Laser Cladding

Laser Cladding

Laser Cladding

Ehsan Toyserkani Amir Khajepour Stephen Corbin

CRC PR E S S Boca Raton London New York Washington, D.C.

Library of Congress Cataloging-in-Publication Data Catalog record is available from the Library of Congress

This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage or retrieval system, without prior permission in writing from the publisher. The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from CRC Press LLC for such copying. Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe.

Visit the CRC Press Web site at www.crcpress.com © 2005 by CRC Press LLC No claim to original U.S. Government works International Standard Book Number 0-8493-2172-7 Printed in the United States of America 1 2 3 4 5 6 7 8 9 0 Printed on acid-free paper

Dedications

Ehsan Toyserkani To my father and mother, who taught me to learn. To my beloved wife and son, Homeyra and Ali, who have truly brought joy in my life.

Amir Khajepour To my son and wife for the joy and comfort you have given to my life, and my family for your support.

Stephen Corbin To the memory of my sister, Pamela Corbin, who was an enthusiastic and passionate teacher for over 20 years. Pamela inspired my own interest in the field of education which ultimately led to my academic career.

© 2005 by CRC Press LLC

Preface

Laser cladding by powder injection has received significant attention in recent years due to its unique features and capabilities in various industries involved in metallic coating, high-value components repair, prototyping, and low-volume manufacturing. This emerging laser material processing technique is an interdisciplinary technology utilizing laser technology, computer-aided design and manufacturing (CAD/CAM), robotics, sensors and control, and powder metallurgy and rapid solidification. Further development of this technique depends on enhancement of the technologies involved and understanding the interconnections among these technologies and the process quality. A good comprehension of the underlying physics of the laser cladding process is key in the development of the process as a reliable coating and manufacturing technology. In addition to a good grasp of feedback control and automation, a strong knowledge of material science, heat transfer, and fluid dynamics is essential to a successful development of an automated laser cladding system. The intent of this book is to address the lack of a comprehensive book dealing with the dierent aspects of laser cladding. The authors have used the results of their own research and experience in the past few years along with the findings of many other researchers in the preparation of this book. The book provides a solid and detailed description of laser cladding in modeling, materials, and control and can be used in both academia and industry. The book begins with a review of the applications of laser cladding, and continues with physical descriptions of the process and the parameters involved, process modeling and control, process applications, and the physical metallurgy of alloying and solidification during laser cladding. We illustrate the general principles of the technique with several case studies based on a number of important common laser cladding applications. Extensive references to the current literature have also been provided to guide the reader to further information on desired topics. We are very much indebted to many colleagues and students for their help in the preparation of this book. Special thanks go to Professors David Weckman, Walter Duley and Jan Huissoon of the University of Waterloo for their valuable comments. We are grateful to Dr. Steen Nowotny of the Fraunhofer Institute for Material and Beam Technology, Dr. Lijue Xue of the Integrated Manufacturing Technologies Institute (IMTI) of National Research Council of Canada (NRC), Dr. Frank Arcella of AeroMet Cooperation, Dr. Joohyun Choi of University of Missouri at Rolla, Mr. David Gill of Sandia National Laboratories, and Mr. Michael Kardos of Optomec Inc. for providing us © 2005 by CRC Press LLC

with copyright permissions for several photos used throughout the book. We would like to express our sincere appreciation and gratitude to Dr. Hamid Niazmand for providing us with coaxial model analysis, Mr. Ian Fraser and the Safety O!ce of the University of Waterloo for allowing us to use their safety materials, Ms. Sarah Mask for editing the manuscript, Ms. Ji-Hyun Kim for drawing the figures, and Ms. Lori Brown for administrative support. Thanks are especially due for the financial support of Materials and Manufacturing of Ontario (MMO) and the National Sciences and Engineering Research Council of Canada (NSERC). Finally, thanks to our families, who make it all worthwhile. Ehsan Toyserkani, Amir Khajepour, Stephen Corbin Waterloo, Ont., Canada, 2004

© 2005 by CRC Press LLC

About the Authors

Dr. Ehsan Toyserkani received his Ph.D in mechanical engineering from the University of Waterloo, Ontario, Canada, in 2003. His early research and industrial interests were on design of mechatronics systems. This interest was expanded to include the development of intelligent controllers for laser cladding technology during his Ph.D program. He was awarded postdoctoral fellowships from the National Sciences and Engineering Research Council of Canada (NSERC) and the Canadian Space Agency (CSA) to conduct projects related to laser cladding technology. In 2004, he joined the Mechatronics Engineering Group of the Department of Mechanical Engineering at the University of Waterloo as an Assistant Professor.

Dr. Amir Khajepour received his Ph.D in mechanical engineering from the University of Waterloo in Ontario, Canada in 4996. In 4997, he joined the Department of Mechanical Engineering, University of Waterloo, where he is currently a Professor in Mechatronics Engineering. The thrust of Dr. Khajepour’s research is in modeling and control of dynamic systems with focus on automated laser cladding, ultra high-speed robotics, and advanced vehicle systems. His extensive research collaboration with industry has resulted in many new technologies, patents, and journal publications.

Dr. Stephen Francis Corbin received his B.Eng and M.Sc. degrees in metallurgical engineering from Dalhousie University in 4986/87 and his Ph.D. in materials engineering from McMaster University in 4993. Following four years as a product and process development specialist with the Westaim Corporation in Fort Saskatchewan, Alberta, Dr. Corbin joined the University of Waterloo, Ontario, Canada in 4997 where he is currently an Associate Professor in the Materials Engineering and Processing Group within the Department of Mechanical Engineering. He has taught introductory materials engineering courses as well as senior and graduate courses in process and physical metallurgy and materials characterization. His areas of expertise include: powder metallurgy, laser processing, solders and brazes, porous materials and metal/ceramic composites.

© 2005 by CRC Press LLC

Contents

1 Introduction 4.4 What is Laser Cladding? 4.2 Dierent Names, Same Technology 4.3 Why Laser Cladding? 4.4 History of Laser Cladding 4.5 Applications and Market Opportunities 4.5.4 Coating 4.5.2 Parts Repair and Refurbishment 4.5.3 Rapid Prototyping and Tooling 4.6 Future Direction of Laser Cladding Technology 4.7 Looking Ahead 2 Background and Basic Overview 2.4 Laser Material Techniques 2.2 Dierences Between Laser Cladding, Alloying and Glazing 2.3 Dierent Methods of Laser Cladding 2.3.4 Two-Step Laser Cladding (Pre-placed Laser Cladding) 2.3.2 One-Step Laser Cladding 2.4 Clad Dimensional Characteristics 2.5 Important Parameters in Laser Cladding by Powder Injection 2.5.4 Dilution 2.5.2 Wetting Angle and Interfacial Free Energies 2.5.3 Laser Pulse Shaping 2.6 Combined Parameters 2.6.4 Aspect Ratio 2.6.2 Combined Energy and Powder Densities’ Parameters 2.7 Comparison Between Laser Cladding and Other Metallic Coating Techniques 2.8 Comparison Between Laser Cladding and Other Prototyping Techniques 3 Laser Cladding Equipment 3.4 Lasers 3.4.4 Laser Types 3.4.2 Laser Beam Characteristics © 2005 by CRC Press LLC

3.4.3 3.2

3.3

Types of Lasers and Laser Beam Characteristics in Laser Cladding Process Powder Feeders and Powder Delivery Nozzles 3.2.4 Powder Feeder Types 3.2.2 Applications of Powder Feeders to Laser Cladding 3.2.3 Nozzles Positioning Devices 3.3.4 CAD/CAM System for Trajectory Generation

