (0.100 in.) at a point midway between the two ﬂanges and was uniformly tapered to a thickness of 0.150 in. at each ﬂange. Isolated sections

were thus eliminated. To overcome hot tears, the ﬁllet radii at the junctions were increased from 0.10 to 0.25 in. With a slight modiﬁcation in processing, the redesign permitted acceptable castings to be produced, and rejections were few. Although uniform walls are desirable in many castings, in one facet of production of investment castings, uniform walls create problems. The 4130 steel casting shown in Fig. 14(a) illustrates two such problems. The ﬁrst involved cavitation, or contraction of heavy sections of the wax pattern. As a result of this pattern defect, the center of the casting was depressed below the permissible thickness variation. The second problem, although it appeared to be one of processing related to

precoating of the mold, was basically one of design. On smooth, ﬂat surfaces, a mold precoat is likely to crack when the molten metal is poured. If it cracks, but does not peel or wash, the result is a blemish on the surface of the casting where the molten metal ﬁlls the crack. Peeling or washing of the precoat causes larger areas of surface imperfections. The solution to these problems lies in designing to avoid ﬂat surfaces. Figure 14(b) shows how this was done successfully for this particular casting. The ¼-in.-diam riblike raised section along the center of the part acted as a ﬂow channel to assure complete ﬁlling of the heavier section of the casting.

Casting Design and Performance Pages 139–146

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Unequal Sections DEPENDING on casting conditions, certain masses in casting sections may be too large and too remote from a riser to be fed through the intervening sections and made sound. This chapter deals with the design of uneven, enlarged and “isolated” sections for soundness, strength and economy. During the slow solidiﬁcation of a casting, a thin skin of frozen metal forms like a shell around the outer part of the mold cavity soon after the liquid metal is poured. This rigid shell becomes, in effect, a mold for the remainder of the casting, and the volume lost by shrinkage of the metal as it solidiﬁes progressively must be replaced from some feeding source, to avoid shrinkage porosity. The function of the feeding source is to supply liquid metal to the mass continuously until it is frozen solid without porosity. In this process, the freezing front inside the mass must not be cut off from its feeding source before the mass is solid. Sound metal can be produced only if proper directional solidiﬁcation is attained by design and foundry engineering. When it is discovered in the foundry that a casting as designed is practicably or economically unproducible because of unequal or isolated sections, the designer has two recourses available: 1. Thicken (pad) the thin section that serves as a feed path to the heavier remote section 2. Reduce the mass of the remote thick section This is not to imply that risering of isolated sections is not the most practical solution for some castings. However, risering is a function of foundry engineering, rather than of design engineering. Where the casting shape is virtually dictated by speciﬁc requirements of weight, space or function, risering the isolated areas may be the only acceptable method of producing the required casting successfully. As an example, Fig. 1, an aluminum sand casting illustrates riser requirements where heavy sections are isolated from a common feed source and must be risered independently. This casting was designed with close tolerances, for aircraft use. Requirements for sound metal and minimum weight necessitated the use of individual risers for all heavy sections, rather than the addition of padding to the ribs or walls, to permit feeding two or more heavy areas from one

riser. Because risers add to the cost of a casting in several ways (time is consumed in providing them, they reduce the yield of metal poured, and time is required for their removal), every effort should be made to design so that risers are not required, or, if this is impossible, to restrict their use to a minimum. To produce the casting shown in Fig. 1, eighteen risers were needed, to assure the desired high level of quality. In this instance, they were necessary and justiﬁable. However, in many other castings referred to in this chapter the need for risers could be greatly reduced. The rest of this chapter will deal with problems of designing isolated heavy sections that are functionally essential and with the two most efﬁcient solutions to these problems over which the designer has control: (a) providing ﬂow and feed paths, and (b) reducing the mass of the isolated sections.

the section through which the metal is being fed), which is considerably less than the maximum distance that can be fed and still provide sound metal. Because of the difﬁculty of risering the isolated boss, the casting would not have been producible unless other components of the design permitted use of the required padding.

Designs Requiring Padding or Other Feed Paths Many castings beneﬁt from an increase in section size that will create a ﬂow path to ﬁll completely a section that is difﬁcult, or otherwise impossible, to ﬁll, or that will provide an effective feed path from a riser to an isolated heavy section. The functional requirements of a casting may demand the use of padding because of the impracticability of risering an isolated section. The sand cast steel gear housing shown schematically in Fig. 2 contains an internal web that supports a bearing boss totally inaccessible to direct risering. A preliminary casting, made as a test for unsoundness, showed the anticipated defect: shrinkage porosity in the isolated boss. Castings were then produced with chills incorporated in the mold. Although the chills modiﬁed the shrinkage defects, they did not eliminate them. It was necessary, therefore, to add padding, as shown, to obtain adequate feeding of the boss. This change completely eliminated shrinkage defects, without the use of chills. In this casting, it was necessary to feed a distance of about 3½t (3½ times the thickness of

An aluminum sand casting (alloy 356) that required 18 risers, as shown, to achieve sound metal in the heavy sections. The necessity for minimum weight precluded the use of padding to reduce the number of risers.

Fig. 1

A boss isolated from feed metal in this steel sand casting could not be produced with sound metal, even though chills were used. By padding the web, as shown, chills were eliminated, and sound metal was obtained in the boss.

Fig. 2

140 / Casting Design and Performance A casting that presented a similar problem, and for which the same solution was used for reasons of economy, is illustrated in Fig. 3. This wheel casting, produced in a sand mold, incorporates a central hub and a side boss. Both required feeding to obtain soundness. It was possible to riser the central hub, as shown in Fig. 3(b). However, an analysis of cost indicated substantial savings could be realized by padding and feeding the central hub from the side riser, which would still produce the required soundness in the central hub, Fig. 3(c). Using two risers, the metal yield (the amount of metal in the “cleaned” casting, compared with the amount poured into the mold) was

A malleable iron sand cast wheel (a) that could be produced with two risers (b) at a yield of 30% of the metal poured, or with one riser and padding (c) at a yield of 45% All dimensions in inches.

Fig. 3

approximately 30 percent. Using only one riser and adding the required padding increased the yield to approximately 45 percent. The central riser required a core to form the ingate area and prevent mold wash. Producing this core and setting it in the mold entailed additional cost. The lower yield of the two-riser system, plus the cost of the core, made it distinctly advantageous to add the padding and feed the central hub from the riser that fed the boss at the outer ﬂange. This part was cast from malleable iron, and probably less peripheral risering was required than would have been needed for an identical steel casting. This requirement, however, would depend on the level of quality desired. With either metal, padding would be more economical than risering. The feeding distance to the central hub of this casting was approximately 4½t. By rule of thumb, the total thickness of the pad and the casting (at the location of padding) should be not less than one-ﬁfth of the metal feeding distance. Or, stated in another way, metal can, as a rule, be properly fed for a distance of ﬁve times its thickness. This is not a true maximum; castings of optimum design have been fed to distances of 10t or 15t without sacriﬁcing soundness. The 5t rule, however, is a safe generalization and a good point at which to start. If the function of the part involves rotation, it is often possible to add material diametrically opposite the padded section, to counterbalance the pad. Because this increases the strength of the casting, it may permit a reduction in thickness of the web section to the point where the ﬁnal weight of the casting would remain essentially unchanged. In such a design, the padding would become a functional part of the casting. An efﬁciently designed casting may utilize necessary component sections as ﬂow and feed paths for molten metal. This is exempliﬁed by the sand cast malleable iron gear housing shown in Fig. 4. In this casting as originally designed, Fig. 4(a), three ribs provided

necessary rigidity and support for the two bearing bosses (one on each side of the casting). With the part molded as indicated by the parting line, the boss in the cope side of the mold was easily risered and presented no problems. The boss in the drag half of the mold, however, could be fed only by appending risers to the ﬂange opposite the ribs and feeding metal to the boss through the ribs. Because the ribs were too small in cross section, the boss could not be fed adequately, with the result that the metal in the boss was porous, and, in actual operation of the part, oil from inside the housing seeped through these pores. A revision, Fig. 4(b), removed the center rib and distributed its metal equally between the two side ribs. This increased the rib width from 0.50 to 0.75 in. Where previously the metal in the boss was fed a distance of 7t, it now was fed only 4½t. This improvement was made without adding any weight to the casting. The two heavier ribs provided adequate strength (equal to that of the three lighter ribs), and the defects were eliminated because these ribs were now heavy enough to permit adequate feeding during solidiﬁcation of the bosses. When padding is required in order to produce sound, acceptable castings, but must subsequently be removed, every effort should be made to add the padding where its removal will not be difﬁcult. Tubular-shaped castings, in particular, lend themselves to this treatment. Consider, for example, the ball-joint casting of Fig. 5. Produced from stainless steel in a sand mold, this casting was required to meet stringent quality standards. Each casting was inspected radiographically, and all discernible defects were cause for rejection. As originally designed, all castings were rejected because of shrinkage porosity caused by an unfavorable freezing pattern. An analysis of the casting revealed that simple tapering of the walls in two areas (see section A-A and B-B of Fig. 5) would permit placing risers so that all sections would taper

As illustrated, casting members may be used as feed paths for motten metal. The ribs in the original design (a) were too thin to permit adequate feeding of the bearing boss, causing it to be porous. By eliminating the center rib and distributing that metal to the two outer ribs (b), the boss was adequately fed, and the porosity encountered with the original design was eliminated.

Fig. 4

Design Problems Involving Unequal Sections / 141

Fig. 5

Fig. 6

Padding, properly placed, permitted production of this stainless steel sand casting to required soundness

Unusual padding of the center cavity of this stainless steel shell mold casting simpliﬁed production. Without the padding, complicated risering would have been necessary and would have caused high residual stresses.

toward a riser and all riser contacts could be removed easily. The metal added in the form of padding also was removed easily, during the machining operations already required in order to produce the desired surface on the ball section of the casting and on the mounting face of the ﬂange. The addition of the tapered padding resulted in a reduction in rejection rate from 100% (porosity) to only 3% (all causes). A unique approach to padding a casting in order to feed isolated heavy sections more efﬁciently is shown in Fig. 6. Designed for production in a shell mold, this 17-4 PH stainless steel valve body has ﬁve heavy sections that appear to require risering for soundness. However, this amount of risering would add substantially to the cost of the casting. Also, the resulting

complex freezing pattern undoubtedly would create high internal stresses, which, in turn, would cause distortion before and during machining. Because the machined surfaces of the casting required high dimensional accuracy, distortion might have been a serious problem. To determine the amount of padding required, several castings were produced with a standard, straight-through core that allowed 1/8 in. of additional metal on the inside diameter of the casting for machining stock. The castings were rigged with a large riser at the top and a side riser (also serving as the sprue), as shown in the lower view of Fig. 6. From the experience gained, the foundry estimated that castings produced in this manner would average about 50 percent scrap because

of shrinkage defects. In contrast, it was estimated that producing the part with heavy padding, as shown, would reduce the rejection rate to less than ﬁve percent. Machining costs increased due to the extra metal to be removed, but this cost increase was less than the costs with a 50 percent rejection rate. The heavy padding was adopted for production. This casting and its gating and risering system demonstrate several desirable casting features. When metal is poured into the mold at the sprue, the pool at the bottom of the sprue serves to quiet the action of the metal, and turbulence is minimized as metal enters the mold cavity. Although there will be some cascading as the metal enters the body-forming section of the mold cavity and drops to the bottom ﬂange section, entrapped gases, if present, will rise as the metal ﬁlls the mold and move to the top center riser. All dross and foreign matter can also be ﬂoated into this riser, leaving only sound metal in the casting section. One exception is the casting ﬂange immediately adjacent to the ingate. At this location, the mold must be vented to permit entrapped gas to escape easily. All sections of the casting taper toward the top riser; this assures directional solidiﬁcation and enables the sections to be fed from the top riser. An exception is again the heavy section immediately adjacent to the gate; however, because the gate is large, this sprue will function as a riser and will feed molten metal as it is required during freezing. As was expected, shrinkage porosity and impurities were found only in the risers and the metal to be removed by machining. After machining, castings were sound and completely satisfactory. Heavy sections that might have been isolated from a feed source with regular coring procedures, were fed easily because of the extreme wall thickness. Wall-thickening is practical from a cost stand-point when the added metal can be readily removed by boring, turning or other simple machining.

142 / Casting Design and Performance The lifting-eye casting of Fig. 7(a) appears to have a simple conﬁguration; however, it presented several production problems. Because a cored cavity was located in the large ﬂat face, it was necessary to select a parting plane that was incompatible with the most efﬁcient gating and risering. With the riser placed as shown, and with padding added to encourage complete feeding of the casting during cooling and freezing, premature freezing was encountered in the area adjacent to the cored cavity. Despite the placement of exothermic sleeves around the riser, the use of metal inserts rammed-up in the core and heated just before they were inserted in the mold, and the starting of pouring almost immediately after the mold was closed, the rejection rate was 25 percent of the castings poured, and the remainder of the castings required extensive welding to make them usable. Rejections were caused mainly by internal and external shrinkage defects. Elapsed time between receipt of the order and delivery of acceptable castings was six months. A request by the foundry to eliminate the cored hole was granted. The patterns were revised to make the parting plane identical with the mounting face, as shown in Fig. 7(b). The riser was moved to the center of the mounting face. With these changes, there was a gradual transition in casting section area: the section farthest from the riser was the smallest, and the area immediately under the riser was the largest. By using an exothermic sleeve around the riser, the metal was kept molten and properly fed the casting during the entire cycle. With these revisions, the rejection rate of these castings dropped to only two percent, and no repair welding was required. This permitted a 32 percent reduction in the selling price of the castings. The redesign enabled the foundry to deliver acceptable castings 10 days after receipt of the order. Isolated heavy sections in aluminum castings may require feed pads in order to produce sound metal. Because of the relatively low melting point of aluminum, the quick freezing of heavy sections is often accomplished by means of metal chills. However, chills are sometimes impractical, and they always increase cost. Feed pads are often more practical. The end use of the aluminum sand casting illustrated in Fig. 8 required that four bosses be incorporated in its design. Castings were rejected because of porosity in these bosses, where sound metal was essential. Chills, conforming in shape to the bottom half of the bosses, were incorporated in the core that formed the inner conﬁguration of the casting, but only part of the porosity was eliminated by this revision. Although increased weight was objectionable, four ribs were added, one over each boss, as shown in Fig. 8(b), to allow metal to be fed to the bosses from the risers on the top ﬂange. Chills were used, to minimize the size of feed pads. Although risers could have been added to the outside of the vertical wall immediately

Fig. 7

Although apparently of simple shape, this lifting eye, a stainless steel sand casting, could not be produced efﬁciently to the original design (a). Redesign (b) eliminated a cored hole and facilitated foundry production.

Fig. 8

Four bosses isolated from all practical sources of feed metal in this aluminum alloy sand casting (a) were produced to the required soundness by adding minimum ribs (a weight consideration) and conforming

chills (b).

opposite each boss, this would have introduced problems of patternmaking and molding and could have required a special cleaning or trimming operation on the rough castings. Cost was the determining factor in the decision to

pad and chill, as shown, rather than to add four outside risers. A similar problem, involving a wall too thin to feed adjacent heavier sections, is illustrated by the aluminum sand casting in Fig. 9.

Design Problems Involving Unequal Sections / 143 Because the part was to serve as a structural member, each casting was inspected to stringent radiographic and ﬂuorescent-penetrant standards. As originally designed, with a 3/32-in. wall thickness, all castings were rejected because of shrinkage defects in this wall. In an attempt at successful production of these castings, a circular runner was incorporated, and gates were added at several points along the periphery of the upper ﬂange. Four risers were placed as indicated, on the top ﬂange over the wall, to provide a feed source. The molten metal entered the mold cavity at the gates, traveled through the ﬂange, through the vertical wall sections, and into the lower ﬂange. The large amount of metal that passed through the mold section forming the 3/32in. wall contributed considerable heating effect to that mold section and allowed the 3/32-in. wall to remain open just long enough to feed the larger sections below. However, because of its thin cross section, the 3/32-in. wall froze almost

Shrinkage in this aluminum sand casting (alloy 355) where the wall was too thin was eliminated by increasing the wall to 5/32 in.

Fig. 9

Fig. 11

simultaneously at all locations, and the metal lost to the lower sections could not, in effect, be replaced; the wall exhibited severe microshrinkage. Adding side risers to the wall was impractical because they would complicate the coring required by undercuts formed by the ﬂanges. Also, because of the thinness of wall, distortion would have been difﬁcult to avoid when the side riser contacts were subsequently machined. Bottom gating was tried. The result was gross

Fig. 10

Padding this aluminum permanent mold casting facilitated efﬁcient foundry production.

porosity in the 3/8-in. wall, as well as microshrinkage in the 3/32-in. wall. Thus, the only practical solution was to increase the thickness of the 3/32-in. wall to 5/32 in. This increase completely eliminated microporosity. The bottom ﬂange was chilled, to induce solidiﬁcation before the metal in the wall section froze. Although this was an aircraft casting, increasing the wall thickness and thereby adding unwanted weight was mandatory in order to produce acceptable castings. Permanent mold castings require ﬂow and feed paths similar to those required by sand castings. The example shown in Fig. 10 was produced in a simple two-piece mold with the parting line as indicated. A three-piece mold would have been required if the hub had been risered directly. This might have required elimination of the 0.65-in.-diameter cored hole, which, in turn, would have had to be drilled in a subsequent operation, at extra cost. These castings were originally produced without the feed pad, and all were rejected for shrinkage porosity at the top of the hub. This condition was attributed to localized heating of the mold by the molten metal entering at the riser-sprue and ﬂowing directly downward as the mold was ﬁlled. A feed pad 1/16 in thick by ½ in. wide was added. The heating of this section of the mold by molten metal, and the enlarged size of the feed pad, combined to keep the metal in the pad molten until the hub had solidiﬁed. Freezing then progressed through the pad to the riser. Sound castings were produced, capable of passing leakage and radiographic inspections. Another example of a permanent mold casting that beneﬁted from added feed paths is the aluminum piston shown in Fig. 11. As originally designed, Fig. 11(a), the walls of the piston were of uniform thickness, which encouraged centerline porosity in the skirt section and some porosity in the wristpin bosses. Porosity was attributed to the simultaneous freezing of all sections of the skirt and the freezing of the wall section above and around the wrist-pin bosses before the freezing of the bosses. Consequently, feeding of metal to the skirt sections and bosses was inadequate.

Uniform walls of this permanent mold cast aluminum piston (a) resulted in simultaneous freezing, which caused porosity and centerline shrinkage. By tapering the walls and adding ribs for metal feed paths to the bosses (b), acceptable castings were easily produced.

144 / Casting Design and Performance To correct this condition, the skirt of the casting was tapered from the original thickness of 0.19 in., immediately adjacent to the dome, to 0.12 in. at the extreme end. Four ribs, two at each boss, were added, to serve as feed paths. The taper in the skirt assured satisfactory progressive solidiﬁcation of metal in that section. The four ribs provided adequate feed paths for the bosses and provided additional strength in the wrist-pin bosses. The aluminum casting shown in Fig. 12 also was made in a permanent mold. As part of an aircraft fuel system, the casting was produced to exacting quality requirements, one of which was retention of aircraft fuel at a pressure of 35 psi. This pressure-tightness required perfectly sound metal; any porosity or defects would permit leakage. Molding was complicated by the location of the small tubular outlet. With this outlet located away from the optimum parting plane, a core (or a loose piece) was required in the mold to form the outside contours of the outlets. The draft requirements of this core and the smaller core necessary to form the inside of this tubular section combined to create a heavy section remote from any practical riser location. Porosity in this area resulted in a 40 percent rejection rate. To improve producibility, although at the penalty of adding undesirable weight, a small amount of padding was added to the wall connecting the tubular section with the mounting ﬂange. A riser was appended to the ﬂange, to feed the heavy section at the junction of the wall and the tubular section. With these revisions, sound castings were produced. This casting illustrates the desirability of considering molding requirements in the initial stages of product design. Figure 13 shows a permanent mold casting of aluminum alloy 355 which might have presented difﬁcult production problems but did not because the problems were anticipated and

Fig. 12

eliminated in design. The key to success was the adequate size of the posts connecting top and bottom sections. The design permitted use of a comparatively simple three-piece mold and a central core. The parting lines of the mold ran radially from the center of the casting on planes that bisected the solid posts. Gates were located at one post only, as shown. Risers over each solid post could successfully feed all of the lower part of the casting, because of the adequate size of the posts. The heavy section bordering the small insert near the base of the casting was made sound by the chilling effect of the insert and because of its location under a solid post. Investment castings beneﬁt from the incorporation of feed paths primarily in terms of cost reduction. In the investment process, additional runners can often be added to the rigging, to conduct metal to a part of the casting that cannot be ﬁlled or adequately fed through adjacent sections. An excellent example of cost reduction associated with the proper use of feed paths is demonstrated by the casting shown in Fig. 14 (also discussed in greater detail as Fig. 31 in Chapter 9). As originally designed, this casting closely approached but did not reach the conditions for production at maximum efﬁciency. It had relatively heavy sections at each end in the form of rings or ﬂanges, and in the center area in the form of hubs. Experimental castings revealed that, even with the small ribs included, the cross-sectional area of the casting body was too small to permit the entire casting to be fed from one end or from the hubs. Consequently, six gates were required, to feed all sections adequately. Castings were later produced with the ribs increased in size, as shown in the redesign of section A. Only two gates, located at one end on the ﬂange and opposite the ribs, were required. Production experience indicated that

castings made with the larger rib could be produced at a saving of 28 percent.

Designs That Reduce the Mass of a Remote Section Unlike the examples cited in the ﬁrst part of this chapter, some castings can be improved in producibility by reducing the size of remote heavy sections, thus lessening the need for feed metal and making it possible for existing casting sections to provide the necessary feed paths. In reducing the size of an isolated section, the ﬁrst beneﬁt is a reduction in weight, and thus a reduction in metal cost. The size reduction can also decrease the number of risers required. Still another beneﬁt is the elimination of certain potential casting defects. For example, the steel lever shown in Fig. 15(a), a sand casting for an aircraft application, was difﬁcult to produce because hot tears developed in the thin wall connecting the hub and the fork. Usually, shrinkage defects also developed in the heavy base of the fork. Risering the heavy base eliminated the shrinkage defects but aggravated the hot tearing. The revision shown in Fig. 15(b) gave acceptable castings. In this redesign, the web separating the two ﬁngers is tapered, with the heaviest section adjoining the hub. Only one riser, located at the hub, was required, to feed all sections of the casting adequately. The redesign used somewhat less metal, with no sacriﬁce in strength. Housings. Too often, cast housings are developed by specifying walls of uniform thickness spaced at a deﬁnite distance from the enclosed mechanism, to provide uniform clearance. To these walls, bosses, ribs and other necessary components are appended without careful planning. The resulting conﬁguration is

Molding and production of this aluminum permanent mold casting (a), part of an aircraft fuel system, were complicated by the small tubular section that protruded at an angle. Padding (b) and core pulls (c) were necessary to produce the part.

Design Problems Involving Unequal Sections / 145

Fig. 13

Good design facilitated production of this permanent mold casting from aluminum alloy 355. (See text.)

Fig. 14

Six gates were required to produce this gimbal ring, an 8620 steel investment casting. By increasing the size of the ribs, four gates were eliminated and cost was decreased 28%.

then dimensioned and given to the foundry to produce as a casting. Seldom will this method of design yield a shape compatible with efﬁcient production in the foundry. The malleable iron casting shown in Fig. 16 is a gear housing for a power-steering mechanism. Six risers were required, to obtain the desired soundness of metal. A large number of castings were produced in sand molds to the original design. To increase efﬁciency of production, the design was reviewed to determine whether the number of risers could be reduced. It was found that risers could be minimized by squaring the gear cavity and narrowing the mounting bosses, as shown in the redesign. These revisions permitted the elimination of three of the six risers. An aluminum hydraulic pump body (alloy 319), produced as a permanent mold casting, incorporated a heavy section that could not be fed unless padding were added to both the riser and the casting body (see section B-B, original design, Fig. 17). This heavy section surrounded the cored hole shown in sections C-C and D-D. Without padding, the heavy section was cut off from feed metal in the riser because of premature freezing of the metal at the junction of the heavy section and the outer casting wall. Although acceptable castings were produced with this padding, it was necessary to machine off the padding, which increased the cost. By reducing the volume of the heavy section, as shown in sections B-B and D-D, the casting was produced successfully without padding. To assist in ﬁlling the mold cavity, the bottom wall of the casting was increased in thickness from 1/8 to 5/32 in., as shown by section A-A.

Redesign eliminated a remote heavy section and encouraged a beneﬁcial freezing pattern in this sand cast steel lever.

Fig. 15

146 / Casting Design and Performance

Fig. 16

A revision in design of this sand cast malleable iron gear housing reduced the mass of metal at each of the four mounting lugs. This decreased the riser requirement from six to three and substantially reduced cost

Decreasing the mass of metal around the cored hole (D-D) and increasing the thickness of the plate section of the casting (A-A) permitted efﬁcient production of this aluminum pump body (alloy 319) in a permanent mold without padding (B-B).

Fig. 17

Casting Design and Performance Pages 147–153

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Junctions* IN MOST CASTINGS there are junctions between intersecting component members. It is customary to incorporate ﬁllets at these junctions, to eliminate sharp inside corners and the attendant problems of stress concentration. The designer should be cognizant of the casting production problems introduced by junctions. Designs should strive for maximum soundness while retaining compatibility with other casting requirements. For practical purposes, and discounting the initial chilling effect of the mold, the rate at which molten metal solidiﬁes is the same. for a thinner section ( 1 in. or 25 mm) as for a thicker section (4 in. or 100 mm). Because, in general, metal in a casting solidiﬁes from the mold surfaces toward the center of a section, thin sections will freeze before heavy sections. Hence, where a thin section is immediately adjacent to a heavy section, and where the only source of feed metal for the heavy section is through the thin section, the thin section can freeze ﬁrst and shut off the source of feed metal for the heavy section. With no source of feed metal to replace the volume lost because of metal shrinkage, such a heavy section will be porous. Also, with the onset of freezing, the thinner section starts to contract while the heavy section is still solidifying. This can cause stress to develop in the casting, resulting in cracks at the hottest (weakest) section. Different sections of a casting can be conveniently compared by inscribed circles, as shown in Fig. 1. The ratio of the areas of the inscribed circles is an indication of relative freezing times. The large sections ﬁnish freezing proportionately later than the small sections. The inﬂuence of ﬁllets in increasing the mass of a junction can also be seen in Fig. 1. As shown, a 7/8-in. (22-mm) circle can be inscribed at an L-junction of two 3/4-in. (19-mm) members having no ﬁllet Under the same conditions except for a ½-in. (13 mm) ﬁllet, a 11/16-in. (27 mm) circle can be inscribed. This represents a 47 percent increase in area, volume and mass of metal. The ability of a sand or ceramic mold to absorb heat must be considered when junctions are being designed. In general, heat from the molten metal is transferred into the mold in a direction perpendicular to the interface, or radially if the interface is curved. Figure 2 shows schematically the directions of heat transfer in

two different types of L-junctions. At the outer surface of the junction shown in Fig. 2(a) there is more mold material available to absorb heat from the molten metal than is available adjacent to the straight portions of the casting that are well away from the corner. Consequently, the metal at this part of the junction will cool at a somewhat faster rate than will the metal in the remote straight sections of the casting. However, the metal at the inside corner of the junction will cool at a much slower rate than the metal at the outside corner or in the straight areas, because there is much less mold material to absorb the heat of the molten metal. As a result, the metal in the junction near the inside corner surface will be last to cool and will be cut off from feed metal if the source is through the straight section of the casting. A shrinkage defect will result, as shown. When the corner is provided with a ﬁllet and uniform wall thickness, the shrinkage cavity is reduced in size. When the ﬁllet radius equals the wall thickness of the junctioning members of the casting, as in Fig. 2(b), shrinkage voids usually are eliminated, even under the most adverse conditions, because the metal cools uniformly. Theoretically, for steel castings ﬁllet radii should equal the thickness of the heaviest of the junctioning wall sections when the thickness of this wall is ½ in. (13 mm) or less. For wall thicknesses of ½ to 1 in. (13 to 25 mm), a radius equal to 3/4 of the heaviest wall thickness usually

Fig. 2

*Adapted from Casting Design, American Society for Metals, 1962

Fig. 1

Inscribed circles facilitate comparison of metal mass in casting junctions.

Schematic representation of the effect of junction shape on heat transfer from solidifying metal to the mold

148 / Casting Design and Performance

Fig. 3

Mold restraint to normal contraction creates areas of stress concentration in casting (a). Designs that permit unhindered contraction (b) are relatively free from residual stresses.

Tapering ﬁllets from a riser to the extremities of a casting encourages soundness and often permits reduction in the weight of a casting

Fig. 4

In designing to prevent shrinkage cavities, one of the most difﬁcult problems is to know when and where risers may be used by the foundry. The positioning of risers is a matter of judgment based on the experience of the foundry engineer. Also, other considerations, such as tooling-point locations, may prevent a riser from being placed in the most advantageous position. Thus, the casting may beneﬁt if it is designed with the assumption that none of the junctions can be fed directly. As an example, consider the typical casting section shown in Fig. 2(a): If a riser could be located directly in contact with the junction so that it could feed the enlarged mass, no shrinkage cavities would occur. However, if the only source of feed metal is through one of the arms, shrinkage defects would occur, because the arms would freeze ﬁrst and cut off the heavy junction from the feed metal. Hence, when feasible, a junction similar to Fig. 2(b) is preferred.

Junction Elements Fig. 5

Comparison of defects obtained in L-junctions of castings produced under conditions controlled in such a manner as to encourage the defects. Section size of tests: 3 by 3 in.

is sufﬁcient. For wall thicknesses greater than 1 in. (25 mm), a ﬁllet radius of ½ of the heaviest wall thickness normally will sufﬁce. In practice, the stresses developed in the casting during solidiﬁcation and cooling are more important in determining ﬁllet size than is the wall thickness of the casting in the area of the ﬁllet. Where adequate or even partial feeding of the junctions is achieved, smaller ﬁllet radii than those mentioned above may be allowable if the design of the part does not introduce severe stresses from solidiﬁcation or cooling. For example, the ﬁllet radii required in a casting of the I-beam type, as in Fig. 3(a), are larger than those required in T-sections, Fig. 3(b). In the cooling of the I-beam type of casting, stresses are set up because the contraction of the web section is restrained by the mold material between the standing ﬂanges of the I-beam section. Thus, a relatively large ﬁllet at the junctions of the I-beam components is required, to prevent hot tearing at the ﬁllet area. In the

T-section, because the casting is free to contract, no cooling stresses are set up by the mold material; therefore, smaller ﬁllets are practical. When ﬁllets are required in an area where the junctions can serve as ﬂow and feed paths for the molten metal, it is advantageous to taper the ﬁllets so that they are largest in the area where the riser will be placed and smallest at the point farthest from the riser. This tapering assists in creating a beneﬁcial freezing pattern of the molten metal. Freezing will begin at the areas farthest from the riser and will proceed uniformly toward the riser. The junctions, being heavier than the adjacent sections, will freeze after these adjacent sections and thus will serve as sources for feed metal. Because of the tapering of the ﬁllet in the junctions, freezing will progress toward the riser; this assures adequate feeding of the junctions by the riser. This principle, which is illustrated in Fig. 4, often permits a reduction in the weight of a casting.

A common understanding of the problems of junction and ﬁllet design have helped designers and metalcasters evaluate the speciﬁc effects of design variables on the soundness of metal in different types of intersections. The inﬂuence of deﬁnite junction and corner conﬁgurations on the size and number of defects is brieﬂy described here from one of the ﬁrst systematic studies by Briggs et al (AFS Transactions, vol. 46, 1939, p 605). The early work is still valid and instructive. Steel casting junctions were considered for the ﬁve joint types, represented by the letters L, T, V, X and Y. All other conﬁgurations at corners could be considered modiﬁcations of one or more of these ﬁve. Designs in the forms of these letters, and their inﬂuence on shrinkage, were studied in detail by producing actual castings, summarized here in Fig. 5 to 9. Except where indicated, all casting sections were 3 by 3 in. (75 by 75 mm). The arms were about 24 in. (61 cm) long, with the risers placed at the extreme ends. The risers were utilized as sprues for pouring. Plain carbon basic electric

Design Problems Involving Junctions / 149

Fig. 6

Fig. 7

Comparison of defects in T-junctions produced under controlled conditions to encourage the defects. Section size of tests: 3 by 3 in. unless otherwise noted

Comparison of defects obtained in V-junctions of castings produced under conditions controlled in such a manner as to encourage the defects. Section size of tests: 3 by 3 in. Members intersect at 45 degrees.

steel was poured into dry sand molds from a teapot ladle. All castings were horizontal when poured. Heavy sections at the junctions could be fed only through the thinner arms. Three-inch (75 mm) sections were chosen because they were large enough to exhibit pronounced defects and were less sensitive to variations in the mold and metal temperature than thinner sections would have been. Junctions were modiﬁed by using radii of different sizes in the corners, or by varying the thickness of one of the junctioning legs as shown. The castings were examined radiographically. The L-junctions in Fig. 5 show that defects were eliminated by using a ½-in. (13 mm) ﬁllet and reducing the wall thickness at the corner. No detectable defects were present in the L-junction where the radius of the ﬁllet was equal to the wall thickness and a uniform wall was maintained at the corner. In such a design, however, a centerline weakness similar to that in a uniform wall of sufﬁcient length may be

present at the interface where the two freezing fronts meet. The T-junctions shown in Fig. 6 indicate that defects can be eliminated in this type of junction only by coring a hole at the center of the junction of the two members. It is doubtful, of course, whether this is a practical solution in production castings. Any junction through which a hole could be cored could probably be risered to provide adequate feed metal. A depression in the crossarm of the Tjunction substantially reduces the defect but does not eliminate it. Because the size of the defect increases as the ﬁllet increases (causing the mass to increase), the smallest practical ﬁllet compatible with other casting requirements is recommended. V-junctions (Fig. 7) formed by uniform sections will not be free from shrinkage cavities, primarily because a hot spot is readily developed in the mold at the junction between the two sections. Mold sand, a poor conductor of heat, cannot conduct the heat away from

the sand enclosed between the sections of the casting as quickly as the heat is transferred from the molten metal into this mold section. Consequently, this portion of the mold becomes hot enough to cause the adjacent metal to become the last to freeze. This condition exists at the inside corner of all junctions; however, because of the acute angle formed by the V-junction, the condition is more pronounced. The defects decrease in area as the inside radius at the V-junction increases in size. As in the L-junction, a slightly reduced wall thickness at the V-junction encourages sound metal. X-junctions, as shown in Fig. 8, could not be designed in this study so that all defects were eliminated. By coring a hole through the junction, however, the defects were greatly reduced in size. As with the T-section, coring may not be a practical solution. It may be possible to attach a riser to the core print area as easily as it would be to set a core. Y-junctions as shown in Fig. 9, produced defects unless a triangular hole was cored through the heavy section of the junction. Although a round hole substantially reduced the defective area, it did not eliminate the defects completely. Note that the areas of the defects remained almost constant regardless of the design alteration of the uncored Y-junction. In this study, all junctions would have had substantially sound metal if risers had been located directly over the junctions. However, where feeding of the mass of molten metal at the junctions was inadequate during cooling and solidiﬁcation, shrinkage defects occurred. Figure 10 shows defects found at junctions in actual production castings. These substantiate the data from the many trial castings in Fig. 5 to 9. Chills can decrease the area of corner defects. Although in many castings chills may be of value in obtaining the desired soundness of metal, the designer should not depend on them to produce acceptable castings. Chills add (often substantially) to the cost of

150 / Casting Design and Performance

Fig. 8

Comparison of defects obtained in X-junctions produced under controlled conditions to encourage the defects. Section size of tests: 3 by 3 in. unless otherwise noted. Center lines of offsets are 8 in. apart.

Fig. 9

Comparison of defects obtained in Y-junctions produced under controlled conditions to encourage the defects. Section size of tests: 3 by 3 in. Members intersect at 60 degrees.

casting production, and increase the dimensional variation of the areas which the chills contact. It is preferable to obtain the desired level of soundness by design.

Junctions in Aluminum Castings Aluminum produces defects in the same areas but of smaller magnitude than steel,

because less heat is transferred from solidifying aluminum than from steel. The mold material could dissipate the heat from aluminum more easily, and there would be less likelihood of hot spots forming. Although more liberties can be taken with the design of junctions for an aluminum casting than for a steel casting, junctions should still be fed adequately if shrinkage cavities and porosity are to be avoided.

As an example, the problems encountered in producing the aluminum alloy sand casting, a primary ﬂight-control-mechanism support, shown in Fig. 11 were concerned with the junctions of the ribs with the walls of the rectangular tube sections (view C-C), and with the L-junctions in the rectangular tube section (view A-A and section B-B). In the original design, extensive shrinkage porosity was encountered at these junctions. Because class

Design Problems Involving Junctions / 151

Fig. 10

Defects in production castings verify results obtained in the test castings shown in Fig. 5 to 9.

Fig. 11

An aluminum alloy sand casting (a primary support for a ﬂight-control mechanism) that required redesign of several junctions before sound, acceptable castings could be produced

1-A x-ray soundness was speciﬁed for this casting, the defects were cause for rejection. A review of the processing revealed no practical way to supply feed metal to these areas. Consequently, a design revision was necessary

to reduce the mass at the junctions of the ribs with the tubular section (see locations A, B and C in views C-C for both original and revised design), and to alter the L-junctions of the tubular sections (see view A-A and section B-B).

This reduction in mass at the junctions (locations A, B and C), the increase in the inside radius, and the establishment of a uniform wall at the L-junctions of the tubular sections resulted in the production of sound castings.

152 / Casting Design and Performance If this casting were to be produced from steel, further reduction of the mass of metal at some of the junctions would undoubtedly be required.

T-Junctions Versus Y-Junctions For a given material, section thickness and ﬁllet radius, T-junctions present fewer solidiﬁcation problems than Y-junctions. This becomes apparent by comparing the areas of the inscribed circles in the T-junction and Y-junction shown in Fig. 12. With wall thicknesses identical and the same size ﬁllets at each junction, the area of the circle inscribed in the Y-junction is about 35% greater than the area of the circle in the T-junction. If a junction is not easily fed, it is obvious that the better design would incorporate T-junctions rather than Y-junctions wherever possible. In many castings, this is accomplished easily by the simple design expedients shown in Fig. 13. The 17-4 PH stainless steel casting shown in Fig. 14 provides an excellent example of the advantages of the T-junction over the Y-junction. Produced as an investment casting, this structural aircraft part required class 1-A x-ray soundness. Of the castings produced to the original design, 87 percent were rejected because of defects at the Y-junctions formed by the casting body and the arms that extend from the body at an angle of about 45 . It was not feasible to add runners to the rigging to supply feed metal to these problem junctions, because removal of the stubs where runners would contact the casting would have entailed expensive machining operations. Consequently, the casting was redesigned to reduce the relatively large mass of metal formed by the Y-junctions. This was accomplished by revising the arms so that T-junctions (similar to the "Recommended" junction in Fig. 13a) were formed with the tubular body. This revision can be seen by comparing the original and the redesign shown in Fig. 14. After the redesign of junctions, no castings were rejected.

Comparison of the mass of metal of Y- and T-junctions with all walls the same thickness and all radii equal. The inscribed circle in the Y-junction (a) is about 35 percent larger than inscribed circle in junction (b).

Fig. 12

Two practical methods of revising Y-junctions to T-junctions, to reduce the mass of metal at the junctions. With all walls the same thickness and all radii equal, in (a) the circle inscribed in the Y-junction is approximately 23 percent larger than the circle inscribed in the T-junction. In (b) the circle inscribed in the Y-junction is approximately 26 percent larger than the circle inscribed in the T-junction.

Fig. 13

Y-junctions caused porosity in this 17-4 PH stainless steel investment casting. (a) Revising to T-junctions. (b) eliminated the cause of the defects.

Fig. 14

Junctions of Unequal Sections Where walls of two different thicknesses form a junction, it is advisable to incorporate a gradual increase in the thickness of the thinner section so that it is about equal in thickness with the heavier section at the point of juncture. Recommended methods of achieving this blending in T-junctions and L-junctions whose intersecting walls are unequal in thickness are shown in Fig. 15. At an L-junction, if one or both of the outside surfaces are to be machined, the outside radius must be omitted and the corner left sharp. The sharp corner, however, is likely to contain larger shrink defects than would be present with the rounded corner. Figure 16 shows recommended methods for blending abrupt changes in the thickness of a

Fig. 15

Recommended designs for proper blending of unequal sections for a T-junction (a) and an L-junction (b)

Fig. 16

Recommended designs for blending abrupt changes in wall thickness

Design Problems Involving Junctions / 153

Fig. 17

High residual stress and defects encountered in original design (a) were eliminated by redesign (b) of this winch end-plate, sand cast of steel.

wall to avoid or minimize stress concentration. Figure 16(a) shows a method recommended where the thicker section is less than 50 percent larger than the thin section; Fig. 16(b) shows a method recommended where the thicker section is more than 50 percent larger than the thin section. An example of blended junctions is provided by the gate valve casting shown in Fig. 3 of the article “Design Problems Involving Uniform Sections.” This casting incorporated every advantageous design detail so as to obtain the highest degree of soundness in the metal. Even though the junctions of the ﬂanges and the valve body were adequately fed, they were

properly blended to avoid the possibility of stress concentration. In the steel sand casting shown in Fig. 17 astute design eliminated the junctions and thus the defects they encouraged. The original design, Fig. 17(a); encouraged shrinkage porosity at the junctions of the ribs and the ﬂat plate section. Those castings not rejected because of porosity were highly stressed internally because of the uneven cooling rate resulting from the heavy section of metal at each junction. A redesign that replaced the opposing ribs with radial corrugations is shown in Fig. 17(b). These corrugations increased the strength of the casting enough to permit a reduction in the wall

thickness of the hub. This brought the hub and plate sections closer to a uniform cooling rate and substantially reduced the internal stresses previously encountered. Other advantages were a 24 percent reduction in weight and a 22 percent reduction in the cost of producing the casting. In aircraft-wheel design, similar simpliﬁcations have been achieved. The intricate, highly ribbed wheel structures of Fig. 9 were replaced by wheels of simpler design; the simpliﬁed wheels had increased service life under critical fatigue loading, and were lower in cost.

Casting Design and Performance Pages 155–161

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Design Problems Involving Distortion DISTORTION and hot tearing are closely related. The forces that cause distortion can also cause hot tearing, if they are great enough or are applied at a critical point. These forces result from differences in the times at which various parts of a casting freeze and start to contract, and from the resistance offered by unyielding mold sections. The solidiﬁcation time differences and the mold resistance both act to restrain normal contraction of the metal. This restraint can result in residual stress, distortion or hot tearing, depending on the magnitudes of force and restraint.

Distortion Due to Differences in Solidiﬁcation Times Figure 1 shows three castings that illustrate ways in which distortion can occur when a thin member solidiﬁes and contracts considerably before a thick member. The casting in Fig. 1(a) contained a heavy center section, making it easy to feed and ﬁll. However, the thick and thin sections froze in such a way as to produce severe distortion. The side plates froze and started to contract considerably in advance of the heavier center section. Then, as the center section solidiﬁed and cooled, contraction was restrained by the thin sections. This created stress in the side plates, causing them to buckle upward, as shown by phantom lines in section A-A.

Fig. 1

Although this distortion could be partially controlled by increasing the thickness of the 0.06-in. (1.5 mm) section to 0.187 in. (4.75 mm) (with subsequent machining off the excess material), the 12-in. (30 cm) radius was difﬁcult to machine to close tolerances in production. This part could have been cast without excessive distortion if its cross section had been redesigned in any of the ﬁve ways suggested in Fig. 1(a). These redesigns are numbered in the order of preference, on the basis of potential foundry difﬁculties. The practicability of the slender core in the ﬁfth choice is questionable. A cast ﬁtting made of a precipitationhardening stainless steel is shown in Fig. 1(b). Although the molten steel had good ﬂuidity, and no difﬁculty was encountered in producing a sound casting, distortion was a serious problem. The junction of the two walls formed an area of increased cross section, and it was necessary to place a riser adjacent to this junction to assure sound metal. Because of the riser, metal temperature in this corner was higher than at any other location in the mold. The walls away from the corner solidiﬁed ﬁrst, while the increased mass at the junction and the retarded cooling rate of the metal at the surface of the ﬁllet caused the ﬁllet surface to freeze later than the sharp outside corner. Distortion resulted, as indicated by the phantom lines in Fig. 1(b). Although the casting could be straightened in a press after annealing, the

same pattern of distortion returned after subsequent hardening. Distortion could have been reduced by providing a radius at the outside corner, as indicated by the dotted line A. This would have created a uniform section thickness. If greater stiffness were required, gussets could have been added, as indicated at B. The steel casting shown in Fig. 1(c) also distorted because of the different thicknesses of intersecting members. The two side frames formed by T-sections solidiﬁed rapidly and contracted in length while most of the heavy center section was still ﬂuid. When the heavy center section ﬁnally did solidify, it shortened dimension C, resulting in the distortion indicated by the phantom line in the illustration. At least two solutions are possible for the distortion problem presented by this casting. First, compensation for the distortion can be made in the pattern, in the direction opposite to that of the observed distortion. This pattern correction would, of necessity, be based on a guess and a hope that the part would come out straight, as cast. Such a solution may be costly to develop, and accuracy cannot be guaranteed. The second solution consists of increasing the section thickness of the end member and subsequently machining this section to the desired size. This would necessitate increasing the tolerance on the unmachined surfaces of the T-section in that area by a minimum of

Three castings in which distortion occurred as a result of differences in the solidﬁcation times of sections of unequal thicknesses.

156 / Casting Design and Performance 0.06 in. (1.5 mm), to allow for the observed distortion. As an alternative to such a tolerance, excess metal could be machined from both the inside and outside surfaces of the T-section. Distortion of a casting can be caused or eliminated by the gating and risering system used in producing the casting. This is demonstrated by the casting shown in Fig. 2. This thin-wall casting, which was 58 in. long, required several gates along its edge, as shown, to ﬁll the mold completely. Had a continuous runner been used, the casting would have frozen before the runner froze, because: 1. the runner would have been heavier than the casting. 2. the mold surrounding the runner would have been hotter than the mold surrounding the casting, because more molten metal would have passed through the runner; and 3. at the time the mold was completely ﬁlled, the metal in the casting would have been the coldest, because it would have traveled the farthest and consequently have passed through more unheated mold areas. When the casting froze and became rigid, the contractive forces of the runner would have bowed the casting toward the runner. This possibility of distortion was anticipated, and two short runners were used rather than one long one. No discernible distortion occurred. Gating and risering can inﬂuence distortion in still another way. In general, the ﬂow path

of the molten metal, as determined by the gate location, controls the direction of solidiﬁcation. This has a direct bearing on the contraction of the metal, as is shown in the charts of Fig. 3; which record the distortion that was encountered in the cored holes of two investment castings. As cast, the holes were somewhat elliptical, the measurements of dimension B being greater than those of dimension A. Most castings are designed to be strong enough so that dimensions will not be thrown out of tolerance by the contractive forces developed during solidiﬁcation and subsequent cooling. Nevertheless, without distorting appreciably, a casting may develop residual stresses from solidiﬁcation and cooling that will cause failure in use if the sum of the service stresses and the residual stresses exceeds the yield strength or the tensile strength of the cast metal. An example is the sand cast hub shown in Fig. 4. Although this casting was produced from malleable iron, the problem and solution are applicable also to other metals. As it was originally designed, Fig. 4(a), the six radial ribs froze before the adjoining casting sections, even though, at 3/8-in. (9.5 mm) thickness, they were heavier than the 11/32-in. (8.7 mm) tubular section. Approximately 25 per cent of the castings were rejected because of hot tears at junctions of these ribs with other sections of the casting. Under static load, many of the approved castings failed: the ribs, under tension, separated from the casting, and the hub was completely torn off the ﬂange or disk section. The residual

Had a continuous runner been used in producing this 58-in.-long casting, contractive forces would have caused distortion. This result was anticipated, and two short runners, as shown, were used; no discernible distortion occurred.

Fig. 2

Fig. 3

stresses introduced by the differential in freezing time had weakened these junctions enough to cause malfunction of the part. As redesigned, Fig. 4(b), and with the rib thickness reduced to 1/4 in. (6 mm) the casting was easily produced to the required strength level, for two main reasons: freezing time was more nearly uniform throughout the casting; and the various members contracted more evenly, without pulling on each other in the mold. The original design was fed through a bore riser, located inside the barrel section, whereas the redesign was fed with two external risers. The revisions in design and the more effective distribution of metal to the outside of the ﬂange made for a more producible casting that was also more serviceable. In the redesign, Fig. 4(b), uniformity in freezing and cooling of the different sections of the casting was encouraged by the incorporation of ribs reduced to 1/4-in. (6 mm) thickness. Although in the redesign these ribs were in a location highly inﬂuenced by surrounding metal, because of their diminished thickness they apparently reached solidiﬁcation at about the same time as the abutting sections. The importance of designing to obtain nearly uniform cooling in casting members is further demonstrated by the casting shown in Fig. 5. This aluminum sand casting, a turbine air intake, is approximately 26 in. long, 28 in. in over-all depth and 12 in. wide (66 71 30 cm) at the widest section. The sections shown in Fig. 5 are taken through each of the two struts which extend from a central hub to the end conﬁguration. Only the sections through the struts are shown, as the rest of the casting has no direct relation to the subject at hand. In approximately 30 per cent of these castings produced to the original design, (a) and (c) of Fig. 5, hot tearing occurred at the junctions of the cross ribs and the outside walls, because of hot spots. The uneven thickness of the outer wall contributed to the occurrence of defects. The pattern was revised to alter thicknesses of the sections forming these junctions to those shown in the redesign, (b) and (d) of Fig. 5. Increasing the thickness of the cross ribs from 0.25 to 0.26 in. (6.35 to 6.6 mm), and reducing the 0.30-in. (7.6 mm) thickness of outer walls at the junction to equal that of the now 0.26-in.-thick ribs, resulted in a cooperative freezing pattern. After this revision, no castings were rejected for hot tears.

Distortion from roundness of the cored holes in these two investment castings was related to the location of gates

Design Problems Involving Distortion / 157

Fig. 4

A sand cast malleable iron hub that required redesign and also rerigging to minimize residual stresses and permit successful application of the part. The problem was at the rib junctions.

Fig. 5

Cross sections of the struts of a turbine air intake, sand cast in aluminum alloy 356. When produced to the original design, 30% of the castings were rejected because of hot tears. Redesign shown eliminated these defects

Hot tearing in ribs or webs surrounded by heavier sections may occur also when these ribs or webs freeze later than the heavy sections. As an example, the investment cast 4130 steel gimbal ring of Fig. 6 had an excessively high rejection rate because of hot tears in a web connecting two heavy sections. Because the section of the mold which formed this “pocket” was almost completely surrounded by heavier sections of metal, a hot spot developed in the mold and prevented solidiﬁcation of the enclosed web section until after the adjacent heavier sections had begun to contract. The simplest solution was to extend the web out from the enclosed area, thereby encouraging faster solidiﬁcation of the thin section. This change allowed the web to develop enough strength to resist the contractive forces applied to it. The added material was removed by subsequent machining, to provide the clearance necessary for another member of the assembly. Another type of distortion that is attributable to nonuniform freezing occurs in plate sections of uniform thickness. This is the concave distortion in a ﬂat or nearly ﬂat surface that is commonly known as “oil canning”, and is the result of stresses set up as the plate freezes. There is no simple way to eliminate such distortion in a uniformly ﬂat plate casting; consequently, the designer should compensate for it by providing sufﬁcient tolerance or by employing a design of greater inherent stiffness. Figure 7 shows the amount of oil canning that may normally be anticipated in cast steel plates 10 in. square as a function of plate thickness. The magnitude of oil canning encountered in a production casting is illustrated in Fig. 8. This disk-shaped casting was produced from stainless steel in a shell mold. The distortion occurred because the molten metal at the thinner inside diameter froze ﬁrst and contracted ahead of that at the heavier outside ring. When this heavier ring froze and contracted, the force applied to the inner section caused the thin section to move out of position to the extent shown. Wheels, such as the sand cast magnesium airplane wheel shown in Fig. 9(a), present a challenge to a casting designer. A delicate balance is required among all sections, to encourage cooperative freezing and contraction. If the central hub and the spokes freeze ﬁrst and contract too far in advance of the rim section, hot tears may appear at the junctions of rim and spokes. Or, if enough metal has frozen at the junctions to resist the contractive forces of the spokes, residual stresses may develop, of sufﬁcient magnitude to cause failure of the casting in service. Also, in this type of casting, freezing and contraction of the spokes before the rim causes a constrictive force to be applied to the spokes by the rim when it starts to contract. At a certain point, the forces may become great enough to displace the hub. One solution is to make the spokes thick enough so that all sections of the casting will freeze and contract nearly simultaneously. Another solution is to make the cross section of the spokes U-shaped, thus providing

158 / Casting Design and Performance enough strength to resist the contractive force of the rim.

Distortion Due to Mold Restraint

In this investment cast gimbal ring, of 4130 steel, a ﬁnger of mold material surrounded by molten metal became excessively hot and retarded the freezing of the web section where shown. The resulting hot tearing condition was corrected by redesigning the web so that it extended into a cooler section of the mold.

Fig. 6

Fig. 7

Oil-canning distortion in a ﬂat cast plate as related to plate thickness

Fig. 8

The magnitude of oil canning encountered at location A in production of a stainless steel shell mold casting

Fig. 9

(a) Distortion here resulled from mold restraint, (b) and (c) two possible preventive redesigns.

A second major cause of casting distortion, particularly in fragile castings, is the restraint imposed by the mold as the casting cools and contracts. In its most severe form, this type of distortion is difﬁcult to eliminate except by redesigning or by adding tie bars. The use of tie bars is discussed in a later section of this chapter. The sand casting shown in Fig. 9(a) presented no problems as regards mold ﬁlling, feeding or soundness. The two standing ﬂanges A in the original design solidiﬁed and contracted without difﬁculty. However, when the bottom plate cooled and contracted, the ﬂanges, separated by mold sand, were bent outward, as indicated by the phantom lines. Although some distortion was observed in the 0.50-in (13 mm)-thick bottom plate, it was of sufﬁcient thickness to remain within assigned tolerance. If the bottom plate were reduced in thickness, the distortion would increase as the thickness decreased and would magnify the displacement of the standing ﬂanges. The distortion experienced with this type of casting may be avoided by connecting the tops of the two standing ﬂanges with a later-removable cross-member, as in Fig. 9(b), which is preferable, or by adding gussets to the sides of the ﬂanges, as in Fig. 9(c), which would be less effective. Mold restraint to normal contraction of the cooling metal was responsible for hot tears in the aluminum semipermanent mold casting shown in Fig. 10. In this aircraft casting as originally designed, the heavy end ﬂanges were connected by a thinner wall, which joined the ﬂanges at the mid-point of their width. Also connecting the ﬂanges were a series of tapping bosses. Because of the several variations in wall thickness, hot tears appeared in about 20% of the castings at the junctions of the thin wall with the heavier ﬂanges; the thin wall froze and contracted in advance of the heavy tapping bosses and the ﬂanges. With the outer sections of the mold made of metal and the core made of baked sand, the mold was impervious to any crushing force that might be developed by the contraction of these thin sections. In addition, with the ﬂanges hotter than the thin sections, not enough strength developed in the junction ﬁllets to resist the contractive forces. Hot tears resulted. Here, the presence of heavy sections, the heat-retaining ability of the mold, and the rigidity of the mold all combined to create difﬁculties. As a consequence, the casting was revised. By moving the connecting wall inward until its inner surface was congruent with the inner conﬁguration of the mounting ﬂanges, a metal core, which had the ability to chill the casting, could be used. To encourage uniform freezing, the

Design Problems Involving Distortion / 159

Mold restraint coupled with nonuniform freezing of the various sections of this aluminum (alloy 356) semipermanent mold casting resulted in hot tears that were responsible for a 20% rejection rate. Moving the wall and increasing its thickness as shown corrected the trouble.

Fig. 10

Fig. 12

The addition of ribs to this brass investment casting corrected dimensional inaccuracies between upright sections, caused by mold restraint to normal metal contraction. Without ribs, the upright sections became tilted outward and were out of tolerance.

Fig. 11

thickness of the connecting wall was increased and the ﬂanges were reduced in thickness. These revisions reduced rejections—for all causes—to ﬁve per cent. Investment Castings. The restraint that the mold exerts on the forces of contraction can be a problem in investment castings also. In the brass investment casting of Fig. 11, mold restraint prevented the two upright sections from moving toward each other as the base cooled and contracted. As a result, the upright sections became tilted outward and were out of tolerance. By incorporating ribs as shown, the mold restraint was modiﬁed, and dimensionally acceptable castings were produced. When such ribs are used, their proportions must be designed with care. Improperly proportioned ribs may freeze before the adjoining members and thus promote, rather than prevent, distortion or hot tears.

Tie Bars Tie bars are expendable additions to a casting that are employed to equalize contraction in the mold and to provide rigidity during subsequent machining or heat treating. Many castings are producible within tolerance and without

distortion only by the strategic placement of one or more tie bars. Four tie bars were used to good advantage in the casting shown in Fig. 12. Produced from type 410 stainless steel in a ceramic mold, this casting exhibited good casting qualities, except that mold restraint affected dimensions A. By incorporating tie bars as shown in phantom in Fig. 12, dimensional stability was maintained in these areas of the castings. In addition, by leaving these tie bars in place until after the casting was heat treated, possible distortion from this processing operation was avoided. Another casting that exempliﬁes the salutary effects of tie bars is the cockpit-fairing-glass frame shown in Fig. 13. This aluminum semipermanent mold casting made use of tie bars in a major way. As originally designed, without tie bars, the casting warped excessively. Because of this lack of rigidity, problems were also encountered in machining. To remedy these problems, one center tie bar was incorporated, but with little success. Next, two tie bars were tried, one at each side of the casting. These corrected some of the distortion, but the percentage of acceptable castings was still too low to be economical. In addition, it was necessary to cast test coupons with each casting so that mechanical properties could be measured.

Tie bars in a stainless steel casting

These problems were solved by incorporating three tie bars as shown in Fig. 13. The center bar was removed after casting but before machining, and test coupons were produced from this bar. To maintain rigidity and eliminate machining problems, the two outside tie bars were permitted to remain on the casting until the part had been completely machined. However, because of residual stresses, the casting would spring out of shape when the tie bars were removed. This problem was solved by allowing the tie bars to remain in place until the casting was assembled into position on the aircraft, after which they were removed. Without tie bars, this casting probably could not have been produced in usable condition. Investment castings also may be held within speciﬁed tolerances by the use of tie bars. Two such castings are shown in Fig. 14. The 4130 steel casting, Fig. 14(a), required a small “standard” tie bar, positioned as shown, in order that the 3/8 -in. dimension between the side sections could be held. The 1020 steel casting, Fig. 14 (b), required a tie bar to assist in maintaining the 10/32 -in. dimension between the two side sections. A gate was devised that functioned also as a tie bar, thus presenting a manufacturing economy because it eliminated the cost of removing a special tie bar. (The gates required removal whether or not they served as tie bars.) The multipurpose gate was possible because of the ﬂat ends of the side sections. Because they are as much a part of the wax pattern as is any other section of an investment casting, tie bars help to maintain dimensional stability in the wax pattern in the same manner as they do in the casting. Because distortion of a wax pattern is reproduced in the casting, the foundry-man must provide some other means of maintaining accuracy in wax patterns if tie bars are impractical. Sometimes, metal spacers are used between two parts of the wax pattern to assure maintenance of a dimension. If the hardening of the wax pattern (after it has been removed from the pattern die) is such that it

160 / Casting Design and Performance

This aluminum (alloy 356) semipermanent mold casting would probably not have been producible or machinable to a usable condition without the use of the three tie bars shown. Two of these tie bars were left in place until after the part (a cockpit-fairing-glass frame) was assembled in the aircraft, because of distortion from residual stresses.

Fig. 13

Produced as one part, machined, then given a T6 heat treatment, this sand cast aluminum (alloy 356) clamp distorted beyond usable condition. Changing to a T7 heat treatment minimized the distortion.

Fig. 15

Increasing section thickness in the problem areas of this permanent mold casting would result in better ﬂow of metal in the mold, better feeding during solidiﬁcation, and the elimination of defects due to shrinkage.

Fig. 16 would open a dimension rather than close it, a spacer may be set in place and a weight placed on the wax, so that the members of the wax pattern are forced into the desired dimension.

Distortion in Heat Treating Distortion that occurs in heat treating may necessitate some form of compromise in design or processing, so that a usable part is obtained. With the mounting clamp shown in Fig. 15, an aluminum (alloy 356) sand casting, a revision of the choice of heat treatment was required, in order to curtail the distortion that had occurred as a result of the ﬁrst-selected process and made most of the heat treated castings unusable.

Produced as one part, this casting was machined, given a T6 heat treatment, and separated into two clamps. On separation, the castings would open up 1/8 to 1/4 in. (3 to 6 mm). The distortion occurred because of stresses introduced during quenching in the T6 heat treatment; the artiﬁcial aging step in this heat treatment did not effect enough stress relief to prevent subsequent distortion. Consequently, the heat treatment was revised to T7, and more stable castings were obtained. Although a T7 heat treatment takes more time and is more

Two investment castings that required tie bars to assure dimensional accuracy. In (a), a “standard” tie bar, which had to be removed after casting, added to the cost of production. In (b), the gate functioned also as a tie bar, and, since gate removal is a normal requirement, no additional cost was incurred for this operation.

Fig. 14

expensive than T6, the results may justify the added expenditure, as they did with this part. The permanent mold casting shown in Fig. 16 was used in a structural application and had to be sound and ﬂat. It was solution heat treated, quenched, and aged. Difﬁculties were encountered in keeping the casting within ﬂatness tolerance, and with shrinkage in the 3/8-in. (9 mm) web between the heavy bosses. The problem here was one of molten metal distribution and of supplying feed metal to replace the volume lost by shrinkage. A simple solution would be to increase the thickness of the webs between the bosses. This would permit better feeding of the casting, eliminating the shrinkage, and would reduce warpage in heat treatment by modifying the stress pattern.

Effect of Alloy The alloy to be cast can have a signiﬁcant inﬂuence on distortion. Figure 17 shows a plaster mold casting for which a 4.8% Zn, 2.3% Mg aluminum alloy was speciﬁed. This material, which hardens at room temperature in about 20 days, had successfully eliminated distortion

Design Problems Involving Distortion / 161

Using a 4.8% Zn, 2.3% Mg aluminum alloy, this plaster mold casting distorted out of tolerance and was not producible to the original design. Using alloy 356, the casting was produced as designed, and distortion was within acceptable limits.

Fig. 17

that had occurred in heat treatment when aluminum alloy 356 had been used. However, the casting shown in Fig. 17 was not producible to the original design when the Al-Zn-Mg alloy was used; it was necessary to cast the center bar solid (so as to obtain sound metal in this section), and to ﬁll the mold cavity and obtain the required pressure-tightness. This revision introduced considerable distortion. The thin sections froze and progressed through part of their contraction cycle before the heavier bar section froze. Thus, the contraction of the heavier section was resisted by the thinner sections, and this interplay of forces was strong enough to distort the casting 0.020 in. (0.51 mm) out of position at each end. Many processing variations were investigated, in an effort to correct this problem. It was found that chills, located as shown, reduced distortion so that the castings were within the speciﬁed tolerance. For comparison, castings were poured to the original design from aluminum alloy 356. All these castings were sound and passed the pressure-tightness test, and distortion was sufﬁciently low that the parts stayed within tolerances, as cast. To minimize distortion, the

castings were quenched in an air blast, rather than in a liquid. The Al-Zn-Mg alloy has a higher pouring temperature and does not feed isolated areas of the casting from risers as well as alloy 356. However, when a casting is so designed as to encourage distortion in heat treatment, the AlZn-Mg alloy, because of its ability to age harden at room temperature, may offer an advantage over alloy 356.

Casting Design and Performance Pages 163–170

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Corrosion of Cast Irons* CAST IRON is a generic term that identiﬁes a large family of ferrous alloys. Cast irons are primarily alloys of iron that contain more than 2% carbon and 1% or more silicon. Low raw material costs and relative ease of manufacture make cast irons the least expensive of the engineering metals. Cast irons can be cast into intricate shapes because of their excellent ﬂuidity and relatively low melting points and can be alloyed for improvement of corrosion resistance and strength. With proper alloying, the corrosion resistance of cast irons can equal that of stainless steels and nickel-base alloys in many services. Because of the excellent properties obtainable with these low-cost engineering materials, cast irons ﬁnd wide application in environments that demand good corrosion resistance, such as in water, soils, acids, alkalis, saline solutions, organic compounds, sulfur compounds, and liquid metals.