4 Laser Cladding Process Modeling 4.4 Physics of the Process 4.2 Governing Equations 4.2.4 Essential Boundary Conditions 4.3 Laser Cladding Models in Literature 4.3.4 Steady-State Models 4.3.2 Dynamic Models 4.4 Lumped Models 4.5 Analytical Modeling 4.6 Numerical Modeling — A Case Study 4.6.4 Thermal Mathematical Model 4.6.2 Solution Algorithm 4.6.3 Numerical Parameters 4.6.4 Numerical Results 4.6.5 Experimental and Numerical Analysis 4.6.6 Comparison Between Numerical and Experimental Results 4.7 Flow Field Modeling at the Exit of Coaxial Nozzle 4.7.4 Laminar Model 4.8 Experimental-Based Modeling Techniques 4.8.4 Stochastic Analysis 4.8.2 Artificial Neural Network Modeling 5 Control of Laser Cladding Process 5.4 Sensors 5.2 Closed-Loop Control of Laser Cladding 5.3 Closed-Loop Control of Laser Cladding, An Example 5.3.4 Equipment and Configuration 5.3.2 Optical CCD-based Detector 5.3.3 Control Strategy 5.3.4 Closed-Loop vs. Open-Loop 5.3.5 Application of the Developed Controller to Fabrication of Two Simple Components 5.4 Application of Knowledge-Based Control to Laser Cladding 5.4.4 Fuzzy Logic Controller © 2005 by CRC Press LLC

6 Physical Metallurgy and Material Systems of Laser Cladding 6.4 Cladability 6.4.4 Processing Parameter Considerations 6.4.2 Metallurgical Considerations 6.2 Solidification Conditions Encounter in Laser Cladding 6.2.4 Process Conditions 6.2.2 Constitutional Supercooling 6.2.3 Rapid Solidification 6.2.4 Microstructure Maps 6.2.5 Microstructural Scale 6.3 Material Systems Used in Laser Cladding 6.3.4 High Temperature Alloys 6.3.2 Composites 7 Safety 7.4 Laser Classification 7.2 Laser Hazards 7.2.4 Eye Hazards 7.2.2 Collateral Radiation 7.2.3 Electrical Hazards 7.2.4 Chemical Hazards 7.2.5 Fire Hazards 7.2.6 Explosion Hazards 7.2.7 Eye Protection 7.3 Powder Hazards References

© 2005 by CRC Press LLC

1 Introduction

With the rapid growth of laser applications and the reduced cost of laser systems, laser material processing has gained increased importance in a variety of industries. Automotive, aerospace, navy, defense, and many other sectors are widely adapting laser technology for welding, cutting, and hardening. Among the applications of laser technology, laser cladding has received significant attention in recent years due to its diversified potential for material processing such as metallic coating, high-value components repair, prototyping, and even low-volume manufacturing.

1.1

What is Laser Cladding?

Laser cladding is an interdisciplinary technology utilizing laser technology, computer-aided design and manufacturing (CAD/CAM), robotics, sensors and control, and powder metallurgy. Laser cladding utilizes a laser heat source to deposit a thin layer of a desired metal on a moving substrate. The deposited material can be transferred to the substrate by several methods: powder injection, pre-placed powder on the substrate, or by wire feeding. Among these methods, laser cladding by powder injection has been demonstrated to be most eective. In this process, the laser beam melts the powder particles and a thin layer of the moving substrate to deposit a layer of the powder particles on the substrate. A great variety of materials can be deposited on a substrate using laser cladding by powder injection to form a layer with thicknesses ranging from 0.05 to 2 mm and widths as narrow as 0.4 mm. In addition to metallic coating applications, the laser cladding process oers a revolutionary layered manufacturing and prototyping technique. Integration of the laser cladding technology with a three-dimensional CAD solid model, which is sliced into many layers, provides the ability to fabricate complex components without intermediate steps. The development of the laser cladding technology depends on enhancement of the technologies involved. Understanding the interconnections between the involved technologies and the process quality is a major step for the development of laser cladding. However, numerous interactions between the © 2005 by CRC Press LLC

technologies involved in laser cladding not only increase the complexity of the process but also increase the number of process parameters.

1.2

Dierent Names, Same Technology

A survey of the literature indicates that many names have been given to laser cladding technology based on its highly diversified applications. For example, in the coating applications, in addition to “laser cladding”, researchers use the terms “laser coating” [4, 2], “laser powder deposition” [3, 4] or “laser surfacing” [5, 6, 7]. In the rapid prototyping or layered manufacturing applications, numerous names have been used. In prototyping by pre-placed powder laser cladding, the technology is called “selective laser sintering of metals” (SLSM) [8, 9], or “direct metal laser sintering” [40]. In the powder injection laser cladding process, which is the focus of this book, a wide variety of names have been used as outlined below. • At Sandia National Laboratories, a process has been developed for rapid prototyping, which is called “Laser Engineered Net ShapingTM ” R ) [44, 42, 43]. (LENS° • At the University of Michigan in Ann Arbor, the developed process is called “direct metal deposition” (DMDTM ). This process incorporates features of laser cladding, CAD/CAM package, vision and control system [44, 45]. The University of Missouri at Rolla also uses the same name for the process [46, 47]. • “Laser direct casting” (LDC) is used at the University of Liverpool to describe the process, in which a coaxial nozzle is utilized to produce 3D components from a selection of metal powders [48]. • The Integrated Manufacturing Technologies Institute (IMTI) of National Research Council of Canada (NRC) uses the term “laser consolidation” for this process [49, 20, 24, 22]. • The term “laser powder fusion” (LPF) is used by some industries that are involved in turbine blade repair [23]. • At the Laser Aided Manufacturing Processes Laboratory at the University of Missouri at Rolla and at Swiss Federal Institute of Technology the process is named “laser metal forming” (LMF) [24, 25]. • “Directed light fabrication” (DLF) is the name used at Los Alamos National Laboratory in which a process has been developed for free forming and prototyping [26, 27, 28]. © 2005 by CRC Press LLC

• “Laser powder deposition” (LPD) has been used by several research groups in China and England [29, 30]. • At the University of Waterloo, an automated laser cladding technology has been developed for coating and prototyping, which is named “automated laser powder deposition” (ALPD) [34]. • At Stanford University, Carnegie Mellon University, and Penn State University, the process is called “solid free-form fabrication” or “shape deposition manufacturing” [32]. • “Laser rapid forming” (LRF) is the name used by a research group at Shanghai Jiaotong University [33]. This name can be confused with another laser-based technology, laser bending, which is also called laser rapid forming. • Finally, “laser additive manufacture” (LAMSM ) is the acronym used by AeroMet Corporation of Eden Prairie at Minnesota, fully owned subsidiary of MTS Systems Corporation [34]. Despite the variety of names, in practice all of the terms describe technology that share several common features: deposition of thin layers of powder particles melted by a laser heat source on a substrate. The process will be termed “laser cladding” throughout this book.

1.3

Why Laser Cladding?

Laser cladding oers many advantages over conventional coating processes such as arc welding and plasma spraying. The laser cladding technique can produce a much better coating, with minimal dilution, minimal distortion, and better surface quality. There are also a number of advantages to use this technique as a rapid prototyping technique. Rapid prototyping can be used to produce a mechanical component in a layer-by-layer fashion, which enables the fabrication of part with features that may be unique to laser cladding prototyping, such as a homogeneous structure, enhanced mechanical properties, and one-step production of complex geometries. Parts fabricated using the technique are near net shape, but will generally require final machining. They also have good grain structure, and have properties similar to, or even better than the intrinsic materials. Both pre-placed and powder injection laser cladding oer these features; however, laser cladding by powder injection has fewer material limitations than pre-placed in which it does not require secondary placing powder laminating operations and it can be used to repair parts as well as to fabricate them. © 2005 by CRC Press LLC

Due to its additive nature, laser cladding can be applied in a variety of ways to parts, tools and advanced manufacturing to overcome the limitations of existing metal fabrication technologies. This results in a number of benefits as follows: 4. Reduction of production time: The length of time required to build a prototype is a problem for new product development. In many cases, both prototype and production tooling are needed; therefore, the length of time to produce a prototype and the necessary tooling can be several months. The laser cladding process can reduce this time by fabricating tools and main prototypes directly from the CAD solid model [44]. 2. Enhancement of thermal control: The laser cladding process oers a well-controlled heat-treated zone due to the nature of the laser beam. A high-power laser beam is well confined and tense and, as a result, the rapid heating and cooling that occur in the process have little eect of heat on the base material. Therefore, the original properties of the base material are aected to a limited extent only. In addition, this thermal zone can be monitored and optimized during the process, which can significantly improve the quality of the tools produced. Laser cladding also oers a controllable energy over the surface of the desired tool to control the rate of solidification, which is the main parameter in the formation of microstructure and mechanical properties [35]. 3. Parts repair: Current tool repair technology relies on destructive, high-temperature welding processes. In addition, machining errors or last-minute engineering changes can aect on-time delivery of tooling, and potentially impact the introduction date of a new product. Laser cladding can be applied as a safe technology to repair tooling, especially on critical contacting surfaces. Laser cladding increases tool life and in many cases can save a high-value tool that would otherwise need to be replaced [20, 36]. 4. Production of a functionally graded part: In conventional metallic fabrication, it is di!cult to produce a part from dierent materials layers. Laser cladding oers a method to produce functionally graded parts by injecting dierent materials during the fabrication of the parts. It is also possible to produce desired alloys by injection of dierent powders through various nozzles around the process zone [37, 38]. 5. Production of smart structure: In conventional metallic fabrication methods, embedding objects into the tools is impossible due to the nature of manufacturing. Laser cladding, with its additive nature, oers the ability to create “smart structure” by embedding objects such as sensors and magnets during fabrication. Encapsulating these objects reduces the potential for damage or failure from temperature and environmental conditions [39]. © 2005 by CRC Press LLC

Despite its obvious benefits, laser cladding is not yet widely utilized in metallic coating or prototyping applications. While laser cladding clearly offers a number of advantages over conventional fabrication technologies, the process can also have some drawbacks. Due to disturbances in the process, the clad quality may vary significantly. Variations of the quality may even be observed between processing cycles performed using the same operating conditions. This poor reproducibility arises from the high sensitivity of laser cladding to small changes in the operating parameters such as laser power, beam velocity and powder feed rate, as well as to process disturbances such as variations in absorptivity. Finding an optimal set of parameters experimentally and using them in an open-loop laser cladding process may not result in a good quality clad due to random or periodic disturbances in the system. Therefore, development of an intelligent closed-loop control system is essential for overcoming the eects of disturbances in the process. High investment cost, low e!ciency of the laser sources, and lack of control over the cladding process are disadvantages of the use of this technology in coating and prototyping processes. However, with continued technological developments in high-power diode lasers (HPDL), fiber lasers, and sophisticated knowledge-based controllers, the laser cladding process shows a great industrial potential for use in metallic coating and prototyping applications.

1.4

History of Laser Cladding

The invention of the first working laser by Maiman [40] in the 4960’s was a breakthrough in science. Immediately after this invention, scientists claimed that the laser was the answer to a multitude of scientific problems that might not have been even known during those years. These problems had arisen in many areas resulting in lasers being adapted to many technologies to dissolve the problems with their unique features. One of the areas that benefited from the lasers technology was material processing, which was rapidly developing in the 4970’s when the e!ciencies and power of commercial lasers became higher and higher. The development of high-power gas lasers (e.g., CO2 lasers) in 4975 made laser welding, cutting and metal hardening possible in that decade. Among the laser material processing technologies, laser cladding was used by Gnanamuthu at Rockwell International Corporation in Thousand Oaks of California in the late 4970’s [44]. A pre-placed laser cladding method was used to investigate the feasibility of the process in applying dense ceramic cladding to metallic workpieces. At about the same time, several research groups around the world began projects to develop apparatus and systems for development and improvement of the process. Among these groups, the project conducted by William M. © 2005 by CRC Press LLC

Steen first at Imperial College of University of London, England and subsequently at Liverpool University, where he moved in April 4988, had a great impact on the development of laser cladding technology [6, 42]. He along with Vijitha Weerasinghe introduced laser cladding by powder injection to academia and conducted a number of projects to evaluate the developed process [43, 44, 45, 46]. The other research group, led by Jyoti Mazumder, at the University of Illinois at Urbana-Champaign, Urbana, USA, contributed many fundamental principles to this area in the 4980’s. Mazumder’s group not only developed models for this process and studied the mechanism of the process [47, 48, 49], but they also applied the technology to many metals and ceramics to investigate their potential for cladability, and also wear and corrosion resistance [50, 54, 52, 53, 54]. A review of the literature shows that the number of papers and patents related to this technology increased significantly in the 4980’s. These papers disclosed the devices for enhancement of the technology, such as development of powder feeders, cooling systems, hemispherical reflecting device for re-absorbing the reflected light, etc. [55]. The applications of the technology for wear and corrosion resistant alloys were also reported by many research groups in the 4980’s [56, 57]. The features of this technology received attention from industry in the 4980’s as well [58]. Laser cladding was identified as a process with a significant edge over the conventional processes for wear and corrosion resistant coating. The research projects being conducted by dierent industries were even ahead of academic projects. The first reported use of the laser cladding by industry was the hard-facing of Nimonic turbine blade interlock shrouds for the RB244 jet engine at Rolls Royce in 4984. In 4983, at Pratt and Whitney, the nickel-base alloy turbines of JT8 and JT9 engines were hard-faced using preplaced laser cladding [59]. Hard-facing of turbine blade shroud tips and znotches by laser cladding continued to gain acceptance by dierent companies. The technology was being accepted by leading engine manufacturers such as General Electric, Pratt & Whitney, Allied Signal, Rolls-Royce, Allison, Solar and MTU. From a commercialization point of view, several companies, such as Avco Everett Metalworking Lasers Inc. and United Technologies Industrial Lasers Inc. were established in the 4980’s to address the needs of industry for metallic coating and repair in North America and Europe. In the automotive industry, laser cladding technology was transferred to the market for the engine valve seat coating by some European and Asian automotive companies, such as Fiat [58], Toyota [60], and Mercedes Benz. In the component repair market, laser cladding brought a huge amount of consideration in the 4980’s. Laser cladding was successfully utilized for rebuilding and coating of the H-dimension (airfoil section thickness) of worn turbine vanes, the tip of the turbine blades, and turbine bolts [58, 64, 36, 62, 63]. A number of dierent companies and research groups around the world have utilized laser cladding technology for turbine blade repair, including Human Corporation and Gorham Technologies in the USA, Starrag in Switzerland, © 2005 by CRC Press LLC