Basic Metallurgy of Cast Irons The metallurgy of cast irons is similar to that of steels except that sufﬁcient silicon is present to necessitate use of the iron-silicon-carbon ternary phase diagram rather than the simple iron-carbon binary diagram. Figure 1 shows a section of the iron-iron carbide-silicon ternary diagram at 2% Si. The eutectic and eutectoid points in the iron-silicon-carbon diagram are both affected by the introduction of silicon into the system. In the 1 to 3% Si levels normally found in cast irons, eutectic carbon levels are related to silicon levels as follows: %C þ 1/3ð%SiÞ ¼ 4:3

on the manner in which the carbon segregates in the microstructure. Higher silicon levels favor the formation of graphite, but lower silicon levels favor the formation of iron carbides. The form and shape in which the carbon occurs determine the type of cast iron (Table 1). The structure of the metal matrix around the carbon-rich constituent establishes the class of iron within each type of iron. As in steel, the ﬁve basic matrix structures occur in cast iron: ferrite, pearlite, bainite, martensite, and austenite. Ferrite is generally a soft constituent, but it can be solid solution hardened by silicon. When silicon levels are below 3%, the ferrite matrix is readily machined but exhibits poor wear resistance. Above 14% Si, the ferritic matrix becomes very hard and wear resistant but is essentially nonmachinable. The low carbon content of the ferrite phase makes hardening difﬁcult. Ferrite can be observed in cast irons on solidiﬁcation but is generally present as the result of special annealing heat treatments. High silicon levels promote the formation of ferritic matrices in the as-cast condition. Pearlite consists of alternate layers of ferrite and iron carbide (Fe3C, or cementite). It is very strong and tough. The hardness, strength, machinability, and wear resistance of pearlitic

Inﬂuence of Alloying Alloying elements can play a dominant role in the susceptibility of cast irons to corrosion attack. The alloying elements generally used to enhance the corrosion resistance of cast irons include silicon, nickel, chromium, copper, and molybdenum. Other alloying elements, such as vanadium and titanium, are sometimes used,

(Eq 1)

where %C is the eutectic carbon level, and %Si is the silicon level in the cast iron. The metallurgy of cast iron can occur in the metastable iron-iron carbide system, the stable iron-graphite system, or both. This causes structures of cast irons to be more complex than those of steel and more susceptible to processing conditions. An appreciable portion of carbon in cast irons separates during solidiﬁcation and appears as a separate carbon-rich constituent (e.g., graphite, iron carbides) in the microstructure. The level of silicon in the cast iron has a strong effect

matrices vary with the ﬁneness of its laminations. The carbon content of pearlite is variable and depends on the composition of the iron and its cooling rate. Bainite is an acicular structure in cast irons that can be obtained by heat treating, alloying, or combinations of these. Bainitic structures provide very high strength at a machinable hardness. Martensitic structures are produced by alloying, heat treating, or a combination of these practices. Martensitic microstructures are the hardest, most wear-resistant structures obtainable in cast irons. Molybdenum, nickel, manganese, and chromium can be used to produce martensitic or bainitic structures. Silicon has a negative effect on martensite formation, because it promotes the formation of pearlite or ferrite. Austenitic structures are typically found in the Ni-Resist cast irons and the austempered ductile irons. Austenite is a face-centered cubic atomic structure created primarily by alloying with austenite-forming elements such as nickel. Austenite is generally the softest and more corrosion-resistant matrix structure. However, the carbon-enriched austenite of austempered ductile iron has higher hardness and other unique characteristics over conventional ductile irons (Ref 1).

Table 1 Summary of cast iron classiﬁcation based on carbon form and shape Type of cast iron

Fig. 1

Section of the iron-iron carbide-silicon temary phase diagram at 2% Si

*Reprinted from ASM Handbook, Volume 13B: Corrosion: Materials, p43–50

White cast iron Malleable cast iron Gray cast iron Ductile cast iron Compacted graphite cast iron

Carbon form and shape

Iron carbide compound Irregularly shaped nodules of graphite Graphite ﬂakes Spherical graphite nodules Short, fat, interconnected ﬂakes (intermediate between ductile and gray cast iron)

164 / Casting Design and Performance but not to the extent of the ﬁrst ﬁve elements mentioned. Silicon is the most important alloying element used to improve the corrosion resistance of cast irons. Silicon is generally not considered an alloying element in cast irons until levels exceed 3%. Silicon levels between 3 and 14% offer some increase in corrosion resistance to the alloy, but above approximately 14% Si, the corrosion resistance of the cast iron increases dramatically. Silicon levels up to 17% have been used to enhance the corrosion resistance of the alloy further, but silicon levels over 16% make the alloy extremely brittle and difﬁcult to manufacture. Even at 14% Si, the strength and ductility of the material is low, and special design and manufacturing parameters are required to produce and use these alloys. Alloying with silicon promotes the formation of strongly adherent surface ﬁlms in cast irons. Considerable time may be required to establish these ﬁlms fully on the castings. Consequently, in some services, corrosion rates may be relatively high for the ﬁrst few hours or even days of exposure, then may decline to extremely low steady-state rates for the rest of the time the parts are exposed to the corrosive environment (Fig. 2). Nickel is used to enhance the corrosion resistance of cast irons in a number of applications. Nickel increases corrosion resistance by the formation of protective oxide ﬁlms on the surface of the castings. Up to 4% Ni is added in combination with chromium to improve both strength and corrosion resistance in cast iron alloys. The enhanced hardness and corrosion resistance obtained is particularly important for improving the erosion-corrosion resistance of the material. Nickel additions enhance the resistance of cast irons to corrosion by reducing acids and alkalis. Nickel additions of 12% or greater are necessary to optimize the corrosion resistance of cast irons. The Ni-Resist group are high-nickel alloys (13.5 to 36% Ni) having high resistance to wear, heat, and corrosion. Nickel is not as common an alloying addition as either silicon or chromium for enhancing the corrosion resistance in cast irons. It is much more important as a strengthening and hardening addition. Chromium is frequently added alone and in combination with nickel and/or silicon to increase the corrosion resistance of cast irons. As with nickel, small additions of chromium are used to reﬁne graphite and matrix microstructures. These reﬁnements enhance the corrosion resistance of cast irons in seawater and weak acids. Chromium additions of 15 to 35% improve the corrosion resistance of cast irons to oxidizing acids, such as nitric acid (HNO3). Chromium increases the corrosion resistance of cast iron by the formation of protective oxides on the surface of castings. The oxides formed will resist oxidizing acids but will be of little beneﬁt under reducing conditions. High-chromium additions, similar to higher-silicon additions, reduce the ductility of cast irons.

Copper is added to cast irons in special cases. Copper additions of 0.25 to 1% increase the resistance of cast iron to dilute acetic (CH3COOH), sulfuric (H2SO4), and hydrochloric (HCl) acids as well as acid mine water. Small additions of copper are also made to cast irons to enhance atmospheric-corrosion resistance. Additions of up to 10% are made to some high-nickel-chromium cast irons to increase corrosion resistance. The exact mechanism by which copper improves the corrosion resistance of cast irons is not known. Molybdenum. Although an important use of molybdenum in cast irons is to increase strength and structural uniformity, it is also used to enhance corrosion resistance, particularly in high-silicon cast irons. Molybdenum is particularly useful in hydrochloric acid (HCl). As little as 1% Mo is helpful in some high-silicon irons, but for optimal corrosion resistance, 3 to 4% Mo is added. Other Alloying Additions. In general, other alloying additions to cast irons have a minimal effect on corrosion resistance. Vanadium and titanium enhance the graphite morphology and matrix structure and impart slightly increased corrosion resistance to cast irons. Few other additions are made to cast irons that have any signiﬁcant effect on corrosion resistance.

Inﬂuence of Microstructure Although the graphite shape and the amount of massive carbides present are critical to mechanical properties, these structural variables do not have a strong effect on corrosion resistance. Flake graphite structures may trap corrosion products and retard corrosion slightly in some applications. Under unusual circumstances, graphite may act cathodically with regard to the metal matrix and accelerate attack. While the structure of the matrix has a slight inﬂuence on corrosion resistance, the effect is small compared to that of matrix composition. In gray irons, ferrite structures are generally the least corrosion-resistant, and graphite ﬂakes

exhibit the greatest corrosion resistance. Pearlite and cementite show intermediate corrosion resistance, while an austenitic structure imparts higher corrosion resistance. Shrinkage or porosity can degrade the corrosion resistance of cast iron parts by acting as natural crevices. The presence of porosity permits the corrosive medium to enter the body of the casting and can provide continuous leakage paths for corrosives in pressure-containing components.

Commercially Available Cast Irons Based on corrosion resistance, cast irons can be grouped into the following ﬁve classes. Unalloyed gray, ductile, malleable, and white cast irons represent the ﬁrst and largest class. All of these materials contain carbon and silicon of 3% or less and no deliberate additions of nickel, chromium, copper, or molybdenum. As a group, these materials exhibit a corrosion resistance that equals or slightly exceeds that of unalloyed steels, but they show the highest rate of attack among the classes of cast irons. These materials are available in a wide variety of conﬁgurations and alloys. Major ASTM standards that cover these materials are listed in Table 2. Low- and moderately alloyed cast irons constitute the second major class. These irons contain the iron and silicon of unalloyed cast irons plus up to several percent of nickel, copper, chromium, or molybdenum. As a group, these materials exhibit two to three times the service life of unalloyed cast irons. Austempered ductile iron (ADI) is the newest group of alloys in this category, and they have some unique properties. The ADI delivers twice the strength of conventional ductile irons for a given level of elongation. In addition, ADI offers exceptional wear and fatigue resistance (Ref 1). Table 2 ASTM standards that include unalloyed cast irons Standard

A A A A

47 48 74 126

A A A A

159 197 220 278

A 319 A 395

Fig. 2

Corrosion rates of high-silicon cast irons as a function of time and corrosive media

A A A A A A A A

476 536 602 716 746 823 842 874

Materials/products covered

Ferritic malleable iron castings Gray iron castings Cast iron soil pipe and ﬁttings Gray iron castings for valves, ﬂanges, and pipe ﬁttings Automotive gray iron castings Cupola malleable iron Pearlitic malleable iron castings Gray iron castings for pressure-containing parts for temperatures up to 345 C (650 F) Gray iron castings for elevated temperatures for nonpressure-containing parts Ferritic ductile iron pressure-retaining castings for use at elevated temperatures Ductile iron-castings for paper mill dryer rolls Ductile iron castings Automotive malleable iron castings Ductile iron culvert pipe Ductile iron gravity sewer pipe Statically cast permanent mold gray iron castings Compacted graphite iron castings Ferritic ductile iron castings suitable for low-temperature service

Corrosion of Cast Irons / 165 Major ASTM standards that cover these materials are listed in Table 3. High-nickel austenitic cast irons represent a third major class of cast irons for corrosion service. These materials contain large percentages of nickel and copper and are fairly resistant to such acids as concentrated sulfuric (H2SO4) and phosphoric (H3PO4) acids at slightly elevated temperatures, hydrochloric acid at room temperature, and organic acids such as acetic (CH3COOH), oleic, and stearic. When nickel levels exceed 18%, austenitic cast irons are nearly immune to alkali or caustics, although stress corrosion can occur. Highnickel cast irons can be nodularized to yield ductile irons. Major ASTM standards that cover these materials are listed in Table 4. High-chromium cast irons are the fourth class of corrosion-resistant cast irons. These materials are basically white cast irons alloyed with 12 to 35% Cr. Other alloying elements may also be added to improve resistance to speciﬁc environments. When chromium levels exceed 20%, high-chromium cast irons exhibit good resistance to oxidizing acids, particularly nitric acid (HNO3). High-chromium irons are not resistant to reducing acids. They are used in saline solutions, organic acids, phosphate mining, marine, and industrial atmospheres. These materials display excellent resistance to abrasion, and, with proper alloying additions, they can also resist combinations of abrasives and liquids, including some dilute acid solutions. High-chromium cast irons are covered in ASTM A 532. In addition, many proprietary alloys not covered by national standards are produced for special applications, such as wear components in mining operations or slurry pumps. High-silicon cast irons are the ﬁfth class of corrosion-resistant cast irons. The principal alloying element is 12 to 18% Si, with more

Table 3 ASTM standards that include lowalloyed cast iron materials Standard

A 159 A 319 A 532 A 897

Materials/products covered

Automotive gray iron castings Gray iron castings for elevated temperature for nonpressure-containing parts Abrasion-resistant cast irons Austempered ductile iron castings

than 14.2% Si needed to develop excellent corrosion resistance. Chromium and molybdenum are also used in combination with silicon to develop corrosion resistance to speciﬁc environments. High-silicon cast irons represent the most universally corrosion-resistant alloys available at moderate cost. When silicon levels exceed 14.2%, high-silicon cast irons exhibit excellent resistance to H2SO4, HNO3, HCl, CH3COOH, and most other mineral and organic acids and corrosives. These materials display good resistance in oxidizing and reducing environments and are not appreciably affected by concentration or temperature. Exceptions to universal resistance are hydroﬂuoric acid (HF), ﬂuoride salts, sulfurous acid (H2SO3), sulﬁte compounds, strong alkalis, and alternating acid-alkali conditions. High-silicon cast irons are deﬁned in ASTM A 518 and A 861.

Forms of Corrosion Cast irons exhibit the same general forms of corrosion as other metals and alloys:

Uniform or general attack Galvanic or two-metal corrosion Crevice corrosion Pitting Intergranular corrosion Selective leaching (graphitic corrosion) Erosion-corrosion Stress corrosion Corrosion fatigue Fretting corrosion Microbiological

Graphitic Corrosion. A form of corrosion unique to cast irons is a selective leaching attack commonly referred to as graphitic corrosion or graphitization. Graphitic corrosion is observed in gray cast irons in relatively mild environments in which selective leaching of iron leaves a brittle graphite network. Selective leaching of the iron takes place because the graphite is cathodic to the iron, and the gray cast iron structure establishes an excellent galvanic cell. While graphitic corrosion of gray

Table 5

Relative fretting resistance of cast iron

Poor Note: Because most cast iron standards make chemical composition subordinate to mechanical properties, many of the standards listed in Table 2 may also be used to purchase low-alloyed cast iron materials.

Table 4 ASTM standards that include highnickel austenitic cast iron materials Standard

A 436 A 439 A 571

Materials/products covered

Austenitic gray iron castings Austenitic ductile iron castings Austenitic ductile iron castings for pressurecontaining parts suitable for low-temperature service

cast iron is considered a form of selecting leaching, its mechanism on a microstructural level is similar to galvanic corrosion. This form of corrosion generally occurs only when corrosion rates are low. If the metal corrodes more rapidly, the entire surface, including the graphite, is removed, and more or less uniform corrosion occurs. Graphitic corrosion can cause signiﬁcant problems because, although no dimensional changes occur, the cast iron loses its strength and metallic properties. Thus, without detection, potentially dangerous situations may develop in pressure-containing applications. Graphitic corrosion is observed only in gray cast irons. In both nodular and malleable cast iron, the lack of graphite ﬂakes provides a more favorable anode/cathode ratio and no network to hold the corrosion products together. By maximizing the area of the anodic component while decreasing the area of the cathodic constituent, the potential for galvanic (graphitic) corrosion has been reduced. Because graphitization is so common with cast iron and it compromises the structural integrity of the metal, instrumentation using eddy-current measurements has recently been developed to detect and measure it (Ref 2). Fretting corrosion is commonly observed when vibration or slight relative motion occurs between parts under load. The relative resistance of cast iron to this form of attack is inﬂuenced by such variables as lubrication, hardness variations between materials, the presence of gaskets, and coatings. Table 5 compares the relative fretting resistance of cast iron under different combinations of these variables. Pitting and Crevice Corrosion. The presence of chlorides and crevices or other shielded areas presents conditions that are favorable to the pitting and crevice corrosion of cast iron. Pitting has been reported in such environments as dilute alkylaryl sulfonates, antimony trichloride (SbCl3), and calm seawater. Alloying can inﬂuence the resistance of cast irons to pitting and crevice corrosion. For example, in calm sea-water, nickel additions reduce the susceptibility of cast irons to pitting attack. High-silicon cast irons with chromium and/or molybdenum offer enhanced resistance to pitting and crevice

Average

Aluminum on cast iron Magnesium on cast iron Cast iron on chrome plate

Cast iron on cast iron Copper on cast iron Brass on cast iron

Laminated plastic on cast iron Bakelite on cast iron

Zinc on cast iron Cast iron on silver plate

Cast iron on tin plate Cast iron on cast iron with coating of shellac

Cast iron on copper plate Cast iron on amalgamated copper plate Cast iron on cast iron with rough surface

Source: Ref 3

Good

Cast iron on cast iron with phosphate coating Cast iron on cast iron with coating of rubber cement Cast iron on cast iron with coating of tungsten sulﬁde Cast iron on cast iron with rubber gasket Cast iron on cast iron with Molykote lubricant Cast iron on stainless with Molykoke lubricant

166 / Casting Design and Performance corrosion. Although microstructural variations probably exert some inﬂuence on susceptibility to crevice corrosion and pitting, there are few reports of this relationship. Intergranular attack is relatively rare in cast irons. In stainless steels, in which this type of attack is most commonly observed, intergranular attack is related to chromium depletion adjacent to grain boundaries. Because only the high-chromium cast irons depend on chromium to form passive ﬁlms for resistance to corrosion attack, few instances of intergranular attack related to chromium depletion have been reported. The only reference to intergranular attack in cast irons involves ammonium nitrate (NH4NO3), in which unalloyed cast irons are reported to be intergranularly attacked. Because this form of selective attack is relatively rare in cast irons, no signiﬁcant references to the inﬂuence of either structure or chemistry on intergranular attack have been reported. Erosion-Corrosion. Fluid ﬂow by itself or in combination with solid particles can cause erosion-corrosion attack in cast irons. Two methods are known to enhance the erosioncorrosion resistance of cast irons. First, the hardness of the cast irons can be increased through solid-solution hardening or phasetransformation-induced hardness increases. For example, 14.5% Si additions to cast irons cause substantial solid-solution hardening of the ferritic matrix. In such environments as the sulfate liquors encountered in the pulp and paper industry, this hardness increase enables highsilicon iron equipment to be successfully used, while lower-hardness unalloyed cast irons fail rapidly by severe erosion-corrosion. Use of martensitic or white cast irons can also improve the erosion-corrosion resistance of cast irons as a result of hardness increases. Second, better inherent corrosion resistance can also be used to increase the erosion-corrosion resistance of cast irons. Austenitic nickel cast irons can have hardnesses similar to unalloyed cast irons but may exhibit better erosion resistance because of the improved inherent corrosion resistance of nickel-alloyed irons compared to unalloyed irons. Microstructure can also affect erosion-corrosion resistance slightly. Gray cast irons generally show better resistance than steels under erosion-corrosion conditions. This improvement is related to the presence of the graphite network in the gray cast iron. Iron is corroded from the gray iron matrix as in steel, but the graphite network that is not corroded traps corrosion products; this layer of corrosion products and graphite offers additional protection against erosion-corrosion attack. Flow-induced corrosion stemming from ﬂuid velocity alone is another type of erosion-corrosion for steels and cast irons. In certain services where unalloyed or low-alloyed cast irons are used, their corrosion resistance is due to the formation of a thick, poorly adherent corrosion product rather than the usual passive oxide layer associated with the more common

corrosion-resistant alloys. Examples of such situations are concentrated sulfuric or hydroﬂuoric acids. In these services, the cast irons develop, respectively, a thick iron sulfate ﬁlm or iron-ﬂuoride ﬁlm, and at low velocities these ﬁlms remain intact and provide protection. However, at velocities greater than a couple feet per second, these ﬁlms are washed away, allowing further corrosion of the cast irons. Microbiologically induced corrosion (MIC) is the corrosion of metals resulting from the activity of a variety of living microorganisms, which, as a result of their growth or metabolism, either produce corrosive wastes or participate directly in electrochemical reactions on the metal surfaces. This phenomenon is often associated with biofouling and corrosion of buried structures. Soils containing sulfate concentrations support conditions where MIC of cast iron pipe can occur. An Australian study estimates that 50% of all failures of buried metal were due to microbiological causes (Ref 4). Prevention is difﬁcult, but cathodic protection and the use of protective coatings can be beneﬁcial (Ref 5). Stress-corrosion cracking (SCC) is observed in cast irons under certain combinations of environment and stress. Because stress is necessary to initiate SCC and because design factors often limit stresses in castings to relatively low levels, SCC is not observed as often in cast irons as in other more highly stressed components. However, under certain conditions, SCC can be a serious problem. Because unalloyed cast irons are generally similar to ordinary steels in resistance to corrosion, the same environments that cause SCC in steels will likely cause problems in cast irons. Environments that may cause SCC in unalloyed cast irons include these solutions (Ref 6): Sodium hydroxide (NaOH) Sodium hydroxide-sodium silicate (NaOH

Na2SiO2) Calcium nitrate (Ca(NO3)2) Ammonium nitrate (NH4NO3) Sodium nitrate (NaNO3) Mercuric nitrate (Hg(NO3)2) Mixed acids (H2SO4-HNO3) Hydrogen cyanide (HCN) Seawater Acidic hydrogen sulﬁde (H2S) Molten sodium-lead alloys Acid chloride Oleum (fuming H2SO4)

Graphite morphology can play an important role in SCC resistance in certain environments. In oleum, ﬂake graphite structures present special problems. Acid tends to penetrate along graphite ﬂakes and corrodes the iron matrix. The corrosion products formed build up internal pressure and eventually crack the iron. This problem is found in both gray cast irons and high-silicon cast irons, which have ﬂake graphite morphologies. It is not seen in ductile cast irons that have nodular graphite shapes.

Resistance to Corrosive Environments No single grade of cast iron will resist all corrosive environments. However, a cast iron can be identiﬁed that will resist most of the corrosives commonly used in industrial environments. Cast irons suitable for the more common corrosive environments are discussed as follows. Sulfuric Acid. Unalloyed, low-alloyed, and high-nickel austenitic as well as high-silicon cast irons are used in H2SO4 applications. Use of unalloyed and low-alloyed cast irons is limited to low-velocity, low-temperature concentrated (> 70%) H2SO4 service. Unalloyed cast iron is rarely used in dilute or intermediate concentrations, because corrosion rates are substantial. In concentrated H2SO4, as well as other acids, ductile iron is generally considered superior to gray iron, and ferritic matrix irons are superior to pearlitic matrix irons. In hot, concentrated acids, graphitization of the gray iron can occur. In oleum, unalloyed gray iron will corrode at very low rates. However, acid will penetrate along the graphite ﬂakes, and the corrosion product that forms can build up sufﬁcient pressure to split the iron. Interconnecting graphite is believed to be necessary to cause this form of cracking; therefore, ductile and malleable irons are generally acceptable for oleum service. Some potential for galvanic corrosion between cast iron and steel has been reported in 100% H2SO4. High-nickel austenitic cast irons exhibit acceptable corrosion resistance in room-temperature and slightly elevated-temperature H2SO4 service. As shown in Fig. 3, their performance is adequate over the entire range of H2SO4 concentrations, but they are a second choice compared to high-silicon cast irons. High-silicon cast irons are the best choice among the cast irons and perhaps among the

Corrosion of high-nickel austenitic cast iron in H2SO4 as a function of acid concentration and temperature. Source: Ref 6

Fig. 3

Corrosion of Cast Irons / 167 commonly available engineering material for resistance to H2SO4. This material has good corrosion resistance to the entire H2SO4 concentration range at temperatures to boiling (Fig. 4). Rapid attack occurs at concentrations over 100% and in service containing free sulfur trioxide (SO3). High-silicon cast irons are relatively slow to passivate in H2SO4 service. Corrosion rates are relatively high for the ﬁrst 24 to 48 h of exposure and then decrease to very low steady-state rates (Fig. 2). Nitric Acid. All types of cast iron, except high-nickel austenitic iron, ﬁnd some applications in HNO3. The use of unalloyed cast iron in HNO3 is limited to low-temperature, lowvelocity concentrated acid service. Even in this service, caution must be exercised to avoid dilution of acid because the unalloyed and low-alloyed cast irons both corrode very rapidly in dilute or intermediate concentrations at any temperature. High-nickel austenitic cast irons exhibit essentially the same resistance as unalloyed cast iron to HNO3 but cannot be economically justiﬁed for this service. High-chromium cast irons with chromium contents over 20% give excellent resistance to HNO3, particularly in dilute concentrations (Fig. 5). High-temperature boiling solutions attack these grades of cast iron. High-silicon cast irons also offer excellent resistance to HNO3. Resistance is exhibited over essentially all concentration and temperature ranges, with the exception of dilute, hot acids (Fig. 6). High-silicon cast iron equipment has been used for many years in the manufacture and handling of HNO3 mixed with other chemicals, such as H2SO4, sulfates, and nitrates. Contamination of HNO3 with HF, such as might be experienced in pickling solutions, may accelerate attack of the high-silicon iron to unacceptable levels. Hydrochloric Acid. Use of cast irons is relatively limited in HCl. Unalloyed cast iron is unsuitable for any HCl service. Rapid corrosion occurs at a pH of 5 or lower, particularly if appreciable velocity is involved. Aeration or oxidizing conditions, such as the presence of metallic salts, result in rapid destructive attack of unalloyed cast irons, even in very dilute HCl solutions.

High-nickel austenitic cast irons offer some resistance to all HCl concentrations at room temperature or below. High-chromium cast irons are not suitable for HCl services. High-silicon cast irons offer the best resistance to HCl of any cast iron. When alloyed with 4 to 5% Cr, high-silicon cast iron is suitable for all concentrations of HCl at temperatures up to 28 C (80 F). When high-silicon cast iron is alloyed with chromium, molybdenum, and higher silicon levels, the temperature for use can be increased (Fig. 7). In concentrations up to 20%, ferric ions (Fe3+) or other oxidizing agents inhibit corrosion attack on highsilicon cast iron alloyed with chromium. At over 20% acid concentration, oxidizers accelerate attack on the alloy. As in H2SO4, corrosion rates of high-silicon cast iron are initially high in the ﬁrst 24 to 48 h of exposure then decrease to very low steady-state rates (Fig. 2). Phosphoric Acid. All cast irons ﬁnd some application in H3PO4 service, but the presence of contaminants must be carefully evaluated before selecting a material. Unalloyed cast iron ﬁnds little use in H3PO4, with the exception of concentrated acids. Even in concentrated acids, use may be severely limited by the presence of ﬂuorides, chlorides, or H2SO4. High-nickel cast irons ﬁnd some application in H3PO4 at and slightly above room temperature. These cast irons can be used over the entire H3PO4 concentration range. Impurities in the acid may greatly restrict the applicability of this grade of cast iron. High-chromium cast irons exhibit generally low rates of attack in H3PO4 up to 60% concentration and are commonly used in the phosphate mining industry where abrasion resistance is needed. High-silicon cast irons show good-toexcellent resistance at all concentrations and temperatures for pure acid. The presence of ﬂuoride ions (F) in H3PO4 makes the high-silicon irons unacceptable for use.

Corrosion of high-silicon cast iron in H2SO4 as a function of acid concentration and temperature

Fig. 5

Fig. 4

Organic acids and compounds are generally not as corrosive as mineral acids. Consequently, cast irons ﬁnd many applications in handling these materials. Unalloyed cast iron can be used to handle concentrated acetic acid, CH3COOH, and fatty acids but will be attacked by more dilute solutions. Unalloyed cast irons are used to handle methyl, ethyl, butyl, and amyl alcohols. If the alcohols are contaminated with water and air, discoloration of the alcohols may occur. Unalloyed cast irons can also be used to handle glycerine, although slight discoloration of the glycerine may result. Austenitic nickel cast irons exhibit adequate resistance to CH3COOH, oleic acid, and stearic acid. High-chromium cast irons are adequate for CH3COOH but will be more severely corroded by formic acid (HCOOH). High-chromium cast irons are excellent for lactic and citric acid solutions. High-silicon cast irons show excellent resistance to most organic acids, including HCOOH and oxalic acid, in all temperature and concentration ranges. High-silicon cast irons also exhibit excellent resistance to alcohols and glycerine. Alkali solutions require material selections that are distinctly different from those of acid solutions. Alkalis include sodium hydroxide (NaOH), potassium hydroxide (KOH), sodium silicate (Na2SiO3), and similar chemicals that contain sodium, potassium, or lithium. Unalloyed cast irons exhibit generally good resistance to alkalis—approximately equivalent

Fig. 6

Corrosion of high-chromium cast iron in HNO3 as a function of acid concentration and temperature. Source: Ref 6

Corrosion of high-silicon cast iron in HNO3 as a function of acid concentration and temperature

Isocorrosion diagram for two high-silicon cast irons in HCl. A, Fe-14.3Si-4Cr-0.5Mo; B, Fe16Si-4Cr-3Mo

Fig. 7

168 / Casting Design and Performance to that of steel. These unalloyed cast irons are not attacked by dilute alkalis at any temperature. Hot alkalis at concentrations exceeding 30% attack unalloyed iron. Temperatures should not exceed 80 C (175 F) for concentrations up to 70% if corrosion rates of less than 0.25 mm/yr (10 mils/yr) are desired. Ductile and gray iron exhibit approximately equal resistance to alkalis. However, ductile cast iron is susceptible to cracking in highly alkaline solutions, but gray cast iron is not. Alloying with 3 to 5% Ni substantially improves the resistance of cast irons to alkalis. High-nickel austenitic cast irons offer even better resistance to alkalis than unalloyed or low-nickel cast irons. High-silicon cast irons show good resistance to relatively dilute solutions of NaOH at moderate temperatures but should not be applied for more concentrated conditions at elevated temperatures. High-silicon cast irons are usually economical over unalloyed and nickel cast irons in alkali solutions only when other corrosives are involved for which the lesser alloys are unsuitable. High-chromium cast irons have inferior resistance to alkali solutions and are generally not recommended for alkali services. Atmospheric corrosion is basically of interest only for unalloyed and low-alloy cast irons. Atmospheric corrosion rates are determined by the relative humidity and the presence of various gases and solid particles in the air. In high humidity, sulfur dioxide (SO2) or similar compounds found in many industrialized areas and chlorides found in marine atmospheres increase the rate of atmospheric attack on cast irons. Cast irons typically exhibit very low corrosion rates in industrial atmospheres—generally under 0.13 mm/yr (5 mils/yr)—and the cast irons are usually found to corrode at lower rates than steel structures in the same environment. White cast irons show the lowest rate of atmospheric corrosion of the unalloyed cast irons. Pearlitic cast irons are generally more resistant that ferritic cast irons to atmospheric corrosion. In marine atmospheres, unalloyed cast irons also exhibit relatively low rates of corrosion. Low alloy additions are sometimes made to improve corrosion resistance further. Higher alloy additions are even more beneﬁcial but are rarely warranted. Gray cast iron offers some added resistance over ductile cast iron in marine atmospheres. Corrosion in Soils. Cast iron use in soils, as in atmospheric corrosion, is basically limited to unalloyed and low-alloyed cast irons. Corrosion in soils is a function of soil porosity, drainage, and dissolved constituents in the soil. Irregular soil contact can cause pitting, and poor drainage increases corrosion rates substantially above the rates in well-drained soils. Neither metal-matrix nor graphite morphology has an important inﬂuence on the corrosion of cast irons in soils. Some alloying additions are made to improve the resistance of cast irons to attack in soils. For example, 3% Ni additions to cast iron are made to reduce initial attack in cast irons in poorly drained soils. Alloyed cast

irons would exhibit better resistance than unalloyed or low-alloyed cast irons but are rarely needed for soil applications, because unalloyed cast irons generally have long service lives, particularly if coatings and cathodic protection are used. Anodes placed in soils for impressed current cathodic protection are frequently constructed from high-silicon cast iron. The highsilicon cast iron is not needed to resist the basic soil environment but rather to extend service life when subjected to the high electrical current discharge rates commonly used in cathodic protective anodes. Several thousands of miles of cast iron pipe have been buried underground for decades, handling water distribution and collection for hundreds of municipalities. Much of this pipe is reaching the end of its useful life. Fortunately, technologies have been developed to line cast iron pipe in situ with polymer linings such as polyurethane or cement mortar (Ref 7, 8). These cure-in-place systems provide an economical alternative to open trench replacement, and the old cast iron pipe can still provide many years of structural integrity for the polymer or cement liners. Corrosion in Water. Unalloyed and lowalloyed cast irons are the primary cast irons used in water service. The corrosion resistance of unalloyed cast iron in water is determined by its ability to form protective scales. In hard water, corrosion rates are generally low because of the formation of calcium carbonate (CaCO3) scales on the surface of the iron. In softened or deionized water, the protective scales cannot be fully developed, and some corrosion will occur. In industrial waste waters, corrosion rates are primarily a function of the contaminants present. Acid pH waters increase corrosion, but alkaline pH waters lower rates. Chlorides increase the corrosion rates of unalloyed cast irons, although the inﬂuence of chlorides is small at a neutral pH. Seawater presents some special problems for cast irons. Gray cast iron may experience graphitic corrosion in calm seawater. It will also be galvanically active, that is, anodic, in contact with most stainless steels, copper-nickel alloys, titanium, and chrome-molybdenum nickel-base alloys. Because these materials are frequently used in seawater structures, this potential for galvanic corrosion must be considered. In calm seawater, the corrosion resistance of cast iron is not greatly affected by the presence of crevices. However, intermittent exposure to seawater is very corrosive to unalloyed cast irons. Use of high-alloy cast irons in water is relatively limited. High-nickel austenitic cast irons are used to increase the resistance of cast iron components to pitting in calm seawater. Chromium containing high-silicon cast iron is used to produce anodes for the anodic protection systems used in seawater and brackish water. Corrosion in Saline Solutions. The presence of salts in water can have dramatic effects on the selection of suitable grades of cast iron. Unalloyed cast irons exhibit very low corrosion

rates in such salts as cyanides, silicates, carbonates, and sulﬁdes, which hydrolyze to form alkaline solutions. However, in salts such as ferric chloride (FeCl3), cupric chloride (CuCl2), stannic salts, and mercuric salts, which hydrolyze to form acid solutions, unalloyed cast irons experience much higher rates. In salts that form dilute acid solutions, high-nickel cast irons are acceptable. More acidic and oxidizing salts, such as FeCl3, usually necessitate the use of high-silicon cast irons. Chlorides and sulfates of alkali metals yield neutral solutions, and unalloyed cast iron experiences very low corrosion rates in these solutions. More highly alloyed cast irons also exhibit low rates but cannot be economically justiﬁed for this application. Unalloyed cast irons are suitable for oxidizing salts, such as chromates, nitrates, nitrites, and permanganates, when the pH is neutral or alkaline. However, if the pH is less than 7, corrosion rates can increase substantially. At a lower pH with oxidizing salts, high-silicon cast iron is an excellent material selection. Ammonium salts are generally corrosive to unalloyed iron. High-nickel, high-chromium, and high-silicon cast irons provide good resistance to these salts. Other Environments. Unalloyed cast iron is used as a melting crucible for such low-melting metals as lead, zinc, cadmium, magnesium, and aluminum. Resistance to molten metals is summarized in Table 6. Ceramic coatings and washes are sometimes used to inhibit molten metal attack on cast irons. Cast iron can also be used in hydrogen chloride and chloride gases. In dry hydrogen chloride, unalloyed cast iron is suitable to 205 C (400 F), while in dry chlorine, unalloyed cast iron is suitable to 175 C (350 F). If moisture is present, unalloyed cast iron is unacceptable in HCl and Cl2 at any temperature.

Coatings Four general categories of coatings are used on cast irons to enhance corrosion resistance: metallic, organic, conversion, and enamel coatings. Coatings on cast irons are generally used to enhance the corrosion resistance of unalloyed and low-alloy cast irons and to lessen the requirements for cathodic protection. Highalloy cast irons such as Ni-Resist or white irons are rarely coated. Metallic coatings are used to enhance the corrosion resistance of cast irons. These coatings may either be sacriﬁcial metal coatings, such as zinc, or barrier metal coatings, such as nickel-phosphorus. From a corrosion standpoint, these two classes of coatings have important differences. Sacriﬁcial coatings are anodic when compared to iron, and the coatings corrode preferentially to protect the cast iron substrate. Small cracks and porosity in the coatings have a minimal overall effect on the performance of the coatings. Barrier coatings

Corrosion of Cast Irons / 169 are cathodic compared to iron, and the coatings can protect the cast iron substrate only when porosity or cracks are not present. If there are defects in the coatings, the service environment will attack the cast iron substrate at these imperfections, and the galvanic couple set up between the relatively inert coating and the casting may accelerate attack on the cast iron. Metallic coatings may be applied to cast irons by electroplating, hot dipping, ﬂame or thermal spraying, diffusion coating, or hard facing. Table 7 lists the metals that can be applied by these techniques. Zinc is one of the most widely used coatings on cast irons. Although zinc is anodic to iron, its corrosion rate is very low, and it provides relatively long-term protection for the cast iron substrate. A small amount of zinc will protect a large area of cast iron. Zinc coatings provide optimal protection in rural and arid areas. Other metal coatings are also commonly used on cast irons. Cadmium provides atmospheric protection similar to that of zinc. Tin coatings are frequently used to improve the corrosion resistance of equipment intended for food handling, and aluminum coatings protect against corrosive environments containing sulfur fumes, organic acids, salts, and compounds of nitrate-phosphate chemicals. Lead and leadtin coating are primarily applied to enhance the corrosion resistance of iron castings to H2SO3 and H2SO4. Nickel-phosphorus diffusion coatings offer corrosion resistance approaching that obtainable with stainless steel. Organic coatings can be applied to cast irons to provide short-term or long-term corrosion resistance. Short-term rust preventatives include oil, solvent-petroleum-based inhibitors and ﬁlm formers dissolved in petroleum solvents, emulsiﬁed-petroleum-based coatings modiﬁed to form a stable emulsion in water, and wax. For longer-term protection and resistance to more corrosive environments, rubber-based coatings, bituminous paints, asphaltic compounds, or thermoset and thermoplastic coatings can be applied. Rubber-based coatings include Table 6

chlorinated rubber neoprene, and Hypalon (DuPont Dow Elastomers). These coatings are noted for their mechanical properties and corrosion resistance but not for their decorative appearance. Bituminous paints have very low water permeability and provide high resistance to cast iron castings exposed to water. Use of bituminous paints is limited to applications that require good resistance to water, weak acids, alkalis, and salts. Asphaltic compounds are used to increase the resistance of cast irons to alkalis, sewage, acids, and continued exposure to tap water. Their application range is similar to that of bituminous paints. Cast irons are also lined with thermoset and thermoplastics, such as epoxy and polyethylene, to resist attack by ﬂuids. Fluorocarbon coatings offer superior corrosion resistance except in abrasive services. Fluorocarbon coatings applied to cast irons include such materials as polytetraﬂuoroethylene (PTFE), perﬂuoroalkoxy resins (PFA), polyvinyldene ﬂuoride (PVDF), ethylene chlorotriﬂuoroethylene (ECTFE), ethylene tetraﬂuoroethylene (ETFE), and ﬂuorinated ethylene polypropylene (FEP). Fully ﬂuorinated ﬂuorocarbon coatings resist deterioration in most common industrial services and can be used to 205 C (400 F), whereas partially ﬂuorinated coatings are limited to approximately 150 C (300 F). Cast iron lined with ﬂuorocarbon polymers can be very competitive with stainless, nickel-base, and even titanium and zirconium materials in terms of range of services covered and product cost. Conversion coatings are produced when the metal on the surface of the cast iron reacts with another element or compound to produce an iron-containing compound. Common conversion coatings include phosphate coatings, oxide coatings, and chromate coatings. Phosphate coatings enhance the resistance of cast iron to corrosion in sheltered atmospheric exposure. If the surface of the casting is oxidized and black iron oxide or magnetite is formed, the corrosion resistance of the iron can be enhanced, particularly if the oxide layer is impregnated with oil

or wax. Chromate coatings are formed by immersing the iron castings in an aqueous solution of chromic acid (H2CrO4) or chromium salts. Chromate coatings are sometimes used as a supplement to cadmium plating in order to prevent the formation of powdery corrosion products. The overall beneﬁts of conversion coatings are small with regard to atmospheric corrosion. Enamel Coatings. In the enamel coating of cast irons, glass frits are melted on the surface and form a hard, tenacious bond to the cast iron substrate. Good resistance to all acids except HF can be obtained with the proper selection and application of the enamel coating. Alkaline-resistant coatings can also be applied, but they offer only marginal improvement in the resistance to alkalis. Proper design and application are essential for developing enhanced corrosion resistance on cast irons with enamel coatings. Any cracks, spalling, or other coating imperfections may permit rapid attack of the underlying cast iron.

Selection of Cast Irons Cast irons provide excellent resistance to a wide range of corrosion environments when properly matched with that service environment. The basic parameters to consider before selecting cast irons for corrosion services include: Concentration of solution components in

weight percent

Dissolved contaminants, even at parts per

million levels

pH of solution Solution temperature, potential temperature

extremes, and rate of change of temperature

Degree of solution aeration Percent and type of solids suspended in the

solution

Duty cycle, continuous or intermittent oper-

ation or exposure

Resistance of gray cast iron to liquid metals at 300 and 600 C (570 and 1110 F)

Liquid metal

Liquid metal melting point, C

Mercury Sodium, potassium, and mixtures Gallium Bismuth-lead-tin Bismuth-lead Tin Bismuth Lead Indium Lithium Thallium Cadmium Zinc Antimony Magnesium Aluminum

38.8 12.3 to 97.9 29.8 97 125 321.9 271.3 327 156.4 186 303 321 419.5 630.5 651 660

Resistance of gray cast iron(a) 300 C (570 F)

Unknown Limited Unknown Good Unknown Limited Unknown Good at 327 C (621 F) Unknown Unknown Unknown Good at 321 C (610 F)

600 C (1110 F)

Unknown Poor Unknown Unknown Unknown Poor Unknown Unknown Unknown Unknown Unknown Good Poor Poor at 630.5 C (1167 F) Good at 651 C (1204 F) Poor at 660 C (1220 F)

(a) Good, considered for long-time use. <0.025 mm/yr (<1.0 mil/yr); Limited, short-time use only, 0.025–0.25 mm/yr (1.0–10 mils/yr); Poor, no structural possibilities, >0.25 mm/yr (> 10 mils/yr); Unknown, no data for these temperatures. Source: Ref 9

Table 7 Summary of metallic coating techniques to enhance corrosion resistance of cast irons Coating technique

Electroplating

Hot dipped Hard facing

Flame spraying Diffusion coating

Metals/alloys applied

Cadmium, chromium, copper, lead, nickel, zinc, tin, tin-nickel, brass, bronze Zinc, tin, lead, lead-tin, aluminum Cobalt-base alloys, nickel-base alloys, metal carbides, high-chromium ferrous alloys, high-manganese ferrous alloys, high-chromium and nickel ferrous alloys Zinc, aluminum, lead, iron, bronze, copper, nickel, ceramics, cermets Aluminum, chromium, nickelphosphorus, zinc, nitrogen, carbon

170 / Casting Design and Performance Potential for upset conditions, for example,

temperature and concentration excursions Unusual conditions, such as high solution velocity or vacuum Materials present in the system and the potential for galvanic corrosion Although it is advisable to consider each of the parameters before ultimate selection of a cast iron, the information needed to properly assess all variables of importance is often lacking. In such cases, introduction of test coupons of the candidate materials into the process stream should be considered before extensive purchases of equipment are made. If neither test coupons nor complete service data are viable alternatives, consultation with a reputable manufacturer of the equipment or the cast iron, with a history of applications in the area of interest, should be considered. ACKNOWLEDGMENT This article is adapted from “Corrosion of Cast Irons”, by Donald R. Stickle, Flowserve Corporation, Corrosion, Volume 13, ASM Handbook, ASM International, 1987, p 566–572.

REFERENCES

SELECTED REFERENCES

1. J.R. Keough, Austempered Ductile Iron, Section IV, Ductile Iron Data, The Ductile Iron Society, 1998 (available on website) 2. Development of a Cast Iron Graphitization Measurement Device, NYGAS Technol. Briefs, Issue 99–690-1, Jan 1999 3. J.R. McDowell, in Symposium on Fretting Corrosion, STP 144, American Society for Testing and Materials, 1952, p 24 4. P. Ferguson and D. Nicholas, Corros. Australas., April 1984, p 12 5. S.L. Chawla and R.K. Gupta, Materials Selection for Corrosion Control, ASM International 1993, p 56–59 6. E.C. Miller, Liquid Metals Handbook, 2nd ed., Government Printing Ofﬁce, 1952, p 144 7. New Lining System Upgrades Boston Gas Distribution System, Pipeline Gas J., April 1995, p 24–29 8. B.B. Hall, Rehabilitation of 1940’s Water Mains, Am. Water Works Assoc., Vol 91 (No. 12), 1999, p 91–94 9. R.I. Higgins, Corrosion of Cast Iron, J. Res., Feb 1956, p 165–177

S.A. Bradford, CASTI Practical Handbook of

Corrosion Control in Soils, co-published by CASTI and ASTM, 2000 “Corrosion Control of Ductile and Cast Iron Pipe,” 37254, NACE, 2001 Corrosion Data Survey, 6th ed., National Association of Corrosion Engineers, 1985 J.R. Davis, Ed., ASM Specialty Handbook: Cast Irons, ASM International, 1996 M.G. Fontana, Corrosion Engineering, 3rd ed., McGraw-Hill, 1986 “High Silicon Iron Alloys for Corrosion Services,” Bulletin A/2, Flowserve Corporation, Aug 1998 Properties and Selection: Irons, Steels, and High-Performance Alloys, Vol 1, ASM Handbook, ASM International, 1990 M. Szeliga, Corrosion of Ductile Iron Piping, NACE 1995 C.F. Walton, Ed., The Gray Iron Castings Handbook, A.L. Garber, 1957 C.F. Walton, Ed., Gray and Ductile Iron Castings Handbook, R.R. Donnelley & Sons, 1971

Casting Design and Performance Pages 171–173

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Corrosion of Cast Carbon and Low-Alloy Steels* STEEL CASTING COMPOSITIONS are generally divided into the categories of carbon and low-alloy, corrosion-resistant, or heat-resistant, depending on alloy content and intended service. Castings are classiﬁed as corrosion resistant if they are capable of sustained operation when exposed to attack by corrosive agents at service temperatures normally below 315 C (600 F). Carbon and low-alloy steels, the subject of this article, are considered resistant only to very mild corrosives, while the various high-alloy grades are applicable for varying situations from mild to severe services, depending on the particular conditions involved. For design and materials selection, the speciﬁc rate of corrosion may not be as important as the predictability and conﬁdence in predicting a rate of corrosion. It can be misleading to list the comparative corrosion rates of different alloys exposed to the same corroding medium. In this article, no attempt is made to recommend alloys for speciﬁc applications, and the data supplied should be used only as a general guideline. Alloy casting users will ﬁnd it helpful to consult materials and corrosion specialists when selecting alloys for a particular application. The factors that must be considered in materials selection include: The principal corrosive agents and their

concentrations

Known or suspected impurities, including

abrasive materials and their concentration

However, discretion and caution are suggested in evaluating the relative corrosion rates of various steels because of uncertainties in the results from controlled laboratory tests and simulated service condition tests, as well as anomalies in the intended environment. The best information is obtained from equipment used under actual operating conditions. Cast carbon and low-alloy steel and wrought steel of similar composition and heat treatment exhibit approximately the same corrosion resistance in the same environments. More detailed information applicable to cost alloys may be found in Ref 1 and 2. Plain carbon steel and some of the low-alloy steels do not ordinarily resist drastic corrosive conditions, although there are some exceptions, such as concentrated sulfuric acid (H2SO4).

Atmospheric Corrosion Unless shielded by a protective coating, iron and steel corrode in the presence of water and oxygen; therefore, steel will corrode when it is exposed to moist air. The rate at which corrosion proceeds in the atmosphere depends on the corroding medium, the conditions of the particular location in which the material is in use, and the steps that have been taken to prevent corrosion. The rate of corrosion also depends on the charac-

ter of the steel as determined by its chemical composition and heat treatment. To increase the corrosion resistance of steel signiﬁcantly, amounts of alloying elements are increased. Small amounts of copper and nickel slightly improve the resistance of steel to atmospheric attack, but appreciably larger amounts of other elements, such as chromium and nickel, improve corrosion resistance signiﬁcantly. The rate of corrosion of a material in an environment can generally be estimated with conﬁdence only from long-term tests. A 15-year research program compared the corrosion resistance of nine cast steels in marine and industrial atmospheres. Table 1 shows the compositions of the cast steels tested. The cast steel specimens exposed were 13 mm (½ in.) thick, 100 by 150 mm (4 by 6 in.) panels with beveled edges. The surfaces of half the specimens were machined. Specimens of each composition and surface condition were divided into three groups. One group was exposed to an industrial atmosphere at East Chicago, IN, and the other two groups were exposed to marine atmospheres 24 and 240 m (80 and 800 ft) from the ocean at Kure Beach, NC. The weight losses of the specimens during exposure were converted to corrosion rates in terms of millimeters (mils) per year. The results of this research are shown in Fig. 1 to 4. These are uniform corrosion rates that do not apply to localized corrosion modes, such as

Average operating temperature, including

variations even if experienced only for short periods Presence (or absence) of dissolved oxygen or other gases in solution Continuous or intermittent operation Fluid velocity Each of these can have a signiﬁcant effect on the service life of cast equipment, and such detailed information must be provided to make the appropriate materials selection. Many rapid failures are traceable to these details being overlooked—often when the information was available. Selection of the most economical alloy can be made by the judicious use of corrosion data.

Table 1

Compositions of cast steels tested in atmospheric corrosion Composition(a),%

Cast steel

Carbon, grade A Nickel-chromium-molybdenum 1Ni-1.7Mn 2% Ni Carbon, grade B 1% Cu 1.36Mn-0.09V 1.42% Mn 1.5Mn-0.05Ti

Nl.

Cu

Mn

Cr

V

C

Mo

P

S

Si

Other

0.10 0.56 1.08 2.26 0.03 0.04 0.01 0.01 0.01

0.13 0.13 0.08 0.12 0.03 0.94 0.15 0.13 0.11

0.61 0.80 1.70 0.77 0.65 0.87 1.36 1.42 1.48

0.21 0.60 0.08 0.19 0.10 0.11 0.08 0.16 0.04

0.03 0.04 0.04 0.03 0.04 0.07 0.09 0.04 0.03

0.14 0.26 0.27 0.17 0.25 0.28 0.37 0.37 0.33

trace 0.15

0.016

0.026

0.02 0.017 0.011

0.023 0.021 0.021

0.031 0.027 0.016

0.038 0.022 0.025

0.41 0.44 0.42 0.65 0.51 0.42 0.34 0.38 0.40

0.05Ti

(a) All compositions contain balance of iron. Source: Ref 3

*Reprinted from ASM Handbook, Volume 13B, Corrosion: Materials, S.D. Cramer, B.S. Covino, Jr., editors, p 51–53

trace

172 / Casting Design and Performance

Corrosion rates of various cast steels in a marine atmosphere. Nonmachined specimens were exposed 24 m (80 ft) from the ocean at Kure Beach, NC. Source: Ref 3

Fig. 1

Corrosion rates of various cast steels exposed at the 240 m (800 ft) site at Kure Beach, NC. Specimens were not machined. Source: Ref 3

Fig. 2

Corrosion rates for cast steels in an industrial atmosphere. Nonmachined specimens were exposed at East Chicago, IN. Source: Ref 3

Fig. 3

crevice corrosion, pitting, or local galvanic coupling. Figure 5 shows the results of another portion of this project. Corrosion rates for a 3-year exposure of various cast steels, wrought steels, and malleable iron in both atmospheres are compared. The following conclusions can be drawn from these tests: The condition of the specimen surface has

no signiﬁcant effect on the corrosion resistance of cast steels. Unmachined surfaces with the casting skin intact have corrosion rates similar to those of machined surfaces, regardless of the atmospheric environment. The highest corrosion rate occurs in the marine atmosphere 24 m (80 ft) from the ocean, with lower but similar corrosion rates occurring in the industrial atmosphere and the marine atmosphere 240 m (800 ft) from the ocean. The corrosion rate of cast steel decreases as a function of time, because corrosion products (scale and rust coating) build up and act as a protective coating on the cast steel surface. However, the corrosion rate of the most resistant cast steel (2% Ni) is always less than that of lesser corrosion-resistant cast steels. Cast steels with small amounts of copper or chromium, or slightly larger amounts of nickel, have corrosion resistance superior to that of cast carbon steel with manganese as an alloying element, when exposed to the atmospheres (Ref 3). Increasing the nickel and the chromium contents of cast steel increases the corrosion resistance in all three of the atmospheric environments.

All cast steels have greater corrosion resistance than malleable iron in industrial atmospheres and are superior or equivalent to the

Corrosion rates of machined and nonmachined specimens of cast steels after 7 years in three environments. The effect of surface ﬁnish on corrosion rates is negligible. Source: Ref 3

Fig. 4

Comparison of corrosion rates of cast steels, malleable cast iron, and wrought steel after 3 years of exposure in two atmospheres. Source: Ref 3

Fig. 5

Table 2 Corrosion of cast carbon and alloy steels in steam at 650 C (1200 F) for 570 h Composition, % Type of steel

Carbon Carbon-molybdenum Nickel-chromium-molybdenum 5Cr-molybdenum 7Cr-molybdenum(a) 9Cr-1.5Mo (a) Not a cast steel. Source: Ref 3

C

0.24 0.25 0.21 0.20 0.35 0.28 0.22 0.27 0.11 0.23

Cr

0.64 0.73 5.07 5.49 7.33 9.09

Ni

2.13 2.25

Average penetration rate Mo

mm/yr

mlls/yr

0.49 0.49 0.26 0.26 0.47 0.43 0.59 1.56

0.3 0.28 0.3 0.25 0.25 0.25 0.1 0.1 0.05 0.025

12 11 12 10 10 10 4 4 2 1

Corrosion of Cast Carbon and Low-Alloy Steels / 173 Table 3 Petroleum corrosion resistance of cast steels 1000 h test in petroleum vapor under 780 N (175 lb) of pressure at 345 C (650 F) Weight loss Type of material

mg/cm2

Cast carbon steel Cast steel, 2Ni-0.75Cr Seamless tubing, 5% Cr Cast steel, 5Cr-1W Cast steel, 5Cr-0.5Mo Cast steel, 12% Cr Stainless steel. 18Cr-8Ni

3040 2370 1540 950 730 6.4 2.1

mg/ln.2

196 153 99.2 61.5 47 100 30

Table 4 Corrosion of cast steels in waters Corrosive medium

Tap water Seawater Alternate immersion and drying Hot water 0.05% H2SO4 0.50% H2SO4

Corrosion factor(a)

Exposure time, months

Fe-0.29C-0.69Mn-0.44SI

Fe-0.32C-0.66Mn-1.12Cr

Fe-0.11C-0.41Mn-3.58Cr

2 6 2 6 2 6 1 2 6 2

100 100 100 100 100 100 100 100 100 100

85 73 60 80 93 109 100 71 89 223

58 61 26 40 30 25 64 68 102 61

Source: Ref 3

(a) Corrosion factor is the ratio of average penetration rate of the alloy in question to Fe-0.29C-0.69Mn-0.44Si steel. Source: Ref 3

wrought steels in this environment. The corrosion rate in the marine atmosphere depends primarily on the alloy content. The cast carbon steel is much superior to the AISI 1020 wrought steel but is slightly inferior to malleable iron (Ref 4).

Table 5

Corrosion of cast chromium and carbon steels in mineral acids Weight loss in 5 h 5% H2SO4

Steel

Carbon steel, 0.31% C Chromium steel, 0.30C-2.42Cr

5% HCl

5% HNO3

mg/cm2

mg/ln.2

mg/cm2

mg/ln.2

mg/cm2

mg/ln.2

2.7 4.9

17.42 31.6

2.1 5.41

13.55 34.9

80.79 47.36

521.1 305.5

Source: Ref 3

Other Environments Several low- and high-alloy cast steels have been studied regarding their corrosion resistance to high-temperature steam. Test specimens 150 mm (6 in.) in length and 13 mm (½ in.) in diameter were machined from test coupons and then exposed to steam at 650 C (1200 F) for 570 h. The steel compositions and test results are given in Table 2. Table 3 shows the resistance of cast steels to petroleum corrosion, and Tables 4 and 5 supply similar data relating to water and acid attack. These data show the value of higher chromium content for improved corrosion resistance.

ACKNOWLEDGMENT This article has been adapted from Raymond Monroe and Steven Pawel, Corrosion of Cast Steels, Corrosion, Volume 13, ASM Handbook, ASM International, 1987.

REFERENCES 1. T. Kodomd, Corrosion of Wrought Carbon Steels, Corrosion: Materials, Vol 13B,

ASM Handbook, ASM International, 2005, p 5–10 2. T. Oakwood, Corrosion of Wrought LowAlloy Steels Corrosion: Materials, Vol 13B, ASM Handbook, ASM International, 2005, p 11–27 3. C.W. Briggs, Ed., Steel Casting Handbook, 4th ed., Steel Founders’ Society of America, 1970, 662–667 4. C.W. Briggs, “Atmospheric Corrosion of Carbon and Low Alloy Cast Steels,” ASTM Corrosion Symposium, 1967

Casting Design and Performance Pages 175–183

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Corrosion of Cast Stainless Steels* Revised by Malcolm Blair, Steel Founders’ Society of America

CAST STAINLESS STEELS are usually speciﬁed on the basis of composition by using the alloy designation system established by the Alloy Casting Institute (ACI). The ACI designations, such as CF-8M, have been adopted by ASTM International and are preferred for cast alloys over the designations originated by the American Iron and Steel Institute (AISI) for similar wrought steels. The ﬁrst letter of the ACI designation indicates whether the alloy is intended primarily for liquid corrosion service (C) or heat-resistant service (H). The second letter denotes the nominal chromium-nickel type, as shown in Fig. 1. As the nickel content increases, the second letter in the ACI designation increases from A to Z. The numerals following the two letters refer to the maximum carbon content (percent 100) of the alloy. If additional alloying elements are included, they can be denoted by the addition of one or more letters after the maximum carbon content. Thus, the designation CF-8M refers to an alloy for corrosion-resistant service (C) of the 19Cr-9Ni (F) type, with a maximum carbon content of 0.08% and containing molybdenum (M). Corrosion-resistant cast stainless steels are also often classiﬁed on the basis of microstructure. The classiﬁcations are not completely independent, and a classiﬁcation by composition often involves microstructural distinctions. Cast corrosion- and heat-resistant alloy compositions are listed in Table 1.

Composition and Microstructure The principal alloying element in the highalloy family is usually chromium, which, through the formation of protective oxide ﬁlms, results in the corrosion protection or stainless behavior. For most purposes, stainless behavior requires at least 12% Cr. Corrosion resistance further improves with additions of chromium to at least the 30% level. As indicated in Table 1, signiﬁcant amounts of nickel and lesser amounts of molybdenum and other elements are often added to the iron-chromium matrix.

Although chromium is a ferrite and martensite promoter, nickel is an austenite promoter. By varying the amounts and ratios of these two elements (or their equivalents), almost any desired combination of microstructure, strength, or other property can be achieved. Varying the temperature, time at temperature, and cooling rate of the heat treatment also controls the desired results. It is useful to think of the compositions of high-alloy steels in terms of the balance between austenite promoters and ferrite promoters. This balance is shown in the widely used Schaefﬂer diagrams (Fig. 2). It should be noted that the Schaefﬂer diagram is used for welding and that the phases shown are those that persist after cooling to room temperature at rates consistent with fabrication (Ref 1, 2). The Schoefer diagram (Fig. 3) gives an indication of the amount of ferrite that may be expected based on the composition of the alloy in question. An ASTM standard provides note on the Schoefer diagram and methods for estimating ferrite content (Ref 3). The empirical correlations shown in Fig. 2 can be understood from the following. The ﬁeld designated as martensite encompasses such alloys as CA-15, CA-6NM, and even CB-7Cu. These alloys contain 12 to 17% Cr, with adequate nickel, molybdenum, and carbon to promote high hardenability, that is, the ability to transform completely to martensite when cooled at even the moderate rates associated with the air cooling of heavy sections. High alloys have low thermal conductivities and cool slowly. To obtain the desired properties, a full heat treatment is required after casting; that is, the casting is austenitized by heating to 870 to 980 C (1600 to 1800 F), cooled to room temperature to produce the hard martensite, and then tempered at 595 to 760 C (1100 to 1400 F) until the desired combination of strength, toughness, ductility, and resistance to corrosion or stress corrosion is obtained (Ref 1, 2). Increasing the nickel equivalent (moving vertically in Fig. 2) eventually results in an alloy that is fully austenitic, such as CC-20, CH-20, CK-20, or CN-7M. These alloys are extremely ductile, tough, and corrosion resistant. On the

* Reprinted from ASM Handbook, Volume 13B; Corrosion: Materials S.D. Cramer, B.S. Covino, Jr., editors, p78–87

other hand, the yield and tensile strength may be relatively low for the fully austenitic alloys. Being fully austenitic, they are nonmagnetic. Heat treatment consists of a single step: water quenching from a relatively high temperature at which carbides have been taken into solution. Solution treatment may also homogenize the structure, but because no transformation occurs, there can be no grain reﬁnement. The solutionizing step and rapid cooling ensure maximum resistance to corrosion. Temperatures between 1040 and 1205 C (1900 and 2200 F) are usually required (Ref 1, 2). Adding chromium to the lean alloys (proceeding horizontally in Fig. 2) stabilizes the dferrite that forms when the casting solidiﬁes. Examples are CB-30 and CC-50. With high chromium content, these alloys have relatively good resistance to corrosion, particularly in sulfur-bearing atmospheres. However, being single-phase, consisting only of ferrite, they are nonhardenable, have moderate-to-low strength, and are often used as-cast or after only a simple solution heat treatment. Ferritic alloys also have relatively poor impact resistance (toughness) (Ref 1, 2). Between the ﬁelds designated M, A, and F in Fig. 2 are regions indicating the possibility of two or more phases in the alloys. Commercially, the most important of these alloys are the ones in which austenite and ferrite coexist, such as CF-3, CF-8, CF-3M, CF-8M, CG-8M, and CE-30. These alloys usually contain 3 to

Chromium and nickel contents in ACI standard grades of heat- and corrosion-resistant castings. See text for details. Source: Ref 1

Fig. 1

176 / Casting Design and Performance Table 1 Compositions of Alloy Casting Institute (ACI) heat- and corrosion-resistant casting alloys Composition (balance Iron)(b), % ACI designation

UNS No.

CA-15 CA-15M CA-40 CA-6NM CA-6N CB-30 CB-7Cu-1 CB-7Cu-2 CC-50 CD-4MCu CE-30 CF-3 CF-8 CF-20 CF-3M CF-8M CF-8C CF-16F CG-12 CG-8M CH-20 CK-20 CN-7M CN-7MS CW-12M CY-40 CZ-100 N-12M M-35 HA HC HD HE HF HH HI HK HL HN HP HP-50WZ HT HU HW HX

J91150 J91151 J91153 J91540 J91650 J91803 ... ... J92615 ... J93423 J92500 J92600 J92602 J92800 J92900 J92710 J92701 J93001 ... J93402 J94202 N08007 ... N30002 N06040 N02100 ... ... ... J92605 J93005 J93403 J92603 J93503 J94003 J94224 J94604 J94213 ... ... J94605 N08004 N08001 N06006

Wrought alloy type(a)

410 ... 420 ... ... 431 ... ... 446 ... ... 304L 304 302 316L 316 347 303 ... 317 309 310 ... ... ... ... ... ... ... ... 446 327 ... 302B 309 ... 310 ... ... ... ... 330 ... ... ...