Sultzer in the Netherlands, SIFCO Turbine Components in Ireland, and many others. Another application of laser cladding, rapid prototyping or layered manufacturing, received a great deal of attention in the 4990’s and continues to be explored in the new millennium. In 4986, a new process for prototyping complex parts called “stereolithography” was patented. This process used ultraviolet lasers to selectively cure photo polymer materials. In 4988, the first commercial stereolithography machine was sold and a new industry in rapid prototyping was established. Stereolithography gave product developers the ability to quickly and accurately visualize, iterate, optimize, and fabricate new designs directly from a three-dimensional CAD solid model. Although most commercial systems used polymers and photopolymers, the industry was looking for features to rapidly fabricate the metallic prototypes that could be directly used in real machines. Global competition was also forcing product manufacturers to look for new ways to reduce new product design time and manufacturing costs. In response to these demands, eorts began to utilize laser cladding technology for the development of machines for direct metal prototyping. At the University of Illinois at Urbana, the Mazumder research group extended their project to the development of systems for rapid prototyping, which was later called “direct metal deposition” (DMDTM ) [64]. The group examined building parts in one and two dimensions, taking into consideration both the time and cost involved in the process compared with traditional methods. Due to the success of their project, the research group moved to the University of Michigan to conduct their research with more emphasis on automotive manufacturing. In the late 4990’s, the developed technology was licensed to Precision Optical Manufacturing Inc. (POM Inc.) at Plymouth, Michigan to supply molds and dies fabricated by the developed technique to the automotive industry of the Michigan area. Many other research and development groups initiated projects to develop methods for prototyping metallic parts based on laser cladding by powder injection. Among them, Sandia National Laboratories, which is a multi-program laboratory operated by the Lockheed Martin Corporation, was funded by the Departments of Energy of the US government to conduct research for development of laser near shape fabrication methods. The University of Liverpool research group, led by Steen, also began projects for laser direct manufacturing, which contributed extensively to this field [65, 66]. In the late 4990’s, a research group under the leadership of Xue and Islam of the Integrated Manufacturing Technologies Institute (IMTI) of National Research Council (NRC) of Canada developed apparatus and methods for layered manufacturing called “laser consolidation” [49, 20, 24, 22]. Their achievements had a great impact in this field due to the unique surface quality of the parts produced using their technique. In the last few years, a research group at the University of Waterloo conducted research for the development of an intelligent laser cladding apparatus. © 2005 by CRC Press LLC

The main focus of the research is in the direction of development of intelligent modeling techniques and knowledge-based controllers for the process. These knowledge-based controllers will be eventually used in an autonomous laser cladding machine, which can not only deposit a wide range of alloys, but can also make the complex shapes without the need for the presence of specialists [34, 67, 68]. The flexibility of laser cladding is beginning to be recognized by many industries and research groups. The potential of this technology is great as research groups continue to contribute to its growth through research programs and training of students in laser cladding techniques technology.

1.5

Applications and Market Opportunities

As mentioned earlier, laser cladding has several diverse applications. In the following sections, various attempts by research groups and industry to adapt the process to dierent applications are explained.

1.5.1

Coating

Coating results in deposition of a thin layer of material (e.g., metals and ceramics) onto the surface of a selected material. This changes the surface properties of the substrate to those of the deposited material. The substrate becomes a composite material exhibiting properties generally not achievable through the use of the substrate material alone. The coating provides a durable, corrosion-resistant layer, and the core material provides the load bearing capability. A number of dierent types of metals, such as chromium, titanium, nickel, copper, and cadmium, can be used in the metallic coating process. There are many coating deposition techniques available. However, selecting the best depends on many parameters such as size, metallurgy of the substrate, adaptability of the coating material to the technique intended, level of adhesion required, and availability and cost of the equipment. Although laser cladding has the potential for utilization in dierent industrial divisions for metallic coating, its application to metallic coating is limited due to the high cost and the low process speed. However, with improvement of laser e!ciency, reduction in the cost of lasers, and the development of new generation of lasers such as high-power diode and fiber lasers, there is a strong potential for laser cladding to be widely used for coating applications in several major industries. Another indication of the potential of laser cladding for coating of a variety of materials is the increase in the number of published papers and reports concerning the technology in the recent years. Based on © 2005 by CRC Press LLC

information from the Compendex search engine, the number of papers dealing with applications of laser cladding to coating of dierent materials has risen from 620 papers in the 4990’s to more than 750 papers from 2000 to 2004. The majority of published papers related to the metallic coating by laser cladding address the use of several major materials in aerospace, medical, and automotive industries. Titanium-based alloys [69, 70, 74, 72], nickel-based superalloys [73, 74, 75, 76, 77, 78], and cobalt-based alloys [79, 80, 84] are some of the important alloys that are deposited on dierent substrates such as unalloyed steels, alloyed steels, hardened steels, stainless steels, aluminum alloys, cast irons, and nickel or cobalt-based alloys. POM Inc. in Michigan performed the deposition of wear-resistant and high-temperature materials (e.g., cobalt-based Stellite and nickel-based alloys) onto tool surfaces that are exposed to harsh high temperature, thermal shock environmental conditions, to increase their lives [82]. Recently, the biocermics coating on titanium alloys was also performed by laser cladding; the coated parts are then used in orthopedic implants with a calcium phosphate layer in order to promote the growth of the bone when the implant is inserted in the body [83]. Laser cladding along with other laser surface treatment methods has also been examined for the production of glassy metallic layers, which provide superior resistance against wear and corrosion [84]. The most leading metallic coatings market for laser cladding is the coating of commercial aircraft gas turbines. The world original equipment manufacturer’s (OEM) market for coatings used in commercial aircraft gas turbine sections is estimated to be approximately $460 million per annum, based on the information released by Gorham technologies. In response to demands for the development of a higher e!ciency, lower cost industrial gas turbine engine, high-strength, high-temperature capability materials, such as nickelbased superalloys, are coated on the turbine bodies to meet the needs of the hot gas path components. The shroud interlock between turbine blades has been also hardfaced with Triballoy to reduce the wear due to sliding between blades during the warm up and cold down the engine. Laser cladding has been recently used in this sector to deposit the mentioned materials on the spacecraft components. With recent technological improvements in the new generation of lasers, it is expected that laser cladding technology will take on an increasingly important role in this market. In addition, laser cladding also has several other coating applications for industrial parts to produce surfaces, that are resistant to abrasive, erosive and adhesive wear; wet corrosion; and high temperature oxidation and corrosion. Some of the products that have received metallic coating by laser cladding are: • Shafts used in drilling tools • Engine valve seats © 2005 by CRC Press LLC

FIGURE 1.1 Coating of oil drilling tools by laser cladding (Source: Courtesy of Fraunhofer Institute for Material and Beam Technology, Germany [85]).

• Tools hardfacing • Hydraulics pump components • Molds Figure 4.4 shows one of the applications of laser cladding for coating of oil drilling tools, which are subjected to significant wear in their operation [85].

1.5.2

Parts Repair and Refurbishment

A major application of laser cladding is in the repair and refurbishment of high-value components such as tools, turbine blades and military components. Laser cladding can be used to rescue high-value components which are overmachined due to the errors in design or machining process. These engineering or machining errors can easily jeopardize the entire eort of the design and manufacturing of high-value tools or components. Conventional methods use welding to retrieve these damaged components; however, these methods are usually destructive due to the highly distributed temperature over the area of repair. This thermal destruction causes a low mechanical quality, crack, porosity and very short life of the component. Laser cladding can provide a permanent structural repair and refurbishment on many alloys (e.g., aluminum alloys) that are generally considered unweldable by conventional methods. The success of the laser cladding technology in this © 2005 by CRC Press LLC

area is due to the small heat zone, rapid solidification, increased cleanness, lower dilution, and increased controllability over the depth of heat-aected zone. An example of the repair by laser cladding of a shell made from high strength aluminum alloys (i.e., unweldable 7075/7475 aluminum alloys) is shown in Figure 4.2. This shell is used in undersea weapon components, which sustain wear and damage as a result of handling, operation and the corrosive nature of the saltwater environment. Laser cladding has been applied to repair of such components. The results are promising such that the repairs are permanent. They stop corrosion and increase structural integrity [86].

FIGURE 1.2 Repair by laser cladding of a shell made from high-strength aluminum alloys (Source: Courtesy of the Naval Undersea Warfare Center (NUWC), USA).