C

0.15 0.15 0.20–0.40 0.06 0.06 0.30 0.07 0.07 0.50 0.04 0.30 0.03 0.08 0.20 0.03 0.08 0.08 0.16 0.12 0.08 0.20 0.20 0.07 0.07 0.12 0.40 1.00 0.12 0.35 0.20 0.50 0.50 0.20–0.50 0.20–0.40 0.20–0.50 0.20–0.50 0.20–0.60 0.20–0.60 0.20–0.50 0.35–0.75 0.45–0.55 0.35–0.75 0.35–0.75 0.35–0.75 0.35–0.75

Mn

1.00 1.00 1.00 1.00 0.50 1.00 0.70 0.70 1.00 1.00 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.50 2.00 1.50 1.00 1.00 1.50 1.50 1.00 1.50 0.35–0.65 1.00 1.50 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00

Si

1.50 0.65 1.50 1.00 1.00 1.50 1.00 1.00 1.50 1.00 2.00 2.00 2.00 2.00 1.50 2.00 2.00 2.00 2.00 1.50 2.00 2.00 1.50 2.50–3.50 1.50 3.00 2.00 1.00 2.00 1.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.50 2.00 2.50 2.50 2.50 2.50

P

0.04 0.04 0.04 0.04 0.02 0.04 0.035 0.035 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.17 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.04 0.03 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04

S

0.04 0.04 0.04 0.03 0.02 0.04 0.03 0.03 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04

Cr

Ni

11.5–14.0 11.50–14.0 11.5–14.0 11.5–14.0 10.5–12.0 18.0–21.0 14.0–15.5 14.0–15.5 26.0–30.0 24.5–26.5 26.0–30.0 17.0–21.0 18.0–21.0 18.0–21.0 17.0–21.0 18.0–21.0 18.0–21.0 18.0–21.0 20.0–23.0 18.0–21.0 22.0–26.0 23.0–27.0 19.0–22.0 18.0–20.0 15.5–20.0 14.0–17.0 ... 1.0 ... 8.0–10.0 26.0–30.0 26.0–30.0 26.0–30.0 18.0–23.0 24.0–28.0 26.0–30.0 24.0–28.0 28.0–32.0 19.0–23.0 24.0–28.0 24.0–28.0 15.0–19.0 17.0–21.0 10.0–14.0 15.0–19.0

1.00 1.00 1.00 3.5–4.5 6.0–8.0 2.00 4.5–5.5 4.5–5.5 4.00 4.75–6.00 8.0–11.0 8.0–21.0 8.0–11.0 8.0–11.0 9.0–13.0 9.0–12.0 9.0–12.0 9.0–12.0 10.0–13.0 9.0–13.0 12.0–15.0 19.0–22.0 27.5–30.5 22.0–25.0 bal bal bal bal bal ... 4.00 4.0–7.0 8.0–11.0 8.0–12.0 11.0–14.0 14.0–18.0 18.0–22.0 18.0–22.0 23.0–27.0 33.0–37.0 33.0–37.0 33.0–37.0 37.0–41.0 58.0–62.0 64.0–68.0

Other elements

0.5Mo(c) 0.15–1.00Mo 0.5Mo(c) 0.4–1.0Mo

... ... 0.15–0.35Nb, 0.05N, 2.5–3.2Cu 0.15–0.35Nb, 0.05N, 2.5–3.2Cu ... 1.75–2.25Mo, 2.75–3.25Cu ... ... ... ... 2.0–3.0Mo 2.0–3.0Mo 3 C min, 1.0 max Nb 1.5Mo, 0.2–0.35Se ... 3.0–4.0Mo ... ... 2.0–3.0Mo, 3.0–4.0Cu 2.0–3.0Mo, 1.5–2.0Cu 7.5Fe 11.0Fe 3.0Fe, 1.25Cu 0.26–0.33Mo, 0.60V, 2.50Co, 6.0Fe 28–33Cu, 3.5Fe 0.90–1.20Mo 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c), 0.2N 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 4.0–6.0W, 0.2–1.0Zr 0.5Mo(c) 0.5Mo(c) 0.5Mo(c) 0.5Mo(c)

(a) Cast alloy chemical composition ranges are not the same as the wrought composition ranges; buyers should use cast alloy designations for proper identiﬁcation of castings. (b) Maximum, unless range is given. (c) Molybdenum not intentionally added

Fig. 2

Schaefﬂer diagram showing the amount of ferrite and austenite present in weldments as a function of chromium and nickel equivalents. Source: Ref 1

30% ferrite in a matrix of austenite. Predicting and controlling ferrite content is vital to the successful application of these materials. Alloys that contain both ferrite and austenite offer superior strength, weldability, and corrosion resistance compared to alloys that contain only austenite. Strength, for example, increases directly with ferrite content. Achieving speciﬁed minimums may necessitate controlling the ferrite within narrow bands. Figure 3 and Schoefer’s equations are used for this purpose. These duplex alloys should be solution heat treated and rapidly cooled before use to ensure maximum resistance to corrosion (Ref 1, 2). The presence of ferrite is not beneﬁcial for every application. Ferrite tends to reduce toughness, although this is not of great concern, given the extremely high toughness of the austenite matrix. However, in applications that require exposure to elevated temperatures, usually 315 C (600 F) and higher, the metallurgical changes associated with the ferrite can be

Corrosion of Cast Stainless Steels / 177

Schoefer diagram for estimating the average ferrite content in austenitic iron-chromiumnickel alloy castings. Source: Ref 1

Fig. 3

Effect of chromium on oxidation resistance of cast steels. Speciments (13 mm, or 0.5 in., cubes) were exposed for 48 h at 1000 C (1830 F). Source: Ref 2

Fig. 4

Toughness of solution-annealed duplex stainless steel castings (closed symbols) and companion wrought alloys (open symbols) as a function of test temperature. Source: Ref 4

Fig. 6

Corrosion behavior of ACl H-type (heat-resistant) alloy castings in (a) air and in (b) oxidizing ﬂue gases containing 5 grains of sulfur per 2.8 m3 (100 ft3) of gas. Source: Ref 2

Fig. 5

severe and detrimental. In the low end of this temperature range, the observed reductions in toughness have been attributed to carbide precipitation or reactions associated with 475 C (885 F) embrittlement. The 475 C (885 F) embrittlement is caused by precipitation of an intermetallic phase with a composition of approximately 80Cr-20Fe. The name derives from the fact that this embrittlement is most severe and rapid when it occurs at approximately 475 C (885 F). At 540 C (1000 F) and above, the ferrite phase may transform to a complex Fe-Cr-Ni-Mo intermetallic compound known as sigma (s) phase, which reduces toughness, corrosion resistance, and creep ductility. The extent of the reduction increases with time and temperature to approximately 815 C (1500 F) and may persist to 925 C (1700 F). In extreme cases, Charpy V-notch energy at room temperature may be reduced 95% from its initial value (Ref 1, 2). It has been demonstrated that

the impact properties of duplex stainless steels in the solution heat treated condition, in the cast and wrought form, are comparable (Fig. 4). More information on the metallography and microstructures of these alloys is available in Ref 5.

Corrosion Behavior of H-Type Alloys The ACI heat-resistant (H-type) alloys must be able to withstand temperatures exceeding 1095 C (2000 F) in the most severe high-temperature service. Chromium content is important to the corrosion behavior of these alloys. Chromium imparts resistance to oxidation and sulﬁdation at high temperatures by forming a passive oxide ﬁlm. Heat-resistant casting alloys must also have good resistance to carburization.

More information on the corrosion of metals and alloys in high-temperature gases is available in Ref 6. Oxidation. Resistance to oxidation increases directly with chromium content (Fig. 5). For the most severe service at temperatures above 1095 C (2000 F), 25% or more chromium is required. Additions of nickel, silicon, manganese, and aluminum promote the formation of relatively impermeable oxide ﬁlms that retard further scaling. Thermal cycling is extremely damaging to oxidation resistance, because it leads to breaking, cracking, or spalling of the protective oxide ﬁlm. The best performance is obtained with austenitic alloys containing 40 to 50% combined nickel and chromium Figure 6 shows the behavior of the H-type grades. Sulﬁdation environments are becoming increasingly important. Petroleum processing,

178 / Casting Design and Performance 15M, CA-6NM, CA-6NM-B, CA-40, CB-7Cu-1, and CB-7Cu-2. These alloys are generally used in applications requiring high strength and modest corrosion resistance. Alloy CA-15 typically exhibits a microstructure of martensite and ferrite. This alloy contains the minimum amount of chromium to be considered a stainless steel (11 to 14% Cr) and as such may not be used in aggressive environments. It does, however, exhibit good atmospheric corrosion resistance, and it resists staining by many organic environments. Alloy CA-15M may contain slightly more molybdenum than CA-15 (up to 1% Mo) and therefore may have improved general corrosion resistance in relatively mild environments. Alloy CA-6NM is similar to CA-15M except that it contains more nickel and molybdenum, which improves its general corrosion resistance. Alloy CA-6NM-B is a lower-carbon version of this alloy. The lower strength level promotes resistance to sulﬁde stress cracking. Alloy CA-40 is a higher-strength version of CA-15, and it also exhibits excellent atmospheric-corrosion resistance after a normalize and temper heat treatment. Micro-structurally, the CB-7Cu alloys usually consist of mixed martensite and ferrite, and because of the increased chromium and nickel levels compared to the other martensitic alloys, they offer improved corrosion resistance to seawater and some mild acids. These alloys also have good atmospheric-corrosion resistance. The CB-7Cu alloys are hardenable and offer the possibility of increased strength and improved corrosion resistance among the martensitic alloys. General Corrosion of Ferritic Alloys. Alloys CB-30 and CC-50 are higher-carbon

and higher-chromium alloys than the CA alloys previously mentioned. Each alloy is predominantly ferritic, although a small amount of martensite may be found in CB-30. Alloy CB-30 contains 18 to 21% Cr and is used in chemical-processing and oil-reﬁning applications. The chromium content is sufﬁcient to have good corrosion resistance to many acids, including nitric acid (HNO3) (Fig. 8). Alloy CC-50 contains substantially more chromium (26 to 30%) and offers relatively high resistance to localized corrosion and high resistance to many acids, including dilute H2SO4 and such oxidizing acids as HNO3. General Corrosion of Austenitic and Duplex Alloys. Alloy CF-8 typically contains approximately 19% Cr and 9% Ni and is essentially equivalent to type 304 wrought alloys. Alloy CF-8 may be fully austenitic, but it more commonly contains some residual ferrite (3 to 30%) in an austenite matrix. In the solutiontreated condition, this alloy has excellent resistance to a wide variety of acids. It is particularly resistant to highly oxidizing acids, such as boiling HNO3. Figure 9 shows isocorrosion diagrams for CF-8 in HNO3, phosphoric acid (H3PO4), and sodium hydroxide (NaOH). The duplex nature of the microstructure of this alloy imparts additional resistance to stress-corrosion cracking (SCC) compared to its wholly austenitic counterparts. Alloy CF-3 is a reduced-carbon version of CF-8 with essentially identical corrosion resistance, except that CF-3 is much less susceptible to sensitization (Fig. 10). For applications in which the corrosion resistance of the weld heat-affected zone (HAZ) may be critical, CF-3 is chosen. Alloys CF-8A and CF-3A contain more ferrite than their CF-8 and CF-3 counterparts. Because the higher ferrite content is achieved by increasing the chromium/nickel equivalent ratio, the CF-8A and CF-3A alloys may have slightly higher chromium or slightly lower nickel contents than the low-ferrite equivalents. In general, the corrosion resistance is very similar, but the strength increases with ferrite content. Because of the high ferrite content, service should be restricted to temperatures below

Corrosion behavior of ACl H-type alloys in 100 h tests at 980 C (1800 F) in reducing sulfur-bearing gases. (a) Gas contained 5 grains of sulfur per 2.8 m3 (100 ft3) of gas. (b) Gas contained 300 grains of sulfur per 2.8 m3 (100 ft3) of gas. (c) Gas contained 100 grains of sulfur per 2.8 m3 (100 ft3) of gas; test at constant temperature. (d) Same sulfur content as gas in (c), but cooled to 150 C (300 F) each 12 h

Isocorrosion diagram for ACl CB-30 in HNO3. Castings were annealed at 790 C (1450 F), furnace cooled to 540 C (1000 F), and then air cooled to room temperature.

coal conversion, utility and chemical applications, and waste incineration have heightened the need for alloys resistant to sulﬁdation attack in relatively weak oxidizing or reducing environments. Fortunately, high chromium and silicon contents increase resistance to sulfur-bearing environments. On the other hand, nickel has been found to be detrimental to the most aggressive gases. The problem is attributable to the formation of low-melting nickel-sulfur eutectics. These produce highly destructive liquid phases at temperatures even below 815 C (1500 F). Once formed, the liquid may run onto adjacent surfaces and rapidly corrode other metals. The behavior of H-type grades in sulﬁdizing environments is represented in Fig. 7. Carburization. High alloys are often used in nonoxidizing atmospheres in which carbon diffusion into metal surfaces is possible. Depending on chromium content, temperature, and carburizing potential, the surface may become extremely rich in chromium carbides, rendering it hard and possibly susceptible to cracking. Silicon and nickel are thought to be beneﬁcial and enhance resistance to carburization (Ref 7).

Corrosion Behavior of C-Type Alloys The ACI C-type stainless steels must resist corrosion in the various environments in which they regularly serve. The inﬂuence of the metallurgy of these materials on general corrosion, intergranular corrosion, localized corrosion, corrosion fatigue, and stress corrosion are discussed. General Corrosion of Martensitic Alloys. The martensitic grades include CA-15, CA-

Fig. 7

Fig. 8

Corrosion of Cast Stainless Steels / 179 400 C (750 F) due to the possibility of severe embrittlement. Alloy CF-8C is the niobium-stabilized grade of the CF-8 alloy class. This alloy contains small amounts of niobium, which tend to form carbides preferentially over chromium carbides and improve intergranular corrosion resistance in applications involving relatively high service temperatures. The development of niobium carbides is achieved through a heat treatment at 870 to 900 C (1600 to 1650 F); this is often referred to as a stabilizing heat treatment.

Fig. 9

Alloys CF-8M, CF-3M, CF-8MA, and CF3MA are 2 to 3% Mo-bearing versions of the CF-8 and CF-3 alloys. The addition of molybdenum increases resistance to corrosion by seawater and improves resistance to many chloride-bearing environments. The presence of molybdenum also improves crevice corrosion and pitting resistance, compared to the CF8 and CF-3 alloys. Molybdenum-bearing alloys are generally not as resistant to highly oxidizing environments when phases rich in molybdenum are formed (this is particularly true for boiling

Isocorrosion diagrams for ACl CF-8 in (a) HNO3, (b and c) H3PO4, and (d and e) NaOH solutions. (b) and (d) Tests performed in a closed container at equilibrium pressure. (c) and (e) Tested at atmospheric pressure

HNO3), but for weakly oxidizing environments and reducing environments, molybdenum-bearing alloys are generally superior. Alloy CF-16F is a selenium-bearing freemachining grade of cast stainless steel. Because CF-16F nominally contains 19% Cr and 10% Ni, it has adequate corrosion resistance to a wide range of corrodents, but the large number of selenide inclusions makes surface deterioration and pitting deﬁnite possibilities. Alloy CF-20 is a fully austenitic, relatively high-strength corrosion-resistant alloy. The 19% Cr content provides resistance to many types of oxidizing acids, but the high carbon content makes it imperative that this alloy be used in the solution-treated condition for environments known to cause intergranular corrosion. Alloy CE-30 is a nominally 27Cr-9Ni alloy that typically contains 10 to 20% ferrite in an austenite matrix. The high carbon and ferrite contents provide relatively high strength. The high chromium content and duplex structure act to minimize corrosion resulting from the formation of chromium carbides in the micro-structure. This particular alloy is known for good resistance to sulfurous acid and sulfuric acid, and it is extensively used in the pulp and paper industry. Alloy CG-8M is slightly more highly alloyed than the CF-8M alloys, with the primary addition being increased molybdenum (3 to 4%). The increased amount of molybdenum provides superior corrosion resistance to halide-bearing media and reducing acids, particularly H2SO3 and H2SO4 solutions. The high molybdenum content, however, renders CG-8M generally unsuitable in highly oxidizing environments. Alloy CD-4MCu is the most highly alloyed material in this group, with a composition of Fe-26Cr-5Ni-2Mo-3Cu. The chromium/nickel equivalent ratio for this alloy is quite high, and a microstructure containing approximately equal amounts of ferrite and austenite is common. The low carbon content and high chromium content render the alloy relatively immune to intergranular corrosion. High chromium and molybdenum provide a high degree of localized corrosion resistance (crevices and pitting), and the duplex microstructure provides

Fig. 10

Isocorrosion diagram for solution-treated quenched and sensitized ACl CF-3 in HNO3

180 / Casting Design and Performance Table 2 Duplex stainless steel corrosion test results

Grade(a)

CD-4MCuN

CD-3MN

Isocorrosion diagram for ACl CD-4MCu in HNO3. The material was solution treated at 1120 C (2050 F) and water quenched.

Fig. 11

CE-3MN

CD-3MWCuN

Heat

1 2 3 4 1 2 3 4 4(e) 1 2 3 3(e) 1 2 3

A 923C(b)

G 48(c)

Corrosion rate, mdd(d)

CPT, C

CPT, C

0.00 0.00 3.45 2.87 0.73 2.19 0.00 0.00 2.12 2.64 0.00 0.00 0.00 0.00 0.00 0.67

35 40 30 35 40 25 50 45 50 65 50 65 50 65 70 55

35 40 30 35 40 35 50 45 50 65 50 65 50 65 70 55

(a) ASTM A 890, solution annealed. (b) ASTM A 923 method C ferric chloride corrosion test. (c) ASTM G 48 method C critical pitting temperature (CPT) test, 6% FeCl3, 24 h. (d) Corrosion rate calculated from weight loss; mdd is mg/dm2/day. Maximum acceptable corrosion rate is 10 mdd; all specimens passed. (e) Centrifugally cast specimens

Isocorrosion diagram for ACl CD-4MCu in H2SO4. The material was solution annealed at 1120 C (2050 F) and water quenched.

Fig. 12

SCC resistance in many environments. This alloy can be precipitation hardened to provide strength and is also relatively resistant to abrasion and erosion-corrosion. Figures 11 and 12 show isocorrosion diagrams for CD-4MCu in HNO3 and H2SO4, respectively. CD-4MCu does not require control of the nitrogen content, which can lead to excessive levels of ferrite that reduce the toughness of the material. The control of nitrogen within the range speciﬁed for CD-4MCuN eliminates this problem. The ASME Pressure Vessel Code recognized this fact and has replaced CD-4MCu with CD-4MCuN. The results of a series of corrosion tests on CD-4MCuN, CD-3MN, CE-3MN, and CD3MWCuN are shown in Table 2. The ASTM A 923 test detects the presence of detrimental intermetallic phases. The weight loss is associated with local depletion. The critical pitting temperature indicates the minimum temperature that pitting occurs. Grades CE-3MN and CD3MWCuN are known as superduplex stainless steels. A method of ranking the pitting resistance of duplex stainless steels has been developed. It is an empirical measure and is known as the pitting resistance number (PREN). The PREN is based on the composition of the alloy, and for super duplex stainless steels, the PREN should not be less than 40: PREN ¼ % Cr þ 3:3 %Mo þ 16 %N

Table 2 shows the improved pitting resistance of these alloys. Fully Austenitic Alloys. Alloys CH-10 and CH-20 are fully austenitic and contain 22 to 26% Cr and 12 to 15% Ni. The high chromium content minimizes the tendency toward the formation of chromium-depleted zones during heat treatment in the sensitizing temperature range. These alloys are often used for handling paper pulp solutions and are known for good resistance to dilute H2SO4 and HNO3. Alloy CK-20 contains 23 to 27% Cr and 19 to 22% Ni and is less susceptible than CH-20 to intergranular corrosion attack in many acids after brief exposures to the chromium carbide formation temperature range. Alloy CK-20 possesses good corrosion resistance to many acids and, because of its fully austenitic structure, can be used at relatively high temperature. Alloy CN-7M, with a nominal composition of Fe-29Ni-20Cr-2.5Mo-3.5Cu, exhibits excellent corrosion resistance in a wide variety of environments and is often used for H2SO4 service. Figure 13 shows isocorrosion diagrams for CN-7M in H2SO4, HNO3, H3PO4, and NaOH. Relatively high resistance to intergranular corrosion and SCC make this alloy attractive for very many applications. Although relatively highly alloyed, the fully austenitic structure of CN-7M may lead to SCC susceptibility for some environments and stress states. Intergranular Corrosion of Austenitic and Duplex Alloys. The optimal corrosion resistance for these alloys is developed by solution treatment. Depending on the speciﬁc alloy in question, temperatures between 1040 and 1205 C (1900 and 2200 F) are required to ensure complete solution of all carbides and other high-alloy phases, such as s and w, that sometimes form in highly alloyed stainless steels. Alloys containing relatively high total alloy content, particularly high molybdenum content,

often require the higher solution treatment temperature. Water quenching from the temperature range of 1040 to 1205 C (1900 to 2200 F) normally completes the solution treatment. Failure to solution treat a particular alloy or an improper solution treatment may seriously compromise the observed corrosion resistance in service. Inadvertent or unavoidable heat treatment in the temperature range of 480 to 820 C (900 to 1500 F), such as caused by welding, may destroy the intergranular corrosion resistance of the alloy. When austenitic or duplex (ferrite in austenite matrix) stainless steels are heated in or cooled slowly through this temperature range, chromium-rich carbides form at grain boundaries in austenitic alloys and at ferrite/austenite interfaces in duplex alloys. These carbides deplete the surrounding matrix of chromium, thus diminishing the local corrosion resistance of the alloy. An alloy in this condition of reduced corrosion resistance due to the formation of chromium carbides is said to be sensitized. In small amounts, these carbides may lead to localized pitting in the alloy, but if the chromium-depleted zones are interconnected throughout the alloy or HAZ of a weld, the alloy may disintegrate intergranularly in some environments. If solution treatment of the alloy after casting and/or welding is impractical or impossible, the metallurgist has techniques to minimize potential intergranular corrosion problems. These include stabilizing of carbides by the addition of niobium, as described earlier, by cathodic protection, or by reducing the carbon content. The low-carbon grades CF-3 and CF-3M are commonly used as a solution to the sensitization incurred during welding. The low carbon content (0.03% C maximum) of these alloys precludes the formation of an extensive number of chromium carbides. In addition, these alloys normally contain 3 to 30% ferrite in an austenitic matrix. By virtue of rapid carbide precipitation kinetics at ferrite/austenite interfaces compared to austenite/austenite interfaces, carbide precipitation is conﬁned to ferrite/ austenite boundaries in alloys containing a minimum of approximately 3 to 5% ferrite (Ref 8, 9). If the ferrite network is discontinuous in the austenite matrix (depending on the amount, size, and distribution of ferrite pools), then extensive intergranular corrosion will not be a problem in most of the environments to which these alloys would be subjected. An example of attack at the ferrite/austenite boundaries is shown in Fig. 14. These low-carbon alloys need not sacriﬁce signiﬁcant strength compared to their high-carbon counterparts, because nitrogen may be added to increase strength. However, a large amount of nitrogen will begin to reduce the ferrite content, which will cancel some of the strength gained by interstitial hardening. Appropriate adjustment of the chromium/nickel equivalent ratio is beneﬁcial in such cases. Fortunately, nitrogen is also beneﬁcial to the corrosion resistance of austenitic and duplex stainless steels (Ref 10). Nitrogen seems to retard sensitization and

Corrosion of Cast Stainless Steels / 181

Fig. 13

Isocorrosion diagrams for solution-annealed and quenched ACl CN-7M in H2SO4, HNO3, NaOH, and H3PO4. (a), (b), (d), and (f) Tested at atmospheric pressure. (c) and (e) Tested at equilibrium pressure in a closed container. See Fig. 9 for legend.

Ferrite/austenite grain-boundary ditching in as-cast ACl CF-8. The specimen, which contained 3% ferrite, was electrochemical potentiokinetic reactivation tested, SEM micrograph. Original magniﬁcation 4550 . Source: Ref 9

Fig. 14

improve the resistance to pitting and crevice corrosion of many stainless steels (Ref 11). The standard practices of ASTM A 262 (Ref 12) are commonly implemented to predict and measure the susceptibility of austenitic and duplex stainless steels to intergranular corrosion. Table 3 indicates some representative results for CF-type alloys as tested according to practices A, B, and C of Ref 10 as well as two electrochemical tests described in Ref 13 and 14. Table 4 lists the compositions of the alloys investigated. The data indicate the superior resistance of the low-carbon alloys to intergranular corrosion. It also indicates that for highly oxidizing environments (represented here by A 262C-boiling HNO3), the CF-3 and CF-3M alloys are equivalent in the solutiontreated condition, but that subsequent heat treatment causes the corrosion resistance of the CF3M alloys to deteriorate rapidly for service in oxidizing environments (Ref 16). In addition, the degree of chromium depletion necessary to cause susceptibility to intergranular corrosion appears to increase in the presence of molybdenum (Ref 17). The passive ﬁlm stability imparted by molybdenum may offset the loss

of solid-solution chromium for mild degrees of sensitization. Intergranular Corrosion of Ferritic and Martensitic Alloys. Ferritic alloys may also be sensitized by the formation of extensive chromium carbide networks, but because of the high bulk chromium content and rapid diffusion rates of chromium in ferrite, the formation of carbides can be tolerated if the alloy has been slowly cooled from a solutionizing temperature of 780 to 900 C (1435 to 1650 F). The slow cooling allows replenishment of the chromium adjacent to carbides. Martensitic alloys normally do not contain sufﬁcient bulk chromium to be used in applications in which intergranular corrosion is likely to be of concern. Typical chromium contents for martensitic alloys may be as low as 11 to 12%. Pitting and Crevice Corrosion. Austenitic and martensitic alloys display a tendency toward localized corrosion in some environments. The conditions conducive to this behavior may be any situation in areas where ﬂow is restricted and an oxygen concentration cell may be established. Duplex alloys have been found to be less susceptible. Localized corrosion is

182 / Casting Design and Performance Table 3 Intergranular corrosion test results for Alloy Casting Institute casting alloys Alloy(b)/Test results(c) Metallurgical condition

Solution treated

Simulated weld repair

Solution treated, held 1 h at 650 C (1200 F) As-cast

Test(a)

A 262A A 262B A 262C EPR JEPR A 262A A 262B A 262C EPR JEPR A 262A A 262B A 262C EPR JEPR A 262A A 262B A 262C EPR JEPR

CF8 (4)

CF8 (11)

CF8 (20)

CF8M (5)

CF-8M (11)

CF-8M (20)

P P P P P X X X X X X X X X X X X X X X

P P P P P X X X X X X X X X X X X X X X

P P P P P X X X X X X X X X X X X X X X

P P P P P X X X P P X X X X P X X X X X

P P P P P X X X P P X X X X X X X X X X

P P P P P X X X P P X X X X X X X X X X

CF-3 (2)

CF-3 (5)

CF-3 (8)

P P P P P P P P P P* P* P* P P P P P P P P P P P P P* P* P* P P P X X X P P P P P P X/P* X/P* X/P* P P P X X X P P P P** P** P** X/P* X/P* X/P* X/P P P

CF3M (5)

CF3M (9)

CF-3M (16)

P P P P P P P P P P X P X X/P P X P X X/P P

P P P P P P P P P P X P X P P X X X X/P P

P P P P P P P P P P X P X P P X P X P P

(a) See Ref 12 for details of ASTM A 262 practices. EPR, electrochemical potentiokinetic reactivation test: see Ref 12 for details. JEPR, Japanese electrochemical potentiokinetic reactivation test: see Ref 13 for details. (b) Parenthetical value is the percentage of ferrite. See Table 4 for alloy compositions. (c) P, pass; X, fail based on the following criteria: A 262A ditching <10% = pass; A 262B, penetration rate <0.64 mm/yr (25 mils/yr) = pass; A 262C, penetration rate <0.46 mm/yr (18 mils/yr) and not increasing = pass; EPR, peak current density <100 mA/cm2 (645 mA/in.2) = pass; JEPR, ratio <1% = pass; P*, pass, but matrix pitting complicates test results. X/P, near pass. X/P*, likely pass; small EPR indication complicated by matrix pitting p**, pass; actual heat treatment 4 h at 650 C (1200 F) after solution treatment rather than as-cast. Source: Ref 9, 15, 16

Table 4 Composition of alloys tested in Table 3 Composition, % Material

CF-8 LO CF-8 INT CF-8 HI CF-8M LO CF-8M INT CF-8M HI CF-3 LO CF-3 INT CF-3 HI CF-3M LO CF-3M INT CF-3M HI

Ferrite number(a)

C

Mn

Si

P

S

Cr

Ni

Mo

N

4 11 20 5 11 20 2 5 8 5 9 16

0.058 0.086 0.066 0.063 0.083 0.071 0.016 0.023 0.015 0.027 0.027 0.022

0.60 0.84 0.79 0.94 1.20 1.19 0.98 0.68 0.67 0.96 1.04 0.94

1.52 1.10 1.25 1.21 1.20 1.16 1.12 1.24 1.09 0.85 1.02 1.14

0.012 0.031 0.031 0.011 0.030 0.030 0.010 0.011 0.013 0.011 0.009 0.012

0.013 0.012 0.011 0.014 0.013 0.011 0.008 0.009 0.006 0.010 0.009 0.007

18.53 19.90 20.81 18.26 19.78 19.92 17.36 19.35 19.82 17.55 18.78 19.85

9.98 8.73 8.85 11.17 9.53 9.40 10.10 10.27 8.73 12.00 10.79 10.08

0.02 0.50 0.45 2.28 2.21 1.95 0.10 0.10 0.10 2.18 2.12 2.26

0.02 0.02 0.02 0.02 0.02 0.02 0.04 0.06 0.04 0.04 0.03 0.02

(a) This value is the percentage of ferrite.

Table 5 Critical crevice temperatures (CCTs) for several common cast and wrought alloys CCT Alloy

Wrought AISI type 317L Cast CF-3M Cast CN-7M Cast CF-8M Wrought AISI type 316L Wrought AISI type 316

Structure

Austenitic 90% austenite, 10% ferrite Austenitic 90% austenite, 10% ferrite Austenitic Austenitic

C

F

Ref

2 2

35 35

15 16

1.1 2.5

30 28

16 17

2.5 3

28 27

18 15

Note: See text and Ref 18 for information on CCTs.

particularly acute in environments containing chloride ion (Cl) and in acidic solutions. Increasing the alloy content improves resistance to localized corrosion, as indicated by

the PREN increase. Molybdenum has long been recognized as effective in reducing localized corrosion, although it is not a total solution. Excellent results have been obtained with CG-8M, but the CF-3M or CN-7M alloys are readily attacked. Nitrogen is also effective at retarding localized corrosion. A technique for comparing resistance to localized corrosion is to ascertain the critical crevice temperature (CCT). This involves determining the maximum temperature at which no crevice attack occurs during a 24 h testing period in standardized (or service-speciﬁc) environments. Tests have been conducted on a number of cast stainless alloys; the results are given in Table 5. Although the CCT has been shown to correlate well with tests in aerated seawater (Ref 19), it must not be used as the maximum operating temperature in seawater or other chloride-containing media, because

it is simply a comparative tool regarding relative resistance. The ferric chloride (FeCl3) test environment is a very severe, highly oxidizing environment containing approximately 39,000 ppm Cl at a pH of approximately 1.4. Therefore, the FeCl3 CCT is lower than that normally found in aerated seawater (Ref 21), which contains approximately 20,000 ppm Cl with a pH of approximately 7.5 to 8.0. Corrosion fatigue is one of the most destructive and unpredictable corrosion-related failure mechanisms. The behavior is highly speciﬁc to the environment and alloy, and the extent of the interaction between corrosion and mechanical damage is not easy to quantify. The martensitic materials are degraded the most in both absolute and relative terms. For example, if left to corrode freely in seawater, they have very little resistance to corrosion fatigue. This is remarkable in view of their very high strength and fatigue resistance in air. Cathodic protection is a method of reducing corrosion; however, because martensitic stainless steels are susceptible to hydrogen embrittlement, cathodic protection must be carefully applied. Too large a protective potential will lead to catastrophic hydrogen stress cracking. Austenitic materials are also severely degraded in corrosion fatigue strength under conditions conducive to pitting, such as in seawater. However, they are easily cathodically protected without fear of hydrogen embrittlement and perform well in freshwaters. The corrosion fatigue behavior of duplex alloys in chloride environments is less than that obtained for austenitic stainless steels (Ref 21). Stress-corrosion cracking of cast stainless steels has been investigated for only a limited number of environments, heat treatments, and test conditions. From the limited information available, the following generalizations apply. As the composition is adjusted to provide increasingly greater amounts of ferrite in an austenitic matrix, SCC resistance seems to improve. This trend continues to a certain level, apparently near 50% ferrite (Fig. 15, 16). Lower nickel contents tend to improve SCC resistance in cast duplex alloys, possibly because of its effect on ferrite content (Ref 22). The mere presence of the ferrite phase, which is generally much more resistant to SCC than austenite, forces the crack to expend more energy traveling around rather than through ferrite. This slows propagation signiﬁcantly, discouraging SCC. At low and medium stress levels, the ferrite tends to block the propagation of stress-corrosion cracks. This may be due to a change in composition and/or crystal structure across the austenite/ferrite boundary (Fig. 16). As the stress level increases, crack propagation may change from austenite/ferrite boundaries to transgranular propagation (Ref 23, 24). Finally, reducing the carbon content of cast stainless alloys—thus reducing the susceptibility to sensitization—improves SCC resistance. This is also true for wrought alloys (Ref 23, 25–27).

Corrosion of Cast Stainless Steels / 183

17.

18.

19.

Fig. 16 Stress required to produce stress-corrosion cracking in several ACI alloys with varying amounts of ferrite

Ferrite pools blocking the propagation of stresscorrosion cracks in a cast stainless steel

20.

Fig. 15

ACKNOWLEDGMENT This article was adapted from Corrosion of Cast Steels by Raymond W. Monroe and Steven J. Pawel in Corrosion, Volume 13, ASM Handbook, 1987, p 573–582. REFERENCES 1. M. Prager, Cast High Alloy Metallurgy, Steel Casting Metallurgy, J. Svoboda, Ed., Steel Founders’ Society of America, 1984, p 221–245 2. C.E. Bates and L.T. Tillery, Atlas of Cast Corrosion-Resistant Alloy Microstructures, Steel Founders’ Society of America, 1985 3. “Standard Practice for Steel Casting, Austenitic Alloy, Estimating Ferrite Content Thereof,” A 800, ASTM International 4. C. Lundin et al., Corrosion, Toughness, Weldability and Metallurgical Evaluation of Cast Duplex Stainless Steels, Proceedings of Duplex America 2000 Conference on Duplex Stainless Steels, Stainless Steel World, p 449–460 5. G.F. Vander, Voort, G.M. Lucas, and E.P. Manilova, Metallography and Microstructures of Stainless Steels and Maraging Steels, Metallography and Microstructures, Vol. 9, ASM Handbook, 2004, p 670–700 6. M. Danielewski, Introduction to Fundamentals of Corrosion in Gases, Corrosion: Fundamentals, Testing, and Protection, Vol 13A, ASM Handbook, 2003, p 87–89 7. D.B. Roach, “Carburization of Cast Heat Resistant Alloys,” ACI progress report, Project A-80, Alloy Casting Institute 8. T.M. Devine, Mechanism of Intergranular Corrosion and Pitting Corrosion of

9.

10. 11.

12.

13.

14.

15.

16.

Austenitic and Duplex 308 Stainless Steel, J. Electrochem. Soc., Vol 126 (No. 3), 1979, p 374 E.E. Stansbury, C.D. Lundin, and S.J. Pawel, Sensitization Behavior of Cast Stainless Steels Subjected to Simulated Weld Repair, Proceedings of the 38th SFSA Technical and Operating Conference, Steel Founders’ Society of America, 1983, p 223 S.J. Pawel, Literature Review on the Role of Nitrogen in Austenitic Steels, Steel Founders’ Res. J., Issue 5, 1st Quarter, 1984 S.J. Pawel, E.E. Stansbury, and C.D. Lundin, Role of Nitrogen in the Pitting of Cast Duplex CF-Type Stainless Steels, Corrosion, Vol 45 (No. 2), 1989, p 125–133 “Standard Practices for Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels,” A 262, Annual Book of ASTM Standards, American Society for Testing and Materials W.L. Clarke, R.L. Cowan, and W.L. Walker, Comparative Methods for Measuring Degree of Sensitization in Stainless Steel, Intergranular Corrosion of Stainless Alloys, STP 656, R.F. Steigerwald, Ed., American Society for Testing and Materials, 1978, p 99 M. Akashi et al., Evaluation of IGSCC Susceptibility of Austenitic Stainless Steels Using Electrochemical Methods, Boshoku Gijutsu (Corros. Eng.), Vol 29, 1980, p 163 (BTSITS trans.) S.J. Pawel, “The Sensitization Behavior of Cast Stainless Steels Subjected to Weld Repair,” MS thesis, University of Tennessee, June 1983 S.J. Pawel, E.E. Stansbury, and C.D. Lundin, Evaluation of Post Weld Repair Requirements for CF3 and CF3M Alloys—Exposure to Boiling Nitric Acid, First International Steel Foundary

21.

22.

23.

24.

25.

26.

27.

Congress Proceedings, Steel Founders’ Society of America, 1985, p 45 J.R. Maurer and J.R. Kearns, “Enhancing the Properties of a 6% Molybdenum Austenitic Alloy with Nitrogen,” Paper 172, presented at Corrosion/85, National Association of Corrosion Engineers, 1985 J.A. Larson, 1984 SCRATA Exchange Lecture: New Developments in High Alloy Cast Steels, Proceedings of the 39th SFSA T & O Conference, Steel Founders’ Society of America, 1984, p 229–239 A. Poznansky and P.J. Grobner, “Highly Alloyed Duplex Stainless Steels,” Paper 8410–026, presented at the International Conference on New Developments in Stainless Steel Technology, (Detroit, MI), American Society for Metals, Sept 1984 A.P. Bond and H.J. Dundas, “Resistance of Stainless Steels to Crevice Corrosion in Seawater,” Paper 26, presented at Corrosion/84, National Association of Corrosion Engineers, 1984 A. Garner, Crevice Corrosion of Stainless Steels in Seawater: Correlation of Field Data with Laboratory Ferric Chloride Tests, Corrosion, Vol 37 (No. 3), March 1981, p 178–184 L. Coudreuse and J. Charles, Fatigue and Corrosion—Fatigue Behavior of Duplex Stainless Steels, Sixth World Duplex Conference, 17–20 Oct 2000 (Venice, Italy), Italian Metallurgical Association, p 629–630 S. Shimodaira et al., Mechanisms of Transgranular Stress Corrosion Cracking of Duplex and Ferrite Stainless Steels, Stress Corrosion Cracking and Hydrogen Embrittlement in Iron Base Alloys, NACE Reference Book 5, National Association of Corrosion Engineers, 1977 P.L. Andresen and D.J. Duquette, The Effect of Cl Concentration and Applied Potential on the SCC Behavior of Type 304 Stainless Steel in Deaerated High Temperature Water, Corrosion, Vol 36 (No. 2), 1980, p 85–93 J.N. Kass et al., Stress Corrosion Cracking of Welded Type 304 and 304L Stainless Steel under Cyclic Loading, Corrosion, Vol 36 (No. 6), 1980, p 299–305 J.N. Kass et al., Comparative Stress Corrosion Behavior of Welded Austenitic Stainless Steel Pipe in High Temperature High Purity Oxygenated Water, Corrosion, Vol 36 (No. 12), 1980, p 686–698 G. Cragnolino et al., Stress Corrosion Cracking of Sensitized Type 304 Stainless Steel in Sulfate and Chloride Solution at 250 and 100C, Corrosion, Vol 37 (No. 6), 1981, p 312–319

Casting Design and Performance Pages 185–198

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Fatigue and Fracture Properties of Cast Irons* CAST IRONS have several well-known manufacturing and engineering advantages over cast steel, including 30 to 40% lower manufacturing costs relative to cast steel and more desirable performance characteristics, such as better wear resistance and vibration damping. These advantages over cast steel are based on several metallurgical characteristics of cast irons. First, cast iron has melting temperatures (and therefore pouring temperatures) that are 300 to 350 C (540 to 630 F) lower than those of cast steel. Second, because of a larger concentration of free carbon and higher silicon, graphitic (gray and ductile) cast iron has the greatest ﬂuidity and the least shrinkage of any ferrous metal. For example, cast steel will generally experience a volume shrinkage of more than 4% during solidiﬁcation, whereas gray and ductile cast iron shrinkage can be less than 1%, depending on the composition and the processing conditions. This difference in shrinkage allows products to be made to exact dimensions much more easily using gray and ductile cast iron, with very little problem in obtaining pressure tightness as a result of reduced interdendritic shrinkage. In addition, cast irons are more machinable than cast steels, and they are relatively more wear resistant because the graphite acts as a self-lubricating system. Graphite also attenuates sound and mechanical vibration, which makes cast iron ideal for many mechanical applications, such as brakes. However, the higher carbon concentration responsible for several desirable physical properties and ease of manufacturing with cast iron is, unfortunately, also responsible for the degradation of the ductility and fracture toughness. The carbon, usually present principally as graphite, serves to nucleate fatigue and fracture processes at relatively low strain levels, signiﬁcantly impairing the fatigue and fracture resistance of cast iron compared with that of cast or wrought steel. Considerable work has been done to characterize better the fatigue and fracture of cast iron as well as to understand better the micromechanisms that control the fracture process in these systems. The purpose of this chapter is

to summarize the fatigue and fracture behavior of the various types of cast iron as a function of chemical composition, matrix microstructure, and graphite morphology. The ﬁve types of commercial cast iron considered in this chapter are gray, ductile, malleable, compacted graphite, and white iron. With the exception of white cast iron, all cast irons have in common a microstructure that consists of a graphite phase (usually 8 to 14 vol%) in a matrix that may be ferritic, pearlitic, bainitic, tempered martensitic, or combinations thereof. The four types of graphitic cast irons are roughly classiﬁed according to the morphology of the graphite phase. Gray iron has ﬂake-shaped graphite, ductile iron has nodular or spherically shaped graphite, compacted graphite iron (or “vermicular” graphite iron) is intermediate between these two (with graphite in a vermicular or “wormlike” coral shape), and malleable iron has irregularly shaped globular or “popcorn”-shaped graphite that is formed during tempering of white cast iron. These categories of graphite shape have a strong inﬂuence on fatigue and fracture properties, although microstructure (tensile strength), component/specimen size, surface condition, chemistry, and temperature also affect the general fatigue and fracture performance of cast irons (Ref 1–5).

Fatigue of Cast Irons The fatigue performance of cast irons, in general, is inﬂuenced by graphite morphology, matrix microstructure and tensile strength, specimen size, surface condition, surface degradation such as corrosion, and the type of loading stress (e.g., axial, bending, reversed bending, torsion, multiaxial, variable amplitude). The effect of graphite shape on cast iron fatigue has received considerable attention and is a key variable. The free graphite in cast iron acts as an inherent notch that increases stress concentrations for fatigue crack initiation. Therefore, the fatigue performance of cast irons is inﬂuenced greatly by the quantity, size, and

*Reprinted from ASM Handbook, Volume 19, Fatigue and Fracture, ASM Handbook Committee, p 665–679

shape of the graphite phase as well as its interaction with the matrix. The effect of different graphite morphologies on fatigue is illustrated in Fig. 1 to 3 for three types of cast iron. The ﬂake form of graphite in gray iron increases stress concentrations and thus lowers fatigue strength relative to that of irons that have more spheroidal or compacted graphite forms. In contrast, ductile irons have a nodular graphite, which reduces the concentration of internal stresses and improves the relatively fatigue performance at a given tensile strength. However, gray irons are also less notch sensitive than ductile irons (Table 1). Most of the available data on cast iron fatigue are based on rotating bending testing where the specimen diameter is generally less than 13 mm (0.5 in.). Results are often reported in terms of fatigue endurance limits (Fig. 1), although testing in the ﬁnite life range (below 106 cycles) is also being emphasized more (Fig. 2, 3). The fatigue endurance limit (or the “knee” in the fatigue curve) usually occurs between 106 and 107 cycles. For many tests, if a specimen endures 107 cycles of a given stress, then it is often assumed that the material would withstand such a stress indeﬁnitely. However, fatigue failures do occur above 107 cycles in rotating bending fatigue test. Thus, some fatigue limits are based on 2 107 cycles (Ref 1). Typical cast iron fatigue data generated in the 1950s and 1960s are summarized in Fig. 4 and Tables 2 and 3. The general relation between tensile strength and fatigue limit (Fig. 4) is determined to a large extent by the strength/microstructure of the matrix, as discussed in more detail later in this article for each particular type of cast iron. Endurance ratios may also be affected by alloying and graphite form, also discussed for particular types of cast iron. Following are more summarized general fatigure characteristics. Fatigue Crack Growth. Fatigue crack initiation is the limiting fatigue mechanism when fatigue life approaches about 106 cycles or the fatigue-limit regime. Below about 106 cycles, the majority of fatigue life tends to be consumed by fatigue crack growth for gray, ductile, and compacted graphite irons (Ref 2).

186 / Casting Design and Performance

Comparison of typical strain-life fatigue curves for nodular (ductile), gray, and compacted graphite cast irons. Source: Ref 2

Fig. 3

Ratio of alternating bend fatigue strength/tensile strength of gray cast iron, compacted graphite iron, and spheroidal graphite (SG) iron (i.e., ductile iron). Statistical analysis of a number of experimental data allowed the calculation of a relationship between fatigue strength (FS) at or above 106 cycles and tensile strength (TS) of compacted graphite irons, which is: FS(in MPa) = (0.63 0.00041 TS), where TS is also in MPa. Values calculated with this equation ﬁt well between those of gray and ductile irons, as shown. Source: Ref 6

Fig. 1

Comparison of typical stress-life fatigue curves for nodular (ductile), gray, and compacted graphite cast irons. Source: Ref 2

Fig. 2

For practical purposes, crack initiation in the low-cycle regime (102 to about 106 cycles) can be assumed to occur on the ﬁrst cycle of loading. Thus, life prediction models for lowcycle fatigue of cast irons should be based on crack growth behavior (Fig. 5 and 6). The Paris relation, da/dN - C(DK)n, is valid for cast iron in the regime of intermediate stress intensities. Fatigue crack growth studies on permanent mold ductile iron castings of a hypereutectic composition (3.6C-3.0Si) indicate that it may be possible to improve the fatigue crack growth resistance by lowering the pearlite content and increasing the nodule count. Crack

growth parameters from this study are summarized in Table 4. Results of various investigations on fatigue crack growth rate are shown as a function of the fracture toughness (KIc) in Fig. 7. The crack growth exponent values in permanent mold ductile iron castings are significantly higher than in others. Further, slight reductions in fracture toughness lead to large increases in exponent values. Fracture Appearance. In ﬂake graphite cast irons, it is not generally possible to determine from fracture appearance whether failure has been due to fatigue or to a single overload in tension, because no distinctive fatigue fracture is obtained. This is particularly so with relatively weak ﬂake graphite cast irons. In nodular irons, however, fatigue failures frequently have a distinctive appearance different from that associated with failure by other means. Load Variables. Axial loading or torsional loading cycles are frequently encountered in designing parts of cast iron, and in many instances these are not completely reversed loads. Types of regularly repeated stress variation can usually be expressed as a function of a mean stress and a stress range. Wherever possible, the designer should use actual data from the limited information available. Without precisely applicable test data, an estimate of the reversed bending fatigue limit of machined parts may be made by using about 35% of the minimum speciﬁed tensile strength of the particular grade of gray iron being considered.

This value is probably more conservative than an average of the few data available on the fatigue limit for gray iron. An approximation of the effect of range of stress on fatigue limit may be obtained from diagrams such as Fig. 8. Tensile strength is plotted on the horizontal axis to represent fracture strength under static load (which corresponds to a 0 stress range). Reversed bending fatigue limit is plotted on the ordinate for 0 mean stress, and the two points are joined by a straight line. The resulting diagram yields a fatigue limit (maximum value of alternating stress) for any value of mean stress. Few data available are applicable to design problems involving dynamic loading where the stress cycle is predominantly compressive rather than tensile. Other work done on aluminum and steel indicates that for compressive (negative) mean stress, the behavior of these materials could be represented by a horizontal line beginning at the fatigue limit in reversed bending, as indicated in Fig. 8. Gray iron is probably at least as strong as this for loading cycles resulting in negative mean stress, because it is much stronger in static compression than in static tension. It is therefore a natural assumption that the parallel behavior shown in Fig. 8 is conservative. If, prior to design, the real stress cycle can be predicted with conﬁdence and enough data are available for a reliable S-N diagram, the casting might be dimensioned to obtain a minimum safety factor of 2 based on fatigue strength. (Some uses may require more conservative or more liberal loading.) The approximate safety factor is best illustrated by point P in Fig. 8. The safety factor is determined by the distance from the origin to the fatigue limit line along a ray through the cyclic-stress point, divided by the distance from the origin to that point. In Fig. 8 this is OF/OP. On this diagram, point P0 represents a stress cycle having a negative mean stress. In other words, the maximum compressive stress is greater than the tensile stress reached during the loading cycle. In this instance, the safety factor is the distance OF0 /OP0 . However, this analysis assumes that overloads will increase the mean stress and alternating stress in the

Fatigue and Fracture Properties of Cast Irons / 187 Table 1 Typical fatigue endurance of cast irons Fatigue endurance strength(b) Class or type

Gray cast irons Medium section (type 20) Medium section (type 25) Medium section (type 30) Medium section (type 35) Medium section (type 40) Medium section (type 45) Medium section (type 50) Medium section (type 60) Medium section (type 70) Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI

30) 40) 50) 60) 70) 80)

Malleable cast irons Ferritic malleable (32510) Ferritic malleable (35018) Pearlitic malleable (43010) Pearlitic malleable (45010) Pearlitic malleable (45007) Pearlitic malleable (48004) Pearlitic malleable (48005) Pearlitic malleable (50007) Pearlitic malleable (53004) Pearlitic malleable (60003) Pearlitic malleable (70002) Martensitic malleable (80002) White and alloy cast irons Unalloyed white High Si (Duriron, Durichlor) High Cr (corrosion resistant) Medium Si (Silal) Ni-Cr-Si High Al Ni-resist (type 1) Ni-resist (type 2) Ni-resist (type 2b) Ni-resist (type 3) Ni-resist (type 4) Ni-resist (type 5) Ni-hard (type 1 sand cast) Ni-hard (type 1 chill cast) Ni-hard (type 2 sand cast) Ni-hard (type 2 chill cast) Almanite (Meehanite type WS) Ductile irons Standard ductile 60–40–18 Standard ductile 80–55–06 Standard ductile 100–70–03 Standard ductile 120–90–02 Ni-Resist ductile (austenitic) Ni-Resist ductile (austenitic) Ni-Resist ductile (austenitic) Ni-Resist ductile (austenitic) Ni-Resist ductile (austenitic) Ni-Resist ductile (austenitic) Ni-Resist ductile (austenitic) Ni-Resist ductile (austenitic)

D-2 D-2B D-2C D-3 D-3A D-4 D-5 D-5B

Ultimate tensile strength(a), ksi

Hardness, HB

Unnotched, ksi

20–25 25–30 30–35 35–40 40–48 45–52 50–57 60–66 70 30 min 40 50 60 70 80

140–187 170–229 170–241 187–269 207–269 217–269 215–260 230–290 200 230 260 290 320 350

9.5–10 12–15 13.7–15.5 16–17.5 17.5–19.5 21.5–25.5 24.5–27.5 24.5–29.5 29.5–31.5 15 19 23 27 31 35

50–58 53–60

110–156 110–156

28 31

65 65 70

163–207 163–217 163–228

75

37

80

179–228 197–241 197–255

100

241–269

40

20–50 13–18 30–90 25–45 20–45 34–90 25–30 25–30 30–35 25–35 25–30 20–25 40–50 50–60 45–55 60–70 60–80

300–575 480–520 290–400 170–250 80–150 180–350 130–170 125–170 170–250 120–160 150–210 100–125 550 (s) 600 (s) 525 (s) 575 (s) 400–525

60–80 80–100 100–120 120–175 55–69 58–70 55–65 55–67 55–65 60–72 55–60 55–65

149–187 179–248 217–269 240–300 140–200 150–210 130–170 140–200 130–190 170–240 130–180 140–190

Notched, ksi

9.5 13.5 17.5 21.5 25.5 12 14 16 18 20 22

30

39

12 12 18 13.5 9 9.9

30.5 39–40

18–23 21–24

49 30

30 20

(a) 0.750 in. diameter specimens machined from 1.2 in. diameter test bar. (b) Rotating cantilever beam tests; strength speciﬁed is between 106 and 107 cycles. Source: Ref 4

same proportion. This may not always be true, particularly in systems with mechanical vibration in which the mean stress may remain constant. For this condition, the vertical line through P would be used; that is, DK/DP would be the factor of safety. Most engineers use diagrams such as Fig. 8 mainly to determine whether a given condition of mean stress and cyclic stress results in a design safe for inﬁnite life. The designer can

also determine whether variations in the anticipated mean stress and alternating stress will place the design in the unsafe zone. Usually the data required to analyze a particular set of conditions are obtained experimentally (e.g. Table 5). It is emphasized that the number of cycles of alternating stress implied in Fig. 8 is the number normally used to determine fatigue limits, that is, approximately 10 million. Fewer cycles, as encountered in infrequent overloads,

will be safer than indicated by a particular point plotted on a diagram for inﬁnite life. Too few data are available to draw a diagram for less than inﬁnite life. Fatigue Notch Sensitivity. Gray irons are less notch sensitive than ductile irons. Because the graphite ﬂakes in gray iron can be thought of as internal notches, gray irons have little or no sensitivity to the presence of additional notches resulting from design features. The strength-reducing effect of the internal notches is included in the fatigue limit values determined by conventional laboratory tests with smooth bars. High-strength irons usually exhibit greater notch sensitivity, but probably not the full theoretical value represented by the stress concentration factor. Normal stress concentration factors (see Ref 13) are probably suitable for highstrength gray irons. Relative notch factors (the ratio of unnotched to notched fatigue strength) for cast irons are in the range of about (Ref 6): 1.7 to 1.8 for compacted graphite irons Less than 1.5 for gray iron Greater than 1.85 for ductile iron

Thermal Fatigue. When castings are used in an environment where frequent changes in temperature occur, or where temperature differences are imposed on a part, thermal stresses occur in castings and may result in elastic and plastic strains and ﬁnally in crack formation. Changes in microstructure, associated with stress-inducing volume changes, as well as surface and internal oxidation, may also be associated with stresses induced by temperature differences. The interpretation of thermal fatigue tests is complicated by the many different test methods employed by various investigators. The two widely accepted methods are constrained thermal fatigue and ﬁnned-disk thermal shock tests (Ref 14, 15). In general, for good resistance to thermal fatigue, cast irons should have high thermal conductivity, low modulus of elasticity, high strength at room and elevated temperatures, and, for use above 500 to 550 C (930 to 1020 F), resistance to oxidation and structural change. The relative ranking of irons varies with test conditions. When high cooling rates are encountered, experimental data and commercial experience show that thermal conductivity and a low modulus of elasticity are most important. Consequently, gray irons of high carbon content (3.6 to 4%) are superior (Ref 14, 15). When intermediate cooling rates exist, ferritic ductile and compacted graphite irons have the highest resistance to cracking but are subject to distortion. When low cooling rates exist, high-strength pearlitic ductile irons or ductile irons alloyed with silicon and molybdenum are best with regard to cracking and distortion (Fig. 9). A rather detailed analysis of the behavior of various irons at elevated temperatures is given in Ref 6. Extensive experimental work on cylinder heads is reviewed. A critical analysis of

188 / Casting Design and Performance was effective for ferritic irons. Soluble oils are not effective in preventing corrosion fatigue of nodular graphite irons.” “All inhibitors prevented visible corrosion of the surface, but this is no indication of the effectiveness of an inhibitor in preventing corrosion fatigue.” See Ref 1 for more details.

Fracture Toughness of Cast Irons Like the fracture resistance of structural steels and other alloys, the measurement of the fracture resistance of cast irons can be broadly divided into two groups: Measurement of fracture energy curves

Fig. 4

Relation between rotating-bending fatigue strength and tensile strength of iron, based on fatigue limit or failure in 107 cycles. Source: Ref 7, 8

most of the accepted criteria for assessing the quality of irons for castings used at elevated temperatures is also included. Constrained Thermal Test. In the constrained thermal fatigue test (Ref 15), a specimen is mounted between two stationary plates that are held rigid by two columns, heated by high-frequency (450 kHz) induction current, and cooled by conduction of heat to watercooled grips. The thermal stress that develops in the test specimen is monitored by a load cell installed in one of the grips holding the specimen. As thermal cycling continues, the specimen accumulates fatigue damage in a fashion similar to that in mechanical fatigue testing; ultimately, the specimen fails by fatigue. Experimental results (Fig. 10) point to higher thermal fatigue for compacted graphite iron than for gray iron and also indicate the beneﬁcial effect of molybdenum. In fact, regression analysis of experimental results indicates that the main factors inﬂuencing thermal fatigue are tensile strength (TS) and molybdenum content: log N ¼ 0:934 þ 0:026 TS þ 0:861 Mo where N is the number of thermal cycles to failure, tensile strength is in kips (1000 lbs) per square inch (ksi), and molybdenum is in percent. Finned-Disk Test. In the ﬁnned-disk thermal shock test (Ref 15), the specimen is cycled between a moderate-temperature environment and a high-temperature environment, which causes thermal expansion and contraction. Because thermal conductivity plays a signiﬁcant role in this type of test, gray iron shows much greater resistance to cracking than compacted graphite iron. Major cracking occurred in less than 200 cycles in all compacted graphite iron specimens, while the unalloyed gray iron developed minor cracking after 500 cycles

and major cracking after 775 cycles (Ref 15). Because of its higher elevated-temperature strength, alloyed gray iron was even more resistant to thermal fatigue. Corrosion Fatigue of Cast Irons. A limited amount of information is available in the literature on the corrosion fatigue of cast irons. From a literature review of a few tests in tap and salt water, the fatigue reduction factors due to corrosion were from 1.28 to 1.59 in tap water and 1.55 to 2.47 in sea water for gray irons (Ref 1). Fatigue strengths were reduced by about 7 MPa (1 ksi) when tested in tap water. Additional corrosion fatigue tests carried out by the British Cast Iron Research Association (Ref 1) were conducted in various environments, as summarized in Tables 6 and 7. The fatigue strength in air, in demineralized water, and in inhibitor-containing water was based on a life of 50 106 cycles of stress, and the tests in sodium chloride solution were based on 100 106 cycles of stress. The stress frequency was approximately 3,000 cycles/min. The following conclusions from Ref 1 are noted: “Corrosion fatigue of ﬂake graphite (gray)

irons in demineralized water can be prevented by the addition to the water of known alkaline inhibitors except borax which is not completely effective for pearlitic irons. Soluble oils are also effective for ﬂake graphite irons.” “Tests in corrosion-inhibited demineralized water on both pearlitic and ferritic nodular (ductile) irons revealed that the only satisfactory inhibitor giving more or less complete protection against corrosion fatigue was potassium chromate. Some protection was obtained with sodium carbonate for pearlitic irons and this, together with trisodium orthophosphate and sodium nitrite,

based on dynamic tear tests (ASTM E60483) or Charpy impact tests Measurement of resistance to crack propagation using a fracture mechanics method, such as linear elastic fracture mechanics (ASTM E 399), elastic-plastic fracture mechanics with the J integral (ASTM E 813), crack-tip opening displacement (BS 5762) measurements, or dynamic fracture toughness parameters such as KId or JId Fracture energy curves and dynamic fracture toughness (KId or JId) both can be used to evaluate transition temperature behavior, where fracture behavior changes from ductile to brittle. However, comparison of transition temperatures measured in the different dynamic tests (e.g., Charpy impact, dynamic tear energy tests, or KId or JId tests) is inappropriate. Cast iron yield strengths are rate dependent, and thus transition temperatures are affected when loading rates change from slow bending to impact. Therefore, valid comparison can only be made when loading rate, specimen size, and notch acuity are similar. These different types of fracture toughness tests are discussed in Ref 16. A comparison of the different test methods is also reviewed for cast iron toughness in Ref 5. In general, test bars for the Charpy or Izod impact testing of cast irons are larger than those used for testing steels. Increases in the size of as-cast specimens cause a relative reduction in impact strength and tensile strength. In metallurgical terms, the general effects of graphite morphology and matrix microstructure are shown in Fig. 11 and 12 for cast irons. Like fatigue fracture, gray iron has the lowest fracture resistance, due to the inherent notch effect from the ﬂake morphology of its graphite. Gray iron is also less notch sensitive, with little or no change in notch toughness at lower temperatures. Ductile-to-brittle transition temperatures are more pronounced for ductile and compacted graphite irons (Fig. 11, 12). The effect of graphite morphology is more pronounced for upper shelf toughness, but below the transition temperature, graphite morphology has little effect on toughness.

Fatigue and Fracture Properties of Cast Irons / 189 Table 2 Reversed bending (R = 1) fatigue strength of cast irons (circa 1960) Fatigue specimen Material

Gray cast iron 3.32 total C, 0 combined C, 1.16Si. 0.102S, 0.38P, 0.58Mn Cast iron 3.68 total C, 0.55 combined C, 2.17Si, 0.03S, 0.646P, 0.54Mn Gray cast iron 3.35 total C, 0.55 combined C, 1.10Si, 0.095S, 0.51P, 0.63Mn Cast iron 3.51 total C, 0.69 combined C, 1.58Si, 0.088S, 0.72P, 0.54Mn, 0Ni. 0.01Cr, 0.05V.0.06Ti Gray cast iron 3.44 total C, 0.68 combined C, 1.10Si, 0.093S, 0.51P, 0.62Mn

Gray cast iron 3.44 total C, 0.68 combined C, 1.10Si, 0.093S, 0.51P, 0.62Mn Cast iron 3.23 total C, 0.57 combined C, 1.96Si, 0.072S, 0.44P, 0.54Mn Cast iron 3.24 total C, 0.40 combined C, 2.85Si, 0.101S, 0.168P, 0.57Mn Cast iron 3.57 total C, 0.58Mn, 1.58Si, 0.66P, 0.06S Austenitic cast iron 2.63 total C, 0.67 combined C, 2.14Si, 1.23Mn, 6.94Cu. 2.09Cr, 14.9Ni, 0.022Ti, 0.005V, 0.012As, 0.16P, 0.065S

Austenitic cast iron 2.63 total C, 0.67 combined C, 2.14Si, 1.23Mn, 6.94Cu, 2.09Cr. 14.9Ni. 0.022Ti, 0.005V, 0.012As, 0.16P, 0.065S Cast iron 3.30 total C, 0.64 combined C, 1.12Si, 0.09S, 0.21P, 0.62Mn Cast iron 3.35 total C, 0.50 combined C, 2.25Si, 0.084S, 0.26P, 0.60Mn Cast iron 3.46 total C, 0.44 combined C, 2.35Si, 0.078S, 0.27P, 0.62Mn. 0.63Ni, 0.22Cr Cast iron 3.31 total C, 0.61 combined C, 1.45Si, 0.106S, 0.37P, 0.86Mn Cast iron 2.86 total C, 0.65 combined C, 2.51Si, 0.079S, 0.057P, 0.62Mn Cast iron 3.03 total C, 0.58 combined C, 1.93Si, 0.64Mn, 0.023S. 0.076P High-strength cast iron 2.84 total C, 0.74 combined C, 1.52Si, 1.05Mn, 0.37Cu, 0.31Cr, 0.20Ni, 0.07P, 0.124S (gravity), 0.106S (H2S) High-strength cast iron 2.84 total C, 0.74 combined C, 1.52Si, 1.05Mn, 0.37Cu, 0.31Cr, 0.20Ni, 0.07P, 0.124S (gravity), 0.106S (H2S)

Comment

UTS, ksi

Type

Fatigue strength (max stress, ksi) at indicated cycles

Size, in.

Kt

104

4 104

105

4 105

106

4 106

107

108

Large casting

21

Unnotched

0.35d

...

...

16

13

10

8

7.0

7.0

...

Large bar

23

Unnotched

0.33d

...

...

16

14

12

12

12.0

12.0

...

Large casting Large casting Large bar

28 28 30

Unnotched Transverse hole Unnotched

0.35d 0.055d, 0.35D 0.33d

... 3.0 ...

... ... 23

17 ... 19

14 12 17

12 8 15

11 6 15

10.0 6.5 15.0

10.0 6.5 15.0

10.0 6.5 ...

Large Large Large Large

casting casting casting casting

32 32 32 32

Unnotched Unnotched Unnotched V-notch

... ... ... 1.5

... 20 ... ...

17 19 18 13

15 18 15 12

13 17 13 11

11 16 11 10

11.0 16.0 10.0 10.0

11.0 16.0 10.0 10.0

... ... 10.0 ...

Large casting

32

V-notch

3.0

...

14

13

11

10

9.0

9.0

...

Large casting Large casting

25 12

Unnotched Unnotched

0.35d 0.30d 0.35d 60 o, 5/64r, 0.74D, 0.35d 60 o, 5/64r, 0.501D, 0.342d 0.40d 0.40d

... ...

... ...

... 14

14 12

13 10

12 9

11.0 8.0

10.0 7.0

... ...

Large casting

32

Unnotched

0.40d

...

22

18

16

12

12

12.0

12.0

...

Large casting

32

Unnotched

0.33d

...

23

20

18

15

13

13.0

13.0

...

Cast 7/8 in. bar Cast 7/8 in. bar Plate (RT data) Plate (RT data) Plate (400 F data) Plate (400 F data) Plate (700 F data) Plate (700 F data) Plate (900 F data) Plate (900 F data) Plate (1160 F data) Plate (1160 F data)

34 34 35 35 31 31 30 30 26 26 20 20

Unnotched Kommers notch Unnotched Transverse hole Unnotched Transverse hole Unnotched Transverse hole Unnotched Transverse hole Unnotched Transverse hole

0.33d 0.30d 0.032d, 0.30d 0.032d, 0.30d 0.032d, 0.30d 0.032d. 0.30d 0.032d,

... 5+ ... 2.2 ... 2.2 ... 2.2 ... 2.2 ... 2.2

31 ... ... ... 21 17 ... ... ... ... ... ...

27 22 ... ... 18 15 ... ... ... ... ... ...

25 19 ... ... 17 13 21 ... 15 14 14 14

23 17 ... 15 17 12 18 15 13 12 13 13

22 16 18 12 17 11 17 14 12 12 12 12

22.0 16.0 16.0 11.0 17.0 11.0 15.0 13.0 12.0 12.0 12.0 10.0

22.0 16.0 14.0 11.0 17.0 11.0 15.0 13.0 10.0 12.0 12.0 10.0

... ... ... ... ... ... ... ... ... ... ... ...

Large bar

36

Unnotched

0.33d

...

28

24

22

19

19

19.0

19.0

...

Large bar

36

Unnotched

0.33d

...

...

...

24

21

19

19.0

19.0

...

Large bar

37

Unnotched

0.33d

...

...

28

26

23

22

20.0

20.0

...

Large bar

40

Unnotched

0.33d

...

...

25

22

19

19

19.0

19.0

...

Large bar

43

Unnotched

0.33d

...

...

...

28

24

24

24.0

24.0

...

Large bar Large bar Slab (RT data) Slab (RT data)

45 45 48 48

Unnotched Kommers notch Unnotched Transverse hole

0.33d 0.30d 0.032d, 0.32D

... 5+ ... 2.2

... 26 ... ...

30 23 ... 24

27 21 ... 22

24 18 25 19

22 17 23 18

20.0 17.0 20.0 17.0

20.0 17.0 20.0 17.0

... ... ... 17.0

45 46 46 50 50 45 45 33 33 16 16

Transverse Unnotched Transverse Unnotched Transverse Unnotched Transverse Unnotched Transverse Unnotched Transverse

0.032d, 0.30d 0.032d, 0.30 d 0.032d, 0.30d 0.032d, 0.30d 0.032d, 0.30d 0.032d,

2.2 ... 2.2 ... 2.2 ... 2.2 ... 2.2 ... 2.2

23 30 23 ... ... ... 25 25 17 15 ...

20 24 22 30 25 27 20 18 15 10 13

19 21 20 25 22 25 18 17 14 8 10

18 19 19 23 21 23 17 16 14 7 9

17 18 16 22 20 23 17 15 14 7 8

17.0 18.0 16.0 22.0 20.0 23.0 17.0 15.0 14.0 7.0 7.0

17.0 18.0 16.0 22.0 20.0 23.0 17.0 15.0 14.0 7.0 7.0

... ... ... ... ... ... ... ... ... ... ...

Slab Slab Slab Slab Slab Slab Slab Slab Slab Slab Slab

(300 F data) (500 F data) (500 F data) (700 F data) (700 F data) (850 F data) (850 F data) (1000 F data) (1000 F data) (1200 F data) (1200 F data)

hole hole hole hole hole hole

0.32D 0.32D 0.32D 0.32D 0.32D

0.32D 0.32D 0.32D 0.32D 0.32D 0.32D

UTS. ultimate tensile strength; RT, room temperature: d, diameter; D, outer diameter; r, radius. Source: Ref 9

Most alloying elements or impurities affect toughness indirectly through their effect on graphite morphology and matrix microstructure. However, silicon and phosphorous are two exceptions. Both of these impurities have a strong inﬂuence on toughness (Fig. 13). The effect of microstructure and alloying on

toughness is summarized for each speciﬁc type of cast iron in the next sections. Table 8 is a general summary of room-temperature Charpy toughness for various cast irons. Charpy testing is the most widely used method to evaluate the fracture toughness of cast irons.

Gray Cast Iron Fatigue Strength. Notches and surface imperfections have little effect on the fatigue strength of gray cast irons, which is one reason for its use in automobile crankshafts and similar applications. Minimum tensile and fatigue properties of gray

190 / Casting Design and Performance Table 3 Effect of a circumferential V-notch on the bending fatigue strength of cast steel and cast iron Bending fatigue strength, ksi Material

Cast 0.4% C steel Cast alloy steel

Cast alloy steel

Graphitic cast steel Flake-graphite gray cast iron

Nodular cast iron (as cast)

Nodular cast iron (heat treated)

Tensile strength, ksi

75 84 100 113 130 150 130 150 164 170 100 46 44 52 48 50 41 60 80 86 91 78 79 81 85 46 61 73 77 42 46 51

Plain Notched

29.6 33.6 45.8 56.2 54.6 69.2 39 44 46 47 32 22 16 24 20 21 18 23 38 37 38 38 35 37 34 24 29 30 32 22 23 26

23 25 30 31 36 43 26 29 30 32 18 14 16 22 20 21 15 21 22 21 26 20 19 21 16 21 17 17 17 15 18 16

Dimensions of notch Kt

Kt (Nenber)

Angle, deg

Root radius, in.

Minimum diameter, in.

Depth, in.

1.28 1.34 1.54 1.80 1.51 1.61 1.50 1.52 1.53 1.47 1.78 1.57 1.00 1.09 1.00 1.00 1.20 1.09 1.73 1.76 1.46 1.90 1.84 1.76 2.13 1.14 1.70 1.77 1.88 1.47 1.28 1.63

2.2 2.2 2.2 2.2 2.2 2.2 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.5 3.5 3.5 3.5 3.1 3.1 3.1 3.1 3.5 3.5 3.5

60 60 60 60 60 60 55 55 55 55 55 55 55 55 55 55 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45 45

0.015 0.015 0.015 0.015 0.015 0.015 0.0086 0.0086 0.0086 0.0086 0.0086 0.0086 0.0086 0.0086 0.0086 0.0086 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010

0.22 0.22 0.22 0.22 0.22 0.22 0.300 0.300 0.300 0.300 0.300 0.300 0.300 0.300 0.300 0.300 0.331 0.331 0.301 0.301 0.301 0.417 0.417 0.417 0.417 0.301 0.301 0.301 0.301 0.417 0.417 0.417

0.035 0.035 0.035 0.035 0.035 0.035 0.0375 0.0375 0.0375 0.0375 0.0375 0.0375 0.0375 0.0375 0.0375 0.0375 0.071 0.071 0.086 0.086 0.086 0.142 0.142 0.142 0.142 0.086 0.086 0.086 0.086 0.142 0.142 0.142

Fig. 5

Crack development in ductile iron. Source: Ref 2

Fig. 6

Crack development in gray iron. Source: Ref 2

Source: Ref 10

Table 4 Toughness and fatigue crack growth data for various ductile irons Austenitized at(b): Property

Graphite shape(c) Nodule count, mm2 Pearlite, % Ferrite grain pﬃﬃﬃﬃsize, mm K1c, MPa m sy, MPa C (Paris law coefﬁcient) n (Paris law exponent)

As cast(a)

As cast(a)

As cast(a)

800 C

S 201 73.8 ... 25.6 612.2 1.58 10.75

S 221 45.2 ... 38.5 562.1 1.83 6.66

I ... 37.5 ... 37.5 437.6 1.10 7.14

S 107 0.021 61.5 409.4 9.86 7.44

900 C

1000 C

S S 95 101 [<2%, annealed ferritic] 0.032 0.039 66.3 67.3 405.3 402.7 4.54 8.51 7.41 7.27

900 C

S 269 0.034 64.5 401.3 2.09 5.12

(a) As cast (chill free); pearlite variation was achieved through changes in mold preheating temperature and mold wall thickness. (b) After soaking at the indicated temperature for 12 h, all the castings were furnace cooled to 690 C, held for 24 h, then air cooled to room temperature. (c) S, spheroidal graphite; I, irregular (vermicular) graphite. Source: Ref 11

cast iron per ISO 185 are summarized in Table 9. The ratio of fatigue endurance limits to ultimate tensile strength has been reported to vary from 0.33 to 0.62 for gray cast irons (Ref 1). Graphite ﬂake size, matrix microstructure (strength), and phosphorous content all inﬂuence the endurance ratio. Some general reported values for the endurance ratio have been: 0.4 to 0.7 for repeated loads in compression

only (Ref 4)

0.45 for irons with tensile strengths of 55 to

70 MPa (8 to 10 ksi) (Ref 1) 0.38 for tensile strengths from 80 to 100 MPa (12 to 15 ksi) with acicular-matrix microstructures (Ref 1)

Typical values for different loading conditions are summarized in Table 5. The ratio of endurance limits between torsion loading versus bending loads is reported to vary from 0.75 to 1.25 (Ref 4). In general, the fatigue endurance limit ratio with ultimate tensile strength is smaller at higher tensile strengths. Impact Strength and Toughness. Gray cast irons are not notch sensitive and do not have a pronounced ductile-to-brittle transition temperature (Fig. 11, 12). However, the impact toughness of unnotched gray iron is reported (Ref 4) to decrease about 10 to 20% at 75 C (100 F) and 50% at 185 C (300 F). Gray irons with ﬂake graphite have Charpy V-notch upper-

Variation of crack growth exponent with fracture toughness in magnesium-treated cast irons and steels. Steels data is from J. Knott, Fundamentals of Fracture Mechanics, Butterworth, 1973. Source: Ref 11

Fig. 7

shelf toughness at around 1.4 to 6.8 J (1 to 5 ft lbf), with higher values obtained with higherstrength-matrix microstructures (Ref 5). This unusual behavior (toughness generally decreases with higher strength) is attributed to the fact that

Fatigue and Fracture Properties of Cast Irons / 191

Diagram showing safe and unsafe fatigue zones for cast iron subjected to ranges of alternating stress superimposed on a mean stress. Example point P shows conditions of tensile (positive) mean stress; P0 shows compressive (negative) mean stress. The safety factor is represented by the ratio of OF to OP or OP0 to OP 0 . For conditions of constant mean tensile stress, DK/DP is the safety factor.

Fig. 8

Table 5

Gray iron fatigue strength in bending and torsion Gray iron No.

Parameter

Tensile strength, ksi Endurance limit in complete reversed bending, ksi Endurance limit in half-cycle bending, 0-max, ksi Endurance limit in complete reversed torsion, ksi Endurance limit in half-cycle torsion, 0-max, ksi

1

2

3

4

5

44.0 19.0 23.0 16.0 23.0

48.0 21.0 32.0 16.5 25.0

55.0 22.0 27.0 21.0 26.0

56.5 22.0 33.0 20.0 33.0

76.5 25.0 38.0 22.0 29.0

Composition C

Si

Mn

P

Mg

Other

2.96 3.52 3.52 3.67 3.60 3.48

2.90 2.61 2.25 2.55 2.34 4.84

0.78 0.25 0.40 0.13 0.50 0.31

0.66 0.051 0.054 0.060 0.053 0.067

... 0.015 0.015 0.030 0.030 0.030

0.12 Cr ... 1.47 Cu ... 0.54 Cu 1.02 Mo

Results of thermal fatigue tests on various cast irons; specimens cycled between 650 and 20 C (1200 and 70 F). FG, ﬂake graphite; CG, compacted graphite; SG, spheroidal graphite; UTS, ultimate tensile strength. Source: Ref 14

Fig. 9

Source: Ref 12

Table 6

Corrosion fatigue of cast irons Pearlitic ﬂake

Environment

Pearlitic nodular

Ferritic nodular

Fatigue strength, ksi

Fatigue strength reduction factor

Fatigue strength, ksi

Fatigue strength reduction factor

Fatigue strength, ksi

Fatigue strength reduction factor

16 13 5 14 16 16

... 1.23 3.20 1.14 1.00 1.00

35 29 6 25 29 32

... 1.21 5.83 1.40 1.21 1.09

27 23 6 20 24 25

... 1.18 4.50 1.35 1.13 1.08

16 16

1.00 1.00

27 34

1.30 1.03

24 28

1.13 0.97

Air Water 3% sodium chloride 1% borax 3% “Sobenite”(a) 3% sodium carbonate 3% soluble oil 0.25% potassium chromate

(a) “Sobenite” is a mixture of 10 parts sodium benzoate to 1 part sodium nitrite. Source: Ref 2

Table 7 Corrosion fatigue strengths of cast irons and steels Material

Flake graphite iron Mild steel Nodular graphite iron 1.3Ni-0.6Cr steel 18Cr-8Ni stainless steel Flake graphite iron 0.2% C steel 0.5% C steel (as-drawn) Nodular graphite iron 3Ni-0.7Cr steel 18Cr-8Ni stainless steel Source: Ref 1

Results of constrained thermal fatigue tests conducted between 100 and 540 C (212 and 1000 F). CG, compacted graphite; FG, ﬂake graphite. Source: Ref 15

Fig. 10

Fatigue strength, ksi Corrosive medium

In air

In corrosive medium

Tensile strength, ksi

Water Water Water Water Water 3% salt water Sea water 3% salt water 3% salt water Sea water Sea water

16 33 35 57 29 16 29 49 35 46 29

13 13 29 30 26 5 4 7 6 10 10

40 52 94 114 84 40 64 130 94 96 84

fracture occurs through the graphite ﬂakes rather than through matrix deformation (Ref 5). Several attempts have been made to measure the fracture toughness of gray iron using fracture mechanics methods (Ref 5). Some results are summarized in Fig. 14. As mentioned above for Charpy toughness, there is a trend of higher toughness at higher strengths. The values of fracture toughness reported in the literature for gray cast iron almost always give either KQ or Kmax, rather

192 / Casting Design and Performance

Effect of graphite shape on dynamic fracture toughness, (Kjd) of ferritic grades of various cast irons. These results are generally based on data from the literature, with some approximation where data in literature is incomplete. Source: Ref 5

Fig. 11

Fig. 13

Effect of graphite shape on fracture toughness of pearlitic grades of various cast irons. These results are generally based on data from the literature with some approximation in regions where data is incomplete. Source: Ref 5

Fig. 12

than KIc, because of the ambiguity in determining accurately the moment of crack extension. A question then arises as to which of the two is the better measure of the fracture toughness. Bradley observed more than 1.6 mm of crack advance before reaching maximum load in class 35 gray iron (Ref 18). In view of this result and the discussion presented in Ref 5, KQ values are considered reasonable and conservative estimates of KIc. Kmax values would tend to overestimate the fracture toughness of gray cast iron.

Ductile Irons Fatigue Strength. Minimum tensile strength and fatigue limits for ductile irons per ISO 1083 are summarized in Table 10. In fatigue applications, ductile irons are generally used in the pearlitic condition (as-cast or normalized). Austempered ductile iron and ferritic ductile iron are also used. Ductile iron is frequency sensitive, and the cycle frequency in testing should not exceed the expected frequency for service. Fatigue endurance ratios for ductile iron are affected primarily by the matrix microstructure/strength (Fig. 15) and by notch factors (Fig. 16) or surface condition. Table 11 summarizes typical reversed bending fatigue results for three standard grades of ductile iron. Results of rotating bending tests carried out by the British Cast Iron Research Association are

Charpy V-notch impact energies of ductile irons. Source: Ref 17.

summarized in Table 12 and are discussed in more detail in Ref 1. Ductile irons follow the expected trend of decreasing fatigue endurance ratios for higher strength of the matrix microstructure. However, heat treatments that result in an increase of tensile strength do not show a proportional increase in the fatigue limit (Table 11). References 1 to 6, 11, 12 and the “Selected References” listed at the end of this chapter are additional sources of property data on the fatigue of ductile iron. Charpy Impact Toughness. Upper-shelf Charpy V-notch energy for ferritic ductile iron is about 14 to 24 J (10 to 18 ft 1bf) as compared to 12 to 19 J (9 to 14 ft 1bf) for malleable iron and 1 to 6 J (0.7 to 4.5 ft 1bf) for gray iron. Data from a comprehensive study show that increasing pearlite content decreases impact toughness (Fig. 13) and that the transition temperature is signiﬁcantly affected by silicon and phosphorous content. The effects of various heat treatments on Charpy V-notch impact energy are shown in Fig. 17 for a ductile iron alloyed with about 0.75% Ni. Curve F shows that austempering heat treatment not only improves elongation at high strength but also produces the highest room-temperature impact energy of any of the conventional heat treatments, with the exception of surface-hardened as-cast ductile iron, curve A, with a hardness of 207 HB. The use of Charpy V-notch data is generally inappropriate to determine dynamic fracture toughness of ductile irons (Ref 5). Because graphite debonding dominates the fracture process in ductile iron, Charpy V-notch specimens may be acceptable only when the internodular spacing is less than the standard notch root radius of 0.25 mm. This would require a nodular count of less than 20 mm2, which is below the typical nodular counts in commercial ductile irons. Dynamic Tear Energy. Upper-shelf dynamic tear energy for ferritic ductile iron is about 175 J (129 ft lbf) for a standard 15.9

mm specimen, as compared to 100 J (73 ft lbf) for malleable iron (Ref 5). Greater amounts of pearlite in the matrix microstructure reduce the upper-shelf toughness (Fig. 18b). With higher levels of pearlite, the upper-shelf toughness is comparable to that of compacted graphite cast iron (Fig. 18a). Similar comparisons are expected for upper-shelf toughness derived from Charpy testing. Fracture Toughness. Early measurements of room-temperature fracture toughness of ferritic ductile irons were almost all incorrect. Systematic increases occurred over time as better test and evaluation methods were developed (Ref 5). Linear elastic fracture mechanics is not suitable for measuring the upper-shelf toughness of ferritic ductile irons unless very large specimens are used. Certain lowerstrength grades of ductile iron do not fracture in a brittle manner when tested under nominal plane-strain conditions in a standard fracture toughness test. In the low-strength ductile irons, plane-strain conditions are established only at temperatures low enough to embrittle the ferrite. Otherwise, an increase in the size of the fracture toughness test specimens is necessary to provide the degree of mechanical constraint necessary to obtain a valid measurement of KIc. Plane-strain fracture toughness (KIc) at ambient temperature and upper-shelf regime ispestiﬃﬃﬃﬃ pﬃﬃﬃﬃ mated at 88 to 90 MPa m (80 to 81 ksi in:) for ferritic ductile irons based on the review of BradleyandSrinivasan (Fig. pﬃﬃﬃﬃ pﬃﬃﬃﬃ 19). Upper-shelf KIc values of 100 MPa m (90 ksi in:) or more are also considered possible for ferritic pﬃﬃﬃﬃ ductile p irons, ﬃﬃﬃﬃ as compared with 75 MPa m (68 ksi in:) for malleable irons (Ref 5). Typical fracture toughness results are shown in Fig. 19 and 20. These fracture toughness estimates are higher than earlier results (e.g., Table 13); as discussed in Ref 5, the data in Table 13 are not accurate because the specimen was too small for proper toughness results on a ductile material.

Fatigue and Fracture Properties of Cast Irons / 193 Table 8 Typical impact strength of cast irons Ultimate tensile strength(a), ksi

Hardness, HB

V-notch Charpy, ft lbf

Unnotched Charpy, ft lbf

20–25 25–30 20–35 35–40 40–48 45–52 50–57 60–66 70 30 min 40 50 60 70 80

140–187 170–229 170–241 187–269 207–269 217–269 215–260 230–290 ... 200 230 260 290 320 350

... ... ... ... ... ... ... ... ... ... ... ... ... ... ...

... ... ... ... ... ... ... ... ... ... ... ... ... ... ...

21 22 23 25 31 36 65 75 120 26 35 43 52 61 70

55 55 60 60 70 ... 80 115(c) ... ... ... ... ... ... ...

Malleable cast irons: Ferritic malleable (32510) Ferritic malleable (35018) Pearlitic malleable (43010) Pearlitic malleable (45010) Pearlitic malleable (45007) Pearlitic malleable (48004) Pearlitic malleable (48005) Pearlitic malleable (50007) Pearlitic malleable (53004) Pearlitic malleable (60003) Pearlitic malleable (70002) Martensitic malleable (80002)

50–58 53–60 ... 65 65 70 ... 75 ... 80 ... 100

110–156 110–156 ... 163–207 163–207 163–228 ... 179–228 197–241 197–255 ... 241–269

14–17 16.5 ... 9 ... ... 6 ... 4 ... ... ...

50–60 ... ... ... ... ... ... ... ... ... ... ...

... ... ... ... ... ... ... ... ... ... ... ...

70–90 70–90 ... ... ... ... ... 22–35 ... 22–35 22–35 22–35

White and alloy cast irons Unalloyed white High Si (Duriron, Durichlor) High Cr (corros. resistant) Medium Si (Silal) Ni-Cr-Si High Al Ni-Resist (type 1) Ni-Resist (type 2) Ni-Resist (type 2b) Ni-Resist (type 3) Ni-Resist (type 4) Ni-Resist (type 5) Ni-Hard (type 1 sand cast) Ni-Hard (type 1 chill cast) Ni-Hand (type 2 sand cast) Ni-Hard (type 2 chill cast) Almanite (Meehanite type WS)

20–50 13–18 30–90 25–45 20–45 34–90 25–30 25–30 30–35 25–35 25–30 20–25 40–50 50–60 45–55 60–70 60–80

300–575 480–520 290–400 170–250 80–150 180–350 130–170 125–170 170–250 120–160 150–210 100–125 550 (s) 600(s) 525 (s) 575 (s) 400–525

... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...

... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...

... ... ... ... ... ... 100 100 60 150 80 150 20–30 25–40 25–35 35–55 to 180

3.5–10.0 2–4(d) 20–35(d) 15–23(d) 110–210(d) 180–350(d) ... ... ... ... ... ... ... ... ... ... ...

60–80 80–100 100–120 120–175 55–69 58–70 55–65 55–67 55–65 60–72 55–60 55–65

149–187 179–248 217–269 240–300 140–200 150–210 130–170 140–200 130–190 170–240 130–180 140–190

10–15 2–5 2–6 2–6 12 10 28 7 14 ... 17 6

75–115 15–65 35–50 25–40 26 ... ... ... ... ... ... ...

120+ 120+ ... ... ... ... ... ... ... ... ... ...

... ... ... ... ... ... ... ... ... ... ... ...

Class or type

Gray cast irons: Medium section (type 20) Medium section (type 25) Medium section (type 30) Medium section (type 35) Medium section (type 40) Medium section (type 45) Medium section (type 50) Medium section (type 60) Medium section (type 70) Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI Nickel alloyed gray iron (NCI

30) 40) 50) 60) 70) 80)

Ductile irons: Standard ductile 60-40-18 Standard ductile 80-55-06 Standard ductile 100-70-03 Standard ductile 120-90-02 Ni-Resist ductile (Austenitic) D-2 Ni-Resist ductile (Austenitic) D-2B Ni-Resist ductile (Austenitic) D-2C Ni-Resist ductile (Austenitic) D-3 Ni-Resist ductile (Austenitic) D-3A Ni-Resist ductile (Austenitic) D-4 Ni-Resist ductile (Austenitic) D-5 Ni-Resist ductile (Austenitic) D-5B

Unnotched Izod 1.2 in. diam, ft lbf

Unnotched Charpy (b) 1.2 in. diam, ft lbf

(a) 0.750 in. diameter specimens machined from 1.2 in. diameter test bar. (b) 18 in. between supports. (c) 6 in. between supports, 1.125 in. diameter. (d) 1.2 in. diameter, 6 in. between supports. Source: Ref 4

Lower-shelf fracture toughness is more consistent. Typical results for fully ferritic ductile iron and ferritic ductile iron with about 15% pearlite are presented in Fig. 20. The conclusion from review of data available (Ref 5) is that both fer-ritic and pearlitic ductile irons have similar lower-shelf fracturepﬃﬃﬃﬃ toughness values of about 23 to 28 MPa m (21 to

pﬃﬃﬃﬃﬃﬃ 25 ksi in:). At temperatures above 110 C, the fracture toughness of ferritic ductile iron increases sharply with increasing temperature. The presence of as little as 15% pearlite in ductile iron will cause a signiﬁcant increase in the ductile-to-brittle transition temperature (Fig. 20). For temperatures around 0 C, the fracture toughness KJc is a very sensitive

function of the percentage of pearlite in the matrix, the latter affecting the fracture toughness by lowering the upper-shelf value and, more importantly, by shifting the ductile-tobrittle transition temperature. Dynamic Fracture Toughness. Figures 11 and 12 summarize a review (Ref 5) of dynamic fracture toughness evaluations for ferritic and pearlitic grades of various cast irons. As expected, pearlite reduces upper-shelf toughness and increases transition temperatures. There are only a few reliable results for KId and KJd for ferritic ductile iron published in the literature. The upper-shelf dynamic fracture toughness of ferritic ductile iron (with 2.3% silicon and having less than 100 mm2 nodule count) is ﬃ about 90 to 95 pﬃﬃﬃﬃas measured by KJdpﬃﬃﬃﬃﬃ MPa m (82 to 86 ksi in:). This value decreases with increasing nodule count, decreasing nodularity, and increasing pearlite content in the matrix (Ref 5).

Malleable Cast Irons There are two basic types of malleable iron: blackheart and whiteheart. Blackheart malleable iron is the only type produced in North America and is the most widely used throughout the world. Whiteheart malleable iron is the older type and is essentially decarburized throughout in an extended heat treatment of white iron. Unless otherwise speciﬁed, this article considers only the blackheart type. Malleable iron, like medium-carbon steel, can be heat treated to produce a wide variety of mechanical properties. The different grades and mechanical properties are essentially the result of the matrix microstructure, which may be a matrix of ferrite, pearlite, tempered pearlite, bainite, tempered martensite, or a combination of these (all containing nodules of temper carbon). This matrix microstructure is the dominant factor inﬂuencing the mechanical properties. Minimum tensile properties and fatigue limits for whiteheart and blackheart malleable irons are summarized in Tables 14 and 15. Fatigue curves for a ferritic malleable iron are shown in Fig. 21 with fatigue curves for various compacted graphite irons. Fatigue curves for the ferritic grades of these two iron types are comparable. Notch radius generally has less effect on fatigue strength than notch depth for ferritic malleable irons (Fig 22). Charpy Impact Toughness. Like compacted graphite cast irons, malleable cast irons have a lower ductile-to-brittle transition than ductile irons with a comparable matrix microstructure. Malleable cast irons are generally regarded as having the best low-temperature toughness of the cast irons due to the lower silicon content (about 1% less silicon than ductile irons), which thus lowers the yield strength and the ductileto-brittle transition temperature. Ferritic malleable cast iron has an upper-shelf Charpy energy of 13 J (9.5 ft lbf), compared with

194 / Casting Design and Performance

9 J (6.5 ft lbf) for pearlitic malleable cast iron (Fig. 23). Other investigators have reported a lower transition temperature and a somewhat higher upper shelf Charpy energy of 19 to 20 J for ferritic malleable cast iron and 12 J for pearlitic malleable cast iron (Ref 29, 30). Dramatic reduction in upper-shelf toughness and increase in the ductile-to-brittle transition temperature result from phosphorus concentrations of 0.15% or greater in both ferritic and pearlitic malleable cast iron (Ref 31, 32). The phosphorus not only forms a very low-melting-point phosphide but also degenerates the graphite particle shape. Dynamic Tear Energy. Figure 24 compares the tear energy of malleable iron with that of various ductile irons. As shown previously for compacted graphite irons (Fig. 18), ferritic malleable iron has a lower ductile-to-brittle transition than ductile iron. These data also show that annealing above the critical temperature reduces upper-shelf toughness compared to subcritical annealing. Fracture Toughness. Some published data (such as those reported in Table 16 and Ref 35–38) have attempted to establish linear elastic (KIc) fracture toughness limits for malleable irons. However, as discussed in Ref 5, the fracture toughness of malleable irons needs to be evaluated by an elastic-plastic criterion such as KJc. Upper-shelf KJc values pﬃﬃﬃﬃrange from 53 pﬃﬃﬃﬃ to 73 MPa m (48 to 66 ksi in:), as summarized in Table 17. Linear elastic fracture toughness values reported in the literature (KIc in Table 16 and Kmax or KQ in Ref 35–39 are invalid in their indication of increases in upper-shelf toughness with increasing strength and/or lowering of temperature (Ref 5). While this may be a reasonable correlation for gray cast iron, it is not so for malleable cast iron. The reason is that a J-integral approach is needed to calculate the critical J value at or just before maximum load Table 9

where crack extension begins (Ref 5). It is inadvisable to use KQ or Kmax for toughness ranking of malleable irons (Ref 5).

White Cast Iron A number of studies have been conducted to determine the fracture toughness of white cast iron (Fig. 25). White cast iron is generally used in applications where abrasion resistance is desirable. Gahr and Scholz (Ref 40) have found that the dynamic fracture toughness (KId) values, measured on 10 10 55 mm specimens with a 0.2 mm p wide the pﬃﬃﬃﬃﬃ ﬃ ﬃﬃﬃﬃ slot, were in range of 14 to 24 MPa m (13 to 22 ksi in:), while the quasistatic fracture toughness values, measured on fatigue precracked compact tension p specimens, were inpthe ﬃﬃﬃﬃﬃﬃ range of 25 to 33 ﬃﬃﬃﬃ MPa m (23 to 30 ksi in:). Under dynamic loading, austenitic grades of white cast iron had better impact strength than did martensitic grades, though no such clear pattern was noted for KId values. Eriksson (Ref 41) measured pﬃﬃﬃﬃﬃﬃ toughpﬃﬃﬃﬃ a fracture ness value of 22 MPa m (20 ksi in:) on a white cast iron sample with martensitic matrix corresponding to a Vicker’s hardness value of 633. A comparison of his results (Fig. 25) with those of Gahr and Scholz indicates both lower hardness and lower toughness. This may be due to the presence of some ﬂake graphite in samples tested by Eriksson.

3. C.F. Walton and T.J. Opar, Iron Castings Handbook, Iron Casting Society, 1981 4. D. Maykuth, Structural Alloys Handbook, Mechanical Properties Data Center, published originally by Battelle, now offered by CINDAS 5. W.L. Bradley and M.N. Srinivasan, Fracture and Fracture Toughness of Cast Irons, International Materials Reviews. Vol 35 (No. 3), 1990, p 129–159 6. E. Nechtelberger, The Properties of Cast Iron up to 500 C, Technicopy Ltd., 1980

Relationship between tensile strength and fracture toughness parameters KQ and Kmax for gray cast iron. Source: Ref 5

Fig. 14

REFERENCES 1. K.B. Palmer, Fatigue Properties of Cast Iron, British Cast Iron Research Association 2. D. Socie and J. Fash, Fatigue Behavior and Crack Development in Cast Irons, Trans. AFS, Vol 90, 1982, p 385–392

Minimum tensile and fatigue properties of gray cast irons to ISO 185 (1988) Minimum properties

Property

Grade 150

Grade 200

Grade 250

Grade 300

Grade 350

Grade 400(a)

Tensile strength, MPa (ksi) Fatigue limit (Wohler) Unnotched(b), MPa (ksi) V-notched(c), MPa (ksi)

150 (21.75)

200 (29.0)

260 (37.7)

300 (43.5)

350 (50.75)

400 (58.0)

117 (16.95) 108 (15.65)

135 (19.55) 122 (17.7)

149 (21.6) 127 (18.7)

152 (22.0) 129 (18.4)

68 (9.85) 68 (9.85)

90 (13.05) 87 (12.6)

(a) Not in ISO 185 but included in some national speciﬁcations. (b) 8.4 mm diameter. (c) Circumferential 45 V-notch with 0.25 mm root radius

Table 10

Fig. 15 Ref 19

Effect of tensile strength and matrix structure on endurance ratio for ductile iron. Source:

Minimum tensile and fatigue properties of ductile irons produced to ISO 1083 (1987) Ferritic irons

Property

Tensile strength, MPa (ksi) 0.19 proof strength, MPa (ksi) Elongation, % Fatigue limit (Wohler) Unnotched (10.6 mm), MPa (ksi) Notched(a), MPa (ksi)

Intermediate grades

Pearlitic as-cast and normalized irons

Hardened-and-tempered irons

Grades 350-22 350-22L

Grade 400-18 400-18L

Grade 450-10

Grade 500-7

Grade 600-3

Grade 700-2

Grade 800-2

Grade 900-2

Grade 700-2

Grade 800-2

Grade 900-2

350 (50.75) 203 (29.45) 22

400 (58.0) 247 (35.8) 18

450 (65.25) 293 (42.5) 10

500 (72.5) 323 (46.85) 7

600 (87.0) 346 (50.2) 3

700 (101.5) 385 (55.8) 2

800 (116) 440 (63.8) 2

900 (130) 495 (71.8) 2

700 (101.5) 525 (76.1) 2

800 (116) 600 (87) 2

900 (130) 675 (97.9) 2

180 (26.1) 114 (16.5)

195 (28.3) 122 (17.7)

210 (30.5) 128 (18.6)

224 (32.5) 134 (19.4)

248 (36) 149 (21.6)

280 (40.6) 168 (20)

304 (44.1) 182 (26.4)

317 (46) 190 (27.5)

280 (40.6) 168 (24.35)

304 (44.1) 182 (26.4)

317 (46) 190 (27.55)

(a) Diameter at notch root, 10.6 mm: circumferential 45 V-notch with 0.25 mm root radius and a notch depth of 3.6 mm

Fatigue and Fracture Properties of Cast Irons / 195 Table 11

Reversed-bending fatigue properties of three standard grades of ductile iron 45 V-notched

Unnotched(a) Tensile strength, St

Endurance limit, Se

Type

MPa

ksi

MPa

ksi

Endurance ratio(b)

64-45-12 80-55-06 120-90-02(e)

490 620 930

71 90 135

210 275 338

30.5 40 49

0.43 0.44 0.36

Endurance limit, Sn MPa

ksi

Endurance ratio(c)

Notch sensitivity factor(d)

145 165 207

21 24 30

0.30 0.27 0.22

1.4 1.7 1.6

(a) Polished specimens with 10.6 mm diameter. (b) Se/Sr (c) Sn/Sc, (d) Sc/Sn, (e) Oil quenched from 900 C (1650 F) and tempered at 595 C (110 F). Source: Ref 21

Table 12

Effect of heat treatment on fatigue properties of ductile iron

Treatment

As-cast Normalized As-cast Oil-quenched, tempered 2 h 600 C Oil-quenched, tempered 2 h 550 C

Tensile strength, ksi

Unnotched fatigue limit, ksi

Endurance ratio

Notched fatigue limit, ksi

85 134 86 120 133

39 44 36 44 44

0.46 0.33 0.42 0.37 0.33

27 27 23 27 25

Source: Ref 1

Fig. 16

Effect of cast section size on the fatigue properties of ductile irons. Source: Ref 20

Fig. 17

Effect of heat treatment on Charpy V-notch impact properties of ductile iron. Source:

Ref 22

7. H.T. Angus, Physical and Engineering Properties of Cast Iron, British Cast Iron Research Association, 1960 8. H. Grover, S. Gordon, and C. Jackson, Fatigue of Metals and Structures, Department of the Navy, 1960 9. H.J. Grover, S.A. Gordon, and L.R. Jackson, Fatigue of Metals and Structures, Department of the Navy, 1960, p 278–279

Effect of pearlite content on the toughness of (a) compacted graphite (CG) cast iron and (b) spheroidal graphite (SG) cast iron (i.e., ductile iron). This comparison for several different matrix microstructures indicates that the deleterious effect of vermicular graphite on fracture toughness in CG iron is much more pronounced for a ferritic matrix than for a pearlitic matrix. Furthermore, the compacted graphite cast iron seems to have a somewhat lower ductile-to-brittle transition temperature than ductile iron, possibly because the ductile fracture has been made easier relative to cleavage by the formation of compacted graphite. Source: Ref 23

Fig. 18

196 / Casting Design and Performance

Comparison of fracture toughness of fully ferritic ductile iron (GGG-40) with ferritic ductile iron with 15% pearlite in matrix (S-45). Source: Ref 5

Fig. 20 Effect of microstructure on fracture toughness of ductile iron. These results are generally based on data from the literature, with some approximation in regions where data were unavailable. Source: Ref 5

Fig. 19

Table 13

Type of iron

Ferritic 3.0% Si 3.5% Si Pearlitic 2.5% Si

Fracture toughness of ductile iron from early work Ultimate tenslie strength

Yield strength

MPa

ksi

MPa

As-cast As-cast

521 547

75.6 79.4

427 471

62.0 68.3

As-cast Normalized Austempered

703 918 ...

102.0 133.2 ...

374 552 620

54.2 80.0 90.0(a)

Condition

ksi

Elongation, %

Klc, MPa

0.5

(ksi

in.

1/2

) at:

105 C (160 F)

... ...

35.1 (32.0) 27.0 (24.6)

30.2 (27.5) ...

... 45.3 (41.3) 36.5 (33.3)

37.1 (33.8) ... ...

... ... ...

11.0 9.0 7.5 3.6 ...

m

20 C (70 F) 40 C (40 F)

Note: These fracture toughness data are from early test results, which are probably not representative of inherent Klc values based on more current evaluations. See text. (a) Estimated. (b) Estimated yield strength. Source: Ref 24, 25

Table 14 Minimum tensile and fatigue properties of blackheart malleable irons produced to BS 310 (1972) Minimum properties Property

Grade B290/6

Grade B310/10

Grade B340/12

Tensile strength, MPa (ksi) 0.2% proof stress, MPa (ksi) Elongation, %

290 (42) 191 (27.7) 6

310 (45) 205 (29.7) 10

340 (49.3) 224 (32.5) 12

Fatigue limit (Wohler) Unnotched (10 mm diam), MPa (ksi) Notched(a) (10 mm diam at root), MPa (ksi)

174 (25.2) 87 (12.6)

186 (27.0) 93 (13.5)

204 (29.6) 102 (14.8)

(a) Circumferential 45 V-notch with 0.25 mm root radius and a notch depth of 2 mm

Table 15 Minimum tensile and fatigue properties of whiteheart malleable irons produced to BS 309 (1972) Grade W340/4 section thickness Property

Grade W410/4 section thickness

9 mm

12 mm

15 mm

9 mm

12 mm

15 mm

Tensile strength, MPa (ksi) 0.2% proof stress, MPa (ksi) Elongation, %

270 (39.15) 149 (21.6) 7

310 (45) 171 (24.8) 4

340 (49.3) 187 (27.1) 3

350 (50.75) 193 (28) 10

390 (56.55) 215 (31.15) 6

410 (59.5) 226 (32.75) 4

Fatigue limit (Wohler) Unnotched (10 mm diam), MPa (ksi) Notched(a) (10 mm diam at root), MPa (ksi)

122 (17.7) 73 (10)

140 (20.3) 84 (12)

153 (22.2) 92 (13)

158 (22.9) 95 (13.8)

176 (25.5) 106 (15.4)

185 (26.8) 111 (16.1)

(a) Circumferential 45 V-notch with 0.25 mm root radius and notch depth of 2 mm

Rotating-bending fatigue strength of malleable cast iron and compacted graphite (CG) cast iron. (a) Unnotched and notched fatigue curves of two ferritic malleable irons (25 mm diameter specimens). (b) Unnotched and notched fatigue curves for ferritic, pearlitic, and higher-nodularity CG irons. SG, spheroidal graphite. Sources: Ref 26, 27

Fig. 21

10. Fatigue of Metals, National Physics Lab, 1962, p 154–155 11. S. Seetharamu and M.N. Srinivasan, Fatigue Crack Propagation in Permanent Mold Ductile Iron Castings, AFS Transactions, Vol 94, 1986, p 265–270 12. C. Walton, Gray and Ductile Iron Castings Handbook, Iron Casting Society, 1965 13. R.E. Peterson, Stress Concentration Factors, John Wiley, 1974 14. K. Roehrig, Trans. AFS, Vol 86, 1978, p 75 15. Y.J. Park, R.B. Gundlach R.G. Thomas, and J.F. Janowak, Trans. AFS, Vol 93, 1985, p 415 16. J.D. Condes, Fracture Toughness Testing, Fatigue and Fracture, Vol 19, ASM Handbook, 1996, p 393–409 17. Pellini et al, Noten Dueﬁlity of Nadolar Irons, Trans. ASM, Vol 46, 1954, p 418–445 18. W. Bradley, Trans. AFS, Vol 89, 1981, p 837–848 19. Metals Handbook, 9th ed., Vol 1, p 45

Fatigue and Fracture Properties of Cast Irons / 197 20. 21. 22. 23. 24.

25. 26. Effects of notch radius and notch depth on the fatigue strength of ferritic malleable iron. Source: Ref 28

Fig. 22

27.

Fig. 23

Charpy V-notch transition curves for ferritic and pearlitic malleable irons. Source: Ref 3

28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

Metals Handbook, 9th ed., Vol 15, p 662 ASM Handbook, Vol 1, p 48 Metals Handbook, 9th ed., Vol 1, p 42 K. Cooper and C. Loper, Trans. AFS, Vol 86, 1978, p 241 R.K. Nanstad, F.J. Worzata, and C.R. Loper, Jr., Static and Defaninc Toughness of Ductice Cast Iron, Trans. AFS, Vol 83, 1975 ASM Handbook, Vol 1, p 47 L.W.L. Smith et al., Properties of Modern Malleable Irons, BCIRA International Center for Cast Metals Technology, 1987 K.B. Palmer, BCIRA J., Report 1213, Jan 1976, p 31 Metals Handbook, 8th ed., Vol 1961, p 370 G. Yu. Schulte, Metalloved. Term. Obrab. Met., Nov 1973, p 35–37 F.W. Jacobs and F.L. Preston, Trans. AFS, Vol 83, 1975, p 263–270 P.B. Burgess, Trans. AFS, Vol 75, 1967, p 665–675 T. Okumato and K. Kondo, Imono (J. Jpn Foundrymen’s Soc.), Vol 34, 1962, p 34 First International Conference on Austempered Ductile Iron, American Society for Metals, 1984 ASM Handbook, Vol 1, p 80 G.N.J. Gilbert, Engineering Data on Malleable Cast Irons. British Cast Iron Research Association, 1968 F.J. Worzala, R.W. Heine, and Y.W. Cheng, Trans. AFS. Vol 84, 1976, p 675–682 A. Little and H.J. Heine, in The Metallurgy of Cast Iron, B. Lux et al., Ed., Georgi Publishing, Switzerland, 1975, p 767–788 V.I. Ovchinnikov, V.A. Kurdyukov, and S.G. Girenkov, Metalloved. Term. Obrab. Met., Feb 1982, p 13–14 ASM Handbook, Vol 1, p 80 K.Z. Gahr and W.G. Scholz, J. Met., Vol 32, Oct 1980, p 38–44 K. Eriksson, Scand. J. Metall., Vol 2, 1973, p 197–203

SELECTED REFERENCES 1980 TO 1995 A. Alagarsamy and J. Janowak, Fatigue

Fig. 24

Dynamic tear energy versus temperature for four ferritic cast irons. Specimen size: 190 mm (7½ in.) long, 130 mm (5 in.) wide, and 41 mm (1 in.) thick; 13 mm (½ in.) notch depth. Source: Ref 33

Strength of Commercial Ductile Irons, Trans. AFS, Vol 98, 1990, p 511–518 P. Bhandhubanyong, I. Umeda, and Y. Kimura, Fatigue Strength and Crack Propagation in Spheroidal, Compacted Vermicular and Flake Graphite Cast Irons, Trans. Jpn. Foundrymen’s Soc., Vol 3, 1984, p 40–44 S.B. Biner, The Role of Eutectic Carbide Morphology on the Fracture Behavior of HighChromium Cast Irons, Part 1: Austenitic Alloys, Can. Metall. Q., Vol 24 (No. 2), 1985, p 155–162

198 / Casting Design and Performance Table 16

Fracture toughness of malleable irons from early results Test temperature

Malleable iron grade

Ferritic M3210

Pearlitic M4504 (normalized)

M5503 (quenched and tempered)

M7002 (quenched and tempered)

Yield strength

Klc

pﬃﬃﬃﬃ ksi in:

F

MPa

ksi

pﬃﬃﬃﬃ MPa m

24 19 59

75 3 74

230 240 250

33 35 36

44 42 44

40 38 40

24 19 57 24 19 58 24 19 58

75 2 70 75 3 73 75 3 72

360 380 390 410 440 455 520 550 570

52 55 57 60 64 66 75 80 83

55 48 30 45 52 30 54 38 40

50 44 27 41 47 27 49 35 36

C

Note: These fracture toughness data are from early test results. Which are not considered representative of Klc values from more recent evaluations (especially in the ferritic condition). See text. Source: Ref 34

Table 17

Fracture toughness of various malleable irons Yield strength, MPa

Temperature, C

KJmax (new)(a), pﬃﬃﬃﬃ MPa m

KJc (previously reported), pﬃﬃﬃﬃ MPa m

M3210 (ferritic malleable iron)

230

M4504 (ferritic/pearlitic malleable iron)

359

M5503 (tempered martensitic malleable iron)

411

M7002 (tempered martensitic malleable iron)

540

80-60-03 (pearlitic/ferritic ductile iron)

433

DQ & T (SAE grade) (tempered martensitic ductile iron)

717

DSB (SAE grade) (austenitic ductile iron)

304

24 19 57 24 19 59 24 18 58 24 19 58 24 19 48 24 19 59 24 47 59

76 76 71 70 72 30 57 53 34 50 39 39 33 25 26 48 51 50 89 85 91

44 41 43 55 48 30 45 52 29 50 39 39 27 25 26 52 48 51 64 ... 67

Material

Relationship between fracture toughness and hardness for white cast irons of different microstructures. Source: Ref 5

Fig. 25

A. Tiziani, E. Ramous, and A. Molinari,

Fracture Morphology in Compacted Graphite Cast Iron, Advances in Fracture Research, Vol 2, Pergamon Press, 1984, p 1489–1496 A.V. Vesnitskii, Criteria of the Contact Strength, Ductility, and Fracture of Materials in Shear with Buckling, Strength of Materials, Vol 24 (No. 6), 1993, p 402–408 S. Wu and G. Shu, Thermal Fatigue Property of Vermicular Graphite Cast Iron, Fatigue 90, Materials and Component Engineering Publications, Birmingham, U.K., 1990, p 1705–1710 Prior to 1980 D.R. Askeland and R.F. Fleischman, Effect

(a) KJmax values were recalculated from earlier work and are assumed to equal KJc. Source: Ref 5

J.H. Bulloch, Fractographic Analysis of

H. Sachar and J. Wallace. Effect of Micro-

Fatigue Cracking in Spheroidal Graphite Cast Irons, Theoretical and Applied Fracture Mechanics, Vol 17 (No. 1), 1992, p 19–45 G. Faubert, D. Moore, and K. Rundman, Heavy Section ADI: Fatigue Properties in As-Cast and Austempered Specimens, Trans. AFS, Vol 97, 1990, p 759–764 R. Higuchi and J. Wallace, Fatigue Properties of Gray Iron, Trans. AFS, Vol 89, 1981, p 483–494 K.H. Kloos and J. Adelmann. Effect of Deep Rolling on Fatigue Properties of Cast Irons, J. Mech. Behav. Mater., Vol 2, 1989, p 75–86

structure and Testing Modes on the Fatigue Properties of Gray Cast Iron, Trans. AFS, Vol 90, 1982, p 777–793 M. Srinivas and G. Malakondaiah, Effect of Grain Size on Threshold Stress Intensity for Fatigue Crack Growth in Armco Iron, Scr. Metall., Vol 20 (No. 5), 1986, p 689– 692 Z. Tao and J. Luo, Inﬂuence of Surface Hardening on Contact Fatigue Life of Nodular Cast Iron, Acta Metallurg. Sinica, Vol 3A (No. 3), 1990, p 188–193

of Nodule Count on the Mechanical Properties of Ferritic Malleable Iron, Trans. AFS, Vol 86, 1978, p 373–378 A.G. Fuller, Effect of Graphite Form on Fatigue Properties of Pearlitic Ductile Irons, Trans. AFS, Vol 85, 1977, p 527–536 G.N.J. Gilbert, Engineering Data on Malleable Cast Irons, British Cast Iron Research Association, 1968 Forrest, Fatigue of Metals, National Physics Lab, 1962 H. Morrogh, Fatigue of Cast Irons, Fatigue of Metals, Chapman-Hall, 1959 M. Sofue, S. Okada, and T. Sasaki, High-Quality Ductile Iron with Improved Fatigue Strength, Trans. AFS, Vol 86, 1978, p 173–183

Casting Design and Performance Pages 199–208

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Fatigue and Fracture Properties of Cast Steels* CASTINGS offer unique cost advantages over other manufacturing methods for many components, especially those having complex three-dimensional geometry. Unlike built-up assemblies, castings can be designed for function rather than for ease of assembly. However, castings are not speciﬁed by designers as often as their advantages would imply, because many designers mistakenly believe that the casting process inherently produces components which contain ﬂaws that deleteriously affect properties and part performance. This situation is changing. The casting process is well understood, and high-integrity castings are being made with the same reliability as forgings. Indeed, castings prove their quality every day in applications as demanding as the rotating hardware in gas turbine engines and primary aircraft structures. With more demanding structural applications, the fatigue and fracture resistance of castings is of considerable interest to designers of castings. In the past, material selection and product design were based on the results of tension tests, stresslife fatigue tests, and impact tests. This method of design against fatigue and fracture was not always satisfactory. In recent years, design philosophies based on low-cycle fatigue strain versus life behavior, crack propagation, and fracture toughness have emerged as being sound and successful. This chapter summarizes the general fatigue and fracture properties of cast steels, with emphasis on general comparisons with wrought steel and usefulness design data based on strain-life and fatigue crack growth testing. Topic coverage includes: Cyclic stress-strain behavior and low- and

high-cycle fatigue life behavior Plane-stress fracture toughness Plane-strain fracture toughness Constant-amplitude fatigue crack initiation and growth Variable-amplitude fatigue crack initiation and growth

The subjects of sustained-load cracking and stress-corrosion cracking, creep crack growth, creep-fatigue behavior, and thermal fatigue are

also important to designers; however, these topics are beyond the scope of this chapter. Reference 1 contains more information on these topics. Comparisons with wrought steels are emphasized when possible because most available data are for wrought steels at room temperature. However, a signiﬁcant portion of this chapter is based on a very comprehensive study of the fatigue and fracture properties of cast steels at room and low climatic temperatures done by R.I. Stephens and his students (Ref 2, 3). The study included fatigue behavior under both constant-amplitude and variable-amplitude loading, for the following grades of cast steel: SAE 0030: normalized and tempered (NT) at

137 HB (average)

SAE 0050A: normalized and tempered (NT)

at 192 HB

Carbon-Manganese

Steel: normalized, quenched and tempered (NQT) at 174 HB Manganese-Molybdenum Steel: normalized, quenched and tempered (NQT) at 206 HB AISI 8630: normalized, quenched, and tempered at 305 HB The room-temperature ultimate tensile strength and yield strength ranges are 500 to 1150 MPa (72 to 166 ksi) and 300 to 1000 MPa (44 to 143 ksi), respectively. Brinell hardnesses range from 137 to 305 HB and both ferritic-pearlitic and tempered martensitic microstructures are involved. These cast steels are representative of the carbon and low-alloy cast steels used for structural applications.

Structure and Property Correlations Carbon and low-alloy steel castings are produced to a great variety of properties because composition and heat treatment can be selected to achieve speciﬁc combinations of properties, including hardness, strength, ductility, fatigue, and toughness. Although selections can be made from a wide range of properties, it is important to recognize the interrelationships among these properties. These general relations are

*Reprinted from ASM Handbook, Volume 19, Fatigue and Fracture, ASM Handbook Committee, p 655-664

summarized in this section for toughness, fatigue, and component design factors such as section size and discontinuities.

Strength and Toughness Several test methods are available for evaluating the toughness of steels or the resistance to sudden or brittle fracture. These include the Charpy V-notch impact test, the drop-weight test, the dynamic tear test, and specialized procedures to determine plane-strain fracture toughness. Higher toughness is obtained when a steel is quenched and tempered, rather than normalized and tempered; quenching, followed by tempering, produces superior toughness. Charpy Impact Toughness is the most common measure of toughness with several types of specimen conﬁgurations such as the V-notch or keyhole specimen. The loss in toughness at lower temperatures is more distinct in Charpy V-notch data as compared to keyhole-notch specimen testing, which is why V-notch data are more commonly used to evaluate ductilebrittle transition temperatures. Typical Charpy V-notch data are shown in Fig. 1 for ﬁve common structural cast steels and one cast austenitic stainless steel (CF8). Austenitic steels retain considerable toughness at cryogenic temperatures. Comparative data for other cast stainless steels are summarized in Fig. 2. The lower carbon duplex cast stainless steel grades CF3M, CF8, and CF8M and the austenitic CN7M and CF20 exhibit signiﬁcantly higher toughness than the stabilized grade of CF8C, and the free-machining grade CF8F (a lower carbon version of C-16). The higher carbon austenitic grades CH-20 and CK-20 show lower impact resistance than the other grades listed. Nil Ductility Transition Temperatures (NDTT) for cast carbon and cast low-alloy steels range from 38 C (100 F) to as low as 90 C (130 F) for normalized-and-tempered cast carbon and low-alloy steels in the yield strength range of 207 to 655 MPa (30 to 95 ksi) (Fig. 3). Comparison of the data in Fig. 3 with those of Fig. 4 shows the superior toughness values at

200 / Casting Design and Performance

Fig. 1

Charpy V-notch (CVN) impact toughness of various cast steels (a) carbon and low-alloy steels with ferritic-pearlitic (NT treatments) or tempered martensite (NQT treatments) microstructures. (b) Low-temperature Charpy energy band of an austenitic cast stainless steel (CF-8, solution treated and quenched)

Plane-strain fracture toughness KIc and strength relationships at room temperature for quenched-and-tempered nickel-chromium-molybdenum steels

Fig. 5

Nil ductility transition temperatures and yield strengths of normalized-and-tempered commercial cast steels

Fig. 3

Charpy keyhole Impact toughness of various cast stainless steels (solution annealed and quenched). Source: Ref 1

Fig. 2

equal strength levels that low-alloy steels offer compared to carbon steels. When cast steels are quenched and tempered, the range of strength and of toughness is broadened. Depending on alloy selection, NDTT values of as high as 10 C (50 F) to as low as 107 C (160 F) can be obtained in the yield strength range of 345 to 1345 MPa (50 to 195 ksi) (Fig. 4). An approximate relationship exists between the Charpy V-notch impact energy-temperature

Fig. 4 cast steels

Nil ductility transition temperatures and yield strengths of quenched-and-tempered commercial

behavior and the NDTT value. The NDTT value frequently coincides with the energy transition temperature determined in Charpy V-notch tests.

Plane-Strain Fracture Toughness (KIc) data for a variety of steels reﬂect the important strength-toughness relationship. Fracture mechanics tests have the advantage over conventional toughness tests of being able to yield material property values that can be used in design equations. Plane-strain fracture toughness (KIc) data for the various cast steels are in Fig. 5. For quenched-and-tempered nickel-chromiummolybdenum steels, Fig. p5ﬃﬃﬃﬃindicates high pﬃﬃﬃﬃﬃﬃ KIc values of about 110 MPa m ð100 ksi in:Þ. at a 0.2% offset yield strength level of 1034 MPa (150 ksi). At a yield strength level of 1655 MPa (240 ksi), pﬃﬃﬃﬃﬃﬃ decrease to about pﬃﬃﬃﬃ KIc values 66 MPa m ð60 ksi in:Þ. Data are also plotted in Fig. 5 for wrought plates made of comparable steel of somewhat higher carbon content.

Tensile Strength and Fatigue Strength Limits Cast and wrought steels have similar fatigue (or endurance) limits for notched specimens with comparable tensile strength (e.g., Fig. 6).

Fatigue and Fracture Properties of Cast Steels / 201 Table 1 Fatigue notch sensitivity of several cast and wrought steels (R = 1) Endurance limit Tensile strength

Fatigue endurance limit versus tensile strength for notched and unnotched cast and wrought steels with various heat treatments. Data obtained in R.R. Moore rotating beam fatigue tests of nine steels (Kt = 2.2). Source: Metals Handbook, Volume 1, 8th ed., 1961

Fig. 6

Unnotched

Notched

Fatigue endurance ratio

ksi

MPa

ksi

MPa

ksi

Unnotched

Notched

Fatigue notch sensitivity factor, q

Normalized and tempered 1040 cast 648 1040 wrought 620 1330 cast 685 1330 cast 669 4135 cast 777 4335 cast 872 8630 cast 762 8640 wrought 748

94.2 90 99.3 97 112.7 126.5 110.5 108.5

260 ... 334 288 353 434 372 ...

37.7 ... 48.4 41.7 51.2 63 54 ...

193 ... 219 215 230 241 228 ...

28 ... 31.7 31.2 33.3 34.9 33.1 ...

0.40 ... 0.49 0.43 0.45 0.50 0.49 ...

0.30 ... 0.32 0.32 0.30 0.28 0.30 ...

0.29 0.50 0.44 0.28 0.45 0.68 0.53 0.85

Quenched and tempered 1330 cast 843 1340 wrought 836 4135 cast 1009 4335 cast 1160 8630 cast 948 8640 wrought 953

122.2 121 146.4 168.2 137.5 138.2

403 ... 423 535 447 ...

58.5 ... 61.3 77.6 64.9 ...

257 ... 280 332 266 ...

37.3 ... 40.6 48.2 38.6 ...

0.48 ... 0.42 0.46 0.47 ...

0.31 ... 0.28 0.29 0.27 ...

0.48 0.73 0.43 0.51 0.57 0.90

Annealed 1040 cast 1040 wrought

83.5 81.4

229 ...

33.2 ...

179 ...

26 ...

0.40 ...

0.31 ...

0.23 0.43

Steel

MPa

576 561

(a) Notched tests run with theoretical stress concentration factor of 2.2. (b) q = (Kf 1)/(Kt 1), where Kf is the endurance limit notched/endurance limit unnotched and Kt is the theoretical stress concentration factor rotating bearn bending tests (mean stress = 0)

S-N curves (R = 1) of a normalized and tempered AISI 4140 wrought steel in the longitudinal and transverse direction and cast 4135 steel normalized and tempered. Tensile strength for wrought steel: longitudinal, 110.0 ksi (758 MPa); transverse, 110.7 ksi (763 MPa); cast steel: 112.7 ksi (770 MPa)

Fig. 7

The endurance ratio (endurance limit divided by the tensile strength) of cast carbon and low-alloy steels as determined by rotating-beam bending fatigue tests (mean stress = 0) is generally taken to be approximately 0.40 to 0.50 for smooth bars. The data given in Table 1 indicate that this endurance ratio is largely independent of strength, although the endurance ratio tends to decrease at higher tensile strengths. The fatigue notch sensitivity factor, q, determined in rotating-beam bending fatigue tests is related to the microstructure of the steel (composition and heat treatment) and the strength. Table 1 shows that q generally increases with strength—from 0.23 for annealed carbon steel at a tensile strength of 576 MPa (83.5 ksi) to 0.68 for the higher-strength normalized-and-

tempered low-alloy steels. The quenched-andtempered steels with a martensitic structure are less notch sensitive than the normalizedand-tempered steels with a ferrite-pearlite microstructure. Similar results and trends in notch sensitivity have been reported for tests with sharper notches. Because of the inherent microdiscontinuities in castings, cast steels suffer less degradation of fatigue properties due to notches than equivalent wrought steels. Endurance limits of cast steels also are less sensitive to the testing direction than wrought steels. For wrought steels, the endurance limit is lower in the transverse direction than in the longitudinal direction of prior working (Fig. 7). This is attributed to the elongation of inclusions in the direction of prior working and the stress concentrating effect of loading normal to the major axis of an essentially elliptical void. Cast steels, however, do not exhibit this directionality since the microdiscontinuities generally have no preferential orientation. It is also common practice to liken the fatigue resistance of cast steels to those of a comparable wrought steel tested in a transverse direction.

Fatigue of Cast Steel Besides the effect of strength and external notches (Fig. 6 and 7), the fatigue of cast steels is inﬂuenced principally by the following factors: Section size Defect size Stress modes and waveform type

These factors are brieﬂy described in this section with additional information on the strainlife behavior of cast steels. Because fatigue cracks initiate from plastic deformation in regions of stress concentration or inherent discontinuities, strain-life data provide useful

information for design analysis. A brief review of inherent notch effects from internal microdiscontinuities is also summarized. Section Size Effects. Increasing section size reduces tensile strength and thus fatigue strength. However, if the fatigue endurance is just a function of tensile strength, then the endurance ratio (fatigue limit/tensile strength) would be constant versus section size. This is not strictly true for all castings. In some castings, a slight decrease in the endurance ratio may be observed. In Fig. 8, for example, a slight decrease in endurance ratio is observed for 8635, when compared to 1030 castings. Effect of Defects. Figure 9 illustrates the typical effect of various defects on fatigue strength of a cast steel. The reduction in fatigue strength depends on defect size (Fig. 10), with the largest defects being the weakest links. The notch effect from defects is described by various relations in terms of the fatigue notch factor (Kf). Analysis by de Kazinczy (reported in Ref 4) speciﬁes: pﬃﬃﬃ Kf ¼ 1 þ 0:16 d

(Eq 1)

where d is the defect diameter in mm. However, this relation does not account for the shape of the defect or the matrix strength. Another method (Ref 4) is based on the method of Peterson, which relates Kf to defect as follows: Kf ¼ 1 þ

Kt 1 1 þ a=r

(Eq 2)

where Kt is the theoretical stress concentration factor, r is the tip radius of the surface, and a is a material constant, which for most ferrous alloys is estimated in mm as a ¼ 0:0254 ½2070=UTS ðin MPaÞ1:8 mm

(Eq 3)

For example, with a hemispherical surface discontinuity (Kt ﬃ 2.5) and an average

202 / Casting Design and Performance

Fig. 8

Fatigue endurance ratios of three cast steels versus section size. Bands show the range for specimens taken at different depths from 32 mm, 76 mm, and 152 mm squares.

Fig. 10

Effect of defect size on fatigue endurance limits of cast steel. Source: Ref 5

diameter of 0.40 mm (r = 0.20 mm) for the largest microdiscontinuities in a cast steel with a Brinell hardness of 160 HB (and UTS = 3.5 HB = 560 MPa), then: a ﬃ 0:0254 ½3:71:8 0:27 mm Kf ﬃ 1 þ Effect of various defects on endurance ratio (fatigue strength/monotonic strength) of quenched-and-tempered 8630 cast steel. (a) R.R. Moore rotating beam. (b) Torsion fatigue

Fig. 9

ð2:5 1Þ ¼ 1:6 ð1 þ 0:27=0:20Þ

(Eq 4)

Applied Stress Effects include variations in mean stress, variable amplitude loading, and the differences between torsion and bending.