One of the other areas, in which laser cladding plays an important role is turbine blades repair and refurbishment. Turbine blades are under very high thermal and mechanical stresses (i.e., centrifugal force and thermal gradient) in an aggressive environment. The blades usually suer a variety of damage during operation, such as creep, life cycle fatigue, hot corrosion, impact of external particles on their surfaces, etc. Therefore, in terms of maintenance requirements, manufacturing di!culties and costs, the blades are the most critical item of today’s gas turbines. As a result, blade manufacturing and maintenance companies are looking for repair technologies that not only repair the defected blades superiorly, but also maintain the original mechanical and © 2005 by CRC Press LLC

metallurgical features of the repaired components. The low heat input property of laser cladding is the most unique characteristic of this technology that makes it highly attractive for jet engine components repair applications, in which metal depositions are required to be applied to superalloys. These superalloys are highly susceptible to strength loss and physical distortion when exposed to excessive temperature variations. Conventional repair techniques, such as tungsten inert gas, metal inert gas, plasma and electron beam welding, usually cause a large amount of heat during weld metal deposition which results in large temperature increases in the body of the component. The temperature increases above certain limits cause the base alloy to be weakened. This weakening along with component distortion can cause irreversible damage to the part. In contrast to conventional weld repair, the laser cladding process transfers heat only to localized areas, typically using a 0.5 mm diameter laser beam. As a result, heat inputs are at least one order of magnitude lower than the heat input incurred during conventional welding, which results in reduced residual stresses and distortion and a substantially smaller heat-aected zone [87]. An even greater repair market potential exists for the application of laser cladding in turbine engines. Advanced gas turbines, are now being fitted with single crystal and directionality solidified components in order to achieve maximum thermal e!ciencies by operating the engines at higher turbine inlet temperatures. In the fabrication and repair of such engines, laser cladding is rapidly being recognized as a critical and essential technology. In some cases, it is also recognized as the only repair technology because the oriented airfoil castings are highly susceptible to re-crystallization when subjected to the intense heat induced during conventional weld repair [36, 20, 62]. Figure 4.3 shows the application of laser cladding to repair a tip of turbine blade performed in Sulzer Elbar, which is one of the constitute companies that form Sulzer Turbomachinery Services [88]. The turbine blade is made from precipitation-hardened CC-superalloy Inconel 738. The size of the repair and refurbishment market is immense. However, there seems to be no concrete data that defines the size of this market. One of the key aspects of this market is aircraft engine components maintenance. Aircraft engine maintenance accounts for 30 percent of the total cost of aircraft maintenance, which is a good indication of the size of available market for the technology. The global market of the repair of aircraft engine turbines and compressor blades, used in civil and military applications, has been estimated to be about 4.2 billion dollars per annum [89].

1.5.3

Rapid Prototyping and Tooling

A new and major application of laser cladding is in rapid prototyping (RP) and rapid tooling (RT) markets for rapid fabrication of complex components and tools. Production of tools such as cutting tools, dies and molds, which have traditionally been fabricated by highly-skilled tool and mold makers © 2005 by CRC Press LLC

FIGURE 1.3 Application of laser cladding to repair of a tip of rotor blade made of precipitationhardened CC-superalloy Inconel 738 (Source: Courtesy of Sulzer Elbar, The Netherlands).

using CNC and electrical discharge machining, has always suered from cost problems and slow turnaround times for manufacturers. If tools are fabricated late, market opportunities will be missed, which is often the death knell for a new product. As a result, rapid tooling has received significant attentions in recent years from manufacturers looking for technologies that are able to produce high-value tools and components with high integrity, high density, and good surface quality, at low manufacturing costs with a short manufacturing time [90]. The market for RP and RT is significantly large with a significant annual growth. The worldwide tooling market is estimated in the tens of billions of dollars per annum. The market size for RP was approximately $800 million in 2002, when 4.5 million parts were produced by the available units in the market [94]. The applications of RP in North America are categorized within the following sectors: 25 percent consumer products, 24 percent automotive, 44 percent business machines, 44 percent medical, 8 percent academic, 8 percent aerospace, 5 percent government and 8 percent other. However, many of the customers for RP units are looking for a reliable metallic prototyping machine that will not only be intelligent enough to prototype the components without the need for highly qualified personnel, but that will also be robust enough to produce the components with high quality [92]. Laser cladding technology has the potential to address the current gap for metallic rapid prototyping. Laser cladding has demonstrated promising capabilities for tools and components fabrication. A recent survey by the National © 2005 by CRC Press LLC

Center for Manufacturing Science (NCMS) revealed that laser cladding could reduce the time of die production by 40 percent, if the process is controllable over the dimensions of the product. It has been reported that for the production of surgical tools, the technology can reduce 62 steps into 7 steps [44]. In recent years, researchers have been working on enhancement of laser cladding to construct prototypes and production tooling, even for precision metal parts made from dierent commercial alloys such as H43 tool steel, 346 and 304 stainless steels, nickel-based superalloys (e.g., Inconel 625, 690, 748, 2024), aluminum, composite, and titanium-based material (e.g., Ti-6Al4V). The feature of the technology provides the functionally graded material deposition capability, which are applicable in many aerospace components in which a light weight but hard external surface are requested.

FIGURE 1.4 R (Source: Courtesy of Sandia National LaboratoFabrication of a blade by LENS° ries).

R ) [42], developed at Sandia Laser Engineering Net ShapingTM (LENS ° National Laboratories, is one of the rapid metal forming processes that has demonstrated the feasibility of laser cladding to produce near-net shape metal parts. The system utilizes a CAD solid model to build an object one layer at a time. The solid model is sliced into a series of layers that are subsequently used to generate the motion to deposit each layer of material. These layers are then deposited in a subsequent fashion to build the entire part. R technology. Also, Figure 4.4 shows the fabrication of a blade by the LENS° R to Figure 4.5 shows a special housing which has been fabricated by LENS° reduce the secondary machining time. © 2005 by CRC Press LLC

FIGURE 1.5 R A housing part fabricated by hybrid fabrication method including the LENS ° technology (Source: Courtesy of Sandia National Laboratories).

Among the materials used for part manufacturing and prototyping, titaniumbased alloys have received significant attention due to their use in the aerospace industry. Many eorts have been made to obtain the required processing parameters to produce high-quality titanium alloy material, in terms of mechanical and metallurgical properties. One of the premiere companies in this field is AeroMet Corporation, which has developed the “laser additive manufacturing” (LAMSM ) technology. Their results demonstrate the capability of producing shaped structures (e.g., ribbed structural) from Ti, Ti-6Al-4V and other alloys such as Ti-5Al-2.5Sn, including extra low interstitial (ELI) grades. Figure 4.6 shows several components made by LAMSM , which were then machined to meet the required dimensions and tolerances. The required production time and the quality of products are significantly better than those for forging methods. AeroMet’s current production facility includes an 48 kW CO2 laser and a five-axis manipulation capable of producing component sizes within 3 × 3 × 1 m. Figure 4.7 shows the AeroMet equipment for performing LAMSM . Another key player in components manufacturing by laser cladding is the IMTI-NRC in Canada. As mentioned earlier, the technology used by the NRC is called “laser consolidation”. This technology can produce metallurgically sound parts from IN-625, 346L Stellite 6, and M4 without porosity and crack. The samples showed an excellent dimensional accuracy and surface finish. The produced parts from LC IN-625 alloy and 346L stainless steel were shown to have stronger strength than the respective cast and even wrought materials [24, 49]. Figures 4.8 and 4.9 depict some of the parts fabricated by the NRC. Figure 4.8 depicts the components of a robot fabricated by laser consolidation to evaluate the potential of the process as a rapid functional prototype manufacturing process for the production of structural components using Ti6Al-4V alloy. The figure also indicates the steps of production. These parts © 2005 by CRC Press LLC

FIGURE 1.6 Dierent components fabricated by laser forming technology (Source: Courtesy of AeroMet Corporation, USA).

FIGURE 1.7 AeroMet equipment including a 48 kW CO2 laser for performing laser forming (Source: Courtesy of AeroMet Corporation, USA).

© 2005 by CRC Press LLC

were used in the Advanced Robotic Mechatronics System (ARMS) project, which was initiated by MD Robotics and supported by the Canadian Space Agency (CSA) [22].

a)

b)

c)

FIGURE 1.8 A robot’s joint, fabricated by laser consolidation made from Ti-6Al-4V: a) original interface on the flat substrate, b) fabrication of housing on the interface, c) joint after final machining (Source: Courtesy of the NRC’s Integrated Manufacturing Technologies Institute).