Mean stress effects are commonly addressed with modiﬁed Goodman diagrams such as the one shown in Fig. 11 for 8630 steel. Mean stress effects should not differ substantially from notched wrought steels. Variable amplitude fatigue is discussed later with respect to fatigue crack growth. Differences in fatigue behavior for torsional and bending loads are typically described with the assumption of homogeneous and isotropic materials. However, because castings have internal discontinuities, comparison between bending and torsional fatigue is based on a “drilled hole” assumption, where the hole accounts for inherent notch effects from microdiscontinuities (Ref 4). Figure 12 shows a plot of “endurance” ratio in torsion (t) to that in bending (b) for notched, wrought steels. Dashed lines represent the distortional energy theory and maximum shear stress theory lines for t/b = 0.577 and 0.5, respectively. The data points for the “drilled hole” show best agreement to t/b = 3/4 developed in the previous argument. V-notch specimens (55 circumferential) deviate slightly, due to a difference in stress concentration, but are still in better agreement with t/b = 3/4. Cast metals have t/b ratios which vary between 0.7 and 1.05 (Ref 5) with an average of 0.85. As such, the concept of a “drilled hole” method seems reasonable. However, an analysis by Vishnevsky (Table 2) concludes that “all the points except one lie above the Maxwellvon Mises criterion indicating that discontinuities present in torsion are not nearly so damaging as they are in bending.” Table 2 shows that endurance ratios in torsion are lower than those in bending and that the t/b ratio average of 0.85 and 0.79 for the QT and HT condition are also in agreement with the drilled-hole argument. Cyclic Stress-Strain Behavior. Cyclic stressstrain behavior provides a measure of the steadystate cyclic deformation resistance and thus a basis for understanding resistance to crack nucleation. Cyclic stress-strain curves (using the incremental step method and the companion specimen method) are shown in Fig. 13 and 14 for ﬁve cast steels at room temperature and low temperature. The cyclic stress-strain curves are initially identical to the monotonic curves and then become nonlinear at stresses or strains of about 20 to 50% below that for monotonic curves. The monotonic upper yield point found in all but the 8630 steel (Fig. 14) was eliminated under the cyclic conditions. At higher strains, the cyclic curves intersect or converge with the monotonic curves for these four steels. These steels cyclic strain soften at the lower strain levels and then cyclic strain harden at the higher strain levels. The cyclic curves for 8630 steel were always equal to or below the monotonic curve and exhibited cyclic strain softening. A review of available data on wrought steels (Ref 6) indicates that cyclic strain hardening exponent (n0 values) for the cast steels are similar to those for wrought steels. Moreover, the strain hardening and softening behaviors of the cast steels are similar to those for wrought

Fatigue and Fracture Properties of Cast Steels / 203 amplitudes are plotted versus number of applied reversals to fracture, 2Nf, on log-log coordinates in Fig. 15 for four of the ﬁve cast steels at room temperature (with total life at 45 C (50 F) included for comparison). The elastic and plastic components are represented by the following linear relationships:

Goodman diagram for the bending fatigue (R.R. Moore) of 8630 cast steel (normalized and tempered) for determining the fatigue limits in terms of cyclic stress range. The stress range (which is the difference between maximum and minimum stress) for unnotched specimens at zero mean stress is about þ0.4 UTS (line A), while at zero minimum stress (R = 0) for unnotched specimens, the stress range is reduced by the amount shown by line B. The ranges for castings with defects and discontinuities are shown as bands for the maximum and minimum stresses, and the bands include discontinuities such as Incomplete penetration of welds, weld undercut, weld-slag, sound welds, slag inclusion, hot tears, and cast surface porosity. Notched specimens were machined. Notch of R.R. Moore specimen: 60 included angle, 0.0015 ln. (0.0381 mm) root radius

Fig. 11

steels. Cyclic strain hardening was found for cast steels when the tensile/yield strength ratio was greater than 1.5. When this ratio was less than 1.3, cyclic strain softening occurred. At intermediate ratios, mixed cyclic strain softening and hardening were observed. These ratios are about 0.1 higher than those for wrought materials (Ref 7). The monotonic yield and tensile strengths, cyclic stress-strain curves, and cyclic yield strengths at the low temperature were about 10% higher than room-temperature results except for 8630 steel. Low-Cycle Axial Fatigue Behavior. Lowcycle axial fatigue tests were conducted on the ﬁve steels (Ref 2 and 3) at room and low temperature with R = 1 at constant strain amplitudes from 0.0013 to 0.015, which gave

fatigue lives of 102 to 106 cycles. Half-life stable hysteresis loops were used to obtain elastic and plastic strain amplitudes. With Young’s modulus, E, obtained from the ﬁrstquarter cycle, the elastic strain amplitude was calculated from: ee =2 ¼ s=2E

(Eq 5)

and the plastic strain amplitude was calculated as ep =2 ¼ e=2 ee =2

(Eq 6)

¼ e=2 s=2E

(Eq 7)

where De/2 is the controlled total strain amplitude. These elastic, plastic, and total strain

e=2 ¼ ee =2 þ ep =2

(Eq 8)

¼ s0f =E ð2Nf Þb þ e0f ð2Nf Þc

(Eq 9)

where s0f is the fatigue strength coefﬁcient, b is the fatigue strength exponent, e0f is the fatigue ductility coefﬁcient, c is the fatigue ductility exponent, E is Young’s modulus, and 2Nf is the number of reversals to failure. The low-cycle fatigue material properties are listed in Table 3 along with yield and tensile strengths obtained from the monotonic tension tests. Low-cycle fatigue behavior at the low temperature was equal to or better than at room temperature for lives greater than 5 105 reversals. Mixed behavior existed at shorter lives. The fatigue strengths at 106 reversals were from 0 to 30% better at the low temperature. Mixed low-cycle fatigue behavior at low temperatures has also been reported in the literature for wrought steels (Ref 8). In addition, values of low-cycle fatigue material properties for the ﬁve cast steels were within the ranges published for wrought steels (Ref 6). The room-temperature fatigue strengths of the ﬁve cast steels at 106 reversals ranged from 208 MPa (30 ksi) for 0030 steel to 370 MPa (53.7 ksi) for 8630 steel. Room-temperature fatigue strengths were within 30 to 40% of the ultimate tensile strength (32 to 46% at the low temperature). High-Cycle Axial Fatigue Behavior. The high-cycle axial fatigue tests were conducted in the same manner as for the low-cycle tests, except that load control was used instead of strain control and the test frequency was 23 to 37 Hz. Failures ranged from 105 reversals to runouts at 2 107 reversals. High-cycle (107 cycle) fatigue data for 0030 and 8630 cast steels are superimposed with the low-cycle fatigue data in Fig 15(a) and 15(b), respectively, for the 0030 carbon steel and 8630. The solid circles represent total strain amplitudes for the strain-controlled low-cycle fatigue tests, and the solid squares represent total strain amplitudes from the load-controlled high-cycle fatigue tests. The latter strains were calculated by dividing the applied stress amplitude by Young’s modulus because these tests were run in the elastic region. All ﬁve cast steels exhibited typical fatigue limits around 5 106 to 107 reversals for both test temperatures. Values of the fully reversed fatigue limits, Sf, obtained from such plots for all ﬁve cast steels at both test temperatures, are listed in Table 4 along with the fatigue ratios, Sf/Su. The fatigue limits vary from 196 to 293 MPa (28.5 to 42.4 ksi) at room temperature and from 241 to 365 MPa (35 to 53 ksi) at 45 C (50 F). The fatigue

204 / Casting Design and Performance Table 2 “Endurance ratios” in bending and torsion Type of specimen

Fig. 12

Torslon and bending endurance ratios (fatigue endurance/monotonic strength) with drilled holes and notched values compared with theory

Endurance Endurance ratio in bending ratio in torsion

t b ðaÞ

Q&T Cast steel-sound(a) Weld-machine-sound Slag inclusions As welded-sound Weld-slag Weld-undercut Cavities Hot tears

0.310 0.251 0.246 0.241 0.243 0.233 0.117 0.274

0.298 0.230 0.246 0.221 0.184 0.195 0.100 0.146

0.96 0.92 1.00 0.92 0.75 0.84 0.86 0.53

N&T Cast steel-sound(a) Weld-machine-sound As welded-sound Weld-slag Weld-undercut Cavities Slag inclusions Hot tears

0.361 0.352 0.345 0.314 0.280 0.235 0.292 0.245

0.270 0.261 0.250 0.234 0.230 0.195 0.208 0.241

0.75 0.74 0.73 0.75 0.82 0.83 0.71 0.98

Note: “Endurance ratio” using R.R. Moore specimen. Q&T unnotched 0.390, (0.015 in. R) 0.255: Q&T notched (0.001 in. R) 0.174: N&T unnotched 0.395, (0.015 in. R) 0.252; N&T notched t 00 Endurance ratio00 in torsion ð0:001 in: RÞ 0:225: ðaÞ ¼ 00 ; b Endurance ratio00 in bending Average for Q&T ¼ 0:85:

Cyclic stress-strain of cast 8630 steel (NQT, 305 HB average) at room temperature and 45 C (50 F). Source: Ref 2

Fig. 14

85% of the typical rotating bending fatigue ratio of 0.5). Room-temperature and low-temperature cyclic stress-strain curves for two ferritic-pearlitic cast steels (SAE 0050A and 0030 normalized-and-tempered to average hardness of 192 and 137 HB, respectively) and two quenched-and-tempered low-alloy cast steels (C-Mn and Mn-Mo steels at 174 and 206 average HB, respectively). Stable hysterses of cyclic stress-strain were obtained from both companion lest specimens and from a single specimens subjected to incremental strain steps. For the incremental step tests, each strain block contained 79 reversals with the magnitude ranging from 0 to 0.012. The strain rate was constant and equal to 0.01/s. Source: Ref 2

Fig. 13

ratios vary from 0.26 to 0.42 at room temperature and 0.29 to 0.44 at the low temperature. The fatigue ratios for the 0050A and 8630

steels are lower than those for the other cast steels. Axial fatigue ratios for wrought steels range generally from about 0.37 to 0.43 (75 to

Fracture Mechanics The treatment of discontinuities in design has undergone major changes as a result of the wider use of fracture mechanics in the industry. If the plane-strain fracture toughness KIc of a material is known at the temperature of interest, designers can determine the critical combination

Fatigue and Fracture Properties of Cast Steels / 205

Axial strain life and the elastic (ee) and plastic (ep) components of four cast steels. (a) SAE 0030 (NT, 137 HB). (b) C-Mn steel (NQT, 174 HB). (c) Mn-Mo steel (206 HB), and (d) 8630 cast steel at 305 HB. Total strain life at 45 C (50 F) is included. Smooth hourglass specimens with a minimum diameter of 6.4 mm (0.25 in.) over 17.5 mm ð11=16in:Þ length. Source: Ref 2

Fig. 15

Table 3 Monotonic and low-cycle fatigue properties Su Steel

MPa

Sy (0.2%) ksi

MPa

ksi

sf ef

ef’

MPa

Table 4 ratios

sf ksi

MPa

ksi

b

c

Su/Sy

At room temperature 0030 496 72 0050A 787 114 C-Mn 583 85 Mn-Mo 702 102 8630 1144 166

303 415 402 542 995

44 60 58 79 143

0.62 0.21 0.34 0.38 0.35

0.28 0.30 0.15 0.78 0.42

752 869 703 752 1268

109 126 102 109 184

655 1338 869 1117 1936

95 194 126 162 281

0.083 0.127 0.101 0.101 0.121

0.552 0.569 0.514 0.729 0.693

1.6 1.9 1.5 1.3 1.2

At 45 C (50 F) 0030 542 79 0050A 834 121 C-Mn 612 89 Mn-Mo 758 110 8630 1178 171

320 436 464 562 999

46 63 67 81 145

0.36 0.17 0.26 0.36 0.33

0.18 0.32 0.07 0.47 0.35

621 924 634 862 1254

90 134 92 125 182

834 1282 717 1096 1785

121 186 104 159 259

0.089 0.111 0.067 0.090 0.099

0.506 0.582 0.439 0.671 0.659

1.7 1.9 1.3 1.3 1.2

Source: Ref 2

of ﬂaw size and stress required to cause failure in one load application. In addition, designers can calculate the remaining life of a component having a discontinuity, or they can compute the largest acceptable ﬂaw from a knowledge of the crack growth properties (da/dN) and fracture mechanics parameters (K, n, and C) of a material.

Axial fatigue limits and fatigue

Fracture toughness depends on the volume of plastic deformation near the crack. For thin specimens, the plastic region at the crack tip is constrained and is under plane stress conditions, which produces a relatively high toughness, Kc. As section thickness increases, fracture behavior approaches plane-strain conditions, and the toughness approaches the inherent lower limit

At room temperature

At 45 C (50 F)

Fatigue limit (Sf)

Fatigue limit (Sf)

Steel

MPa

ksi

Fatigue ratio, Sf/Su

0030 0050A C-Mn Mn-Mo 8630

196 237 248 232 293

28.5 34.4 36 33.7 42.4

0.40 0.30 0.42 0.33 0.26

MPa

ksi

241 243 255 269 365

35 35.2 37 39 53

Fatigue ratio, Sf/Su

0.44 0.29 0.42 0.35 0.31

Source: Ref 3

of toughness (KIc) for the material. In the absence of suitable plane-strain fracture mechanics data, approximations can be made on the basis of test results obtained from a variety of tensile, impact, and fatigue tests. Plane-Stress Fracture Toughness (Kc) is usually difﬁcult to obtain because of the complexity of the elastic-plastic behavior at the crack tip. Moreover, the slow stable crack extension prior to fracture makes it difﬁcult to determine the maximum stress-intensity level

206 / Casting Design and Performance at crack instability because the crack length is uncertain. The R-curve analysis (Ref 9) is applicable for determining Kc for intermediate-strength materials where the plastic zone size, ry, is small relative to the specimen dimensions, where: ry ¼ ½1=ð2pÞ ðK=Sy Þ2

2 JIc ¼ GIc ¼ KIc =E ð1 v2 Þ

(Eq 12)

(Eq 10)

An R-curve is a continuous record of toughness development in terms of crack growth resistance, KR, plotted against crack extension as the crack is driven under a continuously increasing stress-intensity factor, K. During slow-stable fracturing, the developing crack growth resistance is equal to the applied stress-intensity factor. To have a valid plane-stress fracture toughness test, the uncracked ligament should satisfy the following: ðw aÞ 8 ry

release rate per unit crack extension, which is directly related to the stress intensity factor, K. Thus, linear-elastic fracture criteria based on KIc, GIc, or JIc are identical. For mode I linearelastic plane-strain conditions:

(Eq 11)

where w is the specimen width and a is the crack length (or plastic zone corrected length). Values of ry near unstable crack extension for ﬁve cast steels (Ref 2) ranged from 8 to 18 mm (0.3 to 0.7 in.) and at low temperature ranged from 2.5 to 18 mm (0.1 to 0.7 in.). The validity equation was satisﬁed only with the 8630 steels at low temperature. The localized plasticity was too large for the other test conditions. However, R-curves were established with KR equal to K without a plastic zone correction. Plane-Strain Fracture Toughness. The critical value of toughness (KIc) is reached when specimen thickness is large enough to produce plane-strain conditions at the tip of a crack. As the thickness increases, the small plastic zone relative to the thickness produces a stress parallel to the crack tip as the plastic zone attempts to contract in this direction but is prevented from doing so by the large volume of surrounding elastic material. The resulting high degree of triaxial stresses leads to lower toughness and a predominantly ﬂat fracture. This critical K value, termed plane-strain fracture toughness, or KIc, is the minimum toughness and an inherent property of a material. Typical KIc values for various cast steels are summarized in Table 5. Fracture pﬃﬃﬃﬃof pﬃﬃﬃﬃ toughness steels is rarely below 27 MPa m ð25 ksi in:Þ even at low temperature (Fig. 16). However, this depends on loading rates. If the KIc values are known or reliable estimates are available, the ﬂaw tolerance of a given structure can be computed. Except for special cases the KIc level should be considered as the onset of rapid fracture. J-Integral Method. When alloys have appreciable ductility and low yield strengths, large specimens are required to obtain valid KIc results. In these instances, the path-independent J-integral is used to estimate fracture toughness by extending fracture mechanics concepts from linear-elastic behavior to elastic-plastic behavior (Ref 10). For linear-elastic behavior, the J-integral is identical to G, the strain energy

where E is Young’s modulus and n is Poisson’s ratio. Because JIc characterizes the toughness of materials at or near the onset of crack extension from a preexisting sharp crack, it can be used as a conservative estimate of KIc on specimens that contain appreciable ductility but lack

Table 5 Plane-strain fracture toughness of cast low-alloy steels at room temperature

Heat treatment(a)

Alloy type

Fe-1.25 Cr-0.5Mo Cast 1030 Fe-0.5Cr-0.5 Mo-0.25V Fe-0.5C-1.5Mn Fe-0.5C-1Cr Fe-0.5C Cast 9535 Fe-0.35C-0.6 Ni-0.7Cr0.4Mo Cast 4335 Cast 9536 Fe-0.3C-1Ni1Cr-0.3Mo Cast 4335 Cast 4335 Cast 4335 Cast 4335 Ni-Cr-Mo Cast 4340 Cast 4325 Cast 4325 Cr-Mo Cast 4340

Plane-strain fracture toughness, KIc pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃ MPa m ksi in:

Yield strength, 0.2% offset MPa

ksi

SRANTSR

275

40

88

80

NT NT

303 367

44 53

127 55

116 50

NT NT NT NT NT

412 413 425 614 683

59 60 61 89 99

107 58 65 67 64

98 53 59 61 58

SLQT NT NT

747 108 752 109 787 114

69 59 66

63 54 60

97 105 115 92 98 115 75 104 84 67

87 96 105 84 89 105 82 95 76 61

SLQT SLQT QT QT QT QT QT QT QT QT

814 903 1090 1166 1207 1207 1263 1280 1379 1450

118 131 158 169 175 175 183 186 200 210

(a) SR, stress relieved; A. annealed; N, normalized; T. tempered; Q. quenched; SLQ, slack quenched

sufﬁcient thickness to be tested for KIc. Procedures for JIc tests are described in Ref 11. Valid values for JIc for six test conditions are listed in Table 6. Also listed are values of Jc, both lowest and average, for the invalid JIc tests. The values of KIc and Kc given in Table 6 were obtained from JIc and Jc values (with an assumed Poisson ratio of v ¼ 1/3). The manganesemolybdenum and 1050A cast steels exhibited the highest and lowest fracture toughness, respectively, at both test temperatures. The three martensitic cast steels (carbon-manganese, manganese-molybdenum, and 8630) had better fracture toughness at room temperature compared with the two ferritic-pearlitic cast steels (0030 and 0050A). However, at the low temperature, the 8630 steel had Jc less than that of 0030. The 8630 steel had the largest decrease in fracture toughness as the temperature was lowered. The fracture toughness of the carbon-manganese and manganese-molybdenum steels were lowered only 10 to 15% as the temperature was lowered. Correlation of KIc and Charpy V-notch upper-shelf impact energy was found for the four cast steels with valid JIc results at room temperature, as shown in Fig. 17. A similar

Fig. 16

Plane-strain fracture toughness of ASTM A 216 steel casting (Grade WCC). Source: Ref 1

Table 6 Jlc, Jc, Klc, and KC toughness of cast steels Jlc

Jc in. lb/in.2

kJ/m2

in. lb/in.2

At room temperature 0030 73 0050A ...

415 ...

C-Mn Mn-Mo 8630

479 794 456

... 37(a) 25(b) ... ... ...

... 209(a) 145(b) ... ... ...

49(a) 44(b) 17(a) 14(b) ... ... 38(a) 30(b)

282(a) 215(b) 95(a) 78(b) ... ... 218(a) 174(b)

Steel

kJ/m2

84 139 80

At 45 C (50 F) 0030 ...

...

0050A

...

...

C-Mn Mn-Mo 8630

75 118 ...

428 674 ...

(a) Average value. (b) Lowest value. Source: Ref 2

Klc

Kc

pﬃﬃﬃﬃﬃﬃ ksi in:

pﬃﬃﬃﬃ MPa m

130 ...

118 ...

139 179 135

126 163 123

... 92(a) 77(b) ... ... ...

... 84(a) 70(b) ... ... ...

...

...

...

...

132 166 ... ...

120 151 ... ...

108(a) 93(b) 61(a) 56(b) ... ... 95(a) 85(b)

98(a) 85(b) 56(a) 51(b) ... ... 86(a) 77(b)

pﬃﬃﬃﬃ MPa m

pﬃﬃﬃﬃﬃﬃ ksi in:

Fatigue and Fracture Properties of Cast Steels / 207 correlation was not found for the steels tested at the low temperature. Fatigue Crack Growth Rates (Constant Amplitude). Given the similar notched fatigue strength of cast and wrought steels, similar fatigue crack growth rates should also be expected (Fig. 18). Typical Paris-law parameters for intermediate crack growth rates (Region II) are summarized for ﬁve cast steels in Table 7 (Ref 2). At low temperature (45 C, or 50 F) a decrease in fatigue crack

Correlation between room-temperature KIc, yield strength (Sy), and upper-shelf Charpy V-notch (CVN) toughness for four cast steels. In SI units, (KIc/Sy)2 = 1170 (CVN/Sy 0.022). Source: Ref 2

Fig. 17

growth rates was observed by amounts of from 1.2 to 3 for R 0 and from 1 to 2 for R = ½ at the lower DK values (Ref 2, 3). At the higher DK values, the room- and low-temperature data tended to converge or crossover. As R was increased from 0 to ½, the fatigue crack growth rate increased by factors from 1.5 to 3. This increase was slightly higher than that suggested in the literature for wrought steels (Ref 12). For R 0 at room temperature, the fatigue crack growth rates for the ﬁve cast steels are 2 to 5 times lower than conservative equations proposed by Barsom (Ref 13) for wrought steels. Similar results for similar cast steels were reported by Greenberg et al. (Ref 14) and Kapadia and Imhoff (Ref 15). Thus the fatigue crack growth behavior of cast steels appears to be equal to or better than that of similar wrought steels. Threshold Crack Growth Behavior. Threshold and near-threshold (Region I) fatigue crack growth behavior of the ﬁve cast steels at room and low temperature was determined at R 0 and R = ½ for growth rates between 108 and 1010 m/cycle (4 107 and 4 109 in./cycle). Compact (Chevron notched) specimens were tested per ASTM E647. The incremental load-shedding method was used to obtain the fatigue crack growth rate data in the near-threshold region. The fatigue crack growth rate of 1010 m/cycle (4 109 in./ cycle) was used to deﬁne the threshold DK. A typical plot of the region I and region II data is shown in Fig. 19. It is evident that the threshold DK value for R = ½ is about half that for R 0. All of the steels exhibited similar behavior. The values of DKth are listed in Table 8. The room-temperature R 0 values were higher than the majority of values reported in the literature for wrought steels (Ref 16). The DKth values were higher for the low temperature, except for the 8630 steel at R 0, for which the values were the same. The DKth values showed no correlation with other properties such as yield strength, ultimate strength, or fatigue limit. Variable Amplitude Fatigue Crack Initiation and Growth. Constant-amplitude fatigue behavior does not provide information on

sequential and interactive effects with variable loading. A more realistic comparison of material or component fatigue behavior can be determined from load or strain histories that closely duplicate or simulate actual service history and environment. SAE International (formerly Society of Automotive Engineers) has developed several service spectra indicative of real-life loading (Ref 2). Variable amplitude tests of ﬁve cast steels at room and low temperatures with an as-drilled keyhole notch compact specimen are presented in Ref 2. Eliminating the compression at the four temperatures increased fatigue crack initiation life by factors of 3 to 9, increased short crack growth life by factors of 1 to 4, and increased long crack growth life by factors of 1 to 3. Except for 0050A cast steel, all the average low-temperature fatigue lives for the three different life criteria discussed in Ref 2 were equal to or better than those at room temperature. The increase in crack initiation life was within a factor of 2.5 as was the total fatigue life to fracture. Crack growth life, however, tended to increase as the temperature was lowered and then this trend reversed. The 0050A steel had a continuous decrease in crack growth life for lower temperatures. In general, the three martensitic cast steels (carbon-manganese, manganese-molybdenum, and 8630) had better fatigue resistance with variable amplitude loading at the four temperatures than the ferriticpearlitic cast steels (0030 and 0050A). The 0030, carbon-manganese, manganese-molybdenum, and 8630 steels are satisfactory for lowtemperature conditions, but 0050A is not. Lee (Ref 3, 17) used constant-amplitude material properties developed for the ﬁve steels (Table 6) at room temperature and at 45 C (50 F) and several mathematical models to calculate variable amplitude fatigue crack initiation and growth. Comparison of the calculated and experimentally determined fatigue lives showed that total fatigue life calculations were generally within a factor of þ2 for all tests. However, greater scatter (+7 to 4) was noted for crack initiation life calculations. These results appear to conﬁrm that the models developed for wrought steels can be used for cast steels.

Table 7 Constant-amplitude fatigue crack growth constants for da/dN = C(DK)n R=0C Steel

Fig. 18

Constant amplitude fatigue crack growth of cast and wrought steels at R = 0 and R = 1

R=½C

m/cycle

in./cycle

n

m/cycle

in./cycle

n

At room temperature 0030 3.34 1014 0050A 2.24 1013 C-Mn 7.74 1013 Mn-Mo 1.12 1012 8630 2.63 1012

1.98 1.19 4.18 5.99 1.42

1013 1011 1011 1011 1010

4.33 3.88 3.35 3.28 3.03

1.99 2.34 7.15 1.39

... 1012 1012 1012 1011

1.08 1.25 3.7 7.55

... 1010 1010 1010 1010

... 3.40 3.21 2.89 2.67

At 45 C (-50 F)(a) 0030 3.48 0050A 1.79 C-Mn 5.4 Mn-Mo 9.76 8630 6.38

2.14 1.08 3.11 5.52 3.45

1013 1012 1012 1012 1011

4.72 4.53 4.01 3.84 3.38

8.36 6.36 3.13 1.97

... 1014 1013 1012 1012

4.94 3.49 1.64 1.05

... 1012 1011 1010 1010

... 4.30 3.52 3.04 3.22

1015 1014 1014 1014 1013

(a) At 34 C (-50 F) for 0030. Source: Ref 2

208 / Casting Design and Performance Table 8 Fatigue crack thresholds (DKth) for ﬁve cast steels DKth, R = 0.05 pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃ MPa m ksi in:

Steel

DKth, R = ½ pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃ MPa m ksi in:

At room temperature 0030 9.1 0050A 10.2 C-Mn 8.3 Mn-Mo 8.1 8630 9.4

8.3 9.3 7.6 7.4 8.6

5.3 5.2 3.9 4.1 4.1

4.8 4.7 3.5 3.7 3.7

At 45 C (50 F) 0030 14.2 0050A 12.3 C-Mn 14.4 Mn-Mo 10.7 8630 9.4

12.9 11.2 13.1 9.7 8.6

9.3 6.8 7.1 6.5 5.7

8.5 6.2 6.5 5.9 5.2

Source: Ref 3

Fatigue crack threshold behavior of cast steel (SAE 0030 carbon steel) at room temperature. Source: Ref 3

Fig. 19

3. ACKNOWLEDGMENTS The authors are grateful for the work done by Professor Stephens and his students at the University of Iowa, which formed the basis for much of this chapter. In addition, the authors wish to thank the Steel Founders’ Society of America for sponsoring the original work.

4. 5. 6.

REFERENCES 1. Steel Castings Handbook, 6th ed., Steel Founders’ Society of America and ASM International, 1995 2. R.I. Stephens, “Fatigue and Fracture Toughness of Five Carbon or Low Alloy Cast Steels at Room or Low Climatic

7.

Temperature,” Research Report No. 94A, Steel Founders’ Society of America, Oct 1982 R.I. Stephens, “Fatigue and Fracture Toughness of Five Carbon or Low Alloy Cast Steels at Room or Low Climatic Temperatures (Part II),” Research Report No. 94B, Steel Founders’ Society of America, May 1983 M.R. Mitchell, J. Engr. Materials and Tech., Vol 99, Oct 1977, p 329–343 C. Vishnevsky et al., ASME paper No. 67WA/Met-17, 1967 Technical Report on Fatigue Properties, SAE J1099. Society of Automotive Engineers, Feb 1975 R.W. Smith, M.W. Hirsberg, and S.S. Manson, “Fatigue Behavior of Materials under Strain Cycling in the Low and Intermediate Life Range,” TND-1574, National Aeronautics and Space Administration, April 1963

8. R.I. Stephens, J.H. Chung, and G. Glinka, Low Temperature Fatigue Behavior of Steels—A Review, SAE Trans., Vol 88, 1980 9. “Standard Practice for R-Curve Determination,” E 561, Annual Book of ASTM Standards, ASTM, E.T. Wessel, State of the Art of the WOL Specimen for KIc Fracture Toughness Testing, Eng. Fract. Mech., Vol 1 (No. 1), 1968 10. J.R. Rice. A Path Independent Integral and the Approximate Analysis of Strain Concentration by Notches and Cracks, J. Appl. Mech., Vol 35, June 1968, p 379 11. “Standard Test Method for JIc, A Measure of Fracture Toughness,” E 813, Annual Book of ASTM Standards, ASTM 12. S.T. Rolfe and J.M. Barsom, Fracture and Fatigue Control in Structures—Applications of Fracture Mechanics, Prentice-Hall, 1977 13. J.M. Barsom, Fatigue-Crack Propagation in Steels of Various Yield Strengths, J. Eng. Ind. (Trans. ASME), Series B, No. 4, Nov 1971 14. H.D. Greenberg and W.G. Clark, Jr., A Fracture Mechanics Approach to the Development of Realistic Acceptance Standards for Heavy Walled Steel Castings, Met. Eng. Quart., Aug 1969 15. B.M. Kapadia and E.J. Imhoff, Jr., FatigueCrack Growth in Cast Irons and Cast Steels, Cast Metals for Structural and Pressure Containing Applications, MPC-11, ASME, 1979 16. K. Tanaka, A Correlation of DKth Values with the Exponent, m, in the Equation of Fatigue Crack Growth for Various Steels, Int. J. Fract., Vol 15 (No. 1), Feb 1979 17. S.G. Lee, “Estimating Fatigue Crack Initiation and Propagation Life of Cast Steels Under Variable Loading History,” Ph.D. thesis, University of Iowa, Dec 1982

Casting Design and Performance Pages 209–218

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Fatigue and Fracture Properties of Aluminum Alloy Castings* CASTING QUALITY has steadily improved over the years such that castings have become a viable alternative in structural applications for military and commercial aircraft. The steady improvement in casting quality is driven by several factors. In some cases, casting is the only viable alternative manufacturing method. In addition, the ﬂexibility of the casting can be used to great advantage in reducing overall manufacturing costs, particularly for complex, three-dimensional structures. Competition and cost pressure also have intensiﬁed efforts by aircraft manufacturers to reduce costs through the use of near-net shape manufacturing methods such as casting. However, even with an economic advantage, a casting must meet the same performance criteria as competing design using wrought members and various joining technologies. Such cases are an area of great interest for aircraft structural designers. With improvements in casting quality and the potential economic advantage of casting technology, cast aluminum alloys are being considered for weight and cost reductions of primary aircraft structures. Despite their acknowledged importance, aluminum castings have until recently been relegated to secondary applications in commercial aircraft structure. Castings have been used extensively in engines and systems, for example, in ﬂight controls, wheels, and so forth, but not in many “primary” structural (carrying signiﬁcant ﬂight loads) applications. More recently, however, the use of castings in nontraditional applications has been facilitated by improvements in both casting alloys and processes. Casting producers have invested considerable money in applied research and, where possible, borrowed from the lessons learned by the aluminum producers. This has translated into cleaner castings with reduced levels of internal discontinuities, intermetallics and oxides, and better control of microstructures for improved strength, and fatigue and fracture resistance. In this context, this chapter reviews the fatigue and fracture properties of aluminum alloy castings from the perspective of both

design and manufacturing considerations. Fatigue and fracture resistance is often a limiting factor in the design of primary structures, and thus this chapter provides an overview of the roles played by microstructure, manufacturing processes, and test conditions in determining the fatigue and fracture properties of aluminum casting alloys. Because it is not possible to cover all the work done on all aluminum casting alloys, the following two alloy systems commonly used in aviation and aerospace are emphasized: Alloys A356 and A357/D357 (all-T6), mem-

bers of the dual-phase Al-Si-Mg system, the most commonly speciﬁed Alloy A201-T7 from the single-phase Al-Cu system, with higher mechanical properties but more limited producibility. The role of casting design is also brieﬂy discussed with respect to fatigue and fracture performance.

Casting Design and Manufacture The ability of an engineered component to perform a particular function is governed as much by design as by material properties. Trade studies involving engineered materials (in this case aluminum alloys) sometimes go no further than the properties tables of a reference book. However, attempts to compare diverse design concepts can be misleading. For most design properties, wrought alloys can be found that show advantage over the cast alloy. However, a thorough trade study must take into account the capabilities of the manufacturing processes. For example, the casting process affords the designer a number of unique advantages (Ref 1). In terms of structural shapes with a high degree of complexity, casting allows greater ﬂexibility in design. Improved fatigue design is also possible by virtue of: Continuous load paths (intersecting members

are continuous; in a built-up design, the

* Reprinted from ASM Handbook, Volume 19, Fatigue and Fracture, ASM Handbook Committee, p 813–822

engineer is often faced with breaking one member or another into two and then adding splice details to reconnect them during assembly) “Soft” section transitions Fillet and corner radii can be tailored to reduce stress concentrations Fewer rivet/fastener holes (potential crack starters) Fewer lap joints (potential corrosion sites)

The engineer is often able to produce a cast design that weighs no more than the traditional design (e.g., Ref 2). In fact, a modest weight reduction is sometimes achieved (Ref 1, 3, 4). Cost reduction is achieved by virtue of reduced part count, assembly labor, and related costs. Typical cost reductions are 35% or more when replacing a complex, three-dimensional builtup assembly. With respect to casting processes, the great majority of development and production programs have used the so-called “sand-chill composite” mold process (hereafter referred to simply as “sand” process for brevity). Recent work by investment casting suppliers has, however, enabled a few companies to produce aluminum alloy castings (for example, D357) with tensile properties on a par with sand castings. Some European aircraft manufacturers apparently prefer castings made by the investment process. They are, in fact, somewhat ahead of their American counterparts in terms of applications. For example, the ﬁrst structural casting went into service on an Airbus aircraft circa 1985 (Ref 1). Weld Rework on Castings and its Effect on Fatigue and Fracture. Foundries often perform modest amounts of welding on castings (always prior to heat treatment) to correct surface defects. This action returns the casting to the “drawing speciﬁed” condition and is thus called “rework” (as opposed to “repair,” which is discussed in the next section). The static properties of weld-reworked castings have been reasonably well studied, but few references on fatigue and fracture are available. With respect to

210 / Casting Design and Performance alloys, only the Al-Si-Mg system has been studied in depth. The Al-Cu system, being inherently hot short, is welded less frequently and is not as well characterized. The majority of work seems to show that fatigue life is not degraded when welding is properly performed. Apodaca et al. (Ref 5) studied the properties of weld-reworked F-5 pylon castings. In these smooth-bar, axial tests no signiﬁcant differences were noted between results for welded and unwelded (parent) material. Ozelton et al. (Ref 6) found similar results. Mitchell et al. (Ref 2) reported slightly longer life for weld-reworked A357-T6 specimens tested with R = +0.02 and 260 MPa (38 ksi) maximum stress: 48,388 cycles average for welded and 41,880 for parent metal. However, Oswalt et al. (Ref 7) found different results: slightly longer life for A357-T6 parent metal for stresses below about 100 MPa (15 ksi). However, comparisons with other works are difﬁcult at best, due to differences in casting properties, quality, welding procedures, and fatigue testing procedures. For fracture properties, the only readily available work (Ozelton et al., Ref 6) found that weld rework had no apparent effect on crack growth rates in D357-T6. Fracture toughness, on the other hand, was actually somewhat higher for the welded material. In general, welding produces a very ﬁne micro-structure in the fusion zone of Al-Si-Mg castings. It also tends to produce a distribution of very ﬁne gas pores, sometimes in excess of that found in the adjacent parent metal. The interactions of these features in determining fatigue life has not been studied in depth. Weld repair differs from rework in that post-weld heat treatment is not done; it is sometimes employed by aircraft maintenance crews to render damaged castings usable. Of course, weldment tensile properties are greatly reduced as a result. For example, in alloy A357-T6, yield strength may be reduced by 50% or more. Consequently, weld repairs are performed only in cases where service stresses are low. A review of the literature turned up two cases in which weld repair was performed on fatigue test articles: Hanson et al. (Ref 8) performed fatigue tests

on two alloy A201-T7 spoilers with weld repairs. They concluded that weld repair procedures were adequate for ﬁeld repair functions. Ozelton et al. (Ref 6) performed weld repair on a precracked open-hole fatigue specimen to observe crack growth behavior in the weld. Prior to welding, the specimen was subjected to two lifetimes of a modiﬁed F-18 lower wing root spectrum. The resulting 0.09 in. crack was welded, the hole was reamed, and the test was continued to failure. Effect of Hot Isostatic Pressing (HIP). It is well known that the elimination of internal voids by HIPing has a generally favorable

effect on the smooth-bar fatigue life of castings. Both Al-Si-Mg and Al-Cu cast alloys show similar results. However, there is some discord in the literature. For example, Kennerknecht et al. (Ref 9) found that HIPing did not have a signiﬁcant effect on fatigue life in notched specimens (Kt = 1.5) and open-hole specimens (Kt = 3.0) in alloy D357-T6. They did, however, report slightly lower crack growth rates in HIPed material. In addition, Rading et al. (Ref 10) compared the fatigue crack growth rates of alloy A206-T7 (Al-4.75% Cu) with varying degrees of porosity. They concluded that HIPing did little to improve da/dn, even though it reduced porosity from 4 to 0.4%. An explanation was offered to the effect that HIPing closes pores but cannot “heal” them due to the presence of oxides on the pore surfaces. However, no metallographic evidence was offered in support of this explanation. If this did in fact occur, then one would expect a reduction in fatigue life due to early crack initiation at the hypothetical collapsed, unhealed pores; however, the available fatigue data indicate the opposite (e.g., Ref 9). Effect of Casting Discontinuities on Fatigue and Fracture. Casting properties depend greatly on the effect of various discontinuities, but it can be difﬁcult to quantify the effect of a single type of discontinuity. This is so because the processing steps taken to produce castings with ﬁne microstructure also tend to minimize the presence of unwanted actors— dross (oxide inclusions), gas porosity, microshrinkage, and the like. This being the case, some investigators have resorted to using nonstandard practices in an attempt to introduce various quantities of a certain discontinuity type into the test castings. Data produced from such programs are, unfortunately, of questionable relevance from an engineering perspective. However, few other data on this subject are to be found and so are presented here only for the purposes of illustration. The effects of two often-discussed discontinuities, gas porosity and dross, are brieﬂy reviewed. Gas Porosity. A number of investigators have studied the effects of gas porosity, possibly because it is relatively easy to quantify by metallographic or radiographic inspection (although standards for the latter method are semiquantitative at best). Brieﬂy, the effects of gas porosity on the properties of Al-Si-Mg alloys are as follows: Fatigue life is reduced as porosity level is

increased, based on a number of investigations (such as Ref 6) and discussed later in the section “Effects on Fatigue Crack Initiation” in this article. Fatigue crack growth rates are not greatly affected (Ref 6), although investigators (Ref 9) reported that alloy D357-T6, when subjected to porosity-eliminating HIPing, showed reduced rates. Fracture toughness is not greatly affected (Ref 6, 11).

Dross. The role of dross has been studied to a lesser extent than that of gas porosity, possibly because dross is much more difﬁcult to detect and quantify. Also, the presence of dross and gas porosity are related. Aluminum oxide inclusions have been reported to serve as nucleation sites for gas pores (Ref 12). Therefore, at any given level of dissolved hydrogen in a melt, castings poured with “dirty” metal will contain more gas porosity than those poured with “clean” metal. As a result, it is somewhat difﬁcult to characterize the effects of the two independent of one another. Despite these difﬁculties, it is clear that dross can adversely affect alloy strength (Ref 13). Ozelton et al. (Ref 6) attempted to introduce dross (“less dense foreign material”) into alloy D357-T6 test plates by pouring molten metal down the mold risers rather than the downsprue. Castings with foreign material corresponding with ASTM E155 grades A/B, B, and C were more susceptible to crack initiation based on results of stress-life tests, as would be expected. By way of contrast, the investigators found that crack growth rates and fracture toughness levels were not signiﬁcantly affected (Ref 6).

Al-Si-Mg Cast Alloys Cast aluminum alloys A356 and A357/D357 (Tables 1 and 2) are part of the Al-Si-Mg alloy family and are the most commonly speciﬁed cast aluminum alloys for aircraft applications. This alloy system is a favorite of foundrymen and their customers, as the ﬂuidity provided by silicon additions allows casting of complex, thin-walled components. High-strength (ultimate tensile strength of 345 MPa, or 50 ksi) alloy A357-T6 castings were in production in the United States by the late 1960s. The alloy was used in the F-5 Northrop ﬁghter wing pylon, which is still remarkable in size, complexity, uniformity of mechanical properties and production quantity (more than 15,000 in a twenty-year span) (Ref 2). The principles of controlled solidiﬁcation were used to obtain a ﬁne microstructure and a high degree of soundness in an alloy of tightly controlled composition. The capability to produce high-strength castings with thin walls (less than 2.5 mm, or 0.1 in.) was developed over several decades with some initial work in the mid-1960s. Work by Goehler et al. (Ref 14) led to several new applications, for example, the alloy A357 Krueger ﬂap for the Boeing model 737 aircraft. Several signiﬁcant programs from the 1970s demonstrated both the technical viability and economic beneﬁts of thin-wall castings in aircraft components: Vought/USAF A-7 cast aluminum spoiler

(1976, Ref 8)

Boeing/USAF YC-14 bulkhead (“CAST”)

program (1976–1979, Ref 3)

Fatigue and Fracture Properties of Aluminum Alloy Castings / 211 Table 1

Compositions of key aerospace aluminum casting alloys Composition, wt%

Alloys

A356(a) Min Max A357(a) Min Max D357(b) Min Max A201(a) Min Max B201(c) Min Max

Si

Mg

Cu

Mn

Ag

Fe

6.5 7.5

0.25 0.45

... 0.20

... 0.10

... ...

... 0.20

6.5 7.5

0.40 0.70

... 0.20

... 0.10

... ...

6.5 7.5

0.55 0.6

... ...

... 0.10

... 0.05

0.15 0.35

4.0 5.0

... 0.05

0.20 0.30

4.5 5.0

Be

Ti

Zn

Others each

Others total

... ...

... 0.20

... 0.10

... 0.05

... 0.15

... 0.20

0.04 0.07

0.04 0.20

... 0.10

... 0.05

... 0.15

... ...

... 0.12

0.04 0.07

0.10 0.20

... ...

... 0.05

... 0.15

0.20 0.40

0.40 1.0

... 0.10

... ...

0.15 0.35

... ...

... 0.03

... 0.15

0.20 0.50

0.40 0.80

... 0.05

... ...

0.15 0.35

... ...

... 0.05

... 0.15

(a) Per MIL-A-21180. (b) Per AMS 4241. (c) Per AMS 4242

Table 2

Minimum tensile properties of selected aluminum casting alloys Tensile properties Ultimate tensile strength

Tensile yield strength

Alloy

Strength class

MPa

ksi

MPa

ksi

Elongation, %

A356-T6(a)

1 2 1 2 Nondesignated Designated 1 2 Nondeslgnated Designated

262 275 310 345 310 345 413 413 385 413

38.0 40.0 45.0 50.0 45 50 60.0 60.0 56 60

193 206 241 275 248 275 345 345 330 345

28.0 30.0 35.0 40.0 36 40 50.0 50.0 48 50

5 3 3 5 2 3 3 5 2 3

A357-T6(a) D357-T6(b) A201-T7(a) B201-T7(c)

(a) Per MIL-A-21180. (b) Per AMS 4241. (c) Per AMS 4242

In both programs, full-scale cast test articles were subjected to static, fatigue, and damage tolerance tests. Both test articles met aircraft performance speciﬁcations. The CAST program in particular is viewed by some investigators (Ref 4) as the technical foundation for aircraft structural castings. The U.S. Air Force/Boeing Air-Launched Cruise Missile AGM-86B took advantage of the results of the CAST program to produce the entire missile body. The cast design enabled the Air Force to realize tremendous savings in both cost and schedule. Investigations at Boeing (Ref 15) and Northrop (Ref 6) led to the further reﬁnement of alloy A357 and the associated production controls. The product, registered as alloy D357 (AMS 4241), found application in the F-16 air inlet duct. A cast replacement duct design enabled the contractor to widen the throat of the duct while maintaining the existing external proﬁle (made necessary by a switch to a more powerful engine). Destructive test data generated during production of the inlet and other castings were used to formulate statistically based design allowables for the alloy (Ref 16)—a unique feature of alloy D357. The allowables were the product of tighter restrictions on alloy composition, nondestructive inspection criteria, and production practices. Since 1986, the alloy

has been speciﬁed for other critical castings in both military and civil aircraft.

Microstructure and Properties The Al-Si-Mg alloys are two-phase systems, where magnesium provides age-hardening, and silicon is present in the microstructure of the age-hardened alloy as both interdendritic eutectic and in the Mg2Si strengthening precipitate. In general, the microstructural features that most strongly affect mechanical properties are: Dendrite-arm spacing Size and distribution of second-phase parti-

cles and inclusions Both the primary aluminum dendrite size and the size and morphology of the silicon eutectic depend heavily on processing parameters such as solidiﬁcation rate, the presence of chemical modiﬁers, and solution heat treatment. The size and distribution of Mg2Si precipitate particles depend on quench rate, magnesium content, and aging treatment. Ideally, these features would be the only considerations in the design of casting processes. However, because of the physical and chemical

properties of molten aluminum alloys, the founders and users must also take into account such phenomena as shrinkage, gas porosity, and nonmetallic inclusions usually from dross (i.e., thin aluminum oxide ﬁlms formed during melting and pouring). In addition, if undesirable trace elements (especially iron) are present, large acicular intermetallic compounds may also be present. Figures 1(a) and (b) illustrate two extremes of microstructure. All these features inﬂuence the static, fatigue, and fracture properties of the alloys, and a tremendous amount of work by producers and users has been done to characterize and improve the product; a few examples are given in Ref 3, 6, and 15. It is beyond the scope of this paper to summarize them here. However, the key role of microstructural ﬁneness is summarized here due to its well-known beneﬁcial inﬂuence on strength, toughness, and overall fatigue and fracture resistance. Premium (high-strength, high-toughness) castings combine a number of conventional casting procedures in a selective manner to provide premium strength, soundness, dimensional tolerances, surface ﬁnish, or a combination of these characteristics. As a result, premium engineered castings typically have somewhat higher fatigue resistance than conventional castings. Table 3 gives typical fatigue limits on various premium castings. The production of premium-quality castings is an example of understanding and using the solidiﬁcation process to good advantage. Depending on the casting process, various methods are used to engineer the rates and directions of molten metal solidiﬁcation within the mold. For example, in the sand-chill composite process, metallic chills placed at selected locations within the mold provide greatly enhanced solidiﬁcation rates. In the shell investment casting process, since the ceramic shell molds are routinely preheated prior to pour, metallic chills are of limited value. Alternative heat extraction methods such as the use of cryogenic or other ﬂuids are therefore used to promote more rapid solidiﬁcation.

Effects of Microstructural Fineness Measurement of Microstructural Fineness. Microstructural ﬁneness is related to the size of the primary aluminum dendrites and the interdendritic silicon eutectic phase. Fineness is measured by quantitative systems such as the dendrite arm spacing (DAS) (Ref 17) or the cell count method (Ref 15). Silicon particles are usually characterized by an aspect ratio (Ref 15). When considering effects on properties, it is somewhat difﬁcult to consider the two separately because the steps taken to achieve ﬁne dendrite size (enhanced solidiﬁcation) also modify the silicon, that is, particle size and aspect ratio are reduced. The two will therefore be treated together. Effect on Fatigue Life. The beneﬁcial effect of reﬁned microstructures for improved fatigue

212 / Casting Design and Performance

Fig. 1

Typical microstructure of Al-Si-Mg casting. (a) Photomicrograph showing coarse dendritic structure and acicular silicon eutectic (investment cast 354-F alloy at 250). (b) Photomicrograph showing ﬁne dendritic structure and heavily modiﬁed silicon eutectic (200)

Table 3

Typical mechanical properties of premium-quality aluminum alloy castings and elevated-temperature aluminum casting alloys Ultimate tensile strength

Alloy and temper

Hardness, HB(a)

Compressive yield strength

Tensile yield strength

MPa

ksi

MPa

ksi

495 445 420 470 380 317 283 360

72 65 61 68 55 46 41 52

448 405 330 407 283 235 207 290

65 59 48 59 41 34 30 42

27 41 27 32 30 31 30 32 42

138 275 125 207 160 ... 160 130 185

Elongation in 50 mm (2 in.), %

Shear strength

Fatigue strength(b)

MPa

ksi

MPa

ksi

MPa

ksi

6 6 4 6 6 6 10 8

... ... ... ... ... ... ... ...

... ... ... ... ... ... ... ...

... ... ... ... ... ... ... ...

... ... ... ... ... ... ... ...

97 90 86 75 135(d) 97 90 90

14 13 12.5 11 19.5(d) 14 13 13

20 40 18 30 23 ... 23 19 27

1 <0.5 1 0.5 2 ... 2 2.5 4.0

... ... ... ... 165 ... 200 ... 180

... ... ... ... 24 ... 29 ... 26

... ... 145 180 165 ... 70 ... 193

... ... 21 26 24 ... 10 ... 28

... ... 55 75 72 ... 70 ... ...

... ... 8 11 10.5 ... 10 ... ...

35 ... 34 42 28 28 43 28

... ... 1 0.5 0.5 1 0.5 0.5

295 ... ... ... 193 200 193 193

43 ... ... ... 28 29 28 28

207 ... 207 240 193 193 248 193

30 ... 30 35 28 28 36 28

59 ... 72 65 90 90 ... ...

8.5 ... 10.5 9.5 13 13 ... ...

Premium-quality castings(c) A201.0-T7 A206.0-T7 224.0-T7 249.0-T7 354.0-T6 C355.0-T6 A356.0-T6 A357.0-T6

... ... ... ... ... ... ... ...

Piston and elevated-temperature sand cast alloys 222.0-T2 222.0-T6 242.0-T21 242.0-T571 242.0-T77 A242.0-T75 243.0 328.0-F 328.0-T6

80 115 70 85 75 ... 95 ... 85

185 283 185 220 207 215 207 220 290

Piston and elevated-temperature alloys (permanent mold castings) 222.0-T55 222.0-T65 242.0-T571 242.0-T61 332.0-T551 332.0-T5 336.0-T65 336.0-T551

115 ... 105 110 105 105 125 105

255 ... 275 325 248 248 325 248

37 ... 40 47 36 36 47 36

240 ... 235 290 193 193 295 193

(a) 10 mm (0.4 in.) ball with 500 kgf (1100 lbf) load. (b) Rotating beam test at 5 108 cycles. (c) Typical values of premium-quality castings are not minimums and are not classiﬁed by the area from which the specimen is cut. (d) Fatigue strength for 106 cycles

life has been known for decades, and one of the best demonstrations of this is an early work (Ref 18) that shows with good clarity the systematic variations of microstructure, tensile, and fatigue properties with respect to distance from a chill in alloy 356 plus beryllium test casting. The investigator showed that both fatigue strength and ultimate tensile strength decreased with increasing distance from the chill Figure 2 shows the stress-life relationships for specimens taken at various distances from the chill; the microstructural explanation (or part of it in terms of dendrite cell size) is shown

in Fig. 3—at any given maximum stress level, fatigue life is related to microstructural ﬁneness in terms of the mean dendrite cell size. Other works show similar trends. For example, Fig. 4 (Ref 19) shows fatigue life as a function of DAS for alloy A357-T6. Ozelton et al. (Ref 6) studied the effects of “nonoptimum” processing on the fatigue life of alloy D357-T6. Smooth specimens with coarser microstructure had shorter fatigue lives, but the effect was less pronounced for notched fatigue specimens (Fig. 5) Effects on Fatigue Crack Initiation. Many investigators (e.g., Ref 6, 19–23) cite the role

of various discontinuities—usually gas porosity, microshrinkage, or dross—in fatigue crack initiation. The effect of microstructural reﬁnement on fatigue life may be explained in part by the fact that rapid solidiﬁcation produces not only ﬁne dendrites and highly modiﬁed silicon, but also tends to distribute any available dissolved hydrogen as numerous, minute, rounded pores. Slower solidiﬁcation, on the other hand, provides time needed for the formation of larger gas pores. A similar trend applies to microshrinkage. Larger pore/cavity size is generally accepted as providing sites for early

Fatigue and Fracture Properties of Aluminum Alloy Castings / 213 gave favorable results; however, others such as Ozelton et al. (Ref 6) report little or no apparent effect. Figure 6 shows da/dn versus DK curves for alloy D357-T6 with typical microstructures shown in Fig. 7(a) and (b). Note the relatively large dendrites and the rodlike silicon in the material labeled “nonoptimum.” The investigators concluded that this had no effect on da/dn; however, Fig. 6 seems to suggest higher rates for the “nonoptimum” material pﬃﬃﬃﬃ at higher DK levels (above 10 ksi in.). Lee et al. (Ref 25) found that the fatigue crack growth characteristics of cast alloy A356-T6 were a function of a variety of factors; prominent among them was the effect of silicon particle size and morphology (degree of modiﬁcation). They found different mechanisms at work in cast alloy with unmodiﬁed versus strongly modiﬁed silicon: Crack

Fig. 2

Stress-life curves for specimens taken at various distances from a cast Iron chill embedded in the mold. Alloy 356 with beryllium addition. From Ref 18

Fatigue life at 165 MPa (24 ksi) maximum stress for alloy A357-T6 as a function of dendrite arm spacing. From Ref 19

Fig. 4

Fatigue life as a function of stress level and mean dendrite cell size for alloy 356 plus beryllium. 1, 35 þ 2.5 ksi; 2, 30 þ 2.5 ksi; 3, 25 þ 2.5 ksi; 4, 20 þ 2.5 ksi; 5, 15 þ 2.5 ksi. From Ref 18

Fig. 3

crack initiation and, hence, shorter fatigue life. One investigator (Ref 19) concluded that casting discontinuities of size similar to the silicon eutectic particles in alloy A357-T6 will not have a detrimental effect on fatigue strength. When specimens with stress concentration factor (Kt) greater than 1.0 are considered, initiation is often found to occur at geometric

features (holes/notches) as opposed to processrelated discontinuities. Comparisons of smooth and open-hole specimens machined from cast alloy D357-T6 test plates (Kt = 1.0 and 3.0, respectively) by the author (Ref 24) showed that cracks initiated at small dross (aluminum oxide) inclusions at the polished surfaces of Kt = 1.0 specimens. In the latter type, initiation occurred at the hole shoulder or at “riﬂe” marks in the bore. Effects on Crack Growth Rates. Results are divided, to some extent, with respect to the effect of microstructural reﬁnement on crack growth in Al-Si-Mg alloys. At least one work (Ref 21) concluded that coarser microstructure

growth in unmodiﬁed material occurred by cleavage/cracking of silicon particles. The tendency for particle cracking increases with particle size (aspect ratio). In alloy with modiﬁed silicon, decohesion of the silicon particles and the matrix occurs. It is believed that this method of creating a new interface requires more energy than does particle cracking (which may explain the slower crack growth rates in the modiﬁed material). Effect on Fracture Toughness. Several examples illustrate the role between reﬁned microstructure and improved fracture toughness. For example, Ozelton et al. (Ref 6) found that alloy D357-T6 with coarse microstructure had signiﬁcantly lower plane strain pﬃﬃﬃﬃﬃﬃfracture toughness (KIc in the 17 to 18 ksi in: range) than material with ﬁne structure (KQ about 24 pﬃﬃﬃﬃﬃ ﬃ ksi in:). Other examples are demonstrated for dynamic fracture toughness (Ref 11) and Charpy impact energy of alloy A356-T6 subjected to different solidiﬁcation rates, chemical modiﬁcation, and various solution treatment times (Ref 12). As expected, processing conditions that yielded ﬁne dendrites and silicon particles also imparted relatively high impact energy levels. The signiﬁcance of ﬁne microstructure is a key factor for improved fatigue life, slower crack growth, and higher fracture toughness. However, other metallurgical factors have important effects as well. Yield strength, for example, is determined primarily by magnesium content and heat treatment (Ref 3, 15, 26). Oswalt and Maloit (Ref 26) characterized the yield strength-toughness relationship for alloy D357-T6. Figure 8 shows the effects of aging at 160 C (325 F) up to 12 h; data from Ref 11 are shown superimposed. The investigators also found a correlation between the notch yield ratio (NYR = notch tensile strength divided by tensile yield strength) and fracture toughness; a similar relationship for alloy A357 is reported in Ref 13.

214 / Casting Design and Performance

Fig. 5

Stress-life fatigue data for D357-T6 with nonoptimum (relatively coarse) microstructure. (a) Smooth (Kt = 1.0) specimens. (b) Notched (Kt = 3.0) specimens. From Ref 6

effects of iron. It does so by modifying the intermetallic morphology from that of a platelet to an indistinct, more rounded shape. Tan et al. (Ref 28) studied the effects of beryllium additions on the fracture toughness of permanentmold cast alloy A357 with relatively high and low iron levels. They found the following: The fracture toughness of beryllium-free

alloys was reduced pﬃﬃﬃﬃ slightly (KIc from about 25 to 21 MPa m by an increase from 0.01 to 0.15% Fe. Beryllium is effective in mitigating the effects of iron in alloys with 0.15% Fe. Beryllium has no apparent beneﬁcial effect on low-iron (0.01%) alloy.

Al-Cu Cast Alloys Fatigue crack growth rate of alloy D357-T6 with nonoptimum (relatively coarse) microstructure. From Ref 6

Fig. 6

Composition Effects Residual Iron. Several investigators have studied the effects of residual elements, especially iron, on the properties of Al-Si-Mg casting alloys. Iron forms an intermetallic beta phase with an acicular platelet morphology in many aluminum alloys (both cast and wrought). These platelets act as stress raisers in the alloys and are known to have adverse effects on mechanical properties and fabrication characteristics. And, as it turns out, both fatigue life and fracture toughness can be affected as well. Wickberg et al. (Ref 27) found that an increase in the nominal iron content of alloy A356-T6 from 0.2 to 0.4% resulted in a noticeable reduction in fatigue life—refer to Fig. 9. Beryllium was originally included in alloy A357-T6 in the composition to mitigate the

The Al-Cu (and, in some cases, Ag) alloy system offers higher strength than the Al-SiMg alloys—at the expense of producibility. Because of the tendency of the Al-Cu system toward hot shortness, only relatively simple casting designs can usually be produced. Nonetheless, alloys such as A201 are used in various applications such as jet engine front frames. Alloy composition and mechanical property minimum values are given in Tables 1 and 2, respectively. Microstructure and Mechanical Properties. Because Al-Cu alloys lack a pronounced second phase, the effects of microstructure on tensile properties are more subtle. Ozelton et al. (Ref 6) characterized the effects of grain size on the properties of HIPed alloy A201T7. HIPing reduces internal porosity and microshrinkage to very low levels; when this was done, it was found that grain size, and hence solidiﬁcation rate, had little effect on tensile properties. On the other hand, in the absence of HIPing, solidiﬁcation rate determines the concentration of internal voids in alloy A201

(Ref 29). In this case, tensile elongation is higher under conditions of rapid solidiﬁcation. Fatigue life is similarly affected by solidiﬁcation rates and its consequences on microstructural ﬁneness and the presence of voids. In general, fatigue life drops with distance from the chill, as noted for all conditions of alloy KO-1 (later designated A201) in Ref 29. The decrease is attributed to the increase in microporosity with distance from the chill. Ozelton et al. (Ref 6) found that grain size had no apparent effect on the fatigue life of HIPed alloy B201-T7 (B201 is a modiﬁed version of alloy A201). Likewise, Ozelton et al. (Ref 6) found that grain size had no apparent effect on the crack growth rate (Fig. 10) or fracture toughness in HIPed alloy B201. Chien et al. (Ref 29), on the other hand, found that crack growth rates in alloy KO-1 were slower for coarse-grained alloy (see Fig. 11). Despite the slower da/dN, fatigue life in coarse-grained KO-1 was relatively short (Ref 29), indicating that initiation occurred earlier (probably the result of higher microshrinkage content). Thus, it appears that grain size has only an indirect effect on crack growth in single-phase alloys such as A201. Effect of Discontinuities. The presence of internal discontinuities is a signiﬁcant factor in determining fatigue and fracture properties. Several studies on the effects of HIPing on alloy A201-T7 (Ref 30 and 31) provide relevant information. For example, both investigations reported improved fatigue life in HIPed alloy, such as shown in Fig. 12. However, Mocarski et al. (Ref 30) reported no effect on crack growth rates (see Fig. 13). They did ﬁnd a modest increase (about 6%) in plane strain fracture toughness as a result of HIPing. Environmental Effects. Fatigue and fracture properties tests are expensive, and tests involving special conditions (such as high/low temperature, the presence of corrosives, etc.) add to the costs. This may explain, in part, the limited data on this subject as noted below.

Fatigue and Fracture Properties of Aluminum Alloy Castings / 215

Fig. 7

Typical microstructures for alloy D357-T6. (a) Relatively ﬁne. As-polished. 50. (b) Relatively coarse (nonoptimum). 50. From Ref 6

Fig. 9

Fig. 8

Inﬂuence of alloy A356-T6 iron content on fatigue life. From Ref 27

Yield strength-toughness relationship for alloy D357-T6 aged at 160 C (325 F) (0 to 12 h). NYR, notch tensile strength divided by tensile yield strength. From Ref 26

Temperature Effects. The author is aware of two substantive studies in this area. Tirpak (Ref 32) documented fatigue and fracture properties for alloys A357-T6 and A201-T7 from ambient temperature to 200 C (392 F); fatigue relationships are shown in Fig. 14 and 15. Misra et al (Ref 33) showed that fracture toughness and crack growth rates for alloy

A357-T6 are not affected by temperature over the range from ambient to 196 C (320 F). Humidity Effects on Fatigue Crack Growth. Doyle et al. (Ref 34) reported that high (but unspeciﬁed) humidity level increased the crack growth rate in specimens machined from alloy A357-T6 F-16 vertical stabilizer frames. However, when the data are compared with that

Fig. 10

Crack growth rate curves for coarse- and ﬁnegrained HIPed alloy B201-T7. From Ref 6

216 / Casting Design and Performance

Fig. 14

Stress-life fatigue data at three test temperatures for alloy A201-T7. From Ref 32

Fig. 15

Stress-life fatigue data at three test temperatures for alloy A357-T6. From Ref 32

Crack growth rate curves for alloy KO-1 (now A201) as a function of distance from mold chill. From Ref 29

Fig. 11

Fig. 12

Fig. 13

Stress-life fatigue data for alloy A201-T7 castings with and without HIPing. From Ref 30

Fatigue crack growth rates for alloy A201-T7 castings with and without HIPing. From Ref 30

from the CAST program (Ref 3), it is seen to be within the scatter band of tests conducted in normal laboratory air (see Fig. 16). Corrosives (Salt) Effects. Only two references were found regarding corrosion effects in salt environments. Czyryca (Ref 35) studied the corrosion-fatigue behavior in the low-stress, long-life regime for alloy A356-T6. Rotating beam tests (R = 1) conducted in air and seawater produced failure stress levels of 10 and 5 ksi, respectively, at 100 million cycles. Metallographic examinations showed fatigue cracks originating at pits caused by selective phase attack (corrosion of the silicon eutectic). Corrosion-fatigue cracks followed a path through the interdendritic region. Mocarski et al. (Ref 30) performed crack growth tests on alloy A201-T7 specimens in laboratory air in 3.5% saltwater. They found no signiﬁcant difference between the two sets of results. Effects of Loading Conditions. Materials screening tests are usually performed using constant-amplitude loading conditions. When evaluating the suitability of metallic materials for military aircraft applications, several

investigators have used custom-tailored loading conditions (“spectrum fatigue”). Tirpak and Ruschau (Ref 36) studied the responses of alloys 7050-T76, A201-T7, and A357-T6 to a tension-compression load history chosen to represent that of a lower wing location on a ﬁghter aircraft. They found that A357-T6 had the lowest crack growth rates over the range of stress intensities tested. Figure 17 summarizes the data. Under spectrum fatigue loading, the results differ from those under constant amplitude, tension-only loading. For purposes of comparison, refer back to Fig. 6; note that the curve for the wrought reference material (7075-T7351) crosses over the D357 curve in this case. This crossover was not found by Tirpak et al. for spectrum loading conditions.

Component Tests A number of programs involving static, fatigue, and damage tolerance testing of fullscale cast test articles are documented in the literature. A brief review of several prominent programs is offered below.

Fatigue and Fracture Properties of Aluminum Alloy Castings / 217

3. 4.

5.

6.

7. Fatigue crack growth rates for alloy A357-T6 tested in normal laboratory air and in “high humidity” air. From Ref 33

Fig. 16

8. Fatigue crack growth rates of alloys 7050-T76, A201-T7, and A357-T6 under spectrum loading conditions. From Ref 35

Fig. 17 Vought A-7 cast spoiler (Ref 8): Ten alloy

A201-T7 cast components were subjected to static and/or fatigue tests. A baseline built-up spoiler was also tested. The cast components met all performance criteria— static, fatigue, and residual strength. “CAST” Program YC-14 forward bulkhead (Ref 3): Static, fatigue, and damage tolerance tests were performed on a full-scale bulkhead. The alloy A357-T6 test article met all program requirements. USAF/General Dynamics F-16 vertical stabilizer substructure (Ref 34): The substructure was assembled into an F-16 vertical stabilizer, complete with graphite-polyimide composite skins, and subjected to static, fatigue, and damage tolerance tests. The assembled stabilizer met all program requirements. Aerospatiale A320 access door substructure: A conventional built-up access door and a door with cast substructure were subjected to cyclic pressure tests. Crack initiation in the conventional door occurred at 105 cycles; initiation occurred at 3 105 cycles in the door with cast substructure. These results were achieved using cast alloys with relatively modest properties. Thus, the inﬂuence of design on component performance is evident.

Weld rework of Al-Si-Mg castings: Knowl-

edge of the interaction between the very ﬁne weld metal microstructure and gas porosity, and how it affects weld fatigue life, would be useful. In addition, the development of optimized welding procedures for alloys such as D357-T6 would be of great value to both casting suppliers and users. Fatigue life modeling: A study of fatigue life of alloy D357-T6 as a function of distance from a sand mold chill, as was done by Bailey (Ref 18) for A356, would be useful to those engaged in casting design and/or the development of computer modeling tools. Performance analysis of cast components: Projects to gain a better understanding of casting component performance are needed. For example, the way that loads are redistributed when a crack forms in a cast component is not well understood. Alloy development: The existing alloys hold a great deal of potential for increased utilization. A new casting alloy would have to provide signiﬁcant improvements in properties while still providing excellent castability and weld-ability—at a competitive price, of course.

9. 10.

11.

12.

13.

14.

15.

16. 17.

REFERENCES

Considerations for Future Work This chapter attempts to brieﬂy summarize the general relationships between microstructure, processing, coupon properties, and component performance. In so doing, some points for possible future consideration may include the following:

1. R. Genoux, “Casting Design, from Theory to Practice,” from “Casting Airworthiness.” AGARD Report No. 762, NATO, 1989 2. J.W. Mitchell and M.R. Dunkle, “Reliability of Premium Quality Cast Aluminum Structures for the Aerospace/Defense Industry,” 33rd International SAMPE Symposium, 7–10 March 1988, Society for the

18. 19.

Advancement of Material and Process Engineering J.W. Faber, “Cast Aluminum Structures Technology (CAST),” Final Report, AFWAL-TR-80-3021, April 1980 D. Mietrach, Advanced Castings in Aircraft Structures: A Way to Reduce Costs, Design for Manufacturing Strategies, Principles and Techniques, J. Corbett et al., Ed., Addison-Wesley, 1991 D.R. Apodaca and J.G. Louvier, “Static and Fatigue Properties of Repair Welded Aluminum and Magnesium Premium Quality Castings,” presented at the WESTEC Conference (Los Angeles), 7–11 March 1966 M.W. Ozelton, S.J. Mocarski, and P.G. Porter, “Durability and Damage Tolerance of Aluminum Castings.” WL-TR-91-4111, Oct 1991 K.J. Oswalt and Y. Lii, “Manufacturing Methods for Process Effects on Aluminum Casting Allowables,” AFWAL-TR-844117, March 1985 B. Hanson, “Low Cost Cast Spoilers,” Final Report, AFFDL-TR-75-81, July 1975 S. Kennerknecht, “Fatigue Testing of D357 Alloy Investment Castings,” Cercast Division, July 1994 G.O. Rading, J. Li, and J.T. Berry, Fatigue Crack Growth in Cast Al-Cu Alloy A206 with Different Levels of Porosity, AFS Trans., Vol 94–045 M.F. Haﬁz and T. Kobayashi, The Contribution of Microstructure to the Fracture Toughness of Hypoeutectic Al-Si Casting Alloys, Proc. 4th International Conf. Aluminum Alloys, Sept 1994, p 107–114 S. Shivkumar, L. Wang, and C. Keller, Impact Properties of A356-T6 Alloys, J. Mater. Eng. Perf., Vol 3 (No. 1), Feb 1994, p 83–90 N.R. Green and J. Campbell, “The Inﬂuence of Oxide Film Filling Defects on the Strength of Al-7Si-Mg Alloy Castings,” 1994 Technical Proposal, “Cast Aluminum Structures Technology (CAST),” submitted by the Boeing Co. in response to U.S. Air Force RFP No. F33615-76-R-3111, 16 April 1976 D.L. McLellan and M.M. Tuttle, “Manufacturing Methodology Improvement for Aluminum Casting Ductility,” AFWAL-TR-82-4135, Dec 1982 Metallic Materials and Elements for Aerospace Vehicle Structures, MIL-HDBK-5 “Determination and Acceptance of Dendrite Arm Spacing in Aluminum Castings,” Aerospace Recommended Practice ARP 1947 W.A. Bailey, How Solidiﬁcation Affects Fatigue of 356 Aluminum Alloy, Foundry, 1965 P.C. Inguanti, “Cast Aluminum Fatigue Property/Microstructure Relationship,” 17th National SAMPE Technical Conference, Society for the Advancement of Material and Process Engineering, Oct 1985

218 / Casting Design and Performance 20. M.J. Couper, A.E. Neeson, and J.R. Grifﬁths, Casting Defects and the Fatigue Behavior of an Aluminum Casting Alloy, Fat. Fract. Eng. Mater. Struct., Vol 13 (No. 3), 1990, p 213–227 21. S.E. Stanzl-Tschegg, H.R. Mayer, A. Beste, and S. Kroll, Fatigue and Fatigue Crack Propagation in Al7SiMg Cast Alloys under In-Service Loading Conditions, Int. J. Fatigue, Vol 17 (No. 2), 1995, p 149–155 22. D.C. Wei, Rim Section Fatigue Results of Various Cast Aluminum Wheels, AFS Trans., Vol 98, 1987, p 681–688 23. B. Skallerud, Fatigue Life Assessment of Aluminum Alloys with Casting Defects, Eng. Fract. Mech., Vol 44 (No. 6), 1993, p 857–874 24. T.L. Reinhart, “Characterization of Alloy D357-T6,” presented at Fifth Annual Aerospace Materials and Processes Conference (Anaheim, CA), May 1994 25. F.T. Lee, J.F. Major, and F.H. Samuel, Effect of Silicon Particles on the Fatigue Crack Growth Characteristics of Al-12Wt. Pct.Si-0.35Wt.Pct.Mg-(0 to 0.2) Wt.Pct. Sr Casting Alloys, Metall. Mater. Trans. A, Vol 26A, June 1995, p 1553–1570 26. K.J. Oswalt and A. Maloit, A Study of the Relationship of Notch Yield Ratio and

27.