Figure 4.9 depicts two industrial parts: Figure 4.9a shows a part made from wear-resistant Stellite 6; Figure 4.9b shows a part which is a half-part of a complex flextensional transducer shell made from IN-625 [24]. The produced shell has an excellent dimensional accuracy and wall integrity. Direct metal deposition (DMD) is being developed to fabricate molds and dies as well as for parts repair using the laser cladding technology. The wide application of DMD in the aerospace and medical fields is due to its large potential cost savings. Koch et al. [93] demonstrated the application of laseraided DMD to generate components with a dimensional accuracy of 0.25 mm using a closed-loop control of process parameters. The dimensional accuracy of their part depends on the uniformity and repeatability of the process [44]. The commercialization of this technique, which incorporates a closed-loop system, has been successful and a machine called DMD5000 has been introduced to the market by POM Inc., Michigan. This machine provides features for direct metal fabrication. The DMD5000 machine is an integration of a 5-kW CO2 laser integrated with a flying optic, a gantry robot plus a XY table CAD/CAM system, and several special powder feeders. The machine has a workspace with size of 1.5 × 0.5 × 0.45 m. Laser cladding technology has been under investigation at the University of Waterloo with more emphasis on the development of an automated machine for this technology. The developed machine has been used for production of complex parts made from H43 and nickel-based superalloy. The developed © 2005 by CRC Press LLC

a)

b)

FIGURE 1.9 Two fabricated parts by laser consolidation: a) a component made from Stellite 6, b) a complex flextensional transducer shell made from IN-625 (Source: Courtesy of the NRC’s Integrated Manufacturing Technologies Institute).

technology integrates vision-based detectors and knowledge-based controller with conventional laser cladding equipment to control the process parameters such as laser power, process speed, and powder feed rates, which enables control of the clad geometry in real-time. Figure 4.40 shows a logo made from H43, which was fabricated using the system developed at the University of Waterloo.

FIGURE 1.10 A logo fabricated from H43 tool steel at the Mechanical Engineering Department of the University of Waterloo.

© 2005 by CRC Press LLC

1.6

Future Direction of Laser Cladding Technology

Laser cladding technology is in the early stages of commercialization and will oer a revolutionary new manufacturing technique to the industry in the new millennium. Due to its promising features, many industries keep their eyes on the technology. Research groups and companies involved in the development of the technology should canalize their eorts in several main directions to resolve the current shortcomings of this technology. The main focus of research eorts should be the development of autonomous machines for the process, which can not only deposit a wide range of alloys, but can also make complex shapes without the need for the presence of specialists. The development of an automated machine may not be possible without string collaborations between the researchers with dierent disciplines. The development of a knowledgebased controller for such a machine needs expertise of scientists with control, automation, and also metallurgical backgrounds. Additional research eorts should focus on increasing the speed of the process as the current processing speed is slow compared to the competitive techniques such as plasma and thermal spraying. Laser cladding applications to the aircraft industry and the eects of laser cladding on substrate also require further investigation. The investigation of laser cladding and eects on the applications requires an understanding of the process and the relationship between laser energy, process speed, powder feed rate, and mechanical and metallurgical properties. Therefore, eorts should be dedicated to modeling of the process. In the turbine blade repair market, customers are looking for a machine that has the capability to repair the turbine blade in-situ without removing the blades from the rotor. To do this, it is necessary to develop a sophisticated positioning device that will provide a su!cient maneuverability around the rotor. The capability of cladding on an inclined surface is also essential for in-situ turbine blade repair. In the recent years, micro- and nanotechnology have received a significant attention. Scientists say any technology can have applications in micro- and nanotechnology, if it can be miniaturized in the right fashion. Having said that, laser cladding has a great potential to benefit these cutting-edge technologies [94, 95], as laser cladding techniques can provide a revolutionary technique for maskless micro structure fabrication. In recent years, Optomec Inc., a company in the USA, has developed a process called M3 D, which makes it possible to coat dierent substrates in the range of 40 to 50 µm. The 40 µm deposition makes it attractive particularly for space-limited applications in the electronic technology. The concepts of their work are based on the pre-placed laser cladding process; however, the Optomec Inc.’s technology utilizes the spray of liquid droplets mixed either with metals or organic ele© 2005 by CRC Press LLC

ments. Post-heat treatment is then performed by the laser to sinter the metal particles. Figures 4.44 and 4.42 show two samples made using the Optomec Inc.’s technology. Figure 4.44 shows silver lines with 50 µm width deposited over a 500 µm step of a micro-mirror. Figure 4.42 shows a tapered spiral GPS antenna. Due to the fascinating features of this technology, such as deposition on a non-planar micro substrate with 40 µm coating width, it is anticipated that many research groups will concentrate on this field in the Micro Electro Mechanical Systems (MEMS) prototyping development.

FIGURE 1.11 Silver lines with 50-micron width deposited over a 500-micron step by M 3 D (Source: Courtesy of Optomec Inc., USA).

Recently, the NRC of Canada successfully tested laser cladding technology for the use in fabrication of space-related structures, such as the main parts of a robotic arm. These tests resulted in the production of robotic arm components with excellent mechanical properties, which is a significant indication of the potential of laser cladding technology for in-space manufacturing facilities [22]. In June 2003, the NASA Marshall Space Flight Center brought a variety of national engineering and manufacturing specialists to Huntsville, Alabama, for a workshop in “In-Space Manufacturing of Space Transportation Infrastructure”. This workshop addressed strategies for developing a robust in-space transportation infrastructure that may eventually include permanent refueling stations and maintenance platforms in space, as well as cargo vehicles that haul supplies across the shipping lanes of space. A discussion of the characteristics of laser cladding technology identified its potential to lead inspace manufacturing processes, as it would allow space dwellers and explorers to quickly design and produce replacement parts in the space. Therefore, it is anticipated that the laser cladding process will play an important role in the © 2005 by CRC Press LLC

FIGURE 1.12 Tapered spiral GPS antenna by M 3 D technology (Source: Courtesy of Optomec Inc., USA).

development of possible in-space manufacturing methods in the near future. Last but not least, market development should be taken into consideration as a very important issue for exploring the technology by the manufacturing sectors. A number of industries still consider the laser a “fancy” device. This results in a large technological barrier between conventional and advanced manufacturing experts. This barrier has to be brought down through a close dialog between these two groups along with end-users. It is important to expect users to request appropriate features for a particular process; therefore, a partnership between the conventional and advanced manufacturing experts should be conducted to satisfy end-users. The authors believe that breakthroughs in laser cladding will require a hybridization of advanced and conventional techniques, which may not be possible without a close interaction between conventional and advanced specialists on one side and end-users on the other side. Education of highly qualified people in this area can minimize the current technological gap between the advanced and conventional manufacturing sectors. Unfortunately, current laser material processing programs in universities are very limited, most likely due to the lack of high-power laser in institutes. In Canada, just a few universities are equipped with high-power lasers; therefore, the opportunities for education in the field of laser material processing are very limited. In addition, the field is multidisciplinary and requires training in a number of dierent scientific fields, such as laser, optics, automation, control, robotics and material science. © 2005 by CRC Press LLC

1.7

Looking Ahead

This book consists of seven chapters. An introduction to the potential applications of laser cladding technology in industry was explained in Chapter 4. Chapter 2 reviews the background of laser cladding. In Chapter 3, the equipment used in laser cladding (i.e., laser, positioning device, and powder feeder and nozzles) will be explained. In Chapter 4, the physics of the process will be discussed and dierent modeling approaches for the process will be presented. The application of experimental-based modeling techniques to laser cladding including stochastic and artificial neural network techniques will be also addressed in Chapter 4. Chapter 5 addresses the control aspect of the technology, as well as the design, simulation and implementation of several classical and fuzzy controllers applied to laser cladding. Chapter 6 explains the mechanical and metallurgical characteristics of laser cladding for dierent metallic alloys. Also, the eects of process parameters on the clad bead quality and a methodology for evaluation of clad bead quality using the combined parameters will be developed in this chapter. Chapter 7 ends the book by explaining safety issues related to the laser and powder materials.