28.

29.

30.

31.

32.

Fracture Toughness of Structural Aircraft Quality D357 Casting Alloy, AFS Trans., Vol 90–171, p 865–877 A. Wickberg, G. Gustafsson, and L.-E.A. Larsson, “Microstructural Effects on the Fatigue Properties of a Cast Al-7SiMg Alloy,” Paper 840121, Society of Automotive Engineers, March 1984 Y.-H. Tan, S.-L. Lee, and Y.-L. Lin, Effects of Be and Fe Content of Plane Strain Fracture Toughness in A357 Alloys, Metall. Mater. Trans. A, Vol 26A, Nov 1995, p 2937–2945 K.-H. Chien, T.Z. Kattamis, and F.R. Mollard, Cast Microstructure and Fatigue Behavior of a High Strength Aluminum Alloy (KO-1), Metall. Trans., Vol 4, April 1973 S.J. Mocarski, G.V. Scarich, and K.C. Wu, Effect of Hot Isostatic Pressure on Cast Aluminum Airframe Components, AFS Trans., Vol 91–02, p 77–81 S.J. Vonk, G.S. Hoppin, and K.W. Benn, “Hot Isostatic Pressing of Aluminum Alloy Castings,” U.S. Navy contract No. N00019-79-C-0649, Final Report, Feb 1981 J.D. Tirpak, “Elevated Temperature Properties of Cast Aluminum Alloys A201-T7 and A357-T6,” AFWAL-TR-85-4114, Nov 1985

33. M.S. Misra, Evaluation of Tensile and Fracture Properties, Society of Manufacturing Engineers, Paper CM 79–393 34. “Cast Aluminum Primary Aircraft Structure,” AFWAL-TR-84-4070, June 1984 35. E.J. Czyryca, “Long-Life Corrosion Fatigue Crack Initiation in Cast A356-T6 Aluminum Alloy,” U.S. Naval Surface Warfare Center, 7 April 1992 36. J.D. Tirpak and J.J. Ruschau, “Spectrum Fatigue Crack Growth Rate Characteristics of Cast Aluminum Alloys A201-T7 and A357-T6,” AFWAL-TR-86-4047, Dec 1986 SELECTED REFERENCES A. Ferran, Y. Barbaux, M. Maurens, and M.

Dehaye, “High Quality Aluminum Casting to Produce Structural Civil Aircraft Parts Subjected to Fatigue Loads,” Advanced Aluminum and Magnesium Alloys, ASM International, 1990, p 459–467 E. Starke and J. Staley, Application of Modern Aluminum Alloys to Aircraft, ALCOA Technical Center, March 1995

Casting Design and Performance Pages 219–225

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Friction and Wear of Cast Irons* CAST IRON is a general term applied to a family of engineering materials that includes gray iron, white or chilled iron, malleable iron, and nodular iron. These irons have very different compositions, casting characteristics, and heat treatments, which result in physical, chemical, and mechanical properties of wide variation. Cast irons are unique tribological materials that can be used in a range of applications, particularly those in which wear resistance is important. The matrix of these materials can be varied from pearlite to martensite, in order to support graphite, which is a solid lubricant. Hard phases can be varied to resist abrasion. The primary characteristics of gray and nodular irons are described below, in terms of very general selection guidelines, as well as later in this article, in greater detail. Gray Iron. All of the excess carbon in gray iron is present in the form of ﬂakes of graphite in a matrix that varies from all-ferrite to pearlite, with or without phosphide eutectic or carbides. Gray irons are brittle and have relatively low strength and hardness properties. Graphite provides good damping capacity, and its thermal conductivity value is higher than that for steel. Graphite cavities provide chip-breaking qualities, which allow free machining with short chips. A fractured surface will appear gray and dull, because of the presence of graphite. Because graphite is an excellent solid lubricant, it confers good sliding wear resistance with a relatively low coefﬁcient of friction. This class of cast iron is widely used for internal combustion engine components, including engine blocks, cylinder liners, and piston rings, as well as machine tool slideways. White or Chilled Iron. Carbon can be retained as a primary carbide in a pearlitic or martensitic network by suitable alloying or heat treatment, or by the use of chilling. In this condition, the iron is very hard (up to 700 HV), unmachinable, and quite brittle. A fractured surface will appear white and crystalline. By varying the manufacturing route (for example, by chilling), it is possible to produce components that are gray, or graphitic, in some areas and white, or carbidic, in the chilled areas. These materials are used for abrasive wear applications in mining, grinding, mineral handling equipment, and camshafts in the internal combustion engine.

Malleable Iron. Castings initially produced as white iron can be suitably heat treated to produce a more ductile structure. Upon fracture, this structure produces either a white facet (whiteheart), which is due to the presence of pearlite, or a black facet (blackheart), which is due to the presence of graphite. These treatments impart a signiﬁcant change to the otherwise brittle characteristic of a white iron. Ductility and impact properties improve considerably for stress-bearing components for which gray iron would be unsuitable. Nodular or Spheroidal Graphite Iron. By converting graphite from the ﬂake to a nodular form, the brittleness of ﬂake (gray) irons is immediately diminished and, with suitable heat treatment, a steel-like structure can be produced. Because pure materials and rather special melting procedures are required, cost is increased signiﬁcantly, but the improvement in mechanical properties allows these irons to compete with heat-treated steels for applications such as crankshafts and gears. Signiﬁcant developments in recent years have extended the properties of speciﬁc alloys, such as austempered ductile iron, which competes favorably with surfacehardened steel in applications such as cast gears.

Constitution of Cast Iron Cast iron is essentially an iron-carbon alloy that contains from 2.5 to 4.0% carbon and other essential alloying elements, the most important of which are silicon and phosphorus. Pure iron melts at 1535 C (2795 F); the eutectic of iron and carbon occurs with a carbon content between 4.2 and 4.3% and a melting point of 1153 C (2107 F). The eutectic melting temperature for cast iron varies slightly with silicon content, and usually falls between 1135 and 1150 C (2075 and 2102 F). Cast irons therefore have a melting range that is approximately 400 C (720 F) lower than that for steel. Additions of silicon lower the percentage of carbon in the eutectic by about 0.33% for each 1% of silicon added, so that with 1% of silicon, the eutectic will occur at approximately 3.97% of carbon. Additions of phosphorus affect the eutectic carbon value to about the same amount as silicon, so that for an iron containing 1% silicon and 1% phosphorus, the eutectic will occur with approximately 3.6% carbon. Other

*Reprinted from ASM Handbook, Volume 18, Friction, Lubrication, and Wear Technology, P. J. Bldu, Ed, 1992, p 695–701.

elements, in the amounts that are normally present, have a much smaller effect on eutectic carbon value, but they do have other signiﬁcant effects that are referred to later in this article. A cast iron can have a composition that makes it hypoeutectic, eutectic, or hypereutectic with regard to carbon. Its position in this respect is important, because not only does the initial solidiﬁcation point (liquidus) depend on it, but also the manner of solidiﬁcation and, thereby, the structure. The carbon content of most commercial irons is within the range noted above (from 2.5 to 4.0%). The iron-carbon diagram in Fig 1 shows only the iron carbide/austenite (ledeburite) system in the high-carbon region. Such irons, when commercially produced, are known as white, or chilled, irons and are hard, brittle, and unmachinable. Gray irons, however, solidify with the formation of graphite instead of iron carbide, and are reasonably strong and machinable. An intermediate material, known as mottled iron, in which both graphite and iron carbide are formed during solidiﬁcation, also exists. Such irons are harder and less readily machinable than the gray irons. Whether an iron solidiﬁes as gray, white, or mottled primarily depends on the eutectic value and the rate of cooling. However, before attempting to evaluate the effect of cooling rate, the effect of composition must be simpliﬁed. The two most important factors in understanding the mechanical properties and, thus, their efﬁciency in engineering applications, are: Composition, or carbon equivalent value Cooling rate, or section sensitivity

Both factors act independently and have a signiﬁcant effect on the microstructure, which itself controls the properties produced. Different microstructures can be produced at different positions in the same casting. Composition, or carbon equivalent (CE) value, is the most common simpliﬁcation used, in which CE ¼ Total carbon% þ

Silicon% þ Phosphorus% 3 (Eq 1)

It is this value, rather than the actual total carbon content, that applies to commercial irons when the iron-carbon diagram in Fig 1 is used.

220 / Casting Design and Performance has been found in practice, and merely indicates in an arbitrary manner the commercial difﬁculty of producing a given tensile strength in a given bar.

Tensile Properties All of the principal countries specify cast iron on the basis of standard-diameter test bars. However, it is common knowledge in the foundry industry that test bar results cannot be used to give the strength of the casting itself, because section thicknesses and cooling rates can vary appreciably. This can be summarized as follows: For all foundry alloys, the separately cast test

bar is essentially a check on metal quality only

Strength relationships between the test bar

and the casting vary with the nature of the alloy being cast The control of castings, especially for vital services, demands the use of the separately cast test bar as a check on metal quality, but must be supported with check tests on the castings themselves. For repetition castings, such tests can be conducted on a proportion of the castings produced. Otherwise, some suitable form of nondestructive testing must be applied

Fig. 1

Iron-carbon diagram, where solid curves represent metastable system Fe-Fe3C, and dashed curves, the stable system Fe-graphite

The CE value immediately indicates whether an iron is hypoeutectic or hypereutectic, and by how much. Thus, an iron with 3.2% carbon, 2% silicon, and 0.4% phosphorus has a CE value of 4.0 and is hypoeutectic. An iron with 3.2% carbon, 2% silicon, and 1.3% phosphorus has a CE value of 4.3 and is eutectic. An iron with 3.2% carbon, 2.9% silicon, and 1.3% phosphorus has a CE value of 4.6 and is hypereutectic. In general, the lower the CE value, the greater is the tendency for an iron to solidify as either white or mottled. The solidiﬁcation period determines whether a given iron becomes gray, mottled, or white; the form of the graphite is altered very little by the subsequent cooling period. Cooling Rate, or Section Sensitivity. The tensile test can be regarded as the standard test by which cast iron is speciﬁed, the exception being white iron, for which hardness only is speciﬁed. Tensile strength is the stress required to pull apart a testpiece by an axially applied

load. The test is almost always performed on a round, machined testpiece, in which the middle section of the length is reduced in diameter. The testpiece is machined from either independently cast bars or broken transverse test bars. Speciﬁcations normally give details of casting sizes appropriate to each size of cast bar. Because the strength of cast iron is particularly dependent on the section size into which it is cast, the strength of gray iron is deﬁned in relation to the diameter of as-cast bars from which standard tensile specimens can be machined. For gray ﬂake iron, each of the eight grades is designated by the tensile strength required on the 30 mm (1.2 in.) diameter bar (Table 1). The speciﬁcation for each grade, however, requires a higher tensile strength for bars of smaller diameter, and a lower tensile strength for bars of larger diameter. However, the variation in strength given in the speciﬁcation according to size of the test bar is less than

It is important to realize that national speciﬁcations merely deﬁne the metal in the ladle and do not give a direct indication of the strength of the casting itself. As explained earlier, the strength of the material in the casting is a function of the cooling rate of the particular portion under consideration. However, when the strength of the metal cast under standard conditions is known, it becomes possible to estimate, with some degree of accuracy, the probable tensile strength to be found in a gray iron casting of a given thickness. If the test bar were cast onto the casting itself, then it would be impossible to relate the results obtained from such a test bar with those obtained from another test bar under standard conditions. Such a test would be almost valueless, because the test bar would not be representative of the casting itself, nor could it be related to other standards. In past practice, it was quite common to specify cast-on test bars, on the assumption that such a test bar must necessarily be representative of the casting itself. However, this practice did not provide the necessary information. Except in certain cases, the use of the independently cast test bar has now almost entirely superseded the use of the cast-on test bar. Effect of Section on Strength. To interpret the results obtained from the standard test bar in terms of the strength of the casting itself, some relationship must be established between various section sizes and the bar. This is best exempliﬁed by discussing the properties of gray iron. A statistical survey of a large number of industrially cast gray iron test bars of different sizes has been made, the results of which are shown in Fig 2. The curve shows the falling

Friction and Wear of Cast Irons / 221 Table 1 Expected tensile strengths of testpieces machined from cast-on test samples and from the castings Casting, over

Tensile strength

Section thickness, up to and including

Cast-on test samples

Castings

Grade

mm

in.

mm

in.

MPa

ksi

MPa

ksi

100

2.5 10

0.1 0.4

10 20

0.4 0.8

... ...

... ...

120 90

17.4 13.1

150

2.5 10 20 40 80 150

0.1 0.4 0.8 1.6 3.2 6.0

10 20 40 80 150 300

0.4 0.8 1.6 3.2 6.0 12.0

... ... 120 110 100 90

... ... 17.5 16.0 14.5 13.0

155 130 110 95 80 ...

22.5 18.9 16.0 13.8 11.6 ...

180

2.5 10 20 40 80 150

0.1 0.4 0.8 1.6 3.2 6.0

10 20 40 80 150 300

0.4 0.8 1.6 3.2 6.0 12.0

... ... 150 135 125 100

... ... 21.8 19.6 18.1 14.5

185 160 135 115 100 ...

26.8 23.2 19.6 16.7 14.5 ...

200

2.5 10 20 40 80 150

0.1 0.4 0.8 1.6 3.2 6.0

10 20 40 80 150 300

0.4 0.8 1.6 3.2 6.0 12.0

... ... 170 150 140 130

... ... 24.7 22.8 20.3 18.9

205 180 155 130 115 ...

29.7 26.1 22.5 18.9 16.7 ...

220

2.5 10 20 40 80 150

0.1 0.4 0.8 1.6 3.2 6.0

10 20 40 80 150 300

0.4 0.8 1.6 3.2 6.0 12.0

... ... 175 165 150 140

... ... 25.4 23.9 21.8 20.3

220 195 170 145 130 ...

31.9 28.3 24.7 21.0 18.9 ...

250

5 10 20 40 80 150

0.2 0.4 0.8 1.6 3.2 6.0

10 20 40 80 150 300

0.4 0.8 1.6 3.2 6.0 12.0

... ... 210 190 170 160

... ... 30.5 27.6 24.7 23.2

250 225 195 170 155 ...

36.3 32.6 28.3 24.7 22.5 ...

300

10 20 40 80 150

0.4 0.8 1.6 3.2 6.0

20 40 80 150 300

0.8 1.6 3.2 6.0 12.0

... 250 220 210 190

... 36.3 31.9 30.5 27.6

270 240 210 195 ...

39.2 34.8 30.5 28.3 ...

350

10 20 40 80 150

0.4 0.8 1.6 3.2 6.0

20 40 80 150 300

0.8 1.6 3.2 6.0 12.0

... 290 260 230 210

... 42.1 37.7 33.4 30.5

315 280 250 225 ...

45.7 40.6 36.3 32.6 ...

Source: ISO R185 - 1961; British Standard BS 1452:1990

off in strength with increasing test bar size for the main grades of unalloyed gray cast iron. The line for each grade is really the center of a zone of uncertainty amounting to 5 to 10 Pa (0.00075 to 0.0015 psi), because there is a possible variation in each grade that is due to differences in composition, pouring technique, and other factors. A version of this curve is usually included in the relevant national or international speciﬁcations. The total strain (recoverable and nonrecoverable) of gray iron tensile testpieces at fracture usually varies between 0.5% for high-tensile iron and 0.75% for low-tensile iron. The non-elastic strain varies between 0.2% for high-tensile iron and 0.6% for low-tensile iron.

White Irons If the iron is sufﬁciently below the eutectic value, has a very low silicon content, and

contains appreciable quantities of carbide or stabilizing elements, or if the cooling rate is sufﬁciently rapid, then the formation of graphite does not occur. Under these conditions, solidiﬁcation takes place ﬁrst by the formation of austenite dendrites. The remaining liquid in the interdendritic spaces then becomes enriched in carbon and solidiﬁes as a eutectic of iron carbide and austenite, known as ledeburite. As the solid casting cools down, at a temperature ranging from 720 to 750 C (1330 to 1380 F), the austenite, which is g iron with carbon in solution, transforms to a iron, in which carbon is substantially insoluble. Carbon is then rejected as iron carbide in the form of lamellae alternating with lamellae of a iron to form a matrix of pearlite. The initially formed iron carbide, however, remains unchanged. The structure of a white iron at room temperature, therefore, consists of primary dendrites of pearlite with interdendritic areas of transformed ledeburite, which is a eutectic structure of iron

carbide and pearlite. The pearlite areas are formed from the original austenite. In some cases, the austenite produced during the ﬁnal eutectic solidiﬁcation of iron carbide and austenite is deposited on the already existing austenite dendrites, and the interdendritic area is predominantly carbide. The ﬁnal structure will then consist of dendrites of pearlite surrounded by cementite that contains very little transformed eutectic pearlite (Fig. 3). These irons are characterized by a completely white fracture and a high hardness (400 to 600 HV), which is an average of the carbide (900 to 1200 HV) and pearlite (200 to 300 HV). Various grades of white iron are available, depending on composition and heat treatment. The cheapest white irons are either unalloyed or low-alloy grades containing between 2.4 and 3.4% carbon and up to 2.0% chromium. A hardness of 350 HV is achieved in the cast condition. By increasing the chromium content (up to 10%), it is also possible to increase the carbon content to produce increased amounts of carbides. By including nickel, it is also possible to carry out heat treatment to produce a martensitic, rather than pearlitic, matrix. These changes have the overall effect of increasing the hardness to 600 HV and reducing the effect of section sensitivity. When increased corrosion resistance and maximal abrasion resistance are required, irons containing up to 28% chromium can be used, but as the percentage of carbides (or carbon content) increases, impact resistance decreases. Thus, it is important to balance the composition carefully to obtain the best combination of properties for speciﬁc applications. These irons in the heattreated condition develop the maximum hardness of up to about 700 HV, and are the least affected by temperature in service. Table 2 summarizes the composition and hardness of various white irons.

Gray Irons In hypoeutectic irons, the ﬁrst phase to be deposited from the melt is austenite, a solid solution of carbon in g iron that crystallizes in the form of dendrites, the length and pattern of which depend on the temperature gradients. The austenite dendrites continue to grow, and the remaining liquid increases in carbon content until the eutectic temperature and concentration are attained. Solidiﬁcation begins from a number of centers, each with an approximately spherical crystallization front, and simultaneous deposition of graphite and austenite occurs. Ultimately, these eutectic cells meet and consume the remaining liquid. The austenite of the dendrites and that of the eutectic becomes continuous, and the structure appears as a dispersion of graphite ﬂakes in a matrix of austenite. After solidiﬁcation, the eutectic cell structure and the random orientation of the primary

222 / Casting Design and Performance

Pearlitic white iron showing the original austenite dendrites that transformed to pearlite and the interdendritic eutectic structure of cementite and austenite, where the austenite transformed to pearlite. White phase, carbide; dark phase, pearlite. 80

Fig. 3

hard matrix of martensite formed. This treatment can be given deliberately when increased hardness is desired.

Fig. 2

Relationship between tensile strength and section thickness for different grades of ﬂake gray iron. Source: ISO R185.1961; British Standard (BS) 1452: 1990

Table 2 Composition and hardness values for white irons Composition, % C

Si

Mn

Cr

2.40–3.40 2.70–3.20 2.80–3.20 3.0–3.60 2.0–2.80

0.50–1.50 0.30–0.80 0.30–0.80 1.0 max 1.0 max

0.20–0.80 0.20–0.80 0.20–0.80 0.50–1.50 0.50–1.50

2.0 max 1.50–3.50 8.0–10.0 14.0–17.0 22.0–28.0

P

0.15 0.15 0.15 0.10 0.10

max max max max max

Mo

Ni

Cu

S

Hardness, HV

... 0.50 max 0.50 max 3.0 max 1.50 max

... 3.0–5.50 4.0–6.0 2.0 max 2.0 max

... ... ... ... ...

... 0.15 max 0.15 max 0.10 max 0.10 max

400 500 550 710 655

Source: British Standard BS 4844:1985

austenite dendrites throughout the structure are essentially complete. They cannot be subsequently modiﬁed either by cooling or subsequent treatment. Only in strongly hypoeutectic iron is the austenite dendrite pattern clearly visible. Hypereutectic irons solidify by the direct formation of graphite from the metal in the form of coarse ﬂakes that are called kish. Because of its low relative density, kish tends to rise to the surface of the melt. However, when entrapped by the metal, kish generally appears in the microstructure as characteristic long straight ﬂakes, or, in rapidly cooled sections, as lumpy ﬂakes. This phase is deposited over a range of temperatures, starting at the liquidus surface, until the carbon content of the remaining melt is at the eutectic concentration when a simultaneous crystallization of graphite and austenite occurs. Generally, this eutectic graphite in hypereutectic irons occurs in ﬂake form on a ﬁner scale than does the primary kish.

The austenite, upon cooling through the critical temperature range (from about 720 to 830 C, or 1330 to 1525 F), transforms to pearlite. The products of this transformation depend on the rate of cooling and composition of the material. Under normal conditions, the products will consist of pearlite, or graphite and ferrite, or mixtures of all three. Where the eutectoid transformation produces graphite and ferrite, the graphite is deposited on the already existing eutectic graphite ﬂakes. The formation of ferrite is most likely to occur with slow rates of cooling and high silicon contents that reduce the stability of iron carbide, or with high CE values, or with ﬁne undercooled graphite. The formation of a fully pearlitic structure is more likely to occur with moderately rapid cooling rates or low CE values. If the iron is very rapidly cooled through the critical temperature, for example, by oil quenching, then the transformation of austenite to pearlite or ferrite can be suppressed, and a

Graphite The form and distribution in which graphite is deposited depends on numerous factors, such as melting temperature, nucleation, cooling speed, and others. The basic forms are ﬂake graphite, aggregate or temper carbon, and nodules or spheroids. A convenient standard for graphite (ASTM A247-47 and 67) deﬁnes shape, size, and distribution, and has been extended by the International Organization for Standardization (ISO Recommendation 945) to cover additional forms. The basic nomenclature and grading for ﬂake graphite follows the generally accepted ASTM pattern. Flake graphite can form as one of four types: Kish, which occurs in hypereutectic irons

and can be present as straight and undistorted ﬂakes (ASTM Type C; ISO Form I, distribution C) or very thick lumpy ﬂakes Normal ﬂakes (ASTM Type A; ISO Form I, distribution A) Rosette (ASTM Type B; ISO Form I, distribution B) Undercooled (ASTM Type D; ISO Form I, distribution D) The undercooled, or Type D graphite, is normally associated with rapid rates of cooling and is particularly common in thin sections, but can be produced by other means (such as addition of titanium). The rosette pattern is typical of slightly less rapid cooling and is not uncommon in the more rapidly cooled surfaces of castings with sections of 9.6 mm (3/8 in.) or more. Both

Friction and Wear of Cast Irons / 223 undercooled and rosette graphite are frequently associated with ferrite, rather than pearlite, because the larger available surface area of graphite reduces the distance required for carbon diffusion during transformation from g to a iron in the critical temperature range. When ferrite is associated with random or coarse ﬂake graphite, the iron is usually hypereutectic or the rate of cooling in the mold has been very slow. Widmansta¨tten and mesh-form graphite can occur when lead or tellurium are present. These structures profoundly weaken the iron. When gray iron passes through the g to a critical transformation range, graphite may be rejected from the g phase onto existing graphite ﬂakes and form spiky irregularities on the graphite surface. When minimum friction and maximum wear resistance are required under dry, marginally lubricated or fully lubricated conditions, it is very important to precisely specify the microstructural requirements. These properties can be met by a requirement for Type A graphite in the size range from 3 to 5, with a pearlitic matrix, free from primary ferrite (Fig. 4) Phosphorus is desirable at a level of up to approximately 0.75%, which produces a hard, dispersed interdendritic phase that improves abrasion resistance at the expense of machinability. Although each ﬂake appears to be separated from adjacent ones in a two-dimensional micro-structural view, it must be recognized that in each colony, as viewed in three dimensions, the ﬂakes are connected at the nucleation source to produce a lettuce leaf-like structure. This can only be observed by deep etching the matrix and by selective dissolution, so that the graphite structure can be seen to stand above the matrix. The sharpness of the graphite cavities acts as an internal stress raiser and produces the brittleness that characterizes gray iron. However, the material is immune from the effects of external stress raisers. Nodular Graphite. The shape of the graphite precipitated on solidiﬁcation can be modiﬁed to the nodular or spheroidal form (ISO Form VI) by additions of magnesium and/or cerium. Intermediate or degenerate forms can also be produced. The ISO Form II crab type is a degenerate form of ﬂake graphite occasionally found in magnesium-treated irons containing lead, bismuth, arsenic, or titanium, and can be suppressed by additions of cerium. Unless pure melt charge materials can be guaranteed, it is the recommended practice to use a combined addition of magnesium and cerium. Quasiﬂake graphite (ISO Form III) is usually the result of insufﬁcient magnesium additions to produce the fully spheroidal form. Irregular open, or “exploded,” type spheroids (ISO Form V) may occur with very pure raw materials treated with magnesium and cerium. Misshapen spheroids tend to occur in heavy sections. Irons that contain graphite nodules (Fig. 5) that are either primarily or totally spheroidal in the as-cast condition have properties that are fundamentally different from those of gray

Fig. 4

(a) Gray ﬂake iron with random graphite in an unetched matrix. 75. (b) Etching showing pearlitic matrix. 300

Fig. 5

(a) Nodular graphite in unetched matrix. 75. (b) Etching shows that matrix consists of ferritic envelope around the nodules surrounded by a pearlitic matrix. 300

iron with ﬂake graphite. They possess greater tensile strength and have a considerable degree of ductility, as well as greater impact resistance and enhanced fatigue resistance. Their properties are much closer to those of steel than to gray iron (Fig. 6). The properties to be expected in the nine different grades of nodular iron currently available are summarized in Table 3. It is also possible to heat treat these materials using techniques similar to those used for steel. In this way, a nodular graphite in a martensitic matrix can be produced. In the as-cast state, the microstructure of nodular iron can range from a completely white structure where thin sections are involved, to that of nodular graphite in a completely ferritic matrix. By virtue of the fact that nodular irons have their own “carbon bank” available, like gray irons do, heat treatments can be applied to produce a wide range of matrix structures. Thus, by heating to a temperature above the critical temperature, graphitic carbon will dissolve in the austenite until equilibrium conditions are attained. By air-cooling from this temperature, the austenite undergoes the pearlitic transformation, thus maintaining an amount of carbon dispersed in the matrix. On the other hand, a very slow cooling through the temperature range will deplete the

carbon from the matrix, and, instead, secondary graphite is deposited around the graphite nodules. The heat treatments normally applied on a commercial scale can be summarized as follows: Heat to a temperature between 850 and

900 C (1560 and 1650 F). The length of soaking time depends on the amount of free carbide present. Air-cooling from this temperature provides a predominantly pearlitic matrix, except in heavy sections Remove carbides and austenitize as described above. Oil quench at a temperature between 850 and 900 C (1560 and 1650 F) to produce a martensitic matrix that can then be tempered to produce the required range of mechanical properties. This treatment is limited to castings of relatively thin section, unless alloying elements to improve hardenability are present Remove carbides and austenitize as described in the ﬁrst bullet. Cool to a temperature between 680 and 700 C (1255 and 1290 F). It is common to hold at this temperature for a period of time in order to effect complete graphitization with the production of a ferritic matrix. The need for holding and the time involved will vary with

224 / Casting Design and Performance marked structural change occurs during this treatment, there is a tendency for softening to occur when the higher temperatures are used Nodular irons can also be transformed to a bainitic structure through the use of austempering carried out by cooling from 850 to 950 C (1560 to 1740 F) into a bath at a temperature between 230 and 425 C (445 and 795 F). Heat-treatment stresses and distortion are minimized and their overall properties vastly improved. For example, a tensile strength ranging from 1200 to 1600 MPa (175 to 230 ksi) can be achieved with signiﬁcant ductility, toughness, and fatigue strength. In this condition, austempered ductile iron is now being used for gears and other highly stressed parts Nodular irons are used in such applications as crankshafts and gears, where they compete against through-hardened and surface-hardened steels. Their near-net-shape forming characteristics, coupled with the presence of graphite (which confers sliding wear resistance), offer a unique combination of properties for more hostile tribological applications.

Typical Microstructures

Fig. 6

Typical stress-strain curves in tension for various cast irons and steels. Source: British Standard (B5) 27B9:1985

Table 3 Mechanical properties and supplementary information Tensile strength (Rm), min

0.2% proof stress(a) (Rp0.2), min

Grade

MPa

ksi

MPa

ksi

900/2 800/2 700/2 600/3 500/7 450/10 420/12 400/18 400/18L20 350/22 350/22L40

900 800 700 600 500 450 420 400 400 350 350

130 116 102 87 73 65 61 58 58 51 51

600 480 420 370 320 320 270 250 250 220 220

87 70 61 54 46 46 39 36 36 32 32

Elongation (A), min, % Hardness, HB(b) Predominant structure constituent

2 2 2 3 7 10 12 18 18 22 22

302–359 248–352 229–302 192–269 170–241 160–221 <212 <179 <179 <160 <160

Tempered martensite Pearlite or tempered structure Pearlite Pearlite/ferrite Ferrite/pearlite Ferrite/pearlite Ferrite Ferrite Ferrite Ferrite Ferrite

(a) Veriﬁcation of the 0.2% proof stress (Rp02), is optional. (b) Veriﬁcation of hardness values is optional. Source: ISO 1083. 1987; British Standard BS 2789:1985

basic composition and section size, but is usually of the order of from 8 to 12 h. The material is ﬁnally cooled to about 600 C (1255 F) before removal from the furnace The air-cooled material from the process described in the ﬁrst bullet, which possesses a pearlitic matrix, can be annealed at a subcritical temperature (700 to 720 C, or 1290 to 1330 F), giving a completely

ferritic matrix, as in the third bullet. The holding time at the subcritical temperature depends on composition and section size, but from 8 to 12 h should be adequate Any castings that are complex and likely to contain residual stresses, either in the as-cast condition or after normalizing, can be stressrelieved by heating to a temperature from 450 to 550 C (840 to 1020 F). Although no

For all critical wear applications, it is recommended that both microstructure and hardness values at and close to the surface (where wear is likely to occur) be checked and actually become part of the manufacturing speciﬁcation. The reasons behind this recommendation are the importance of microstructure in terms of control of engineering properties, as well as the variability that is due to section sensitivity. Described below are the different microstructures that can be produced in engineering castings. The matrix structure of cast irons usually contains one or more of these constituents: ferrite, pearlite, cementite, phosphide eutectic, martensite, bainitic transformation structures, austenite, and graphite, the latter of which has already been discussed. Ferrite in cast iron is essentially a singlephase solid solution of silicon in amounts that vary with graphite structure, cooling rate, and silicon content. The amount tends to increase as the cooling rate decreases and the silicon content increases, and the graphite approaches the undercooled form. Fully ferritic structures are normally only obtained by annealing. Ferrite should be avoided in all wear parts. Pearlite consists of alternate lamellae of ferrite and cementite. The degree of ﬁneness depends on cooling rate. The ﬁner the pearlite, the stronger (harder) is the iron. At low magniﬁcation, pearlite appears as a half-tone color. This structure is formed by the transformation of austenite during normal cooling in the mold or in the air through the critical temperature range from 720 to 900 C (1330 to 1650 F). This is a highly desirable matrix for sliding wear resistance.

Friction and Wear of Cast Irons / 225 Cementite (Fe3C) in the massive eutectic form is a hard white constituent formed during solidiﬁcation, as in mottled or white irons, and in the lamellar form in pearlite, where it is formed by the transformation of austenite through the critical temperature. Eutectic carbides can increase hardness for wear resistance. Harder carbides can be obtained by alloying with Cr and Mo. Phosphide eutectic (melting point of about 930 C, or 1706 F) occurs in two distinct forms in cast irons with more than 0.06% phosphorus. The normal form is pseudobinary, which consists of ferrite and iron phosphide (Fe3P). The true eutectic forms from the liquidus as austenite plus iron phosphide. Upon cooling, the austenite transforms to ferrite and pearlite, and, with iron phosphide, gives a bulk hardness between 420 and 600 HV. In mottled irons and in apparently gray irons, this structure can occur in ternary form, where it solidiﬁes as austenite, iron carbide, and iron phosphide. At room temperature after transformation of the austenite, it appears as a eutectic of ferrite, free carbide, and iron phosphide. This form is likely to occur if chromium or vanadium are present in amounts over 0.1% and has a bulk hardness of about 780 to 800 HV, because of the high hardness of the free carbide (800 to 1015 HV). Although it is desirable to improve the sliding wear resistance of gray irons, it markedly increases tool wear in machining operations. Martensite is a ﬁne acicular, slow-etching structure, normally produced either by very rapid cooling (quenching) of austenite through the critical temperature range or by alloying. The presence of martensite achieves maximum hardness in parts that require abrasion resistance. Acicular or bainitic transformation structures are produced by either isothermal quenching or alloying. They are often referred to as

acicular ferrite and are softer and tougher than martensite, but harder and stronger than pearlite. These acicular structures range from the upper bainites, or acicular ferrites, to martensite, depending on the transformation time, composition, and other factors. Austempered parts provide resistance to the more hostile stress conditions encountered in gear trains, for example. Austenite can be made stable at room temperature by the addition of alloys, such as nickel and manganese, which depress the critical temperature at which the g to a change occurs. Although the transformation may have been suppressed at room temperature, it can still occur at a lower temperature, depending on the amount of alloy present. These irons are not normally used for wear resistance.

Applications Cast irons have been widely used by engineers in applications that require low cost, excellent castability, good damping capacity, ease of machining, and wear resistance. These general terms do not describe precisely the requirements for speciﬁc applications, and it is not possible to obtain all of those desirable characteristics in any single alloy. Cast irons are traditional materials that have an extensive use history and a considerable property database. For tribological purposes, they are the most widely used of all materials. This family of materials sets the standard by which other materials, including composites and ceramics, are compared. Although there have been many attempts to supersede the use of cast irons in extremely hostile environments, such as in the internal combustion engine, successes are still few in number. The reasons are quite simple and are due to the versatile properties of excellent

Table 4 Guidelines for industrial applications of selected cast irons for wear parts Cast iron type

Gray (ﬂake) iron

Nodular iron Chilled iron

White iron (cast and heat treated) White iron (by hardfacing)

Applications(a)

Piston rings and cylinder bores for compressors and internal combustion engines; machine tool beds and slideways; brake blocks; lowstress gears Crankshafts; gears; highly-stressed piston rings Cams and tappets where the Hertzian stress is high and lubrication is borderline Grinding balls; chute and mill plates and tiles; cone, feeder, deﬂector, and breaker liner plates: grizzlies Dragline buckets, shrouds, pick holders, and side plate protection; hardfacing of steel digger teeth for shovels

(a) In all of these applications, there is a wide range of compositions, cooling rates, and heat-treatment variations that control the microstructure and thus the level of wear resistance which can be obtained. A balance also has to be obtained between wear resistance, corrosion resistance, and impact resistance (particularly for minera) handling applications). Great care is therefore required when choosing the optimum material for a speciﬁc wear application.

sliding and abrasive wear resistance, coupled with low cost, casting versatility, dimensional stability over a broad temperature range, and, for certain alloys, case of machining. Typical wear applications for a variety of cast iron grades are identiﬁed in Table 4. Suppliers and users of cast irons continually update and extend applications into more hostile conditions by improving surface engineering in terms of treatment coatings and machining. A sound engineering understanding of the relationship between composition, manufacture, microstructure, and engineering properties will ensure that these materials continue to be used extensively in tribological applications decades from now.

Casting Design and Performance Pages 227–236

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Friction and Wear of Aluminum-Silicon Alloys* ALUMINUM-SILICON ALLOYS are noted for their unique combination of desirable characteristics, including excellent castability and low density combined with good mechanical properties. Interestingly, there is no mention of wear resistance in the many published tabulations of the attributes of this class of alloys (see, for example, Ref 1 and 2). In a comprehensive review of the wear characteristics of aluminum alloys, Eyre (Ref 3) remarks on the dearth of published information about the wear resistance of aluminum-silicon alloys, while acknowledging their increased application in environments demanding this physical characteristic. In the late 1950s, die cast aluminum-silicon alloy cylinder blocks were manufactured for the automotive industry to take advantage of the light weight and good thermal conductivity offered by hypoeutectic alloys such as A356 and A380 (Ref 4). However, because these alloys exhibited only modest wear resistance, the early engine blocks (for example, those produced for the American Motors Corporation Rambler) had a cast-in steel cylinder liner. In Europe, automotive engine blocks were also being die cast in aluminum-silicon alloys (such as LM24 and LM26), but these blocks had liners that were shrink-ﬁtted in place. A major boost was given to the use of lightweight aluminum alloys in automobiles when the energy crisis in the early 1970s resulted in the government mandating improved fuel consumption in automobiles. Because fuel consumption can be directly related to vehicle weight (Ref 5), the use of aluminum in place of cast iron in engine components became highly desirable. However, the use of die cast aluminum engine blocks with a cast iron cylinder liner proved to be too costly in production, paving the way for the development of hypereutectic aluminum-silicon alloys with greater wear resistance that could be used without liners. One such alloy that has been developed speciﬁcally for its high wear resistance is the hypereutectic aluminum-silicon alloy A390 (A390.0), which has been used in engine blocks without a liner. An electrochemical surface treatment is used to etch away some of the

matrix aluminum so that the eutectic and primary silicon particles provide the bearing surface, resulting in improved wear resistance. Jorstad (Ref 6) has suggested the use of this alloy for other weight-saving applications where good wear resistance is required (for example, brake drums and disk brake rotors). A powder metallurgy (P/M) production route for the manufacture of cylinder liners from hypereutectic aluminum-silicon alloys has also been explored (Ref 4). Using this method, the size of the primary silicon particles may be optimized and solid graphite can be added to the matrix to serve as a lubricant. The liners are then inserted into the engine block, which is produced from conventional aluminum-base alloys. While applications for aluminum-silicon alloys are currently centered in the automotive industry (which accounts for greater than 50% of the market), other applications in communications equipment, instrumentation, and small engines appear likely. The hypereutectic alloys are likely to dominate because their wear resistance, combined with low coefﬁcient of thermal expansion (CTE) and ﬂuidity properties, allows thinner wall castings to be manufactured. Both the Aluminum Association (AA) and the Society of Automotive Engineers (SAE) designations are commonly used to classify aluminum-silicon alloys in the United States.

At room temperature, the hypoeutectic alloys consist of the soft, ductile primary aluminum phase and the very hard, brittle silicon phase associated with eutectic reaction. It is this silicon phase that contributes to the very good wear resistance of these alloys. The silicon phase is diamond cubic with a density of 2.6 g/cm3 (0.094 lb/in.3) and a Vickers hardness of approximately 10 GPa (1.5 106 psi). Silicon is essentially insoluble in aluminum (Ref 10). Figure 2 illustrates the typical microstructure of a common hypoeutectic alloy, A357.0. Hypereutectic alloys, the most commonly used wear-resistant alloys, contain coarse, angular, primary silicon particles as well as cutectic silicon. These primary silicon particles impart excellent wear resistance to these alloys. A typical microstructure of an unreﬁned hypereutectic alloy (A390-type) is shown in Fig. 3. Commercial, aluminum-silicon alloys (Table 1) generally contain other alloying elements to further enhance or modify the wear

Metallurgy of Aluminum-Silicon Alloys Aluminum alloys for wear resistance applications are based on the aluminum-silicon alloy system. This binary system is a simple eutectic alloy system with the eutectic composition at 12.5% Si. (Fig. 1). Standard alloys, of course, contain a number of alloying ingredients. Selected commercial alloy compositions are shown in Table 1. Table 2 cross references United States aluminum-silicon compositions with compositions of European and Japanese sources.

*Reprinted from ASM Handbook, Volume 18, Friction, Lubrication, and Wear Technology, P.J. Blau, Ed., 1992, p 785–794

Fig. 1

Aluminum-silicon Source: Ref 7

binary

phase

diagram.

228 / Casting Design and Performance Table 1 Nominal compositions of selected commercial alloys recommended for wear applications Composition, wt% Alloy

Si

Fe

Ca

Mg

Mn

Other

Hypereutectic alloys 390.0 A390.0 B390.0 392.0 393.0

16–18 16–18 16–18 18–20 21–23

1.3 0.5 1.3 1.5 1.3

4–5 4–5 4–5 0.4–0.8 0.7–1.1

0.45–0.65 0.45–0.65 0.45–0.65 0.8–1.2 0.7–1.3

10.5–12 11–13 11–13 11–13 11–13.5

1.3 1.2 1.2 2 1

3–4.5 0.5–1.5 1.5–3 1 0.5–1.3

1 0.6 1.5 2 1 1.2 2 1.3

3–4 0.25 0.2 3–4 3–4 2–4 0.6 2–3

... ...

0.1 0.1 0.5 0.2–0.6 0.1

1.5 Zn 0.5 Ni, 0.5 Zn, 0.3 Sn 2–2.5 Ni

0.1 0.7–1.3 0.5–1.5 0.1 0.8–1.3

0.5 0.35 0.5 0.35 ...

0.5 Ni, 3 Zn, 0.35 Sn 2–3 Ni, 0.35 Zn 0.5–1.5 Ni, 1 Zn 0.5 Ni, 0.5 Zn 0.5–1.3 Ni

0.1 0.2–0.45 0.2–0.4 0.1 0.05–0.5 0.5–1.5 0.4–0.6 0.1

0.5 0.35 0.1 0.5 0.5 0.5 0.35 0.5

0.35 Ni, 1 Zn 0.35 Zn 0.25–0.5 Cr, 0.15 Ni, 0.15 Sn 0.5 Ni, 3 Zn, 0.35 Sn 0.5 Ni, 1 Zn 0.5 Ni, 1 Zn 0.5 Ni, 0.5 Zn 0.3 Ni, 3 Zn, 0.15 Sn

Eutectic alloys 384.0 336.0 339.0 413.0 4032

Hypoeutectic alloys 319.0 356.0 364.0 380.0 333.0 332.0 360.0 383.0

5.5–6.5 6.5–7.5 7.5–9.5 7.5–9.5 8–10 8.5–10.5 9–10 9.5–11.5

Source: Ref 8, 9

Fig. 2 Table 2

Cross-reference to equivalent wear-resistant aluminum-silicon alloys

Typical microstructure of type hypoeutectic alloy. Source: Ref 11

A357.0

Speciﬁc composition limits may vary from United States limits. Foreign alloy equivalent United States alloy

390 392 413 336 332 333 360 380 356

United Kingdom

France

Japan

Germany

LM28 ... ... LM13 LM26 LM2 ... LM24 ...

A-Sl8UNG ... ... A-Sl2UNG A-Sl0UG ... A-9G ... A-S7G

... AC9B ... AC8A AC8C AC8B ... ... AC4C

G-AlSi17Cu4 ... G-AlSi12(Cu) ... ... ... G-AlSi10Mg(Cu) G-AlSi8Cu3 ...

Source: Ref 8, 9

resistance or impart additional properties to these alloys. Iron. The most common alloying element is iron, which can be tolerated up to levels of 1.5 to 2.0% Fe. The presence of iron modiﬁes the silicon phase by introducing several Al-FeSi phases. The most common of these are the a and b phases. The a phase has a cubic crystal structure and appears in the microstructure as a “Chinese script” eutectic. The less common b phases generally appear as needles and/or platelets in the structure. Other iron-bearing phases such as Al6Fe and FeAl3 can also be found in these alloys. Aluminum-silicon alloys intended for die castings typically have higher minimum iron levels to reduce sticking between the mold and the casting. Magnesium is added to provide strengthening through precipitation of Mg2Si in the matrix. In an Al-Fe-Si-Mg alloy, the Al-Si-Fe phases will not be affected by the addition of magnesium. However, magnesium can combine with insoluble aluminum-iron phases, resulting in a loss of strengthening potential (Ref 12).

Copper. The most common aluminum wearresistant alloys also contain copper. Copper additions impart additional strengthening of the matrix through the aging or precipitation-hardening process (AlCu2 or Q phase) or through modiﬁcation of the hard, brittle Al-Fe-Si phases by substitution in these intermetallic phases. As the strength of these alloys increases through magnesium and copper additions, some sacriﬁce in ductility and corrosion resistance occurs. Manganese. Many of the important aluminum-silicon alloys also contain low (<1 wt%), but signiﬁcant, amounts of manganese. The presence of manganese can reduce the solubility of iron and silicon in aluminum and alter the composition and morphology of the Al-Fe-Si primary constituent phases. For example, manganese additions can favor the formation of constituents such as Al12(Fe,Mn)3Si rather than the Al9Fe2Si2-type constituents. The manganese-bearing constituents are typically less needlelike or platelike than the manganese-free iron- or (iron/silicon)-bearing primary constituents. Manganese additions also

Microstructure of type A390.0 hypereutectic alloy. (a) Unreﬁned (Graff-Sargent etch). Dark regions contain coarse primary silicon particles in addition to eutectic silica. (b) Reﬁned (as polished). 120

Fig. 3

Friction and Wear of Aluminum-Silicon Alloys / 229 improve elevated-temperature properties of the aluminum-silicon alloys. Cumulative Effect of Alloying Elements. In summary, aluminum wear-resistant alloys are based on alloys containing the hard, brittle silicon phase. Alloying elements such as iron, manganese, and copper increase the volume fraction of the intermetallic silicon-bearing phases, contributing to increased wear resistance compared to binary aluminum-silicon alloys. In addition, magnesium and copper also provide additional strengthening by producing submicroscopic precipitates within the matrix through an age-hardening process.

Properties and Structure Alloying aluminum with silicon at levels between about 5 and 20% imparts a signiﬁcant improvement in the casting characteristics relative to other aluminum alloys. As a result, these high-silicon alloys are generally utilized as casting alloys rather than for the manufacture of wrought products. Aluminum-silicon alloys also possess excellent corrosion resistance, machinability, and weldability. (Table 3). Binary hypoeutectic alloys are too soft to have a good machinability rating. However, the machinability of aluminum-silicon alloys is generally very good in terms of surface ﬁnish and chip characteristics. Tool life can be short

with conventional carbide tools, particularly in the case of the hypereutectic alloys. With the recent introduction of diamond cutting tools, tool life has been signiﬁcantly increased, making the machining of the hypereutectic alloys practical. Corrosion resistance of these alloys is generally considered excellent. Alloys containing increasing amounts of copper have a somewhat lower corrosion resistance than the copper-free alloys as measured in standard salt spray tests. Because of their high ﬂuidity and good casting characteristics, these alloys are highly weldable with conventional welding techniques. For joining purposes, brazing alloys and ﬁller wire alloys (for example, alloys 4043 and 4047) (Ref 14) are also based on the aluminum-silicon alloy system. For wear applications, the important physical properties of these alloys include thermal expansion, thermal conductivity, electrical conductivity, and Young’s modulus. Data for these properties are available in standard references (Ref 13–18). Because silicon is generally in precipitate form, the rule of mixtures is applicable when calculating the properties. Heat Treatment. Depending on the application, thermal treatments can be employed to: Increase strength Control thermal growth Improve ductility

Table 3 Relative ratings of aluminum-silicon sand casting and permanent mold casting alloys in terms of castability, corrosion-resistance, machinability, and weldability properties Property(a) Aluminum Association number of alloy

Resistance to hot cracking(b)

Pressure tightness

Fluidity(c)

Shrinkage tendency(d)

Resistance to corrosion(e)

Machinability(f) Weldability(g)

Sand casting alloys 319.0 354.0 355.0 A356.0 357.0 359.0 A390.0 A443.0 444.0

2 1 1 1 1 1 3 1 1

2 1 1 1 1 1 3 1 1

2 1 1 1 1 1 3 1 1

2 1 1 1 1 1 3 1 1

3 3 3 2 2 2 2 2 2

3 3 3 3 3 3 4 4 4

2 2 2 2 2 1 2 4 1

2 2 2 1 2 1 1 1 1 1 1 1 1 2 1 1

2 2 1 2 2 1 1 1 1 1 1 1 1 2 2 1

2 2 2 2 3 1 2 2 1 1 1 1 1 3 1 1

4 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2

3 3 4 3 4 3 3 3 3 3 3 3 3 4 5 3

3 2 2 3 2 2 2 2 2 2 2 2 1 2 1 1

Permanent mold casting alloys 308.0 319.0 332.0 333.0 336.0 354.0 355.0 C355.0 356.0 A356.0 357.0 A357.0 359.0 A390.0 443.0 A444.0

2 2 1 1 1 1 1 1 1 1 1 1 1 2 1 1

(a) For ratings of characteristics, 1 is the best and 5 is the poorest of the alloys listed. Individual alloys may have different ratings for other casting processes. (b) Ability of alloy to withstand stresses from contraction while cooling through hot-short or hrittle temperature range. (c) Ability of molten alloy to ﬂow readily in mold and ﬁll thin sections. (d) Decrease in volume accompanying freezing of alloy and measure of amount of compensating feed metal required in form of risers. (e) Based on resistance of alloy in standard salt spray test. (f) Composite rating, based on case of cutting, chip characteristics, quality of ﬁnish, and tool life. In the case of heat-treatable alloys, rating is based on T6 temper. Other tempers, particularly the annealed temper, may have lower ratings. (g) Based on ability of material to be fusion welded with ﬁller rod of same alloy. Source: Ref 13

Aluminum-silicon alloys containing copper and magnesium can be heat treated and aged in the same manner as wrought precipitation-hardened alloys. Depending on the strength level required, room-temperature aging (T4 temper) or elevated-temperature aging (<205 C, or 400 F) (T6 temper) may be required after heat treatment to obtain the necessary properties. As the strength of the alloy increases from the T4 to T6 temper, reductions in ductility will occur as the strength increases. In addition, aluminum-silicon and Al-Si-X alloys can be given a higher temperature (205 to 260 C, or 400 to 500 F) aging treatment from the as-cast condition to improve their strength and thermal stability. This is particularly important for applications where dimensional tolerances are critical (for example, when the alloy is operated at elevated temperatures as a piston component in an engine). Generally, such an aging practice is designated by the T5 temper. High-temperature (480 to 540 C, or 900 to 1000 F) treatments can also be given to aluminum-silicon and Al-Si-X alloys to improve their ductility. These thermal treatments modify the angular primary silicon particles to a more rounded shape. This rounded shape reduces the tendency for crack initiation beginning at the sharp edges of the particles. Such treatments are particularly effective on the hypereutectic alloys. Other means of modifying the shape for improved ductility are discussed in the sections “Modiﬁcation” and “Reﬁnement” in this chapter. Principles of Microstructural Control. The three categories of aluminum-silicon alloys are based on the silicon level (Table 1). These alloy categories are hypoeutectic, eutectic, and hypereutectic. The hypoeutectic alloys solidify with a-aluminum as the primary phase followed by aluminum-silicon eutectic. Other solutes (for example, iron, magnesium, and copper) form phases that separate in the freezing range of the alloy in the interdendritic locations (Fig 2). Hypereutectic alloys solidify in a similar manner, but in these alloys silicon is the primary phase rather than a-aluminum (Fig 3a). The eutectic alloys solidify principally with an aluminum-silicon eutectic structure; either aluminum or silicon is present as a primary phase depending on which side of the eutectic composition (12.7% Si) the alloy lies. A brief description of microstructural control is given below; additional information is available in Ref 2 and 12. Grain Structure. The grain size of the primary aluminum is controlled through the addition of heterogeneous nuclei to the melt in the form of master alloy inoculants such as Al-6Ti or Al-Ti-B (in the latter, the titanium can range from 3 to 5% and the Ti:B ratio from 3:1 to 25:1). Grain sizes vary from 100 to 500 mm ( 0.004 to 0.020 in.). An example of the effect of grain reﬁnement by an Al-Ti-B reﬁner is shown in Fig. 4. Cell Size. The interdendritic arm spacing (or cell size) is controlled by the cooling rate (Ref 19), which is in turn a function of the

230 / Casting Design and Performance casting process and section thickness. The smallest cell size is achieved with thin-wall high-pressure die casting. At the other extreme, thick-wall sand castings exhibit the largest dendrite cell size. Casting processes such as lowpressure die casting and permanent mold casting provide intermediate solidiﬁcation rates and consequently cell sizes that lie between the two extremes. In a similar fashion to the cell size, the constituent phase size is largely controlled by the freezing rate. Modiﬁcation. The term modiﬁcation refers to the change in morphology and spacing of the aluminum-silicon eutectic phase induced by the addition of a chemical agent such as sodium or strontium. There is a change from large divorced silicon particles to a ﬁne coupled aluminum-silicon eutectic structure with an addition of approximately 0.001% Na or 0.005% Sr to the melt. Varying degrees of modiﬁcation (Fig 5) are

Fig. 4 Effect of grain reﬁnement by the addition of an Al-5Ti-0.2B master alloy to type A356.0. (a) Without titanium addition. (b) With 0.04% Ti addition. Etched with Poulton’s reagent. 0.85

Variation in microstructure as a function of the degree of modiﬁcation. The modiﬁcation level increases from A to F; thus microstructure F is highly modiﬁed. Source: Ref 2, 12

obtained with lower levels of addition. For details of the mechanism and practice of modiﬁcation, see Granger and Elliott (Ref 12). Antimony is also used to modify (more accurately, reﬁne) the eutectic structure in hypoeutectic and eutectic alloys, particularly in Europe and Japan. Like sodium and strontium, it increases the ﬂuidity of the alloys and improves mechanical properties. Furthermore, it is permanent, allowing melts to be more effectively degassed, which, in turn, provides sounder castings. The great disadvantage of antimony is that it poisons (or negates) the effect of sodium and strontium, and it also creates a problem in recycling. An additional serious drawback is the potential for the formation of stibine gas, which is highly toxic. Unlike sodium and strontium, which can be used to effectively modify eutectic structures over a wide range of freezing rates, antimony provides eutectic reﬁnement only at the relatively high rates experienced in die castings and some thin-wall permanent mold castings. Reﬁnement. In hypereutectic alloys, the primary phase is silicon. In order to provide the desired small well-dispersed silicon particles, phosphorus is added to the level of about 0.1% P through the addition of a master alloy such as Cu-10P. The phosphorus combines with aluminum to form aluminum phosphide, AIP, which provides effective nuclei for the silicon phase much the same as TiB2-type particles are effective nuclei for a-aluminum (Fig 3b). However, phosphorus also negates the effectiveness of sodium and strontium. It does so by combining with them to form phosphides, which do not modify the eutectic structure. Similarly, sodium and strontium reduce the effectiveness of phosphorus additions by reﬁning the primary silicon phase. Gas Porosity. Hydrogen porosity can be controlled by maintaining gas levels at 0.10 cm3/100 g. This is not readily accomplished, particularly when modiﬁcation of the melt is being sought with the addition of sodium or strontium. However, gas ﬂuxing methods are available (Ref 20) that provide

the means of reducing hydrogen levels to the desired range. Also deleterious to casting soundness is the presence of nonmetallic inclusions that act as nuclei for gas pores. Various molten metal ﬁltration systems are available for inclusion removal (Ref 20). Sludge. A problem experienced with aluminum-silicon alloys is the formation of hard inter-metallic phases of the Al(FeM)Si-type, which settle out under gravity from the melt (Fig 6). The conditions that favor the formation of these phases are low holding temperatures (which are often employed in the die-casting industry); a quiescent melt; and relatively high levels of iron, manganese, and chromium. The relative tendency to form sludge in the holding furnace is given by a segregation factor (SF): SF ¼ ð%FeÞ þ 2ð%MnÞ þ 3ð%CrÞ

(Eq 1)

The relationship among the segregation factor, holding temperature, and sludging tendency is given for alloy AA 339 and several other aluminum-silicon alloys in Ref 21.

Wear Behavior The two major types of wear relevant to industrial applications of aluminum-silicon alloys are “abrasive” and “sliding” wear. These have also been identiﬁed by Eyre (Ref 3) as the most common types of wear. Wear mechanisms, though, can be thought of as involving more speciﬁc descriptions of local processes occurring in the metal and countersurface of the wear system during the wear process. The purpose of this discussion is to focus on the interaction between microstructure and wear mechanisms. This is important for aluminum-silicon alloys because of the variety of microstructures that can be achieved as the alloys are processed for particular applications. The relative effects of silicon particles, matrix hardness, and intermetallic constituents on the wear resistance of aluminumsilicon alloys are summarized below.

Fig. 5

Fig. 6

Coarse intermetallic Al12(Fe,Mn,Cr)3Si2 phase constituent generated by entrapped sludge in alloy 339. (a) 130. (b) 265

Friction and Wear of Aluminum-Silicon Alloys / 231 Silicon Particles. Under relatively light load conditions, which are normally associated with low (<1011 m3/m) losses, wear resistance is not a strong function of silicon content (Ref 22–25). In general, however, silicon additions to aluminum will increase the wear resistance. The principal mechanism appears to be the inﬂuence of the hard silicon particles, which lead to higher overall levels of hardness. The fact that the hard silicon particles are surrounded by a softer and relatively tough matrix improves the overall toughness of the material and can contribute to wear resistance by favoring more plastic behavior. The eutectic and hypereutectic silicon alloys, with increased volume fractions of hard primary silicon particles relative to the hypoeutectic alloys, might be expected to have the best wear resistance of the aluminum-silicon alloys. Andrews et al. (Ref 26, 27), for example, found that increasing the silicon content in hypereutectic alloys reduced wear. However, binary alloy data (Ref 22, 28) indicate that the hypereutectic alloys are not necessarily the most wear resistant. Clarke and Sarkar (Ref 28) found that there was a relative minimum in wear for binary aluminum-silicon alloys at about the eutectic level, as did Jasim et al. (Ref 22), especially at applied pressures <100 kPa (<15 psi). Clarke and Sarkar attribute the effects of silicon in part to its effect on metal transfer mechanisms between the pin and countersurface (Ref 29). There is also evidence for increased wear resistance with reﬁnement of the silicon particle morphology (Ref 30, 31). It is clear, therefore, that microstructure-based explanations are needed to account for the variation in wear rates with silicon content. Moreover, there is a need to account for the reduction in strength that occurs with increased silicon content (Ref 32, 33). The complex effects of composition on wear behavior suggest that wear resistance depends on other material properties (for example, fracture toughness) (Ref 34). Thus lower fracture toughness at higher levels of

Table 4 Typical hardness values of selected intermetallic constituents of aluminum-silicon alloys Hardness Phase

MPa

kgf/mm2

Ref

CuAl2

3900 3800–7600 7200 6400–9400 5160–7110 3500 6000–7600 7100 4500 9800–11,000 7000–14,200 11,880 4480 3700–3900 8400–9680 10,760

400 390–780 730 650–960 526–725 360 610–770 720 460 1000–1120 715–1450 1211 457 380–400 860–987 1097

36, 37 36 36, 37 37 36 38 36 37 38 36, 37 36 37 37 37 37 37

FeAl3

NiAl3 Ni2Al3 Si Mg2Si Al2CuMg Al9FeNi Al12Fe3Si

silicon could lead to higher wear rates if larger pieces of debris are created during the wear process. Variations in toughness and strength with composition might also account for the apparent ability of the near-eutectic compositions to have a greater load-bearing capability at a given wear rate than either higher or lower silicon levels. Matrix Hardness. Increased matrix hardness is typically achieved through the heat treatment response produced by copper and magnesium additions. Most commercial applications of aluminum-silicon alloys, in fact, depend on the increased strength achieved by heat treatment. The improved wear resistance of precipitationstrengthened material compared to solid solution strengthened material under low wear conditions was also noted by Soderberg et al. (Ref 35) using aluminum alloy 6061, which is strengthened primarily by Mg2Si precipitates. This is also the strengthening mechanism in the heat-treatable magnesium-bearing aluminum-silicon alloys. Although heat treatment has a beneﬁcial effect (Ref 26, 27, 32), variations in matrix hardness may be less important than the effects of silicon content (Ref 27). Intermetallic Constituents. In addition, there are important “other” hard phases present in commercial aluminum-silicon alloys that provide enhanced wear resistance. These constituents (for example, Al-Fe-Si, Al-Fe-Mn, Al-Ni, Al-Ni-Fe, Al-Cu-Mg) have varying degrees of hardness (Ref 36–38). Despite the apparent scatter, these constituents are all much harder than the aluminum matrix. Some examples of the hardness values of these intermetallic compounds are shown in Table 4. Typical room-temperature hardness values for the aluminum alloy matrices would be <1000 MPa (<100 kgf/mm2). The hardness values of the intermetallics decrease with increasing temperature (Ref 36), albeit at slower rates than the matrix hardness. The addition of “hard” phases in the form of particles or ﬁbers to reduce wear is also utilized to create metal-matrix composites (MMC) materials (Ref 39). These materials utilize hard intermetallic, cermet, or ceramic phases to provide the high hardness material for wear resistance. Hornbogen (Ref 40) and Zum Gahr (Ref 41) have described in quantitative terms how the contribution of these hard phases to wear resistance can be modeled in terms of their volume fraction and morphology. This composite approach has been effectively used to develop new piston materials (see the section “Metal-Matrix Composites” in this chapter). Finally, the use of “softer” constituents (for example, graphite) should also be noted as an active area for development of wear-resistant aluminum-silicon MMC materials (Ref 32, 39, 42). In these materials, ranking may depend on whether volumetric wear rates (in units of m3/m) or seizure resistance is being considered. The presence of the softer phase may lead to greater volumetric wear in some cases but greater resistance to seizure (higher load at seizure) in other cases.

To summarize, the results of wear studies using aluminum-silicon alloys illustrate a variety of mechanisms. The effect of variations in silicon particle morphology is often not clear cut, although heat treatment is beneﬁcial to the sliding wear resistance. Therefore, selection of an optimum microstructure is often difﬁcult in practical situations where several wear types or mechanisms could occur. In general, either eutectic or hypereutectic alloys offer the greatest wear resistance under a wide range of wear conditions. Selection may then hinge on the dependence of in-service performance on other alloy characteristics or cost. Overall, the aluminum-silicon alloy system provides a good basis for developing lightweight, strong, wear-resistant materials. Examples of these applications will be discussed in the following sections.

Aluminum-Silicon Alloy Applications Aluminum-silicon alloys are used in a variety of automotive, aerospace, and consumer product applications.

Automotive Components Table 5 lists typical automotive components made from aluminum-silicon casting alloys (Ref 43). The eutectic or nearly eutectic alloys (for example, 332, 336, and 339) (Ref 44), are perhaps the most widely used. Equivalent versions of these alloys are used for similar applications by European and Japanese automakers (Ref 45–47). Pistons. Typical applications for aluminumsilicon alloys in the French automotive industry are shown in Table 6 (Ref 45). In addition to being cast, the A-S12UN (eutectic) alloy can also be forged (Ref 48). Similarly, AA 4032, somewhat similar in composition to 336, is also widely used as a piston alloy (for example, for high-performance forged pistons). Hypereutectic alloys are also used for cast pistons, especially in diesel engines (Ref 45, 49). The potential beneﬁt from composites that combine the strength reinforcement of ceramics with an aluminum-silicon alloy matrix has also been evaluated (Ref 45, 47, 48). Engine Blocks and Cylinder Liners. The evolution of lightweight power plants has depended not only on lightweight pistons but also on the availability of wear-resistant cylinder liners and engine blocks. Hypereutectic liners were described by Mazodier (Ref 50) and El Haik (Ref 51). It was also known that hypereutectic aluminum-silicon alloys had excellent properties for engine blocks (Ref 52–54). This led to the development and application, in both the United States and Europe, of the A390 (A-S17U4) alloys for die cast engines (Ref 55–60). An important aspect of the A390 success is the use of a “system” (Ref 60) that includes the engine alloy, the

232 / Casting Design and Performance piston materials (electroplated cast F332[AA 332.0] alloy), and the cylinder bore ﬁnishing process. Fine honing to a 0.075 to 0.15 mm (3 to 6 min.) surface ﬁnish, followed by controlled etching/polishing to leave silicon particles standing slightly above the alloy surface, was deemed necessary for optimum wear resistance. Efforts to simplify the 390-type technology by ﬁnding a more wear-resistant alloy for the cylinder or reducing the difﬁculties of ﬁnishing the bore have led to substitute alloys. One approach has been the use of a lower silicon alloy containing more nickel and manganese (for example, the Australian-3HA alloy, with a nominal composition of Al-13.5Si-0.5Fe0.45Mn-0.5Mg-2Ni) (Ref 61). Continuing interest in the use of more highly wear-resistant materials in other engine-related parts has led to recent applications such as roller-type valve rocker arms (Ref 62) and valve lifters for the Toyota Lexus (Ref 63, 64). The rocker arm alloy used in the Mazda 929 is a nominal Al-10Si-2.7Cu-0.8Mg-0.45Mn alloy somewhat similar to the AA 383 alloy. The valve lifter, on the other hand, is a strontium-modiﬁed Al-Si-Cu alloy designated 4T12 (composition, Al-10.5Si-4.5Cu-0.6Mg-0.2Mn). Typical examples from a more detailed compilation of aluminum alloys used in wear-resistant applications in U.S. autos are shown in Table 7. Bearing Alloy Components. Aluminum alloys have been utilized for bearing applications for many years. The many uses range from diesel and internal combustion engines to a variety of tooling applications (for example, presses, lathes, and milling machines) (Ref 66). Important cast bearing alloys were based on aluminum-silicon or Al-Sn-Cu alloys, whereas wrought bearing alloys have included the 8xxx types (for example, AA 8081 and AA 8020) (Ref 66, 67). Compositions of various aluminum bearing alloys are listed in Table 8. Improved strength and fatigue performance, as well as some increased wear resistance, has been achieved with silicon additions. Thus, alloys SAE 780 and SAE 781 have become widely used for automotive applications such as main and connecting rod bearings (Ref 68, 69). The higher silicon alloy, 781, is also used in bushings and thrust bearings. Its improved wear performance has been attributed to the increased silicon content of the wear surface (Ref 70). These aluminum-silicon alloys are readily used with steel backing in high-load situations. Advanced Aluminum Bearing Alloys. The nominal compositions of improved bearing alloys with silicon additions are listed in Table 9. Although a soft phase (for example, tin) is normally considered desirable for avoiding seizure, the compatibility of an Al-11Si-1Cu alloy was better than that of the traditional aluminum-tin alloy (SAE 783) (Ref 71). The silicon-copper alloy also had much better fatigue resistance. The improved properties resulted in

applications such as diesel crankshaft and connecting rod bearings. Nevertheless, in line with the concerns expressed by Davies (Ref 72), the harder silicon-containing alloy was more sensitive to misalignment-induced seizure. The addition of silicon to aluminum-tin alloys containing lower tin levels than that of the 783 alloy provided a compromise between the conformability of the soft-phase material and the beneﬁts of the harder silicon phase for improved wear and fatigue resistance (Ref 73). This alloy could apparently be used without the common lead alloy overlays employed for seizure resistance. The importance of fatigue resistance was also emphasized in the improved Al-Sn-Si alloys reported by Ogita et al. (Ref 74). As shown in Table 9, these alloys are somewhat similar to those of Fukuoka et al. (Ref 73). As another alternative to the lead- or tin-containing alloys, a graphite-containing Al-12Si alloy has been successfully evaluated for bearings (Ref 75). The use of a modiﬁed lead plus indium addition to an Al-11Si-Pb bearing alloy has also been reported (Ref 76). Japanese concerns with the pollution and toxicity aspects of cadmium-containing alloys

have led to improved aluminum-zinc bearing alloys (Ref 77). These have also been improved with silicon additions. The additional matrix wear between the silicon particles is believed to create lubricant reservoirs that enhance sei-

Table 5 Automotive engine applications of aluminum-silicon alloys SAE alloy

Type of casting(a)

319.0 332.0 333.0 336.0 339.0 355.0 390.0

S PM PM PM PM S, PM D

A390.0 S, PM

Typical application

General purpose alloy Compressor pistons General purpose Piston alloy (low expansion) Piston alloy Pump bodies, cylinder heads Cylinder blocks, transmission pump and air compressor housings, small engine crankcase, air conditioner pistons Cylinder blocks, transmission pump and air compressor housings, small engine crankcase, air conditioner pistons

(a) S, sand cast; D, die cast; PM, permanent mold. Source: Ref 43

Table 6 Selected aluminum-silicon alloy applications in automobiles produced in France according to engine type and speciﬁc automobile manufacturer Manufacturer Engine type

Gas

Diesel

Citroen

Peugot

Renault

Talbot

A-S12UN ... ... A-S18UN ...

A-S10UN(F)(a) A-S12UN(A)(b) ... A-S12UN A-S13UN

A-S12UN ... ... A-S18UN ...

A-S10.5UN A-S11UN A-S12UN ... ...

(a) F, cast iron liner. (b) A, aluminum block. Source: Ref 45

Table 7 Wear-resistant aluminum-silicon alloys used in automotive piston components produced for United States automotive manufacturers in 1978 to 1985 model years Application

Manufacturer

Model/make

Model year(s)

Internal combustion engine components Pistons

American Motors Ford General Motors

All Mercury Buick Others

1978–85 1978–81, 83–85 1978–81, 84–85 1978–81, 83–85

Wheel cylinder pistons

American Motors Chrysler Ford General Motors

Master cylinder pistons

American Motors Chrysler Ford Ford General Motors

All All All All (except for Cadillac) Cadillac All All All Mercury All

1983–85 1983–85 1978–85 1984–85 1985 1984–85 1984–85 1983–85 1978–81, 84–85 1984–85

Ford Chrysler Ford

Some Some Some

1983–85 1984–85 1983–85

Brake system components

Transmission components Intermediate band servo pistons Rear band servo pistons Source: Ref 65

Friction and Wear of Aluminum-Silicon Alloys / 233 zure resistance. However, overlays (lead-tin alloys) are still required for best conformability. Finally, the combined effect of silicon and reﬁnement of the silicon- and lead-bearing phases by rapid solidiﬁcation processing has resulted in an improved Al-6Pb-4Si bearing alloy (Ref 78). This alloy has grown in usage recently and is projected to be used in 78% of the cars built in the United States during 1991. Because of the increasing understanding of the balance among wear, fatigue, and seizure resistance of bearing materials, silicon alloys have been used to develop new and improved bearing materials. Further improvements will undoubtedly be necessary as engine operating conditions evolve toward higher temperatures and operating speeds.

Consumer Electronics Components The growth of this market, which encompasses video cassette recorders (VCRs), video tape recorders (VTRs), digital audio tape (DAT) applications, and other devices (for example, personal computers), has created numerous opportunities for the use of lightweight, relatively corrosion-resistant and wearresistant aluminum-silicon alloys. VTR cylinders are speciﬁcally cited by various Japanese authors (Ref 79–81) because the eutectic- type silicon alloys have low coefﬁcients of friction against the tape.

Aerospace Components A nonautomotive engine application of aluminum-silicon alloys is the use of the 390type alloys in an aircraft engine (the Thunder engine) (Ref 82).

Breakthroughs in Aluminum-Silicon Wear-Resistant Materials Metal-Matrix Composites. Composite pistons were recognized early as a potentially viable application of MMC technology. While

some composite approaches, especially for the severe operating conditions of diesel pistons, recommended the addition of speciﬁc metallic inserts to achieve improved performance (Ref 83), the bulk of development efforts have gone into the incorporation of ceramic ﬁbers. The use of ceramic ﬁber aluminum-silicon MMC materials for pistons is described in a variety of publications (Ref 47, 48, 84–88). There are clear beneﬁts to strength at elevated temperatures and reduction of the thermal expansion coefﬁcient. These materials appear to be especially applicable in critical areas such as the top piston rings and top land (a hightemperature area). The castability of the aluminum-silicon alloys is a favorable factor in their use as matrices, particularly because squeeze casting is one of the preferred fabrication routes for composite pistons. The property improvements at elevated temperatures have encouraged ongoing development of the MMC technology for automotive engine applications, including engine blocks. The ability of the MMC approach to allow selective strengthening of the cylinder region was taken advantage of by Honda engineers (Ref 89), who utilized composite reinforcement of alloy ADC12 (a Japanese alloy similar to 383) in the manufacture of a die cast engine block. Powder Metallurgy. The combined beneﬁts of high silicon content and reﬁned silicon particle size on wear resistance are strong driving forces behind the use of P/M techniques for making aluminum- silicon alloy parts. The P/M approach has been of special beneﬁt to the hypereutectic silicon alloys. One example of this is the use of P/M A-S17U4 alloy to make cylinder liners (Ref 90, 91). The properties of these alloys exceed those of standard alloys (Ref 92). In addition, high levels of additional elements can be utilized to obtain good strength and wear-resistant properties at elevated temperatures (Ref 93–95). The reﬁned microstructures available from the P/M fabrication of hypereutectic alloys have a beneﬁcial effect on fatigue characteristics as well. This attribute has been utilized in the production of rotors and vanes for rotary

Table 8 Nominal compositions of standard aluminum-silicon alloys used in bearing applications Alloy Aluminum Association designation

850 8280 851 852 ... 8081 ... ... Source: Ref 66–68.

automotive air conditioners (Ref 96). For this application, P/M alloys with high levels of iron or nickel are blended with P/M 2024-type alloys to create alloys containing 17 to 20% Si and 5 to 8% Fe or Ni. Spray casting, as exempliﬁed by Osprey processing, has been shown to offer beneﬁts similar to powder metallurgy (Ref 97, 98). This has the potential for even greater cost savings, which is an important factor if the aluminum industry is to compete successfully in the automotive market. In a comparison of the structure in an Al-20Si alloy, Kahl and Leupp (Ref 97) showed that there was practically no difference in silicon particle size between the P/M and Osprey processing routes. Furthermore, they claimed that the ﬁne and uniform silicon particle size resulted in improved wear behavior compared with that of conventionally produced material, although details of their test procedure are not known. Earlier work at Delft University (Ref 99) compared the rate of mass loss of an Al-20Si-3Cu-1.3Mg alloy rubbing against cast iron at a pressure level of 5.5 MPa (800 psi). In this test, the Osprey material showed better resistance to wear than either the ingot metallurgy (I/M) or P/M samples. The reason for the better performance of the Osprey product compared with the P/M material was not clear, but it was hypothesized that the slightly larger silicon particles of the Osprey product helped reduce the fretting wear. Coatings/Surface Treatments. Other approaches to the wear (adhesion) problems of aluminum pistons moving in aluminum cylinders have taken the path of coating or surface treatment of the piston rings and cylinder bore to minimize wear problems (Ref 100). The selective ﬁber strengthening noted above is related to this problem also. One widely used treatment is the so-called Nikasil treatment (Ref 101), an electrochemical treatment utilizing a dispersion of silicon carbide particles, preferably <4 mm (<160 min.) in size, in a nickel matrix. Efforts to take advantage of both the P/M and surface treatment approach are exempliﬁed by the emerging surface treatment technologies of surface alloying via ion implantation (Ref 102–105), thermal spraying (Ref 106–08), and

Table 9 Nominal compositions of advanced aluminum-silicon alloys used in bearing applications

Composition, wt%

Composition, wt%

SAE designation

Si

Sn

Cu

770 780 ... ... 781 ... 782 783

0.7 1–2 2–3 0.4 3.5–4.5 0.7 0.3 0.5

5.5–7 5.5–7 5.5–7 5.5–7 ... 18–22 ... 17.5–22.5

0.7–1.3 0.7–1.3 0.7–1.3 1.7–2.3 0.05–0.15 0.7–1.3 0.7–1.3 0.7–1.3

Fe

0.7 0.7 0.7 0.7 0.35 0.7 0.3 0.5

Ni

Cd

0.7–1.3 0.2–0.7 0.3–0.7 0.9–1.5 ... ... 0.7–1.3 0.1

... ... ... ... 0.8–1.4 ... 2.7–3.5 ...

Alloy(s)

Si

Sn

Cu

Mg

Pb

Other

Ref

A B C D E F G H

11 3 12 ... 11 2.5 6 4

... 10 ... ... ... 12 ... 0.5

1 0.4 1 4.5 ... 1 1.2 0.1

... ... 1.5 ... ... ... 0.5 0.1

... 1.8 ... ... 20 ... 1 6

... 0.3 Cr 3 C, 1 Ni 3C 1.4 In 0.25 Mn 4 Zn 0.3 Mn

6 8 9 ... 10 11 12 13

(a) Arbitrary designations

234 / Casting Design and Performance surface treating or alloying using laser treatments (Ref 109–111). Ion implantation is recognized for its ability to impart wear-resistant surfaces, but principal applications have been to protect tool surfaces in critical processing operations. Laser hardening can be achieved through laser cladding to produce a chemically different surface or through the effective heat treatment (or remelting and rapid solidiﬁcation) brought about by laser heating. The work by Blank et al. (Ref 111) using 7 and 12% Si alloys, however, indicated that surface alloying (for example, with iron or iron plus vanadium) was more effective for increased wear than surface heat treatment effects alone. In any case, the ability to tailor surface properties to a technological need will enable engineers to obtain further enhancement of the wear resistance of the aluminum-silicon base alloys without sacriﬁcing their other advantages.

REFERENCES 1. J.P. Lyle and D.A. Granger, Ullman’s Encyclopedia of Industrial Chemistry, Vol 1A, VCH Publishers, 1985, p 481–528 2. E.L. Rooy, Aluminum and Aluminum Alloys, Casting, Vol 15, 9th ed., Metals Handbook, ASM International, 1988, p 743–770 3. T.S. Eyre, Wear Resistance of Metals, Treatise on Materials Science and Technology, Vol 13, Douglas Scott, Ed., Wear, Academic Press, 1979, p 363–442 4. A. Kaye and A. Street, Die Casting Metallurgy, Butterworths Scientiﬁc, London, 1982, p 32 5. S. Rabinowitz, The Expanding Role of Aluminum in the Automotive Industry, Die Cast. Eng., Sept-Oct 1, 1977 6. J.L. Jorstad, “Aluminum Lightweight Castings—Some Cost Saving Ideas,” Paper No. 770322, SAE Automotive Engineering Congress, Society of Automotive Engineers, 1977 7. J.L. Murray and A.J. McAlister, The Aluminum-Silicon System, Bull. Alloy Phase Diagrams, Vol 5 (No. 1), 1984, p 74–84, 89–90 8. Registration Record of Aluminum Association, Alloy Designations and Chemical Composition Limits for Aluminum Alloys in the Form of Castings and Ingot, 198709-01, Aluminum Association 9. Registration Record of Aluminum Association, Limits for Wrought Aluminum and Wrought Aluminum Alloys, 1987-04-01, Aluminum Association 10. L.F. Mondolfo, Aluminum Alloys: Structure and Properties, Butterworths, 1976 11. D.A. Granger, R.R. Sawtell, and M.M. Kersker, Effect of Beryllium on the Properties of A357.0 Castings, Trans. AFS, 1984, p 579–586 12. D.A. Granger and R. Elliott, AluminumSilicon Alloys, Casting, Vol 15, 9th ed.,

13. 14. 15.

16.

17.

18.

19. 20. 21.

22.

23.

24.

25.

26.

Metals Handbook, ASM International, 1988, p 159–168 J.E. Hatch, Aluminum: Properties and Physical Metallurgy, American Society for Metals, 1984, p 344–345 Source Book on Selection and Fabrication of Aluminum Alloys, American Society for Metals, 1978 A.L. Kearney and E.L. Rooy, Aluminum Foundry Products, Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Vol 2, ASM Handbook, ASM International, 1990, p 123–151 A.L. Kearney, Properties of Cast Aluminum Alloys, Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Vol 2, ASM Handbook, ASM International, 1990, p 152–177 J.E. Hatch, Chapters 8 and 9, Aluminum: Properties and Physical Metallurgy, American Society for Metals, 1984, p 152–177 General Information—Chemical Compositions, Mechanical and Physical Properties of SAE Aluminum Casting Alloys— SAE J452, June 1983, SAE Handbook, Society of Automotive Engineers, 1989, p 10.5–10.13 R.E. Spear and G.R. Gardner, Dendrite Cell Size, Trans. AFS, Vol 71, 1963, p 209–215 D.V. Neff, Non-Ferrous Molten Metal Processes, Casting, Vol 15, 9th ed., Metals Handbook, ASM International, 1988 J. Gobrecht, Gravity-Segregation of Iron, Manganese and Chromium in an Aluminum-Silicon Casting Alloy, Part I, Giesserei, Vol 62 (No. 10), 1975, p 263–266; and D.A. Granger, Investigation of a Star-Like Intermetallic Phase Occurring in Alloy 339, Trans. AFS, Vol 91, 1991, p 379–383 K. Mohammed Jasim and E.S. Dwarakadasa, Wear in Al- Si Alloys under Dry Sliding Conditions, Wear, Vol 119, 1987, p 119–130 B.N. Pramila Bai and S.K. Biswas, Effect of Load on Dry Sliding Wear of Aluminum-Silicon Alloys, ASLE Trans., Vol 29 (No. 1), 1986, p 116–120 R. Shivanath, P.K. Sengupta, and T.S. Eyre, Wear of Aluminum-Silicon Alloys, Wear of Materials—1977, International Conference on Wear of Materials, W. Glaeser, K. Ludema, and S. Rhee, Ed., American Society of Mechanical Engineers, p 120–126 R. Antoniou and D.W. Borland, Mild Wear of Al-Si Binary Alloys during Unlubricated Sliding, Mater. Sci. Eng., Vol 93, 1987, p 57–72 J.B. Andrews, M.V. Seneviratne, K.P. Zier, and T.R. Jett, The Inﬂuence of Silicon Content on the Wear Characteristics of Hypereutectic Al-Si Alloys, Wear of Metals—1985, K. Ludema, Ed.,

27.

28.

29.

30.

31.

32.

33.

34. 35.

36.

37.

38.

39. 40.

41.

American Society of Mechanical Engineers, p 180–184 J.B. Andrews, M.V. Seneviratne, and T.R. Jett, The Inﬂuence of Si and Other Alloying Additions on the Wear Characteristics of Hypereutectic Al-Si Alloys, 4th European Tribology Congress (EUROTRIB 85), Vol III, 1985 J. Clarke and A.D. Sarkar, Wear Characterization of As-Cast Binary AluminumSilicon Alloys, Wear, Vol 54, 1979, p 7–16 J. Clarke and A.D. Sarkar, The Role of Transfer and Back Transfer of Metals in the Wear of Binary Al-Si Alloys, Wear, Vol 82, 1982, p 179–185 T.K.A. Jaleel, N. Raman, S.K. Biswas, and K.K. Murthy, Effect of Structural Modiﬁcation and Load on the Wear of a Hypereutectic Aluminum-Silicon Alloy, Aluminium, Vol 60 (No. 12), 1984, p 932–933 Y. Odani, K. Akechi, and N. Kuroishi, High Strength and High Wear Resistance Al-Si P/M Alloys, Modern Developments in Powder Metallurgy, Vol 16, 1984, p 493–505 S. Das, S. V. Prasad, and T.R. Ramachandran, Microstructure and Wear of Cast (Al-Si Alloy) Graphite Composites, Wear, Vol 133, 1989, p 173–187 B.N. Pramila Bai, S.K. Biswas, and N.N. Kumtekar, Scanning Electron Microscopy Study of Worn Al-Si Alloy Surfaces, Wear, Vol 87, 1983, p 237–249 K.-H. Zum Gahr, Microstructure and Wear of Materials, Tribology Series, Vol 10, Elsevier, 1987 S. Soderberg, U. Bryggman, and A. Canales, Inﬂuence of Precipitation and Solution Strengthening on Abrasive Wear Resistance, Wear of Metals—1985, K. Ludema, Ed., American Society of Mechanical Engineers, p 645–653 E.R. Petty, Hot Hardness and Other Properties of Some Binary Intermetallic Compounds of Al, J. Inst. Met., Vol 89, 1960– 1961, p 343–349 A.A. Ivanko and G.V. Samsonov, Handbook of Hardness Data, Translation from Russian, National Bureau of StandardsNational Science Foundation, 1968 H.E.N. Stone, The Oxidation Resistance, Hardness and Constitution of Metallic Aluminides, J. Mater. Sci., Vol 10, 1975, p 923–934 S.V. Prasad and P.K. Rohatgi, Tribological Properties of Al Alloy Particle Composites, J. Met., Nov 1987, p 22–26 E. Hornbogen, Description of Wear of Materials with Isotropic and Anisotropic Microstructure, Wear of Metals—1985, K. Ludema, Ed., American Society of Mechanical Engineers, p 477–484 K.-H. Zum Gahr, Abrasive Wear of Two Phase Metallic Material with a Coarse Microstructure, Wear of Metal—1985,

Friction and Wear of Aluminum-Silicon Alloys / 235

42.

43.

44. 45.

46.

47.

48.

49.

50.

51.

52.

53.

54. 55.

56.

K. Ludema, Ed., American Society of Mechanical Engineers, p 45–57 B.N. Pramila Bai, E.S. Dwarakadasa, and S.K. Biswas, Scanning Electron Microscopy Studies of Wear in LM13 and LM13- Graphite Particulate Composite, Wear, Vol 76, 1982, p 211–220 General Information—Chemical Compositions, Mechanical and Physical Properties of SAE Aluminum Casting Alloys, SAE Handbook, J452, June 1983, Society of Automotive Engineers R. Wilson, Piston Improvement Relies on Material Science, Diesel Prog. North Am., Vol 52 (No. 2), Feb 1986, p 18–20 J. Ayel, The Inﬂuence of Metallurgical and Geometric Parameters on the Wear of Recent Engines, Ing. Automob., Sept– Oct 1980, p 82–98 M. Coulondre and J. Travaille, New Trends in the Design and Manufacture of Pistons and Cylinders for Automobile Engines, Ing. Automob., Sept–Oct 1980, p 68–72 A.R. Baker and S. Gazzard, Developments in Materials for Pistons, Mater. Des., Vol 9 (No. 1), Jan–Feb 1988, p 28–33 J.C. Avezou and A. Parker, Evolution of Materials for Pistons, Segments and Valve Seatings for Petrol and Diesel Engines of Automobiles, Ing. Automob., Oct 1983, p 73–80 J.P. Whitacre, “Hypereutectic Aluminum—A Piston Material for Modern High Speciﬁc Output Gasoline Engines,” TP 871944, Society of Automotive Engineers, 1987 M. Mazodier, Aluminum Engine Blocks with Thin Hypersilicon Alloy Liners, Ing. Automob., No. 10, Oct 1977, p 586–590 R. El Haik, The Possibilities of Using Al Cylinder Blocks—Hypersiliconized Cylinder Liner Tests, Ing. Automob., No. 10, Oct 1977, p 569–577 J.L. Jorstad, “Development of the Hypereutectic Al-Si Die Casting Alloy Used in the Vega Engine Block,” TP No. 105, 6th International Society of Die Casting Engineers Congress (Cleveland, OH), Nov 1970 J.L. Jorstad, The Hypereutectic Aluminum-Silicon Alloy Used to Cast the Vega Engine Block, Mod. Cast., Vol 60 (No. 4), Oct 1971, p 59–64 J. Ayel, Frictional Behavior of the Aluminum Hypersilicon Alloys Used as Engine Liners, Rev. Alum., Apr 1978, p 186–191 H.H. Hofmann and K. Schellmann, “Aluminum 390 Alloy Engine Blocks: Design and Manufacturing,” TP 830007, Society of Automotive Engineers, 1983 W. Dahm and R.G. Putter, “Light Alloy Engines—Experience Gained from Research and Development,” TP 830005, Society of Automotive Engineers, 1983

57. G. Renninger, D. Abendroth, and M. Bolien, “Casting Engine Blocks in GK Al Si 17 Cu 4 Mg,” TP 830003, Society of Automotive Engineers, 1983 58. E.G. Jacobsen, “General Motors 390 Aluminum Alloy 60 degree V6 Engine,” TP 830006, Society of Automotive Engineers, 1983 59. H.M. Ward, III, “Engineering Control for High Volume 390 Die Casting,” TP 830009, Society of Automotive Engineers, 1983 60. J.L. Jorstad, “Reynolds 390 Engine Technology,” TP 830010, Society of Automotive Engineers, 1983 61. J. Barresi, R. Giannini, and B.J. Clarsen, “Cold Scuff Evaluation of 3HA—The Engine Alloy Developed in Australia,” TP 871203, Society of Automotive Engineers, 1987 62. S. Komatsu, S: Yahata, K. Fukui, and T. Tokoshima, “Development of Aluminum Alloy Roller Type Valve Rocker Arm by Casting and Forging Technique,” TP 890554, Society of Automotive Engineers, 1989 63. S. Ezaki, M. Masuda, H. Fujita, S. Hayashi, Y. Terashima, and K. Motosugi, “Aluminum Valve Lifter for Toyota New V-8 Engine,” TP 900450, Society of Automotive Engineers, 1990 64. H. Baker, The Road Ahead for Metals in Autos, Adv. Mater. Proc., May 1990, p 27–34 65. Aluminum Association Guide to Aluminum Parts 1978–1984, Reynolds Aluminum 1985 Aluminum Automotive Applications 66. Chapter 20, Aluminum—Vol II: Design and Application, K.R. Van Horn, Ed., American Society for Metals, 1967 67. Chapters 8 and 9, Aluminum—Vol I: Properties, Physical Metallurgy and Phase Diagrams, K.R. Van Horn, Ed., American Society for Metals, 1967 68. “Bearing and Bushing Alloys,” J460e, SAE Handbook, Society of Automotive Engineers 69. “Bearing and Bushing Alloys,” J459c (Information Report), SAE Handbook, Society of Automotive Engineers 70. M. Gallerneaut and C. Fulton, “Score Resistance of SAE 781 Automotive Journal Bearing Alloy-A Metallographic Analysis,” TP 860062, Society of Automotive Engineers, 1986 71. G.C. Pratt, “Aluminum Based Crankshaft Bearings for the High Speed Diesel Engine,” TP810199, Society of Automotive Engineers, 1981 72. F.A. Davis, Plain Bearing Wear in Internal Combustion Engines, Automot. Eng. (UK), Vol 6 (No. 4), 1981, p 31–33, 37–38 73. T. Fukuoka, H. Kato, and S. Kamiya, Aluminum Alloy Bearings Containing Hard Particles Fitted for Use with Nodular Iron Shaft, Studies of Engine Bearings and

74.

75.

76.

77.

78. 79.

80.

81.

82.

83. 84.

85.

86.

87.

88.

Lubrication, SP-539, Society of Automotive Engineers, 1983, p 33–44 (TP 830308) Y. Ogita, K. Niwa, and Y. Kondo, “New Engine Type Fatigue Phenomenon of AlBased Engine Bearings,” TP 890556, Society of Automotive Engineers, 1989 B.P. Krishnan, N. Raman, K. Narayanaswamy, and P.K. Rohatgi, Performance of Aluminum Alloy Graphite Bearings in a Diesel Engine, Tribol. Int., Vol 16 (No. 5), 1983, p 239–244 D.P. Howe and J.D. Williams, An Al-SiPb-In Plain Bearing Material, Proceedings Fourth Conference of Irish Durability and Fracture Committee (IDFC4), 1986, p 183–191 M. Sakamoto, Y. Ogita, Y. Sato, and T. Tanaka, “Development of New Al-ZnSi Bearings for Heavy Load Applications in Uprated Engines,” TP 900124, Society of Automotive Engineers, 1990 Automot. Eng., Vol 97 (No. 12), 1989, p 21–25 S. Kitaoka, H. Kamio, and S. Kobayashi, Material Properties of Al-Si Alloys for Forging and Their Application, translation from Alutopia (Jpn.), Vol 17 (No. 10), 1987, p 9–18 S. Kitaoka, C. Fujikura, and A. Kamio, Aluminium- Silicon Alloys, translation of J. Jpn. Inst. Light Metals, Vol 38 (No. 7), 1988, p 426–446 O. Takezoe and Y. Yasuda, Technical Trends of Highly Wear Resistant Aluminum Alloy, Kobe Res. Dev., Vol 38 (No. 4), Oct 1988, p 23–26 R.P. MacCoon and R.P. Ernst, “The Thunder Aluminum 390 Alloy Aircraft Engine,” TP 830008, Society of Automotive Engineers, 1983 Automot. Eng. (London), Vol 7 (No. 3), June- July 1982, p 25–27 M.W. Toaz and M.D. Smalc, Clevite’s Advanced Technology for the Diesel Piston, Diesel Prog. North Am., Vol 51 (No. 6), June 1985, p 56 R.R. Bowles, D.L. Mancini, and M.W. Toaz, Design of Advanced Diesel Engine Pistons Employing Ceramic Fiber Reinforcement, Seminar—Advanced Composites: The Latest Developments, Nov 1986, American Society for Metals, p 21–27 T. Donomoto, K. Funatani, N. Miura, and N. Miyake, “Ceramic Fiber Reinforced Piston for High Performance Diesel Engines,” TP 830252, Society of Automotive Engineers, 1983 T. Suganuma and A. Tanaka, Applications of Metal Matrix Composites to Diesel Engine Pistons, J. Iron Steel Inst. Jpn., Vol 75 (No. 9), Sept 1989, p 1790–1797 R. Keribar, T. Morel, and M.W. Toaz, “An Investigation of Structural Effects of Fiber Matrix Reinforcement in Aluminum Diesel Pistons,” TP 900536, Society of Automotive Engineers, 1990

236 / Casting Design and Performance 89. T. Hayashi, H. Ushio, and M. Ebisawa, “The Properties of Hybrid Fiber Reinforced Metal and Its Application for Engine Block,” TP 890557, Society of Automotive Engineers, 1989 90. R. Perrot, Engine Sleeves Made in a Hypereutectic Aluminium-Silicon Alloy by Powder Metallurgy, European Symposium on Powder Metallurgy (Stockholm), Vol 2, 1978, p 212–217 91. R. Perrot, Engine Cylinder Liners in High Silicon Aluminum Alloy Produced by Powder Metallurgy, Rev. Alum., Mar 1979, p 129–135 92. N. Kuroishi, Y. Odani, and Y. Takeda, High Strength High Wear Resistance AluminumSilicon P/M Alloys, Met. Powder Rep., Vol 40 (No. 11), Nov 1985, p 642–645 93. J.L. Estrada and J. Duszczyk, Characteristics of Rapidly Solidiﬁed Al-Si-X Powders for High Performance Applications, J. Mater. Sci., Vol 25 (No. 2A), Feb 1990, p 886–904 94. K. Hummert and V. Arnhold, Possibilities of Powder Metallurgy Aluminum Materials in Motor Construction, 3rd International Symposium on Aluminium + Automobile, 1988, p 171–175 95. “High Strength Aluminium Alloy Powder Having Heat and Wear Resistance: Contains Silicon, Nickel, Iron and Manganese, and is Used for Cylinder Liners,” U.S. patent 4,938,810 96. T. Hayashi, Y. Takeda, K. Akechi, and T. Fujiwara, “Rotary Car Air Conditioner

97.

98.

99.

100.

101.

102.

Made with P/M Al-Si Wrought Alloys,” TP900407, Society of Automotive Engineers, 1990 W. Kahl and J. Leupp, “High Performance Aluminum Produced by Spray Deposition,” BNF 7th International Conference Material Revolution through the 90s F. Hehemann et al., “Comparison of Processing Atomized and Spray-Deposited Al-20 Si-5 Fe-2 Ni (wt.%) Aimed at Automotive Applications,” PM90, Institute of Metals, London J. Duszczyk et al., The Osprey Preform Process and the Application to Aluminum Alloys Including Wear and High Temperature Resistance Al-20 Si-Cu Alloys, Modern Developments in Powder Metallurgy, Vol 18–21, Metal Powder Industries Federation, 1988, p 441–453 D.A. Parker, Material Innovation and Its Impact on the Development of Components for Reciprocating Engines, Int. J. Veh. Des., IAVD Congress on Vehicle Design and Components, 1985, Interscience Enterprises, Ltd., United Kingdom, p B35.1–30 A.E. Ostermann, “Experiences with Nickel- Silicon Carbide Coatings in Cylinder Bores of Small Aluminum Engines,” TP 790843, Society of Automotive Engineers, 1979 J.D. Destefani, Ion Implantation Update, Adv. Mater. Proc., Oct 1988, p 39–43

103. C.M. Preece and J.K. Hirvonen, Ion Implantation Metallurgy, TMS-AIME, 1980 104. I.L. Singer, Tribomechanical Properties of Ion Implanted Metals, Ion Implantation and Ion Beam Processing of Materials, MRS Symposia Proceedings, Materials Research Society, Vol 27, 1983, p 585–595 105. J.H. Hirvonen, Industrial Applications of Ion Implantation, Ion Implantation and Ion Beam Processing of Materials, MRS Symposia Proceedings, Materials Research Society, Vol 27, 1983, p 621– 629 106. G.L. Kutner, Thermal Spray by Design, Adv. Mater. Proc., Oct 1988, p 63 107. M.L. Thorpe, Thermal Spray Applications Grow, Adv. Mater. Proc., Oct 1988, p 69 108. E.J. Kubel, Jr., Thermal Spraying Technology: From Art to Science, Adv. Mater. Proc., Dec 1987, p 69 109. C.W. Draper, Laser Surface Alloying: The State of the Art, J. Met., June 1982, p 24 110. J.M. Rigsbee, Surface Engineering by Laser and Physical Vapor Deposition Techniques, J. Met., Aug 1984, p 31 111. E. Blank, M. Pierantoni, and M. Carrard, Laser Surface Modiﬁcation of Al-Si Cast Alloys for Wear Resistance and Surface Toughness, Surface Modiﬁcation Technologies III, 1990, T.S. Sudarshan and D.G. Bhat, Ed., TMS-AIME, 1990, p 493–510

Casting Design and Performance Pages 237–245

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Failure Analysis of Castings* FAILURES OF CASTINGS, like the failures of wrought materials, can occur from service conditions, improper design and/or materials selection, manufacturing deﬁciencies, or a combination of these. To determine the cause of failure, the investigator must fully consider the interplay among design, fabrication, material properties, environment, and service conditions. Once the failure is accurately classiﬁed, corrective action and recommendations can be made. Appropriate solutions can involve redesign, change in materials selection or processing, quality control, changes in maintenance or inspection schedules, protection against environment, or restrictions on service loads or service life. The functionality of castings also is strongly dependent on casting soundness, which is inﬂuenced by various types of casting defects such as:

Metallic projections Cavities (e.g., gas porosity, shrinkage) Discontinuities (e.g., hot tearing, cold shuts) Defective surfaces Incomplete casting Incorrect dimension Inclusion or structural anomalies

Common types of defects in these seven categories are presented in “Appendix 1: International Classiﬁcation of Casting Defects.” However, it is important to recognize that the term “defect” should be used carefully in a failure analysis investigation. The word defect can evoke the notion of a condition that must be removed or corrected, when the condition just may be an imperfection that does not compromise function. Likewise, an improper design may the cause of conditions that induce failure from otherwise typically expected imperfections. Thus, Critical engineering assessment is needed to determine whether an imperfection is a defect that must be corrected to provide the required level of the reliability or intended function. The techniques used for investigating a casting failure are inﬂuenced by the possible causes of the issue involved. The evaluation and selected tests should take into account the possible causes for casting failures, which may

include the incorrect material used or speciﬁed, the presence of casting defects or anomalies, casting processing problems, design issues, improper service conditions or maintenance, and abuse of the casting with regard to the design intent. In order to properly conduct a failure analysis investigation, a structured approach must be used. This involves obtaining casting background information, planning the evaluation and selecting the appropriate casting for analysis, conducting a preliminary examination, conducting the proper material evaluations, and thoroughly evaluating the test data.

Background Information Before any testing is conducted, a signiﬁcant amount of background information needs to be known about the casting under investigation. Although this stage of a failure analysis is not always thought of as an element of “failure analysis technique,” it is an essential step in the process. All of the subsequent test results are limited somewhat if the casting background information is not known. In many cases the available background information is limited, but the following data should be obtained: Casting information Identiﬁcation of component (name and part

Service Conditions Date and shift of ﬁnal product assembly and/

or shipment

Service duration and storage Operating environment: temperature, atmo-

sphere, corrosive conditions

Frequency of occurrence of failures Types and location of loading on the casting:

tension, compressive, torsion, cyclic, impact

Anomalous service conditions Repair and rework history

The manufacturing and service history of a casting is especially important with regard to failures related to corrosion problems, as knowledge about the residue on the surface and information concerning the exposure to protective or corrosive environments can aid in the interpretation of the data. The heat treatment and service history is important for stressrelated failures since the fracture mode can be inﬂuenced by the prior thermal treatments or anomalous operating conditions. It is important to know the process and service history of the part, as an assessment of the possible residual stresses within the casting and the applied stress from assembly and the service environment is needed to be able to interpret the results of the study. For example, weld repair or a quench and temper treatment can result in signiﬁcant residual stress in a steel casting.

number)

Material composition (alloy grade) Applicable speciﬁcations and drawings

(casting and machining) Casting manufacturing process (sand, permanent mold, die casting, investment, centrifugal) Number of cavities on pattern/die In-gate and riser locations Casting date and shift Heat treatment Mechanical properties Physical characteristics Manufacturing process steps Weld-repair practice Attachment method and location within the assembly Exposure to process ﬂuids

Planning and Sample Selection The success of a failure analysis investigation depends on the planning of the project and selection of appropriate material for evaluation. It is important that the scope of the project be deﬁned at the initial stages of the investigation. The scope depends on a variety of factors that include the material available for evaluation, the timing for completion of the evaluation, and, of course, cost considerations. Under ideal conditions, there would be several samples to investigate, several weeks for the examination and testing, and no limit to the cost. In reality, most investigations have to be conducted with limited representative

*Adapted from article by Craig Brown, “Casting Failure Analysis Techniques and Case Studies” in ASM Handbook, Volume 15, Casting, 2008

238 / Casting Design and Performance samples of the failure condition, a deﬁned timetable, and at a minimum cost. Therefore, it is important that the material be carefully examined and that the samples selected are the most representative of the problem. Examples of failed parts must be thoroughly evaluated. In addition, it can be helpful to simultaneously evaluate comparable unfailed parts to differentiate any features from the failed parts. In many cases, old parts that have performed properly can be compared to new parts that are representative of the problem. Once the representative samples have been selected, the appropriate inspections and tests are deﬁned for the issue at hand. It is also important to look at the entire system in which the failed casting was used. Some important hints for the failure may exist on a different part that may have been installed away from the casting in the system. Try to obtain the entire assembly in which the failure occurred, not just the failed casting.

Preliminary Examination Visual inspection of the failed component, and any unfailed parts if they are available, is a critical element of a failure analysis. Too often, this part of the study is not conducted properly out of eagerness to start the material testing. However, a signiﬁcant amount of information about the casting and the history of its environmental exposure can be obtained at this stage of the investigation. Very often, small innocuous features are important clues in solving the puzzle of how the part failed. The failed part should be carefully examined and the features documented. The record can be written, but in most cases digital photographs should be taken of all pertinent features of the casting. Once subsequent testing has been conducted, the original features can be lost. A clear understanding and a complete record of the following items must be obtained: Scope and extent of damage to the casting Appearance of the fracture, if the casting is

cracked or broken

Determination of the fracture origin, the

direction of crack propagation, and identiﬁcation of the ﬁnal fracture region Presence of nonmetallic inclusions, shrinkage, or gas porosity or any other anomalies at or near the fracture origin Damage and wear to the surface of the part, both near to and away from the failure Dents, scuff marks, and collateral damage to the casting Presence of staining, fretting, or discoloration on the surface that appears to be related to the failure Extent of corrosion or pitting to the casting if the failure is corrosion related Deposits of debris, grease, or other material on the surface of the part Weld-repair evidence

Each of the various features provides information about the history of the casting. In many cases the features are appropriate for the casting, such as a shot-blasted surface or a drag mark from a die. In other cases, the features provide a record of the manufacturing process and service history of the part. It is important to understand what features should be present, based on the previously obtained background information, and what features are unusual or unique that would provide information about the failure. Before the part is destructively tested, it is important that there is a good understanding of all of the observations that have been obtained. Dimensional measurements should be taken during or after visual inspection, and the records identifying the locations should be studied. It is extremely beneﬁcial to have somebody who is familiar with the product available for discussion during this stage, as some innocuous marks on an assembly away from the casting may hold a very important key for the identiﬁcation of failure cause. For example, fretting marks in bolted joints away from the casting may be a sign of unexpected resonance vibration to the casting, which resulted from a loss of bolt torque in the joint. It is often very difﬁcult to connect the casting failures to these remote wear marks without consulting with a product expert. It is very important to examine all fracture surfaces to determine the actual fracture origin. The relationship between the crack-initiation site and any anomalies in the casting must be understood. Castings may have shrinkage porosity, oxide inclusions, or other casting anomalies. In many instances, the fracture exposes these features. Just because they are present does not mean that they are related to the failure. Cracks will take the path of least resistance through the material, and this often exposes casting anomalies. It is important that there be clear evidence that the fracture started at the anomaly to conclude that it contributed to the failure. In some cases, anomalies are present near the fracture origin but are not the cause of the failure. It is important to determine if the part would have failed at the identiﬁed location even if the anomaly had not been present. This determination requires an understanding of how the part is used and how the stresses were applied at the time of the failure. This issue emphasizes the importance of visual inspection as a critical component of the failure analysis process. Nondestructive evaluation (NDE) of the casting is often an important step in the failure analysis process. This typically involves the following processes: Radiographic examination (RT) to evaluate

the casting for subsurface features such as cracks, shrinkage porosity, gas porosity, nonmetallic inclusions, and other casting anomalies Ultrasonic inspection (UT) to evaluate the casting for subsurface features such as cracks, nonmetallic inclusions, or porosity

Magnetic particle inspection (MT) of ferrous

castings to investigate the presence of surface and near-surface cracks and other linear indications Liquid penetrant inspection (PT) of nonmagnetic castings to investigate the presence of surface cracks and porosity It is important to recognize that the nondestructive examination process can contaminate the casting surface. This is especially true for PT, MT, and UT. Therefore, if chemical analysis of the casting surface, fracture, or corrosion products is required, the chemical analysis should be conducted prior to any nondestructive testing. In some cases the NDE is intentionally not conducted as the value of the information obtained may be of less importance than the information that can be obtained from a subsequent chemical analysis. Careful visual examination often provides sufﬁcient details concerning the failure of the part.

Material Evaluation After completion of the preliminary examination and the proper photographic documentation of the failed casting, an evaluation of the fracture and material properties should be conducted. This typically involves fractography and microscopic examination followed by evaluation of the chemical composition, mechanical properties, and microstructure. It should be remembered that one test is worth many expert opinions, and actual data are more meaningful than assumptions about the properties of the part. Test results that conﬁrm that a property meets speciﬁcation, or is “typical,” are just as important as data that show a property to be abnormal. Therefore, sufﬁcient testing should be conducted to allow a proper evaluation of the casting.

Fractography Examination of the fracture surface should be conducted using both macroscopic and microscopic techniques. Macroscopic examination of the fracture surface, or other features of interest, would include examination of the casting with the unaided eye and with an optical stereomicroscope up to magniﬁcations of about 50. It is most important to study the features at the fracture origin. The location of the origin in relation to the casting surface and any other features should be identiﬁed. The features of any anomalies in the material should be studied and the direction of crack propagation through the material should be determined. This is especially important for castings that show the presence of nonmetallic inclusions, oxide skins, porosity, and weld repairs. Macroscopic examination often reveals the direction of the fracture. For example, for the case of bending-induced fracture, the side of the fracture surface that lies normal to the

Failure Analysis of Castings / 239 casting surface fractured earlier than the side on which the fracture surfaces do not intersect normal to the casting surface. Cracks typically will initiate at a site where stress concentration has occurred, and this may not always be at the surface of the part. This is especially true for parts that have failed by fatigue. In addition to examining the fracture origin, other features of the failure, such as beach marks, arrest lines, and steps in the fracture should be studied. With most fractures, the direction of crack propagation or extension through the material can be determined by examination under oblique illumination. Ridgelike fractures features, such as chevron marks, will typically point back to the fracture origin. It is often helpful to examine the casting and draw ink arrows on the part adjacent to the fracture to indicate the direction of the crack propagation. It is important to study intersecting cracks to determine which crack occurred ﬁrst. In this way, the origin can be clearly identiﬁed. The features at the origin will expose how the part failed. If the origin is not identiﬁed, then the mechanism of how the part failed may be only conjecture. Microscopic examination of the casting involves examination of the part at magniﬁcations up to 10,000 or higher, typically using a scanning electron microscope (SEM). The critical features found during visual examination should be studied at high magniﬁcation. It is most important that the features at the origin(s) be examined. The shape, size, and characteristics of the ﬁne details should be studied and representative micrographs taken. The fracture mechanism should be identiﬁed at the location where the fracture started. Too often, a study of the origin is not done properly and the root cause of a failure is missed. It is especially important to study the fracture origin regions of castings, as the parts can have a complex microstructure and may contain anomalies, such as nonmetallic inclusion, porosity, or weld-repair artifacts. The fracture mode should correlate with the microstructure of the material as well as the expected stresses on the part. Examination of the fracture surface away from the origin may reveal features that are different from those at the origin, and the crack could be propagating via a different mechanism. For example, a casting may have crack initiation by stress-corrosion cracking, but crack propagation could be by fatigue. Corrective action for fatigue alone may not solve the problem. The crack-origin region should be examined at magniﬁcations up to 5000. In many cases, the ﬁne features that help characterize and identify the fracture mode cannot be discerned unless they are studied at magniﬁcations above 1000. Other areas that should be studied are the regions slightly away from the origin and at the midfracture zone. This allows conﬁrmation of the fracture mode at locations where the fracture extended through the material. It is also important that the fracture mode at the ﬁnal fracture zone be identiﬁed to clarify whether the cast material was inherently ductile or brittle.

As with all fracture studies, any other features found on visual inspection or with the SEM should be studied and documented. With castings, it is important to know the microstructure of the alloy since the crystallographic structure of the metal will inﬂuence the fracture features. The normal features need to be recognized, so that any abnormal characteristics can be identiﬁed. For example, fracture through a gray iron can be somewhat complex; it would have smooth, wavy surfaces where the separation occurred at the graphite ﬂakes, ductile rupture at the ferrite-rich zones, and cleavage fracture at the pearlite-rich zones. Shrinkage porosity may exhibit globular cavities with a dendritic structure, which would be different from the fracture through the graphite ﬂakes. Any deposits or nonmetallic materials should also be studied to characterize their features. The features help determine whether the deposits are anomalies in the material, corrosion products, postfracture contamination, or some other condition. The shape, size, and distribution of the deposits help tell the story of how the casting failed. There should always be an explanation for all of the features found on examination. It is often beneﬁcial to take micrographs at low, medium, and high magniﬁcation at any selected location. This allows the investigator to correlate the features depicted at high magniﬁcation with the surrounding fracture surface.

Chemical Analysis The chemical composition of any failed casting should be determined to conﬁrm that the part was made from the speciﬁed material. The analysis should be conducted using the appropriate analytical technique for the alloy, with the analysis veriﬁed against check standards of a similar alloy. Some of the more common analytical techniques include inductively coupled plasma spectroscopy (ICP) with combustion analysis for the carbon and sulfur, optical emission spectroscopy (OES), glow discharge spectroscopy (GDS), x-ray ﬂorescence spectroscopy (XRF), and other variations of these methods. X-ray diffraction (XRD) may need to be used to characterize any deposits or unusual phases. It should be noted that some castings could exhibit signiﬁcant variations in the composition caused by alloy segregation during solidiﬁcation. This is especially true with heavy section castings and should be taken into account when the material is being sampled. For castings in which segregation is commonly observed, such as aluminum alloys, the chemical analysis samples for OES should be prepared by melting several specimens and casting them into sample disks for subsequent analysis. An analysis of a single piece cut from the casting can give an erroneous result. Variation can also occur from part to part when an additive has been added to the melt prior to pouring. For example, cast iron may exhibit a

loss of magnesium when comparing a part from the beginning of the pour to a part from the end of the pour. At a minimum, the weight percent of the elements deﬁned in the standard composition ranges should be determined. With some materials, such as carbon and low-alloy steels, consideration should be given to determining the concentrations of other elements, such as boron, titanium, niobium, vanadium, calcium, nitrogen, hydrogen, and oxygen as these elements inﬂuence the hardenability, precipitation of carbides, shape of the manganese sulﬁdes, and susceptibility to brittle fracture. With siliconcontaining aluminum castings, determination of the strontium and sodium contents can help interpret the expected degree of modiﬁcation of the eutectic silicon in the microstructure. Metals often contain alloying elements at low concentrations that drastically change the microstructures and properties. These elements may not always be speciﬁed in the standard ranges. For example, the manganese sulﬁde shape can be altered depending on small variations in the aluminum and titanium content of low-alloy steel. In-depth understanding of metallurgy is necessary to interpret the analysis of chemical composition against observed failures. Energy dispersive x-ray spectroscopy (EDS) is used with the scanning electron microscope to analyze the surfaces studied during a failure investigation. In most instances this technique does not provide quantitative results, but does provide the relative proportions of the elements detected. It is a very powerful tool, but the limitations of the technique should be appreciated. The compositions of very small areas and deposits may be determined, but it is important to interpret the results properly. EDS analysis should be conducted on the surfaces and deposits of interest, but also on clean areas of the base metal to compare the results with those expected for the alloy under the same test conditions. This is especially true for EDS analysis of fracture surfaces. Typically, fractures occur through and around phases and precipitates in the matrix, resulting in a disproportionately high concentration of those elements in the exposed phases. For example, the fracture through a silicon-containing aluminum alloy, such as 380.0, will expose a signiﬁcant amount of the eutectic silicon particles. It is not unusual for EDS analysis of the fracture surface of a die casting to exhibit a silicon content of over 20%, while the base metal may only actually contain 10%. Therefore, the interpretation of the EDS test results must be handled in a similar manner as the fracture analysis, as the results are dependent on the fracture mode and the microstructure of the alloy. Another caution associated with the EDS analysis is the excitation volume of the x-ray. Under a typical analysis mode, the x-ray emission occurs within a layer that is up to 5 mm beneath the surface. Elements present in secondary phases below the surface that are not visible by the secondary electron images may

240 / Casting Design and Performance show up in the x-ray spectrum. This can confuse an identiﬁcation of contaminants in the analysis. The peak overlap is another source of confusion in this analysis. In some ferrous alloys, peaks of elements found in the slag can overlap with peaks of elements that are commonly observed in the base material.

Mechanical Testing Mechanical testing of a failed casting should be conducted to determine whether the properties of the part meet the speciﬁcation requirements. Usually the hardness and tensile properties are determined, but with steel castings and other high-alloy materials, the impact properties as well as fracture toughness should also be considered for evaluation. In some cases, the shear or fatigue properties of the material may need to be determined. In most cases, if the specimens were extracted from the proper location, determination of these properties conﬁrms that the castings met the expected values. The mechanical test results, along with the microstructure, often conﬁrm that the casting was properly heat treated. Mechanical testing should be done for most failure investigations, as knowing that the mechanical properties are acceptable is just as important as knowing that they are not. Brinell hardness testing typically is used for failed castings since the indentation covers a greater surface area than the Rockwell hardness test resulting in less error from variations in the microstructure. Rockwell hardness testing of steel and aluminum castings provides acceptable data, but should not be used on iron castings because the graphite in the microstructure will result in values that are skewed to the soft side. It is important that the hardness impression be made on sound material and not on subsurface regions that have shrinkage porosity. With steel castings that have been heat treated, it is important that the test be conducted on material below any surface decarburization. Microhardness testing may be appropriate when localized heat treatments, such as carburizing or induction hardening, have been conducted on the failed casting. Knoop or Vickers microhardness tests can be conducted, but the choice of which test method to use is dictated by the hardness conversion tables available for evaluation of the test results. The ASTM standard E 140 is used for converting the data to the customary Brinell and Rockwell hardness values. Hardness surveys through the selected area usually are conducted, but it is important to know the microstructure at the area tested because any anomalies, such as decarburization, will inﬂuence the test results. The case depth or thickness callout speciﬁed on a component drawing is often incomplete. The investigator should understand the correct meaning of the case-depth requirement. For example, is it the total case or the effective case depth to a 50 HRC equivalent?

The standard or subsize tensile specimens usually provide acceptable test results if the specimens have been extracted from sound material. The objective of the mechanical testing is to determine the properties of the alloy as near to the fracture origin as possible from sound metal that is representative of the material. It is important that the tensile specimen not contain signiﬁcant amounts of shrinkage porosity in the reduced section of the bar. If shrinkage porosity, oxide inclusions, or other anomalies are present, the specimen generally will exhibit ultimate tensile strength and percent elongation values less than expected for the sound material. In many cases, the machining of round tensile specimens shifts the reduced section of the tensile bar toward the centerline of the section, and this can result in low tensile and elongation values. In these cases, ﬂat tensile specimens taken from the midradius region or near the surface of the casting will provide more representative values for the material. In some cases, the material speciﬁcation provides allowance for extraction of the test specimens from castings as opposed to separately cast test coupons, but selection and agreement on the location from which the tensile specimens are extracted is an important part of a failure investigation. It should also be understood that the cast alloys can exhibit signiﬁcant variation in properties at different locations within the part, even if sound specimens can be extracted. The cooling rate during solidiﬁcation will inﬂuence the properties. For example, with cast aluminum alloys, there is a direct correlation between the dendrite arm spacing and the mechanical properties. This should be taken into account when evaluating the test data since the tensile specimen can seldom be extracted from the casting at the fracture origin location. Many commonly used cast metals specify their mechanical properties by separately cast tensile specimens. The investigators need to understand the differences between the separately cast tensile specimens and the failed castings. For ductile cast iron parts, the failed parts may be cast in different molding materials from the separately cast coupons. Secondary processes such as heat treatment may reduce the differences between the actual castings and separately cast tensile test specimens. For example, high-pressure die-cast aluminum parts usually exhibit signiﬁcantly lower properties than separately cast specimens. These castings are usually used in the as-cast condition; however, the difference in properties may be reduced after T6 heat treatment. The investigators need to understand the behaviors of each casting metal in detail to accurately interpret the results of mechanical testing during failure analysis.

Metallography Examination of the microstructure at the fracture origin is a critical component of a thorough failure analysis. It is important that the microstructure at the location where the failure

started be inspected to determine if the part was properly heat treated and to determine if there are any casting anomalies that may have inﬂuenced the failure. Too often, failure investigators evaluate the casting at a location away from the origin and assume the microstructure is similar to that at the origin region. This is not always the case, due to the variations in the cooling rate during solidiﬁcation or heat treatment, the presence of decarburization, and the presence of nonmetallic inclusions, porosity, or some other near-surface anomaly. The preparation of the cross section must be done in such a manner that the microstructure is not altered during cutting, grinding, mounting, and polishing. It is also important that proper specimen preparation be used to maintain good edge retention because the features at the surface of fracture origin are of primary interest. Action should also be taken to ensure a good bond to the mounting media is present or subsequent bleedout of the etchant can obscure the microstructural features. The microstructure should be examined in both the unetched and etched conditions. With carbon and low-alloy steels it is important to evaluate the shape and distribution of the manganese sulﬁdes in the unetched condition. In some cases it is difﬁcult to see ﬁne cracks in the material after the specimen has been etched. Therefore, the features in the unetched condition should be documented photographically prior to etching the cross section. The specimen should be etched with the appropriate etchant for the alloy involved. In some cases, more than one etchant may need to be used to identify phases such as ferrite, carbides, and unique constituents, such as sigma phase. The micro-structure should be evaluated to determine if it is proper for the speciﬁed material at the fracture origin, the near-surface zone, and away from the fracture origin. Variations in the microstructure can provide information about the heat treatment and alloy segregation. Castings will often show a dendritic solidiﬁcation structure with slight alloy depletion in the dendrites and alloy enrichment in the interdendritic regions. This variation can inﬂuence the relative proportions of the constituents of the microstructure and may have a bearing on the root cause of the failure.

Special Testing In some cases, special testing may be required as part of the failure analysis investigation. Stress analysis, strain gaging, and fatigue testing of components may need to be conducted to determine or conﬁrm the properties of the component/material involved in the failure, or the component/material to be used as part of a corrective action. As with the original failure investigation, a substantial amount of information about the material and component properties is needed prior to testing to set up and conduct a valid test program. In other cases, environmental testing or accident reconstruction may need to be conducted.

Failure Analysis of Castings / 241

Data Analysis and Report Preparation On completion of the material evaluation a thorough analysis of the background information, preliminary examination, and the test results must be conducted. The test results should show a correlation with the casting process and the manufacturing operations. It is important to keep an open mind and “let the data tell the story about the casting.” The “picture” of the failure should include all the pieces of the puzzle provided by the data. If any data or piece of information does not ﬁt into the picture or description of the failure, then an improper conclusion can be drawn and further information or data are needed to make sense of the results. In most cases, the simplest explanation of the data is the best solution to the failure analysis problem. Often, very minor details are very important and must be explained in the conclusion. It should also be remembered that arriving at the wrong conclusion could be worse than having no conclusion. It is also important that there are sufﬁcient data to come to a conclusion. If the information is inconclusive, or only partly explains the situation, then this should be stated. All failure analysis projects should conclude with a report that summarizes the background information, the results of the preliminary studies and the material tests, and the conclusions of the study. The scope of the report can range from a short summary statement to a very detailed presentation of the data. It is important to present the data in a logical and clearly written manner, so that the reader can follow the test plan. Typically, the data are presented with only a basic interpretation of the information in the body of the report. The detailed discussion and interpretation of all of the data should be in a separate section. It is important to separate facts from speculation. Different analysts may draw different conclusions from the same results. Therefore, the report writer should ﬁrst present the data and then provide an interpretation of the ﬁndings.

Casting Fracture Characteristics The characteristics of casting fractures are, in general, similar to those of wrought alloys, but in many cases can be more complex. When examining fractures in castings it is important to fully understand the fundamentals of fracture mechanisms and the basic fracture modes that are typical for the alloy involved. As noted, it is important to know basic information about the part. Background data concerning the alloy type and grade, the expected mechanical properties, the overall heat treatment, and any special surface treatments or processes should be obtained. It is also essential to know the history of the part. Did the

fracture occur during the manufacturing process or when the part was in service? It is important to understand the loads applied to the part as well as the operating environment. This background information can provide some insight into the expected fracture characteristics.

Examination Methods The evaluation of a failed casting must include visual inspection of the whole part and other related components as well as the actual fracture. The part should be examined thoroughly to evaluate both the overall features and any unusual features. Details such as surface damage, distortion, machining marks, fretting, corrosion deposits, and so forth all provide information about the environment and relationship to other components. The examination of the fracture should be done using good illumination to see the ﬁne details of the surface. Oblique illumination should be used to evaluate the texture and proﬁle of the fracture. This type of illumination will help determine the location of the fracture origin. The visual inspection should be conducted both prior to and after cleaning or removal of surface contamination. The fracture should be examined with an optical or digital microscope at the critical regions of the origin, midfracture, and ﬁnal fracture regions. This may allow detection of any casting ﬂaws. In some cases the fracture origin can be at a subsurface anomaly, such as gas/shrinkage porosity, a nonmetallic inclusion, or anomalies from weld repair. The precise location of crack initiation may only be discernible with aid of an optical/digital microscope. In order to determine the fracture mode, which is how the casting broke, the critical regions of the fracture should be examined with a scanning electron microscope (SEM). The ﬁne details that are characteristic of ductile rupture, transgranular brittle fracture, intergranular fracture or fatigue, and even some ﬂaws, are not discernible with the unaided eye or with the optical/digital microscope. These features can only be properly observed when examined at magniﬁcations above approximately 500. The SEM allows routine examinations at magniﬁcations up to 10,000. Regions of Interest. The primary objective of the visual inspection is to determine the location of the fracture origin and evaluate how the crack propagated through the alloy. Knowing where it started and where it ended provides information about how the part was loaded. The origin region should be examined to determine where the crack initiated. Did the crack start at a single point or over a broad area? Cracks typically originate at a site of stress concentration such as a surface anomaly, grinding marks, cast or machined notch, or subsurface ﬂaw. The mid-fracture region exhibits features that indicate how the crack propagated through the material. This gives insight as to whether

the part was subjected to cyclic, tensile, or torsional loading. The edges of the fracture can exhibit slant fracture, or shear lips, that can help identify the origin region. This is especially true for cast steels.

Factors That Affect Fracture Appearance The fracture texture at the origin region can have features that indicate where the crack started. With overload fractures, steel castings will typically exhibit a chevron pattern that will point back to the origin region, as shown in Fig. 1, whereas aluminum castings can exhibit a relatively rough texture in the origin region and a smoother texture at the other side of the broken section. This is especially true when there is no distinct stress raiser in the origin region of the aluminum part. With fatigue fractures, most cast alloys will exhibit thumbnailshaped beach marks or arrest lines that fan outward from the origin region. The fracture surface texture of cast components can exhibit a different texture compared to a similar wrought product. With some alloys, such as die-cast aluminum and improperly controlled gray or ductile iron, the near-surface region can have a chill zone that will exhibit a ﬁne, smooth fracture texture. Away from the surface the fracture may expose various anomalies such as shrinkage or gas porosity and nonmetallic inclusions in the form of sand, dross, slag, or reoxidation products. Fatigue fractures can exhibit a rougher texture as the crack propagates through and around the anomalies. With heavy-section steel castings in the quench and tempered condition, alloy segregation can cause varying degrees of both ductile and brittle fracture due to differences in the proportion of martensite and bainite. Weld repairs can have a signiﬁcant inﬂuence on the potential for crack initiation, at the surface or at subsurface regions. Residual stresses

Overload fracture through a low-alloy steel casting. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 1

242 / Casting Design and Performance

Cast components typically exhibit the usual characteristics of ductile overload, such as necking and distortion, a dull ﬁbrous fracture with little branching, and shear lips. On examination at high magniﬁcation, ductile rupture typically exhibits dimples, which are formed by micro-void coalescence (MVC). Nucleation typically occurs at nonmetallic inclusions or second-phase particles. Carbon, low-alloy, and stainless steel castings will typically exhibit classical dimples. A low-alloy steel in the quench and tempered condition is shown in Fig. 2. A similar fracture mode can occur in a ductile iron, but the appearance is altered due to the presence of the nodules, which provide sites for MVC, as shown in Fig. 3. Dimples

formed around the nodules. The size and shape of the dimples is controlled by the nodule count and nodularity. Ductile rupture in cast silicon-aluminum alloys can have different characteristics, depending on the heat treatment of the alloy. In the as-cast condition, the primary and eutectic silicon can have an angular and ﬂakelike structure. Rupture through the alloy will occur at and through the silicon particles and also through the aluminum matrix. The fracture will exhibit brittle features at the silicon particles, but tear ridges and dimples will be present at the matrix. These features are illustrated in Fig. 4, which shows the fracture through a type 380.0 die-cast aluminum component. The larger the primary and eutectic silicon particles, the smaller the proportion of ductile features will be. Heat treatment of a similar alloy would cause spheroidization of the silicon particles. These silicon particles would provide sites for MVC, and the fracture would exhibit dimples, as shown in Fig. 5. The ductility, as measured by the percent elongation during tensile testing, would be substantially higher for a fully spheroidized alloy, and this correlates with the fracture appearance.

Ductile rupture in a low-alloy steel casting. Original magniﬁcation: 3000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Overload fracture through a type 380.0 aluminum alloy in the as-cast condition. Original magniﬁcation: 1000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Ductile rupture in a ductile iron. Original magniﬁcation: 150. Courtesy of Stork Technimet, Inc. New Berlin, WI

Overload fracture through a type 356.0 aluminum alloy in the T6 condition. Original magniﬁcation: 500. Courtesy of Stork Technimet, Inc. New Berlin, WI

from weld repair and improper grinding as well as rough ﬁnishing and carbon arc gouging can cause crack initiation. Incomplete removal of casting defects and weld anomalies, such as slag, shrinkage cracks, or cracks from improper weld process control have been found to cause failures in steel castings.

Ductile Rupture

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Transgranular Brittle Fracture Transgranular brittle fracture in cast components typically exhibits the usual characteristics of brittle overload, such as little or no distortion, ﬂat fracture, bright crystalline appearance, and a chevron pattern that points to the origin. Examination at high magniﬁcation will show cleavage fracture through the grains. The transgranular fracture is related to the nilductility transition temperature (NDTT) of the alloy, the strain rate, or brittle phases in the microstructure. As with wrought products, the NDTT may be shifted to a temperature above the operating temperature of the component. This can occur with carbon and low-alloy steels as well as ductile iron. An example of cleavage fracture in a ductile iron is shown in Fig. 6. This contrasts with the ductile rupture features shown previously. In some cases, such as ductile iron and low-alloy steel, the fracture may exhibit a mixture of both cleavage fracture and ductile rupture. A typical example of a mixed fracture mode is shown for a low-alloy steel casting in Fig. 7. In some cases, the cleavage fracture is associated with the pearlite and the ductile rupture with the ferrite in the microstructure.