© 2005 by CRC Press LLC

2 Background and Basic Overview

In this chapter, a basic overview on laser cladding is presented. The laser cladding classification, laser cladding process parameters, and comparison of laser cladding process with the competitive techniques are some of the topics addressed in this chapter.

2.1

Laser Material Techniques

A laser beam provides unique characteristics for material processing. The electromagnetic radiation of a laser beam is absorbed by the surface of opaque materials (e.g., metals). The interaction time between the laser and material leads to dierent processes as shown in Figure 2.4. The relative velocity of the laser beam with respect to the substrate causes the thermal cycle in the surface layer. The figure also shows a schematic representation of the physical phenomena that occur during various laser material processes. These processes are due to dierent combinations of absorption, heat conduction, melting, powder addition, and rapid solidification. However, the common physical phenomenon of all laser material processing techniques is rapid solidification, which causes a superior and fine metallurgical structure. Laser cladding is one of the important types of laser material processing, in which a laser beam irradiates powder particles and the surface of the substrate moved by a positioning device. As a result of additive powder particles, a thin layer called a “clad” is produced on the substrate.

2.2

Dierences Between Laser Cladding, Alloying and Glazing

Adding powder to the melt pool may create three dierent products depending on the type and the amount of material added. Figure 2.2 depicts schematic cross sections of the coating-substrate couple for these three processes. These © 2005 by CRC Press LLC

Absorption v Laser Beam

Heat Conduction

Rapid solidification v Laser Beam

Transformation Hardening & Annealing

v Laser Beam

Adding Powder Melting Rapid Solidification

Remelting, Welding & Shock Hardening

Rapid Solidification

Alloying, glazing, cladding

FIGURE 2.1 Schematic of physical phenomena during dierent laser material processing techniques.

are classified as laser cladding, glazing, and alloying. Included on the right is a schematic of the compositional profile across the coating-substrate couple. In laser alloying, a small amount of powder is fed into the melt pool. As such, homogeneous mixing throughout the melt region may be obtained [96]. Laser cladding resembles laser alloying except that dilution by the substrate is kept to a minimum and more addition of material to the surface is required [35]. In laser glazing, a metallic glass coating is deposited in order to provide an environmentally eective surface in terms of wear and corrosion [97]. The principle advantage of laser glazing is that it alters microstructures without changing the composition [97, 98]. Using laser cladding, the following advantages can be obtained compared to other surface material processing [99, 400, 404]: 4. Reduced dilution, which is the mixing percentage of the substrate to the clad region (compared to laser alloying) © 2005 by CRC Press LLC

2. Improved wear resistance of a part 3. Reduced thermal distortion 4. Reduced porosity, particularly in laser cladding by powder injection 5. Improved controllability of the process 6. Reduced post-cladding machining time and cost

Laser Alloying

A+ B

A 100%

Clad Composition

B 100%

B

A 100%

Clad Composition

B 100%

A

B Laser Glazing

Laser Cladding

A

A

A+B

FIGURE 2.2 Dierent microstructures of laser alloying, glazing, and cladding.

2.3

Dierent Methods of Laser Cladding

Basically, there are two dierent techniques for laser cladding as follows: 4. Two-step process (pre-placed laser cladding) 2. One-step process In the two-step process, the first stage consists of a layer of coating material being placed before laser irradiating. It is then melted with the substrate material by the laser beam in the second stage (see Figure 2.3a) [402]. In the one-step process, an additive material is fed into the melt pool. The additive material may be supplied in the following dierent forms: © 2005 by CRC Press LLC

Laser Beam Paste

Clad Substrate b1) Laser Beam

Clad

Laser Beam

Inert Gas

Placed Powder

Powder & Inert Gas

Clad Substrate

Substrate b2)

a)

Laser Beam

Wire Clad Substrate b3)

FIGURE 2.3 Dierent methods of laser cladding: a) two-step laser cladding, b) one-step laser cladding, including b4: paste laser cladding, b2: powder injection laser cladding, b3: wire feeding laser cladding.

© 2005 by CRC Press LLC

• Powder injection • Wire feeding • Paste Figure 2.3b depicts one-step laser cladding techniques where an inert gas shrouds the laser material interaction zone to prevent oxidation of the surface at the high processing temperature. Powder injection cladding is a more robust method than wire and paste cladding, because there is no direct contact with the melt pool, and the laser beam can pass through the stream of powder particles instead of being obstructed by the wire or paste. In the following, we briefly address several issues involved in two-step laser cladding; then we concentrate on laser cladding by powder injection for the rest of this chapter and the book.

2.3.1

Two-Step Laser Cladding (Pre-placed Laser Cladding)

Pre-placed laser cladding is a simple method used for coating and prototyping. Several issues are involved in this process. The pre-placed powder particles on the substrate must have not only enough bonding to the substrate, but also enough cohesion to each other. It is necessary to prevent the powder particles on the substrate from removing due to the gas flow during the melting in the second step of the process. To overcome this problem, the powder is usually mixed with a chemical binder to ensure its cohesion with the substrate during the process. The side eect of a chemical binder is porosity in the clad due to its evaporation during the process. In the second step of the process the following phenomena occur: 4. Creation of a melt pool in the top surface of the pre-placed powder due to the radiation of laser beam 2. Expansion of melt pool to the interface with the substrate due to the heat conduction 3. Penetration of heat to the substrate causing a fusion bond The control of heat is a very important issue in this method to prevent the high dilution. Dilution is considered to be the mixing percentage of the substrate to the clad region. This problem is one of the important shortcomings of two-step laser cladding that usually limits the process only to single track cladding. Powell et al. [402] developed a theoretical model and analysis technique for pre-placed laser cladding. Their theoretical calculation resulted in a plot, which shows the eect of interaction time on the position of the melt pool front. This result is shown in Figure 2.4. The results indicate that the dilution increases with increasing interaction time (decreasing the process speed). © 2005 by CRC Press LLC

Melt Pool Front Position

Original Powder Surface

0

50W

Depth of Melting (mm)

0.4

100W (3.18x107 W/ m2)

0.8

Cladding Powder

1.2 1.6

200W 400W

2.0 2000W

800W

2.4

Substrate

Increasing Dilution

2.8 0

0.1

0.2

0.3

0.4

0.5

Time (s)

FIGURE 2.4 Displacement of melted front with respect to interaction time at various average laser powers, when the beam radius is 4 mm [402].

2.3.2

One-Step Laser Cladding

As they are shown in Figures 2.3b4 to 2.3b3, one-step laser cladding can be categorized into three methods: powder injection, wire feeding, and paste laser cladding. The common feature of all three methods is the feeding of deposited material in the presence of the laser. 2.3.2.1

Laser Cladding by Powder Injection

In laser cladding by powder injection, powder particles are fed into the heat zone to produce a layer of clad as shown in Figure 2.3b2. We will address this process in depth throughout this book. 2.3.2.2

Laser Cladding by Wire Feeding

In laser cladding by wire feeding, the main idea is to use wire instead of powder as shown in Figure 2.3b3. The wire is usually fed through a ceramic drum containing the desired material wire. Due to the nature of feeding mechanisms, it is essential to use a wire that has been straightened and stored without plastic deformation to ensure stable transport without vibration [403, 404]. Compared to laser cladding by powder injection, it has been claimed that laser cladding by wire feeding has some special advantages [405]. One of its most important advantages is its adaptation to the cladding position. Metal wires © 2005 by CRC Press LLC

are cheaper than metal powders, and also wire feeding wastes less material than powder feeding. In contrast, low surface quality, low bonding strength, porosity, cracks and drop transfer are some problems of wire cladding. The melted liquid at the end of the wire does not flow smoothly and continuously onto the workpiece, which is called drop transfer phenomena. Kim et al. [404] conducted research to find the best process parameters that prevent the drop transfer phenomena. They showed that by selecting correct wire feeding direction and position, the splashing of molten drop for laser cladding with wire feeding can be solved. In this case, wire can be plunged into the melt pool and be melted by the heat of the molten metal. However, a successful process strongly depends on the process parameters, and in the presence of disturbances the quality of clad drops dramatically. 2.3.2.3

Laser Cladding by Paste

In laser cladding by paste, a stream of paste-bound material is deposited on a point of the substrate that is usually a little bit ahead of the laser beam [406], as shown in Figure 2.3b4. The paste consists of the hardfacing powder with a suitable binder. However, the binder must be dried in a short period of time while the hardfacing material in a compact form is still kept; otherwise powder particles are blown away by the shielding gas.