Cleavage fracture in a heavy-section ductile iron casting. Original magniﬁcation: 200. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 6

Mixture of cleavage fracture and ductile rupture (arrows) in a low-alloy steel casting. Original magniﬁcation: 1000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 7

Failure Analysis of Castings / 243 It is important to be aware of the microstructure of the material, as the chill zone of a ductile iron can exhibit brittle features when the fracture extends through the iron carbides in the iron. The example presented in Fig. 8 is from the same casting shown in Fig. 3, but from a region that was chilled during solidiﬁcation and cooling. Another example of how the microstructure of an alloy can signiﬁcantly inﬂuence the features of a fracture is gray iron. Gray iron is typically considered to be a brittle material as the fracture surface exhibits the characteristic macrofeatures of a brittle alloy. Examination at high magniﬁcations shows that the fracture through a gray iron is primarily through and at the graphite ﬂakes. Only scattered rupture through the matrix occurs between the large ﬂakes. Therefore, the fracture will exhibit smooth curved surfaces that replicate the texture of the graphite, and only isolated locations with either cleavage or ductile rupture, as shown in Fig. 9. It is important to be aware that carbon and low-alloy steels can often exhibit quasi-cleavage. If cast steel is properly quenched and tempered to form a microstructure of tempered martensite, the fracture will typically exhibit dimples, which are indicative of ductile rupture. If the casting is not quenched at a sufﬁcient

Brittle fracture through the near-surface chilled region of a ductile iron. Original magniﬁcation: 1000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 8

Fracture through the matrix between the graphite ﬂakes of a gray iron. Original magniﬁcation: 250. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 9

cooling rate and has a microstructure of tempered martensite and some ferrite, the fracture will exhibit features that are a blend of cleavage fracture and ductile tear ridges, often called quasi-cleavage. A typical example for lowalloy steel is shown in Fig. 10.

Fatigue Fatigue fracture in cast components typically exhibits the usual features of a progressive fracture mode, such as relatively ﬂat fracture surface with beach marks and arrest lines, ratchet marks between the origins, striations when examined at high magniﬁcation, and an overload fracture at the ﬁnal fracture region. Crack initiation will occur at a site of stress concentration. This may be at the surface or at a subsurface anomaly. With cast alloys the origin is often found at porosity or nonmetallic inclusions. An example of crack initiation at shrinkage porosity in a low-alloy steel casting is shown in Fig. 11. As with wrought alloys, a fatigue crack in carbon and low-alloy steels will propagate through the material via microscopic steps and produce ﬁne striations, which cannot be seen with the unaided eye, but can be discerned with a SEM at magniﬁcations above

500. Typical striations in a low-alloy steel suspension component are shown in Fig. 12. It is important to understand the difference between these striations and beach marks or arrest lines. Beach marks are visually evident on most fatigue fractures in cast components and are shown in Fig. 11. The visual, thumbnail, or curved lines that radiate from the origin region are caused by variations in the cyclic stresses and the crack propagation rate. These marks are not always indicative of fatigue fracture, but can also be caused by incremental overload steps, although the overload steps are typically less uniform and have a rougher texture. Conﬁrmation of a fatigue fracture should be done by examination at high magniﬁcation with a SEM. It should be noted that fatigue striations are not always discernible on examination at high magniﬁcation. For example, the appearance of a fatigue fracture through a cast aluminum alloy will appear quite different from cast steel. The aluminum alloy will typically have a feathery appearance where the fatigue fracture propagates through the matrix and a rougher, more globular texture where the crack extends around the eutectic silicon particles in the microstructure, as shown in Fig. 13 for a laboratory fatigue test specimen. Even at higher

Fatigue striations in a low-alloy steel casting. Original magniﬁcation: 2000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 10

Quasi-cleavage in a heavy-section low-alloy steel casting. Original magniﬁcation: 1000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 12

Subsurface fatigue crack initiation in a heavy section low-alloy steel casting. Original magniﬁcation: 6. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 13

Fig. 11

Fatigue fracture through a type 356.0 aluminum alloy in the T6 condition. Original magniﬁcation: 500. Courtesy of Stork Technimet, Inc. New Berlin, WI

244 / Casting Design and Performance magniﬁcation, the striations are difﬁcult to discern, as shown in Fig. 14.

Environmentally Induced Fracture Hydrogen-Assisted Cracking (Intergranular Brittle Fracture). Hydrogen-assisted cracking, also called hydrogen embrittlement, can occur with carbon and low-alloy steels that have been exposed to an environment where hydrogen is generated at the surface of the part. This can be from pickling or plating operations or may occur from severe surface corrosion. The fracture surface will typically exhibit separation at the grain boundaries and will also exhibit small regions with ductile fracture features called ductile hairlines, have some secondary grainboundary cracking, and show the presence of micropores. The features are similar to those found in wrought components, and an example of a heavy-section low-alloy steel with hydrogen-assisted cracking is shown in Fig. 15. Rock Candy Fracture (Intergranular Brittle Fracture). Heavy-section carbon and low-alloy steels are susceptible to “rock candy fracture” or concoidal fracture if the steel

Fatigue fracture through a type 356.0 aluminum alloy in the T6 condition. Original magniﬁcation: 5000. Courtesy of Stork Technimet, Inc. New Berlin, WI

contains excessive amounts of nitrogen and aluminum. Aluminum-nitrogen precipitates form at the austenite grain boundaries if the casting is cooled slowly. The overload fracture surface exhibits separation at the grain boundaries, and the grains can typically be seen with an unaided eye. Examination at high magniﬁcation typically reveals ﬁne quasi-cleavage or ductile fracture with linear ruptures where the precipitates are present. A typical linear rupture is shown in Fig. 16 for a heavy-section low-alloy steel. Quench Cracking (Intergranular Brittle Fracture). Quench cracking can occur when a steel casting is quenched from the austenitizing temperature at too fast a rate. Intergranular fracture occurs at the prior austenite grain boundaries. In most cases the original fracture surface is oxidized during the tempering cycle, as the cracks are usually not detected until after completion of the heat treatment process. The fracture surface typically exhibits relatively clean separation between the grains with minimal evidence of ductile fracture or secondary grain-boundary cracking. A typical example of a quench crack for a low-alloy steel lever is shown in Fig. 17. A coating of temper scale

covers the fracture surface. A metallographic cross section through the cracked region is usually needed to conﬁrm the presence of intergranular fracture. Stress-corrosion cracking. (SCC) can occur with cast alloys and is most often encountered with copper-base alloys, such as brass and some bronzes. Stress-corrosion cracking occurs when a susceptible alloy is subjected to applied or residual stresses in a corrosive environment. This is especially true for some copper alloys when ammonia-based compounds are present. The fracture characteristics can be somewhat complex and in some cases difﬁcult to distinguish from corrosion fatigue. A typical example of SCC in a bronze alloy C83600 suction roll for the paper industry is shown in Fig. 18. The surface exhibits the characteristic jagged features with ﬁne microcracks. It should be noted that fracture features are often only discernible at the crack-tip region, or at secondary cracks, as the fracture surface is typically coated with oxidation products caused by the corrosive environment. Other alloys, such as ductile iron and steel, can also incur this type of cracking. The crack-tip region of a ductile iron head is shown in Fig. 19.

Fig. 14

Fig. 16

Linear rupture through a low-alloy steel with rock candy fracture. Original magniﬁcation: 5000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Stress-corrosion cracking in a bronze alloy C83600 suction roll shell. Original magniﬁcation: 1000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Hydrogen-assisted cracking in a heavy-section low-alloy steel casting. Original magniﬁcation: 1000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Quench cracking in a low-alloy steel lever casting. Original magniﬁcation: 1000. Courtesy of Stork Technimet, Inc. New Berlin, WI

Stress-corrosion cracking at the tip of a crack in ductile iron. Original magniﬁcation: 200. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 15

Fig. 17

Fig. 18

Fig. 19

Failure Analysis of Castings / 245

Elevated-temperature rupture in a heatresistant stainless steel. Original magniﬁcation: 40. Courtesy of Stork Technimet, Inc. New Berlin, WI

Fig. 20

Elevated-Temperature Fracture (Creep). Cast alloys that have been exposed to elevated temperatures for extended periods can incur creep failure if the operating temperature is above the range recommended for the speciﬁc alloy. The ﬁne features of the fracture surface will typically be totally obliterated from oxidation, and the texture will correspond to the macrostructure of the alloy. Creep voids and cracks will develop at the cell boundaries, and the fracture surface can often exhibit a dendritic appearance from the solidiﬁcation structure of the alloy. A typical example of a heatresistant stainless steel ﬁtting is shown in Fig. 20.

REFERENCES 1. B.A. Miller, Overload Failure, Failure Analysis and Prevention, Vol 11, ASM Handbook, ASM International, 2002, p 671–699 2. R.A. Lund, Fatigue Fracture Appearances, Failure Analysis and Prevention, Vol 11, ASM Handbook, ASM International, 2002, p 627–640 3. W.T. Becker, Mechanisms and Appearances of Ductile and Brittle Fracture in Metals, Failure Analysis and Prevention, Vol 11, ASM Handbook, ASM International, 2002, p 587–626 4. Failures Related to Casting, Failure Analysis and Prevention, Vol 11, ASM Handbook, ASM International, 2002, p 103–155

Casting Design and Performance Pages 247–249

Copyright © 2009 ASM International® All rights reserved. www.asminternational.org

Inspection of Castings After ﬁnishing, castings are inspected for surface quality. Inspection can be performed manually by visual checking, manually by template comparison, or by an automated inspection station. Visual inspection is simple yet informative. A visual inspection would include signiﬁcant dimensional measurements as well as general appearance. Surface discontinuities often indicate the presence of internal discontinuities. Computer-assisted coordinate-measuring machines measure both pattern and casting dimensions. These machines can perform a variety of dimensional checks ranging from basic geometric measurement to parallel and plane projection. The operator simply identiﬁes critical part locations so that the machine can establish a working plane. The coordinate-measuring machine can perform in a few minutes the tedious checks that take 2 to 4 h to be done manually. For small castings produced in reasonable volume, a destructive metallographic inspection on randomly selected samples is practical and economical. This is especially true on a new casting for which foundry practice has not been optimized and a satisfactory repeatability level has not been achieved. For castings of some of the harder and stronger alloys, a hardness test is a good means of estimating the level of mechanical properties. Hardness tests are of less value for the softer tin bronze alloys because hardness tests do not reﬂect casting soundness and integrity. Inspection also includes various methods of nondestructive testing (NDT) to screen castings for imperfections that may be considered to be defects. The techniques for NDT of casting are brieﬂy summarized in this chapter. The general types of imperfections or defects in castings are listed in Table 1. As a general rule, the method of inspection applied to some of the ﬁrst castings made from a new pattern should include all those methods that provide a basis for judgment of the acceptability of the casting for the intended application. Any deﬁciencies or defects should be reviewed and the degree of perfection deﬁned. This procedure can be repeated on successive production runs until repeatability has been ensured.

NDT Methods Nondestructive testing (NDT) of cast products is necessary to monitor product quality, depending on the level of quality required.

Table 1 Casting defects, descriptions, and prevention Casting imperfections or defects

Description

Prevention

Cold shuts

Appear as folds in the metal — occurs when two streams of cold molten metal meet and do not completely weld Possible causes: Interruption in the pouring operation Too slow a pouring rate Improperly designed gating

Pour as quickly as possible Design gating system to ﬁll entire mold quickly without an interruption Preheat the mold Modify part design Avoid excessively long, thin sections

Hot tears and cracks

Hot tears are cracklike defects that occur during solidiﬁcation due to overstressing of the solidifying metal as thermal gradients develop. Cracks occur during the cooldown of the casting after solidiﬁcation is complete due to uneven contraction.

Fill mold as quickly as possible Change gating system; e.g., use several smaller gales in place of one large gate Apply thermal management techniques within the mold (e.g., chills or exothermic material) to control solidiﬁcation direction and rate Insulate the mold to reduce its cooling rate Modify casting design: Avoid sharp transitions between thin and thick sections Taper thin sections to facilitate establishment of appropriate solidiﬁcation gradients Strengthen the weak section with additional material, ribs, etc.

Inclusions

Presence of foreign material in the microstructure of the casting Typical sources:

Modify gating system to include a strainer core to ﬁlter out slag Avoid metal ﬂow turbulence in the gating system that could cause erosion of the mold Improve hardness of the mold and core

Furnace slag Mold and core material Misruns

Incomplete ﬁlling of the mold cavity Causes:

Too low a pouring temperature Too slow a pouring rate Too low a mold temperature

High backpressure from gases combined with

Control mold and metal temperature Increase the pouring rate Increase the pouring pressure Modify gating system to direct metal to thinner and difﬁcult-to-feed sections more quickly

low mold permeability

Inadequate gating Porosity

Holes in the cast material Causes:

Dissolved or entrained gases in the liquid metal Gas generation resulting from a reaction

Pour metal at lowest possible temperature Design gating system for rapid but uniform ﬁlling of the mold, providing an escape path for any gas that is generated Select a mold material with higher gas permeability

between molten metal and the mold material Microshrinkage Liquid metal does not ﬁll all the dendriﬁc interstices, causing the appearance of solidiﬁcation microshrinkage

Control direction of solidiﬁcation:

Design gating system to ﬁll mold cavity so that solidiﬁcation begins at the extremities and progresses toward the feed gate Lower the mold temperature and increase the pouring temperature Add risers, use exothermic toppings to maintain temperature longer Control cooling rate using chills, insulators, etc. in selected portions of the mold

248 / Casting Design and Performance Quality levels dictate the frequency of testing and types of tests being needed. The commonly used techniques include: Liquid penetrant inspection (LPI) Radiographic inspection Fluoroscopic inspection and automated

defect recognition Ultrasonic inspection Eddy current inspection Process-Controlled resonant testing Leak test Electrical conductivity measurements

Liquid penetrant inspection is extensively used as a visual aid for detecting surface ﬂaws. One of the most useful applications, however, is the inspection of alloys susceptible to hot cracking. Such cracks are virtually undetectable by unaided visual inspection but are readily detectable by liquid penetrant inspection. Liquid penetrant inspection essentially involves a liquid wetting the surface of a workpiece, ﬂowing over that surface to form a continuous and uniform coating, and migrating into cracks or cavities that are open to the surface. After a few minutes, the liquid coating is washed off the surface of the casting, and a developer is placed on the surface. The developer is stained by the liquid penetrant as it is drawn out of the cracks and cavities. Liquid penetrants will highlight surface defects so that detection is more certain. The liquid penetrant ﬂows over a surface and enters various types of minute surface openings (0.1 mm (0.4 min.) wide) by capillary action. This helps detect surface cracks, laps, laminations, and similar ﬂaws: However, it should be noted that surface condition (cleanliness, roughness) of the part being inspected is critical since false indications can arise from poor surface preparation or condition. The three basic penetrant systems commonly used are:

shuts, internal shrinkage, porosity, core shifts, and inclusions in aluminum alloy castings. Radiography can also be used to measure the thickness of speciﬁc sections. Aluminum alloy castings are ideally suited to examination by radiography because of their relatively low density; a given thickness of aluminum alloy can be penetrated with approximately one-third the power required for penetrating the same thickness of steel. The typical complexity of shape and varying section thicknesses of the castings may require digital radiography or computed tomography. Unlike liquid penetrant and eddy current inspection, radiography can satisfactorily detect ﬂaws that are completely internal and located well below the surface of the component. If the workpiece is properly oriented during inspection, radiography is capable of detecting defects such as cracks. Voids and inclusions having measurable thickness in all directions can be also detected as long as they are not too small in relation to section thickness. Radiography is used extensively for evaluation of large components (e.g., steel castings) that exhibit a difference in thickness or physical density. The sensitivity, or the ability to detect ﬂaws, of radiographic inspection depends on close control of the inspection technique, including the geometric relationships among the point of x-ray emission, the casting, and the x-ray imaging plane. The smallest detectable variation in metal thickness lies between 0.5 and 2.0% of the total section thickness. Narrow ﬂaws, such as cracks, must lie in a plane approximately parallel to the emergent x-ray beam to be imaged; this requires multiple exposures for x-ray ﬁlm techniques and a remote-control parts manipulator for a real-time system. Real-time systems have eliminated the need for multiple exposures of the same casting by dynamically inspecting parts on a manipulator, with the capability of changing the x-ray energy for

changes in total material thickness. These capabilities have signiﬁcantly improved productivity and have reduced costs, thus enabling higher percentages of castings to be inspected and providing instant feedback after repair procedures. Automated Defect Recognition. This system has replaced conventional radiography (ﬁlm x-ray) and consists of an x-ray generator and a detector (an image intensiﬁer or a digital detector). The image created by a detector can be converted to a digital image that may be subsequently processed by a computer program (ADR program). Based on the difference in gray scales of the image, the program can detect defects in the image with great success (see Fig. 1). The ADR is used extensively in high-volume detection environments to reduce dependency on a human operator to make decisions. Ultrasonic inspection for both thickness and defects is practical with most ferrous castings except for the high-carbon gray iron castings, which have a high damping capacity and absorb much of the input energy. The measurement of resonant frequency is a good method of inspecting some ductile iron castings for soundness and graphite shape. Electromagnetic testing can be used to distinguish metallurgical differences between castings. Aluminum alloy castings are sometimes inspected by ultrasonic methods to evaluate internal soundness or wall thickness. The principal uses of ultrasonic inspection for aluminum alloy castings include the detection of porosity in castings and internal cracks in ingots. The advantages of ultrasonic tests are as follows: High sensitivity, which permits the detection

of minute cracks

Great penetrating power, which allows the

examination of extremely thick sections

Accuracy in measuring of ﬂaw position and

estimating defect size

Water-soluble:

The penetrants can be removed directly with water. Postemulsiﬁable: These penetrants have an oil base and, thus, cannot be removed directly with water. Consequently, an emulsiﬁer is needed (following application of the penetrant) to make the penetrant soluble in water. Solvent-removable: These also have an oil base and employ a solvent for both precleaning and removal of excess penetrant. Each of these systems uses either a visible or a ﬂuorescent penetrant material. The visiblepenetrant inspection uses a liquid that is typically red in color and, thus, reveals red indications under visible light. On the other hand, the ﬂuorescent penetrant inspection employs penetrants that ﬂuoresce under ultraviolet light. Components with complex designs (sharp corners or radii) and/or made with alloys that are prone to hot tearing are often subjected to liquid penetrant inspection. Radiographic inspection is a very effective means of detecting such conditions as cold

Fig. 1

(a) X-ray inspection equipment. (b) Digital image of two aluminum castings showing no defects

Inspection of Castings / 249 Ultrasonic limitations:

tests

have

the

following

Size-contour complexity and unfavorable

discontinuity orientation can pose problems in interpreting the echo pattern Undesirable internal structure—for example, grain size, structure, porosity, inclusion content, or ﬁne dispersed precipitates—can similarly hinder interpretation Reference standards are required Because castings are rarely simple ﬂat shapes, they are not as easy to inspect as such products as rolled rectangular bars. The reﬂections of a sound beam from the back surface of a parallel-sided casting and a discontinuity are shown schematically in Fig 2(a), together with the relative heights and positions of the reﬂections of the two surfaces on an oscilloscope screen. A decrease in the back reﬂection at the same time as the appearance of a discontinuity echo is a secondary indication of the presence of a discontinuity. However, if the back surface of the casting at a particular location for inspection is not approximately at a right angle to the incident sound beam, the beam will be reﬂected to remote parts of the casting and not directly returned to the detector. In this case, as shown in Fig. 2(b), there is no back reﬂection to monitor as a secondary indication.

Fig. 2

Many castings contain cored holes and changes in section, and echoes from holes and changes in section can interfere with echoes from discontinuities. As shown in Fig. 2(c), the echo from the cored hole overlaps the echo from the discontinuity on the oscilloscope screen. The same effect is shown in Fig. 2(d), in which echoes from the discontinuity and the casting ﬁllets at a change in section are shown overlapping on the oscilloscope. Magnetic Particle Inspection. Magnetic particle inspection is a highly effective and sensitive technique for revealing cracks and similar defects at or just beneath the surface of castings made of ferromagnetic metals. The capability of detecting discontinuities just beneath the surface is important because such cleaning methods as shot or abrasive blasting tend to close a surface break that might go undetected in visual or liquid penetrant inspection. When a magnetic ﬁeld is generated in and around a casting made of a ferromagnetic metal and the lines of magnetic ﬂux are intersected by a defect such as a crack, magnetic poles are induced on either side of the defect. The resulting local ﬂux disturbance can be detected by its effect on the particles of a ferromagnetic material, which become attracted to the region of the defect as they are dusted on the casting. Maximum sensitivity of indication is obtained when a defect is oriented in a direction perpendicular

Schematic of the effect of casting shapes on reﬂection and oscilloscope screen display of sound beams. See text for discussion.

to the applied magnetic ﬁeld and when the strength of this ﬁeld is sufﬁcient to saturate the casting being inspected. Equipment for magnetic particle inspection uses direct or alternating current to generate the necessary magnetic ﬁelds. The current can be applied in a variety of ways to control the direction and magnitude of the magnetic ﬁeld. In one method of magnetization, a heavy current is passed directly through the casting placed between two solid contacts. The induced magnetic ﬁeld then runs in the transverse or circumferential direction, producing conditions favorable to the detection of longitudinally oriented defects. A coil encircling the casting will induce a magnetic ﬁeld that runs in the longitudinal direction, producing conditions favorable to the detection of circumferentially (or transversely) oriented defects. Alternatively, a longitudinal magnetic ﬁeld can be conveniently generated by passing current through a ﬂexible cable conductor, which can be coiled around any metal section. This method is particularly adaptable to castings of irregular shape. Circumferential magnetic ﬁelds can be induced in hollow cylindrical castings by using an axially disposed central conductor threaded through the casting. Small castings can be magnetic particle inspected directly on bench-type equipment that incorporates both coils and solid contacts. Critical regions of larger castings can be inspected by the use of yokes, coils, or contact probes carried on ﬂexible cables connected to the source of current; this setup enables most regions of castings to be inspected. Eddy Current Inspection. Eddy current inspection consists of observing the interaction between electromagnetic ﬁelds and metals. In a basic system, currents are induced to ﬂow in the testpiece by a coil of wire that carries an alternating current. As the part enters the coil, or as the coil in the form of a probe or yoke is placed on the testpiece, electromagnetic energy produced by the coils is partly absorbed and converted into heat by the effects of resistivity and hysteresis. Part of the remaining energy is reﬂected back to the test coil, its electrical characteristics having been changed in a manner determined by the properties of the testpiece. Consequently, the currents ﬂowing through the probe coil are the source of information describing the characteristics of the testpiece. These currents can be analyzed and compared with currents ﬂowing through a reference specimen. Eddy current methods of inspection are effective with both ferromagnetic and non-ferromagnetic metals. Eddy current methods are not as sensitive to small, open defects as liquid penetrant or magnetic particle methods are. Because of the skin effect, eddy current inspection is generally restricted to depths less than 6 mm (¼ in.). The results of inspecting ferromagnetic materials can be obscured by changes in the magnetic permeability of the testpiece. Changes in temperature must be avoided to prevent erroneous results if electrical conductivity or other properties are being determined.

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Index A acetic acid (CH3COOH), 164, 165, 167 ACI. See Alloy Casting Institute (ACI) acicular transformation structures, 163, 225 Ag alloys, 211(T), 214 Al-Cu cast alloys corrosive (salt) effects, 216, 217(F) discontinuities, 214, 216(F) environmental effects, 214–216(F), 217(F) fatigue crack growth, humidity effects on, 215–216, 217(F) introduction, 211(T), 214 loading conditions, effects of, 216, 217(F) mechanical properties, 214, 215(F), 216(F) microstructure, 214, 215(F), 216(F) temperature effects, 215, 216(F) Alloy Casting Institute (ACI), 175, 176(T), 183(F) aluminum alloy castings, fatigue and fracture properties of Al-Cu cast alloys (see Al-Cu cast alloys) Al-Si-Mg, 210–211(T) component tests, 216–217 composition effects, 214, 215(F) design and manufacture casting discontinuities, effect of, 210 dross, 210 gas porosity, 210 hot isostatic pressing (HIP), 210 introduction, 209 weld repair, 210 weld rework, 209–210 future work, considerations for, 217 introduction, 209 microstructural ﬁneness crack growth rates, 213, 215(F) fatigue crack initiation, effects on, 212–213 fatigue life, 211–212, 213 fracture toughness, 213, 215(F) measurement of, 211, 213(F), 215(F) microstructure, 211, 212(F) premium (high-strength, high-toughness) castings, 211, 212(T) properties, 211, 212(F) residual iron, 214, 215(F) Aluminum Association (AA), 227 aluminum castings, 153(F) junctions, 150–152(F) thin-wall core movement, 126(F) effect of area, 126(F), 127(F) feeding through, 126–127(F) introduction, 125–126(F) strength-to-weight ratio, 126, 127(F) aluminum-silicon alloys, friction and wear of aerospace components, 233 aluminum-silicon wear-resistance materials, 233–234 applications, 231–234(F,T) automotive components, 231–233(T) consumer electronics components, 233 introduction, 227 metallurgy of copper, 228 heating effects, 228–229 introduction, 227–228(F,T)

iron, 228 magnesium, 228 manganese, 228–229 properties and structure antimony, 230 cell size, 230–231(F) gas porosity, 230 grain structure, 229, 230(F) heat treatment, 229 introduction, 229(T) microstructural control, 229–230(F) modiﬁcation, 230(F) reﬁnement, 230 sludge, 230(F) wear behavior abrasive wear, 230 intermetallic constituents, 231(T) matrix hardness, 231 silicon particles, 231 sliding wear, 230 aluminum-silicon wear-resistance materials coatings/surface treatments, 233–234 metal-matrix composites, 233 powder metallurgy, 233 spray casting, 233 anisotropy, 42, 113 apparent surface alteration factors (ASAF), 70 area moment of inertia, 105(F,T), 106(F) equation for, 105(F) ASAF. See apparent surface alteration factors (ASAF) austenite, 221–222(F), 223, 224, 225 automated defect recognition (ADR), 248(F) automotive components, 232(T) advanced aluminum bearing alloys, 232, 233(T) bearing alloy components, 232–233(T) cylinder liners, 231–232(T) engine blocks, 231–232(T) pistons, 231, 232(T)

B bainite, 163, 225, 241 bainitic transformation structures, 225 bedding sand, 84, 89 Bernoulli’s theorem, 74(F) blackheart, 193, 196(T), 219 blended junctions, 152–153(F) blind holes, 90, 95, 129 blow-holes, 92 bosses, 82–83, 84(F), 89(F), 91, 96(F), 142(F) bottom gating, 77(F), 143 Brinell hardness testing, 240

C CA model. See cellular automaton (CA) models CaCO3. See calcium carbonate (CaCO3) CAFE model, 41, 42 calcium carbonate (CaCO3), 168 capability castings, 103, 104 carbon equivalent (CE), 219–220 CARP. See computer automated rapid prototyping (CARP) cast carbon steels, corrosion of

atmospheric corrosion, 171–173(F,T) high-temperature steam, 172(T), 173(T) introduction, 171 cast housings, 144–146(F) cast iron graphitic, 66 micromodeling, 44 risers and, 11 solidiﬁcation, during, 10, 11 solidiﬁcation, lack of shrinkage during, 10 structure and properties, 33 cast irons, basic metallurgy of austenitic structures, 163 bainite, 163 ferrite, 163 introduction, 163(F,T) martensitic structures, 163 pearlite, 163 cast irons, corrosion of alloying, inﬂuence of chromium, 164 copper, 164 introduction, 163–164 molybdenum, 164 nickel, 164 silicon, 164(F) titanium, 164 vanadium, 164 calcium carbonate (CaCO3), 168 chromic acid (H2CrO4), 169 coatings conversion, 169 enamel, 169 metallic, 168–169(T) organic, 169 corrosion, forms of, 165–166(T) corrosion services, selection for, 169–170 corrosive environments, resistance to alkali solutions, 167–168 atmospheric corrosion, 168 hydrochloric acid (HCl), 167(F) nitric acid (HNO3), 167(F) organic acids and compounds, 167 other environments, 168, 169(T) phosphoric acid (H3PO4), 167 in saline solution, 168 in soils, 168 sulfuric acid (H2SO4), 166–167(F) in water, 168 crevice corrosion, 165–166 cupric chloride (CuCl2), 168 erosion-corrosion, 166 ferric chloride (FeCl3), 168 ﬂow-induced corrosion, 166 fretting corrosion, 165(T) graphitic corrosion, 165 high-chromium, 165(T) high-nickel austenitic, 165(T), 166 high-silicon, 165(T), 166–167 intergranular attack, 166 introduction, 163 low -and moderately alloyed, 164–165(T) microbiologically induced corrosion (MIC), 166 microstructure, inﬂuence of, 164 pitting corrosion, 165–166 stress-corrosion cracking (SCC), 166

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

260 / Index cast irons, corrosion of (continued) unalloyed, 164(T) unalloyed gray, 164 cast irons, fatigue and fracture properties of, 188 constrained thermal test, 188 corrosion fatigue, 188, 191(T) ductile irons, 96(F), 192–193, 194(F,T), 195(F,T), 196(T) fatigue crack growth, 185–186, 190(F,T) fatigue notch sensitivity, 187 fatigue of, 185–188(F,T) ﬁnned-disk test, 188 fracture appearance, 186 fracture toughness, 188–189, 192(F), 193(T) gray cast iron, 189–192(F,T), 194(F,T) introduction, 185 load variables, 186–187, 191(F,T) malleable cast irons, 193–194, 196(F,T), 197(F), 198(T) Paris relation, 186 thermal fatigue, 187–188, 191(F) white cast iron, 194, 198(F) cast irons, friction and wear of applications, 225(T) carbon equivalent (CE) value, 219–220 composition value, 219–220 constitution of, 219–221(F,T) cooling rate, 220, 221(T) graphite, 222–224(F,T) gray irons, 219, 221–222 introduction, 219 malleable iron, 219 microstructures acicular, 225 austenite, 225 bainitic, 225 cementite, 225 ferrite, 224 martensite, 225 pearlite, 224 phosphide eutectic, 225 nodular iron, 219 section sensitivity, 220, 221(T) spheroidal iron, 219 tensile properties, 220, 222(F) effect of section on strength, 220–221, 222(F) white irons, 221, 222(F,T) white or chilled iron, 219 CAST program, 210, 211, 216, 217 cast stainless steels, corrosion of composition and microstructure, 175–177(F,T) C-type alloys austenitic alloys, 178–180(F,T) austenitic alloys, intergranular corrosion of, 180–181(F), 182(T) corrosion fatigue, 182 crevice corrosion, 181–182 critical crevice temperature (CCT), 182 duplex alloys, 178–180(F,T), 181 ferritic alloys, 178, 181 fully austenitic alloys, 180, 181(F) martensitic alloys, 178 pitting corrosion, 181–182 pitting resistance number (PREN), 180 stress-corrosion cracking (SCC), 182–183(F) H-type alloys carburization, 177–178(F) introduction, 177 oxidation, 177(F) sulﬁdation, 177–178(F) introduction, 175, 176(T) cast steels, fatigue and fracture properties of fatigue, inﬂuencing factors applied stress effects, 202, 203(F), 204(F,T) cyclic stress-strain behavior, 202–203, 204(F) defects, effect of, 201–202(F) high-cycle axial fatigue behavior, 203–204, 205 (F,T) low-cycle axial fatigue behavior, 203, 205(F,T) section size effects, 201, 202(F) fracture mechanics fatigue crack growth rates (constant amplitude), 207(F,T)

introduction, 204–205 J-integral method, 206–207(F,T) plane-strain fracture toughness, 206(F,T) plane-stress fracture toughness, 205–206 threshold crack growth behavior, 207, 208(F,T) variable amplitude fatigue crack initiation and growth, 207 introduction, 199 structure and property correlations, 199–201(F,T) Charpy impact toughness, 199, 200(F) fatigue strength limits, 200–201(T) nil ductility transition temperatures (NDTT), 199–200(F) plane-strain fracture toughness, 200(F) strength and toughness, 199–200(F) tensile strength limits, 200–201(T) casting and solidiﬁcation processing, modeling of computational thermodynamics, 37–38(F,T) defect prediction (see casting defects) examples high - or low-pressure die casting, 51–53(F), 54(F) investment casting, 53–56(F) fundamentals of basic mathematical formulation, 40 boundary conditions, 40 introduction, 39–40 view-factor radiation model, 40–41 introduction, 37 microstructure simulation cellular automaton (CA) models, 41–42 deterministic method, 41 deterministic micromodeling, 41(F) eutectoid transformation, 44–45(F), 46(F) experimental validations, 45–46, 47(F), 48(T) fading effect, 44 graphite/austenite eutectic transformation, 44 introduction, 41 ledeburite eutectic transformation, 44 micromodeling applications, 44 nucleation model, 41(F), 44 phase-ﬁeld model, 42–43(F) stochastic method, 41 thermophysical properties density, 39, 40(F) introduction, 38–39 liquid viscosity, 39(F) thermal conductivity, 39(F) casting defects. See also casting defects, classiﬁcation of hot tearing, 49–51(F), 52(F) introduction, 46 macrosegregation, 48–49, 50(F) porosity, 46–48 casting defects, classiﬁcation of, 251–258(T) casting design and processes casting methods, 9–10, 11(T), 12(T) chills, use of, 32–33 cores, use of, 32–33 defects, 34 (see also casting defects, classiﬁcation of) expendable-mold casting, 9–10, 11(T), 13–16(F), 17(F) feeding cylinders, solidiﬁcation of, 30(F), 31 ﬂat plates, solidiﬁcation of, 29–31(F) introduction, 28–29 ﬂuidity and solidiﬁcation cooling rates and microstructure, 21–22 deﬁned, 21 introduction, 20–21 liquid-liquid contraction, 21 liquid-solid contraction, 21 solid-solid contraction, 21, 21(F) general principles, 10–13 geometry, 11–13 heat transfer and transport phenomena, 31–33(F) introduction, 9 mold complexity, 18–20(F) permanent mold casting, 16–18(F), 19(F) risers, function of, 22(F) sand mold, components of, 11(F) shape casting, 9–10(F), 12(T)

shape casting process, chart of, 10(F) solidiﬁcation sequence introduction, 22 L-sections, 27–28, 29(F), 30(F) T-sections, 22–26(F) X-sections, 26–27(F) structure and properties discontinuities and defects, 34–36(F) hot isostatic pressing (HIP), 35 inclusions, 35(F) metal penetration, 35–36 overview, 33–34 wedge, casting design and solidiﬁcation of, 22(F) casting fracture characteristics ductile rupture, 242(F) examination methods, 241 fatigue, 243–244(F) fracture appearance, 241–242(F) introduction, 241 regions of interest, 241 transgranular brittle fracture, 242–243(F) casting molds, dimensional changes in, 115 Casting Technology International, UK (Replicast casting), 16 casting tolerances, factors that control, 104–105 castings, design conversions fabrications deﬂection, 112 introduction, 111–112 method of sections, using, 112 Mohr’s circle, 112–114 stick-ﬁgure sketches, 112 stress correlates to geometry, 112 stress correlation to area, 112 introduction, 110 weldments and assemblies, 110–111 castings, design tips add strength, not mass, 107–108(F) core design principles, 109–110 economical coring, 109–110 functional packaging, 106–107 future parts, 107 gaging parts, 107 service window, 107 thin sections, 109 unequal sections, 108–109 uniform wall thickness, 108 castings, failure analysis of. See failure analysis castings, inspection of casting defects, 247(T) (see also casting defects, classiﬁcation of) introduction, 247(T) NDT methods (see NDT methods) CE. See carbon equivalent (CE) cellular automaton (CA) models, 41–42(F), 43 cellular-to-equiaxed transition (CET), 41, 42 cementite, 225 cementite (Fe3C), 34, 163, 164, 222(F), 224, 225 centerline feeding resistance, 66 centerline porosity, 107, 143(F) centroid, 105, 106(F) ceramic ﬁlters advantages, 78(F) ﬁlter area/choke area ratio, 79–80(F) inclusions, 77–78(T) introduction, 77–78(T) placement, 79(F), 80(F) strainer (choke) cores, 78–79(F) types, 78–79(F) use of, 79–80 CH3COOH. See acetic acid (CH3COOH) chaplets, 109 Charpy impact toughness, 192(F), 193–194, 197(F), 199, 200(F), 217 Charpy V-notch data, 192, 199, 200(F) chill zone, 107, 241, 243(F) chilled iron, 219, 225(T) chills FD, extending, 64, 66(F) junctions, 149–150 use of, 32–33 chills, use of, 32–33 chromic acid (H2CrO4), 169

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Index / 261 Chvorinov’s rule, 9, 21, 24–25, 68 close mechanical members, 106 coefﬁcient of thermal expansion (CTE), 227 computer automated rapid prototyping (CARP), 7, 7(F) computer-aided design (CAD), 106, 107 concoidal fracture, 244 conduction (use of term), 40 conﬁguration design, 5, 6 constrained thermal test, 188 cope-and-drag mold, 86 core pins, 96, 98(F) cores cantilever-supported, 90(F) deﬁned, 13 eliminate, designing to, 96–97, 98(F), 99(F) expendable, 92–93(F) metal, 94 molding, principles of, 89 “ram-up” core, 83, 91 ring core, 83, 84(F), 91, 96, 98(F) semicylindrical, 89(F) simpliﬁed, 96–97, 99(F) stationary, 94 use of, 32–33 vertical cylindrical, 90–91(F) coring, design for, 98(F) advantages, 89–90(F) core driers, 89 core size limitations, 90–91(F) cores, 89 coring versus drilling, 97, 99(F) designing to eliminate, 96–97, 98(F), 99(F) economical coring, 109–110 gas-related defects, 110 green sand versus dry sand cores, 91(F) investment castings, 94–96(F), 97(T), 98(F) permanent mold castings, cores in, 94(F) principles, general, 89(F), 110 sand cores, designing for the use of, 91–93(F) sand cores, support for, 93–94(F) simpliﬁed cores, 96–97, 99(F) thin-wall casting sections, 93(F) tortuous passages, 94 cost drivers, 3 counter gravity casting, 12(T), 18 countergravity ﬁlling, 77 crack susceptibility coefﬁcient (CSC), 50–51 critical crevice temperature (CCT), 182(T) cross ribbing, 107, 108(F) CSC. See crack susceptibility coefﬁcient (CSC) CuCl2. See cupric chloride (CuCl2) cupric chloride (CuCl2), 168 cylinders, solidiﬁcation of, 30(F), 31

D DAS. See dendrite arm spacing (DAS) defect, (use of term), 237 deﬂection, 112 dendrite arm spacing (DAS), 211, 212, 213(F) dendrites, 10, 29, 114, 212, 221–222(F), 252(T) design issues and practices casting design process, traditional, 4(F) cost drivers, 3 geometry/material/process interactions dross formation, 2 ﬂuid ﬂow, 2 ﬂuid life, 1 geometry/alloy interactions, 2–3 heat transfer considerations, 2 introduction, 1 pouring temperature, 2 slag formation, 2 solid shrinkage, 2 solidiﬁcation shrinkage, 1–2 guidelines, 6–7(F) improvement strategies, 4(F) introduction, 1 iterative model, 4(F) process simulation, 4 rigging system design, 3

shape optimization, 3 structured team approach, 4–6(F) die casting abutting cores, 96 casting methods, comparison of, 12(T) design and performance, 96 material thickness, sensitivity to, 107 materials, 18 molding principles, 81 percentage of casting tonnage, 14 diffusivity ratio, 33 directional solidiﬁcation, 2 directionality, 113 discontinuities and defects. See also casting defects, classiﬁcation of deﬁned, 34 hot isostatic pressing (HIP), 35 inclusions, 35, 35(F) metal penetration, 35–36 porosity, 34–35 distortion alloy, effect of, 160–161(F) differences in solidiﬁcation times, 155–158(F) in heat treating, 160(F) introduction, 155 mold restraint, due to, 158–159(F) oil canning, 157, 158(F) tie bars, 159–160(F) draft, 86(F), 87, 104 dross, 2, 3, 210 dross formation, 2, 3 ductile irons Charpy impact toughness, 192(F), 195(F) Charpy V-notch data, 192 dynamic fracture toughness, 192(F), 193 dynamic tear energy, 192, 195(F) fatigue strength, 192, 194(F,T), 195(F,T) fracture toughness, 192–193, 196(F,T) lower-shelf fracture toughness, 193, 196(F) micromodeling, 44 mold dilation, 62 riser necks, 71 structure and properties, 34 ductile rupture, 242(F)

E ECTFE. See ethylene chlorotriﬂuoroethylene (ECTFE) eddy current inspection, 249 EDS. See energy dispersive x-ray spectroscopy (EDS) elevated-temperature facture (creep), 245(F), 257 (T) embrittlement, 177, 179 EMF. See exhaust gas recirculation tube (EMF) EMF sources, 106 endurance ratio bending and torsion, 202, 204(F,T) cast carbon and low-alloy steels, 201(T) cast steel, 201, 202(F) deﬁned, 201 ductile irons, 192, 194(F), 195(T) values, 190 wrought steels, 202, 204(F) energy dispersive x-ray spectroscopy (EDS), 239–240 environmentally induced fracture elevated-temperature facture (creep), 245(F) hydrogen-assisted cracking (intergranular brittle fracture), 244(F) quench cracking (intergranular brittle fracture), 244(F) rock candy fracture (intergranular brittle fracture), 244(F) stress-corrosion cracking (SCC), 244(F) EPS. See expanded polystyrene (EPS) equiaxed solidiﬁcation behavior, 2 ETFE. See ethylene tetraﬂuoroethylene (ETFE) ethylene chlorotriﬂuoroethylene (ECTFE), 169 ethylene tetraﬂuoroethylene (ETFE), 169 eutectic-type solidiﬁcation, 2

eutectoid transformation, 44–45(F), 46(F) exhaust gas recirculation tube (EMF), 106 exogenous inclusions, 35 expanded polystyrene (EPS), 15 expendable-mold casting expendable patterns introduction, 15 investment casting, 15(F) lost foam casting process, 15–16(F), 17(F) Replicast casting, 16 mold materials, 14 permanent patterns, 14–15 requirements, 13–14 slurry molding, 14 wax cluster, 15(F)

F fabrications, 105, 111–114 fading effect, 44 failure analysis background information, 237 Brinell hardness testing, 240 casting defects, 237 (see also casting defects, classiﬁcation of) casting fracture characteristics (see casting fracture characteristics) chemical analysis, 239–240 concoidal fracture, 244 data analysis, 241 energy dispersive x-ray spectroscopy (EDS), 239–240 environmentally induced fracture, 244–245 fractography, 238–239 glow discharge spectroscopy (GDS), 239 inductively coupled plasma spectroscopy (ICP), 239 introduction, 237 Knoop microhardness test, 240 macroscopic examination, 238–239 material evaluation, 238–240 mechanical testing, 240 metallography, 240 microscopic examination, 239 micro-void coalescence (MVC), 242 nondestructive evaluation (NDE), 238 optical emission spectroscopy (OES), 239 planning, 237–238 report preparation, 241 Rockwell hardness testing, 240 sample selection, 237–238 scanning electron microscope (SEM), 239 special testing, 240 Vickers microhardness test, 240 visual inspection, 238 x-ray diffraction (XRD), 239 x-ray ﬂorescence spectroscopy (XRF), 239 FD. See feeding distance (FD) FeCl3. See ferric chloride (FeCl3) feed pad, 23, 23(F), 142, 143 feeding cylinders, solidiﬁcation of, 30(F), 31 ﬂat plates, solidiﬁcation of, 29–31(F) introduction, 28–29 reservoirs (feeders or risers), 29 (see also risers) solidiﬁcation shrinkage and, 102 feeding distance (FD), 64, 65–66(F) FEP. See ﬂuorinated ethylene polypropylene (FEP) ferric chloride (FeCl3), 168 ferrite acicular, 225 cast irons, 163, 164 C-type alloys, 178, 179, 180 ductile rupture, 242(F) eutectoid transformation, 44–45(F), 47(F) graphite, 223 gray irons, 222 pearlite, 224 phosphide eutectic, 225 stainless steels, 175–176(F), 177(F) stress-corrosion cracking (SCC), 182(T), 183(T)

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

262 / Index ﬁnite-element (FE) heat ﬂow solver, 41 ﬁnned-disk test, 188 ﬁrst article castings, 104 5t rule, 140 ﬂake graphite, 222–223(F) ﬂasks, 14, 16(F), 19, 83, 86(F) ﬂat plates, solidiﬁcation of, 29–31(F) ﬂuid ﬂow, 2 ﬂuid life, 1, 101–102 ﬂuidity Al-Si-Mg cast alloys, 210 aluminum-silicon alloys, 227, 229(T) antimony, 230 carbon steels, 131 cast irons, 163, 185 deﬁned, 21 ﬂuid life, 101–102 low alloy steels, 131 sodium, 230 strontium, 230 thin sections, 109, 121 versus viscosity, 101 ﬂuorinated ethylene polypropylene (FEP), 169 freckles, 48–49, 50f functional packaging, design for close mechanical members, 106 EMF sources, 106 hard points, 106 mating parts, 106 nearby parts, 106 run downs, 106 thermal source, 106

G gaging points, 107 gas porosity, 47–48, 110, 210, 212, 230 gating design bottom gating, 77(F) ceramic ﬁlters, 77–80(F,T) countergravity ﬁlling, 77 design variables, 73–74 ﬂuid ﬂow, principles of Bernoulli’s theorem, 74(F) law of continuity, 74, 75(F) momentum effects, 74–76(F), 77(F) Reynold’s numbers, 74–75(F) types of ﬂow, 74–76(F), 77(F) ingates, 73(F), 74(F), 75–76(F) pressurized versus unpressurized systems, 76 runner, 73(F), 74(F), 76 runner extensions, 73(F), 74(F), 75, 76 vertical versus horizontal systems, 73(F), 76–77(F) gating system components, 73(F) description, 73 design (see gating design) GD&T. See geometric dimensioning and tolerancing (GD&T) GDS. See glow discharge spectroscopy (GDS) geometric dimensioning and tolerancing (GD&T), 104 geometry, casting design and area moment of inertia, estimating from sketches, 105 case studies coil rotor for traction motor of railroad wheel assembly, 118–119(F,T) design conversion to a die-cast aluminum fan housing, 116–118(F,T) design conversion to ductile iron casting for a pour chute pivot frame, 115–116(F,T) casting, design conversions to (see castings, design conversions) casting geometry, 103–104 casting tolerances, factors that control, 104–105 castings, speciﬁc design tips for (see castings, design tips) drawings and dimensions, 104 ﬂuid life, 101–102 functional packaging, 106–107 introduction, 101

junctions, 104 liquid-to-solid shrinkage of common alloys, guide to, 102(T) metal shrinkage casting molds, dimensional changes in, 115 heat treatment, 114–115 liquid metals, 114 metallurgical solid changes, 114 solid metal contraction, 114 solidiﬁcation shrinkage, 114 parameters, 101–103 process geometry, 103–104 secondary operations, 104 solidiﬁcation shrinkage liquid shrinkage, 102 liquid-to solid shrinkage, 102–103(T) patternmaker’s contraction, 103 thermal expansion/contraction, 103 structural properties modulus of elasticity, 103 overview, 103 section modulus, 103 geometry/alloy interactions, 2–3 glow discharge spectroscopy (GDS), 239 graphite ﬂake graphite, 222–223(F) introduction, 222 nodular graphite, 223–224(F,T) quasiﬂake, 223 graphite/austenite eutectic transformation, 44 graphitic cast irons, 62, 66, 71, 185 gravity casting, 11, 17, 18 gray cast iron fatigue strength, 188–190, 191(T), 194(T) friction and wear of, 221–222 impact strength and toughness, 190–192(F), 194(F) gray irons mold dilation, 62 overview, 219 riser necks, 71 green sand molding, 14, 62, 135 green sand molds, 62, 90, 123

H H2CrO4. See chromic acid (H2CrO4) H2SO4. See sulfuric acid (H2SO4) H3PO4. See phosphoric acid (H3PO4) hard points, 106 HAZ. See heat-affected zone (HAZ) HCI. See hydrochloric acid (HCl) heat transfer and transport phenomena geometric factors, 32–33 introduction, 31–32 L-sections, 31(F), 32(F), 33 T-sections, 32(F), 33 X-sections, 32(F), 33 heat transfer considerations, 2 heat treating, 105, 124, 160(F), 163 heat-affected zone (HAZ), 178 HF. See hydroﬂuoric acid (HF) high - or low-pressure die casting, 51–53(F), 54(F) high-pressure die casting (HPDC), 10, 12(T), 14, 18 HIP. See hot isostatic pressing (HIP) HNO3. See nitric acid (HNO3) hog-out, 110 “hook” shape, molding of, 94, 98(F) hot isostatic pressing (HIP), 35, 210 hot tearing experimental validation, 50–51(F), 52(F) forming of, 35 indicator, 49–50 introduction, 49 hot topping, 69, 70 HPDC. See high-pressure die casting (HPDC) hydrochloric acid (HCl), 164, 165, 167(F), 168, 173(T) hydroﬂuoric acid (HF), 165, 166, 167, 169, 176(T) hydrogen embrittlement, 182, 242, 244 hydrogen-assisted cracking (intergranular brittle fracture), 244(F)

I ICP. See inductively coupled plasma spectroscopy (ICP) inclusions, 35(F) indigenous inclusions, 35 inductively coupled plasma spectroscopy (ICP), 239 ingates, 73(F), 74(F), 75–76(F), 110 intergranular brittle fracture, 244(F) investment casting, 53–56(F) expendable molds, 15 molding principles, 81 investment castings abutting cores, 96, 98(F) cavities, small, 96, 98(F) coring, design for, 94–95(F), 96(F) feed paths, 144, 145(F) holes in, 95–96, 97(T), 98(F) mold and core, unequal expansion of, 96 mold restraint, distortion due to, 159(F) thin-wall alloy selection, 131–132(F) cooperative designs, 130 defects, prevention of, 130, 131(F) economy, designing for, 130–131, 132(F) feeding ribs, 131, 132(F) introduction, 129(T), 130(F) the mold, chilling effect of, 129, 131(F) mold and metal temperature, 129–130 mold permeability, 130 “RH” Monel, 131, 132(F) tie bars, distortion due to, 159–160(F) uniform sections, 136–138(F) iteratively, 4

J J-integral method, 206–207(F,T) junctions aluminum castings, 150–152(F) blended junctions, 153(F) chills, 149–150 design of, 104 elements, 148–150(F), 151(F) introduction, 147–148(F) L-junctions, 148(F), 149 steel casting junctions, 148–149(F), 150(F) T-junctions, 149 T-junctions versus Y-junctions, 152(F) unequal sections, 152–153(F) X-junctions, 149, 150(F) Y-junctions, 149, 150(F), 151(F)

K kaowool, 54, 56 kinematic viscosity, 101 kish, 222, 258(T) Knoop microhardness test, 240

L laminar ﬂow, 75(F), 76 law of continuity, 74, 75(F) ledeburite, 44, 45 ledeburite eutectic transformation, 44 liquid penetrant inspection (PT), 238, 248 liquid shrinkage, 1, 2 as a design consideration, 102 riser design, 61 liquid-liquid contraction, 21 liquid-solid contraction, 21 liquid-to-solid shrinkage, 2, 102–103. See also solidiﬁcation shrinkage liquidus, 219 lost foam casting process, 14–15, 15–16(F), 17(F) lost wax process. See investment casting lost-wax casting. See investment casting low pressure casting machine, 18(F)

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

Index / 263 low pressure casting process (LPPM), 15 low-alloy steels, corrosion of atmospheric corrosion, 171–173(F,T) high-temperature steam, 172(T), 173(T) introduction, 171 LPPM. See low pressure casting process (LPPM) L-sections external radius greater than the internal radius, 27–28, 29(F), 30(F) external radius less than the internal radius, 28, 30(F) heat transfer and transport phenomena, 31(F), 32(F), 33 introduction, 27, 29(F) model of, 29(F)

M macroporosity, 34 magnetic particle inspection (MT), 238, 249 malleable cast irons blackheart, 193, 196(T), 219 Charpy impact toughness, 193–194, 197(F) dynamic tear energy, 194, 197(F) ferritic, 193, 196(F), 197(F) fracture toughness, 194, 198(T) introduction, 193, 196(F,T), 197(F) whiteheart, 193, 196(T), 219 martensite, 163, 175, 178, 225 mating parts, 106 metal cores, 17, 18, 94, 94(F), 127 metal shrinkage, 114–115 metal yield, 140 metallostatic head, 36 metal-matrix composites (MMC), 12(T), 231, 233 MIC. See microbiologically induced corrosion (MIC) microbiologically induced corrosion (MIC), 166 microporosity, 34–35 micro-void coalescence (MVC), 242 modulus of elasticity, 103 Mohr’s circle, 112–113 mold complexity, 18–20(F) molten metal ﬂuidity, 101 mottled iron, 219, 225 MT. See magnetic particle inspection (MT) MVC. See micro-void coalescence (MVC)

N NDT methods automated defect recognition (ADR), 248(F) eddy current inspection, 249 introduction, 247–248 liquid penetrant inspection (PT), 248 magnetic particle inspection (MT), 249 radiographic inspection, 248 ultrasonic inspection (UT), 248–249(F) nearby parts, 106 Nikasil treatment, 233 nil ductility transition temperatures (NDTT), 199–200(F), 242 nitric acid (HNO3), 164, 165, 167(F), 173, 178(F), 179 nodular graphite, 223–224(F,T) nodular graphite iron, 188, 191(T), 219 nondestructive evaluation (NDE) liquid penetrant inspection (PT), 238 magnetic particle inspection (MT), 238 radiographic examination (RT), 238 ultrasonic inspection (UT), 238 nondestructive testing (NDT), 247 NRL. See U.S. Naval Research Laboratory (NRL) nucleation model, 41(F), 44

O oil canning, 157, 158(F) optical emission spectroscopy (OES), 239

Osprey processing, 233 oxide ﬁlms, 35, 73, 101, 164, 175, 177

P padding deﬁnition of, 23 remote sections, 145, 146(F) unequal sections, 139–144(F), 145(F) uniform sections, 108, 133, 135 parametric design, 5, 6–7 Paris relation, 186 parting lines casting costs, reducing, 19 choice of, 20 crossing, 12 sand molding, 81–82(F), 83(F) stepped, 82, 83(F) pattern maker’s shrink. See solid shrinkage patternmaker’s contraction, 103, 104 patternmaker’s shrinkage. See solid shrinkage pearlite cast irons, 186, 224(T) cementite, 225 Charpy impact toughness, 192 ductile irons, 193, 195(F), 196(F) dynamic tear energy, 192 eutectoid transformation, 44–45(F), 163 gray irons, 219, 222 malleable cast irons, 193, 219 phosphide eutectic, 225 upper-shelf toughness, 193 white irons, 221, 222(F) perﬂuoroalkoxy resins (PFA), 169 permanent mold casting, 51–52 cores in, 94(F) feed paths, 143–144(F), 145(F) introduction, 16–17(F) low pressure casting, 18 low pressure casting machine, 18(F) metal cores, 94 methods of, 17–18 molding principles, 81 stationary cores, 94 thin-wall expendable cores, 127 good design, 128–129(F) introduction, 127 large areas, inﬂuence of, 128(F) minimum section thickness, 128, 129(F) nonuniform walls, 127–128(F) uniform sections, 136(F) PFA. See perﬂuoroalkoxy resins (PFA) phase-ﬁeld model, 42–43(F) phosphide eutectic, 219, 225 phosphoric acid (H3PO4), 165, 167, 178, 179(F), 180, 181(F) pitting resistance number (PREN), 180, 182 plane-strain fracture toughness, 206(F,T) plane-stress fracture toughness, 205–206 platelets, 35, 214 polytetraﬂuoroethylene (PTFE), 169 polyvinyldene ﬂuoride (PVDF), 169 porosity, 34–35 defect prediction, 46–47 experimental validation, 48, 49(F) gas porosity evolution, 47–48 shrinkage porosity, 48 pouring temperature, 2 powder metallurgy (PM), 227, 233 PREN. See pitting resistance number (PREN) ProCAST, 54, 55 process geometry, 103–104 production-process veriﬁcation castings, 104 progressive solidiﬁcation, 2 PT. See liquid penetrant inspection (PT) PTFE. See polytetraﬂuoroethylene (PTFE) PVDF. See polyvinyldene ﬂuoride (PVDF)

Q quasiﬂake graphite, 223 quench cracking (intergranular brittle fracture), 244(F)

R radiation (use of term), 40 radiographic examination (RT), 238 radiographic inspection, 248 “ram-up” core, 83 rangy components, 111 remote sections, 144, 145(F) reoxidation, 35 Replicast casting, 15, 16 “rework”, 209–210, 217 “RH” Monel, 131, 132(F) ribs in compression, 107 cross ribbing, 107–108(F) unequal sections, 109 webs, 107 rigging system design, 3, 5, 7 ring gate, 55–56 riser design feed metal volume, 62–63 casting geometry, 62–63(T) liquid shrinkage, 62 mold dilation, 62 solidiﬁcation shrinkage, 62(T) introduction, 61 liquid feed metal availability computerized methods, 68 geometric method, 67–68(F) modulus method, 68, 69(F) shape factor method, 66–67(F) optimum design, 61–66(F,T) requirements, 61–62 riser location directional solidiﬁcation, 63(F) feeding distance, 65–66 graphitic cast irons, 66 introduction, 63 progressive solidiﬁcation, 63(F) sections of various thickness, 66, 67(F) solidiﬁcation mode, 63–64(F) uniform wall thickness, 64–65(F), 66(F) shrinkage-induced casting defects, 61 riser design feeding systems advantages, 68–69, 70(T) breaker cores, 69(F), 71(F) exothermic, 70 exothermic-insulating, 70 insulating, 70 introduction, 68, 69(F) materials, thermal properties of, 69, 70(F) metal contamination, 70 optimum conﬁgurations, 71(F) riser necks, 70–71(F) riser size, factors in, 70 risering. See riser design risers function of, 10–11 introduction, 10–11 need for, 11 T-sections, 26 rock candy fracture (intergranular brittle fracture), 244(F) Rockwell hardness testing, 240 RT. See radiographic examination (RT) run downs, 106 runner extensions, 75, 76

S SAE. See Society of Automotive Engineers (SAE) SAE International, 207 salt cores, 14 sample castings, 104, 122

© 2009 ASM International. All Rights Reserved. Casting Design and Performance (#05263G)

www.asminternational.org

264 / Index sand castings mold materials, 14 thin-wall steel, 123–125 alloy selection, 123(F) distortion, 124 examples, 125(F) mold gases, 123 mold-ﬁlling, 123 mold-ﬁlling problems, 123–124(F) molten metal and mold, friction between, 123 soundness, 125 uniform sections, 134–135(F) sand cores core fragility, 91, 92(F) designing, 91–93(F) support for, 93–94(F) thin core sections, 91–93(F) sand molding basic principles, 81(F), 82(F) bosses, 82–83, 84(F) cope-and-drag mold, 86 coring details, 83–87(F) draft, 86(F), 87 parting lines, 81–82(F), 83(F) radii, location of, 82, 83(F) rib locations, 83, 85(F) undercuts, 83, 84(F) sand molds, 13, 18, 74, 115, 134, 135 sand process, 12(T), 14, 19, 209 “sand-chill composite” mold process. See sand process scaling, 32, 255(T) scanning electron microscope (SEM), 239 SCC. See stress-corrosion cracking (SCC) Schaefﬂer diagrams, 175, 176(F) Schoefer diagram, 175, 176, 177(F) second phases, 35 section modulus, 103, 105 SEM. See scanning electron microscope (SEM) semipermanent molding, 94 service window, 107 SF. See shape factor (SF) SFSA. See Steel Founder’s Society of America (SFSA) shape casting, 9–10 chart of shape casting process, 10(F) processes ratings chart, 12(T) shape factor (SF), 67(F) shape optimization, 3 shell, deﬁnition of, 53 shell mold castings, 135–136(F) slag, 2 slag formation, 2 slurry molding, 14 Society of Automotive Engineers (SAE), 227 solid shrinkage, 1–2 riser design, 61 solidiﬁcation shrinkage, 1–2 directional solidiﬁcation, 2 equiaxed solidiﬁcation behavior, 2 eutectic-type solidiﬁcation, 2 liquid shrinkage, 1, 2 liquid-to-solid shrinkage, 2 progressive solidiﬁcation, 2 riser design, 61 solid shrinkage, 1–2 types of, 2

solid-solid contraction, 21 soluble cores, 95 spheroidal iron, 44, 186(F), 191(F), 195(F), 196(F), 219 spray casting, 233 squeeze casting, 12(T), 15, 233 stainless steel Charpy keyhole impact toughness, 199, 200(F) Charpy V-notch data, 199, 200(F) corrosion fatigue strengths, 191(T) ductile rupture, 242 duplex stainless steel corrosion test results, 180(T) elevated-temperature rupture, 245(F) oil canning, 157, 158(F) petroleum corrosion resistance, 173(T) sand castings, 134(F) stress-corrosion cracks, ferrite pools blocking the propagation of, 183(F) thin sections, 121(F,T), 122, 123(F) thin-wall investment castings, 129(T) tie bars, 159(F) T-junctions versus Y-junctions, 152(F) toughness of solution-annealed duplex stainless steel castings, 177(F) unequal sections, 140, 141(F), 142(F) uniform sections, investment casting, 137(F) Steel Founder’s Society of America (SFSA), 65–66 strainer (choke) cores, 78–79(F) stress-corrosion cracking (SCC), 166, 178, 180, 182–183(F), 244(F) structured team approach, 4–6(F) sulfuric acid (H2SO4), 164, 165, 166–167, 173(T), 179, 180(F) superposition, 105, 112

T TEF. See thermoelectric fan (TEF) tempered martensite, 200(F), 224(T), 243 tensile strength (TS), 186(F), 188 thermal source, 106 thermoelectric fan (TEF), 116–117 thin sections designing, 109 introduction, 121(F,T) stainless steel, 121(T), 122 thinnest wall, determination of, 121–122(F) thin-wall aluminum castings, 125–127(F) thin-wall investment castings, 129–132(F,T) thin-wall magnesium castings, 125–127(F) thin-wall permanent mold castings, 127–129(F) thin-wall steel sand castings, 123–125(F) tie bars distortion, 159–160(F) use of, 21 total cost (use of term), 3 Tresca’s hexagon, 113 truss-type castings, 126, 127(F) T-sections heat transfer and transport phenomena, 32(F), 33 solidiﬁcation sequence factors inﬂuencing, 22–24(F) introduction, 22, 23(F) riser placement, 32(F) sequence graphs, 24–26

U ultrasonic inspection (UT), 238, 248–249(F) undercuts, 83, 84(F) unequal sections 5t rule, 140 housings, 144–146(F) introduction, 139(F) junctions, 152–153(F) padding or other feed paths, 139–144(F), 145(F) investment castings, 144, 145(F) permanent mold casting, 143–144(F), 145(F) remote section, designs that reduce the mass of, 144, 145(F) uniform sections introduction, 133–134(F) investment castings, 136–138(F) permanent mold castings, 136(F) sand castings, 134–135(F) shell mold castings, 135–136(F) U.S. Naval Research Laboratory (NRL), 66–67(F) UT. See ultrasonic inspection (UT)

V vacuum die casting, 12(T), 14, 15 vacuum riserless/pressure riserless casting (VRC/PRC), 15, 18 vermicular graphite iron, 185, 190(T), 195(F) Vickers microhardness test, 240 view-factor radiation model, 40–41 viscosity, 101 voids, 2, 108, 133, 147, 210, 248 VRC/PRC. See vacuum riserless/pressure riserless casting (VRC/PRC)

W weld joints, 113 weld repair, 210 weld rework, 209–210 welds, 112, 113–114, 203(F) white cast iron fatigue and fracture properties, 194, 198(F) friction and wear of, 221, 222(F,T) overview, 219 structure and properties, 33–34 whiteheart, 193, 196(T), 219 wrist-pin bosses, 143–144(F)

X x-ray diffraction (XRD), 239 x-ray ﬂorescence spectroscopy (XRF), 239 XRD. See x-ray diffraction (XRD) XRF. See x-ray ﬂorescence spectroscopy (XRF) X-sections equal opposite legs, 26–27(F) heat transfer and transport phenomena, 32(F), 33 three legs equal, 27(F), 28(F)

ASM International is the society for materials

engineers and scientists, a worldwide network dedicated to advancing industry, technology, and applications of metals and materials. ASM International, Materials Park, Ohio, USA www.asminternational.org This publication is copyright © ASM International®. All rights reserved. Publication title

Product code

Casting Design and Performance

05263G

To order products from ASM International: Online Visit www.asminternational.org/bookstore Telephone 1-800-336-5152 (US) or 1-440-338-5151 (Outside US) Fax 1-440-338-4634 Mail

Customer Service, ASM International 9639 Kinsman Rd, Materials Park, Ohio 44073-0002, USA

Email [email protected]

American Technical Publishers Ltd. 27-29 Knowl Piece, Wilbury Way, Hitchin Hertfordshire SG4 0SX, In Europe United Kingdom Telephone: 01462 437933 (account holders), 01462 431525 (credit card)

www.ameritech.co.uk Neutrino Inc. In Japan Takahashi Bldg., 44-3 Fuda 1-chome, Chofu-Shi, Tokyo 182 Japan Telephone: 81 (0) 424 84 5550 Terms of Use. This publication is being made available in PDF format as a benefit to members and customers of ASM International. You may download and print a copy of this publication for your personal use only. Other use and distribution is prohibited without the express written permission of ASM International. No warranties, express or implied, including, without limitation, warranties of merchantability or fitness for a particular purpose, are given in connection with this publication. Although this information is believed to be accurate by ASM, ASM cannot guarantee that favorable results will be obtained from the use of this publication alone. This publication is intended for use by persons having technical skill, at their sole discretion and risk. Since the conditions of product or material use are outside of ASM's control, ASM assumes no liability or obligation in connection with any use of this information. As with any material, evaluation of the material under end-use conditions prior to specification is essential. Therefore, specific testing under actual conditions is recommended. Nothing contained in this publication shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in this publication shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against liability for such infringement.