Paste Track Before Laser Processing

Laser Beam

Laser Beam

Laser Beam

Paste Volume

Paste Volume Clad

Clad

Lost Paste Substrate

Ideal Situation

a)

Substrate

Substrate

High Process Speed

Low Process Speed

and/or

and/or

Low Paste Volume

High Paste Volume

b)

c)

FIGURE 2.5 Eect of process speed and paste feed rate on the quality of clad [406].

For this process, a special paste feeding system should be designed. Lugscheider et al. [406] designed and implemented a paste feeding system along with a cooling system to protect the paste from thermal emissions from the process zone. The shape of paste on the substrate is controlled by paste feed rate and substrate speed. To have a good clad quality, it is essential to optimize these parameters. A poor paste supply or too high process speed causes high dilu© 2005 by CRC Press LLC

tion and low track height if the laser energy is kept constant. An oversupply of paste on the substrate increases pores formation since evaporation of the binder is inhibited, and it increases the loss of hardfacing material. Figure 2.5 shows the eect of substrate traverse speed and paste feed rate on the quality of the final clad layer. High porosity, extreme sensitivity of the process to disturbances, and di!culties in paste feeding mechanism are troublesome conditions of paste laser cladding.

2.4

Clad Dimensional Characteristics

Several parameters are associated with the clad geometry, which are shown in Figure 2.6. In this figure, h is the clad height, w is the clad width,  is the angle of wetting, and b is the clad depth representing the thickness of substrate melted during the cladding and added to the clad region.

w Clad Bead

θ

h

b Substrate

FIGURE 2.6 A typical cross section of a clad bead.

2.5

Important Parameters in Laser Cladding by Powder Injection

A large variety of operating parameters and physical phenomena determine the quality of laser cladding. Figure 2.7 summarizes these parameters grouped © 2005 by CRC Press LLC

as inputs, processes, and outputs. Generally, the inputs or operating parameters are the laser, motion device, powder feeder set points, and also the material and ambient properties. The outputs of the process which represent the clad quality are the geometry, microstructure, cracks, porosity, surface roughness, residual stresses and dilution [407, 35, 408, 409, 440]. In the following, some of the major associated parameters with the process are explained.

Inputs Laser

Motion Device

Average power Spot size Wave length Pulsed/CW Beam profile Laser pulse shaping

Material

Relative velocity Relative acceleration System accuracy

Powder Feeder

Substrate geometry Composition Metallurgical, thermo physical & optical properties Powder size

Powder feed rate Inert gas flow rate Nozzle specification Powder stream profile

Surface tension

Processes Physical phenomena

Outputs Clad quality

Absorption

Geometry

Conduction

Microstructure

Diffusion

Hardness

Melt pool dynamics

Cracks

Fluid convection

Pores

Gas/melt pool interaction Laser attenuation by powder Rapid solidification

Residual stresses Surface roughness Microstructure Dilution

Ambient Properties Preheating Shield gas velocity Kind of shield gas

FIGURE 2.7 Inputs, outputs and process parameters of laser cladding by powder injection.

2.5.1

Dilution

One of the properties of the clad layer is called dilution. Dilution has two definitions: geometrical and metallurgical [407]. The geometrical definition of dilution is illustrated in Figure 2.6. According to the specified parameters in the figure, the dilution is b (2.4) dilution = h+b where b is the thickness of substrate that was melted during the cladding process [mm], and h is the height of the clad bead [mm]. © 2005 by CRC Press LLC

Alternatively, dilution may be defined as the percentage of the total volume of the surface layer contributed by melting of the substrate [407]. According to the composition, dilution is defined as dilution =

c (Xc+s  Xc ) s (Xs  Xc+s ) + c (Xc+s  Xc )

(2.2)

where c is the density of melted powder alloy [kg/m3 ], s is density of substrate material [kg/m3 ], Xc+s is weight percent of element X in the total surface of the clad region [%], Xc is the weight percent of element X in the powder alloy [%], and Xs is the weight percent of element X in the substrate [%]. Of interest is the fact that dilution increases with increasing laser power but decreases with increasing travel speed.

2.5.2

Wetting Angle and Interfacial Free Energies

In laser cladding, either pre-placed or powder injection, wetting angle and interfacial free energies are important parameters that indicate the quality of the clad. In general, three types of clad cross section may be produced by laser cladding as shown in Figure 2.8. These cross sections represent the amount of dilution, corresponding wetting angle , and interfacial free energies  [J/m2 ]. Three interfacial energies for laser cladding can be considered as solid-liquid interfacial free energy  SL , solid-vapor interfacial energy  SV , and liquid-vapor interfacial energy  LV .

γ LV γ SL

θ

a)

γ LV γ SV

γ SL

θ

γ SV

b)

γ SL

θ

γ LV γ SV

c)

FIGURE 2.8 Laser cladding cross sections, associated wetting angle and interfacial free energies: a) high dilution, well wetting, b) ideal clad, c) no dilution, non-wetting.

A balance between the mentioned energies governs the shape of the clad bead. This balance is expressed by  SV   SL =  LV cos()

(2.3)

The liquid will wet the substrate as cos() $ 1 or equivalently, if  SV  SL > © 2005 by CRC Press LLC

 LV , which corresponds to Figure 2.8b. A large positive spreading factor S =  SV   SL   LV causes spreading, whereas a lower number causes a non-wetting system as shown in Figure 2.8c. When the laser energy is high, dilution increases and wetting angle decreases as shown in Figure 2.8a. In laser cladding, oxidation is a serious problem at higher processing temperatures, which are required for melting the metals. The oxidation causes a low quality clad as shown in Figure 2.8c. This is due to the poor wetting of an oxide substrate by a liquid metal and also much lower surface energies of metal oxides.

2.5.3

Laser Pulse Shaping

The laser beam can be in the form of continuous wave (CW) or pulsed wave. In the form of pulse wave, several parameters associated with shape of pulses are defined: laser pulse energy E, laser pulse width (laser pulse duration) W , laser pulse frequency (laser pulse repetition rate) F , average power Pl and duty cycle C. These parameters which are shown in Figure 2.9 can be expressed by C = FW

(2.4)

Pl = EF

(2.5)

1-FW F

Energy (J)

W

E

0

Time (s)

1

FIGURE 2.9 Laser pulse shaping, including pulse width W , pulse energy E and pulse frequency

F.

© 2005 by CRC Press LLC

2.6

Combined Parameters

Due to the numerous parameters involved in the process, it is essential to use the combined parameters to address the clad quality of the process by finding the correlation between the combined parameters and the clad quality. Dierent eective parameters have been reported, which are categorized for continuous and pulsed laser energy. In the following, these combined parameters are introduced.

2.6.1

Aspect Ratio

Aspect ratio AR is a ratio between width and height of the clad represented by w AR = h

2.6.2

Combined Energy and Powder Densities’ Parameters

There are dierent approaches to show the correlation of the clad aspect ratio and the eective energy. However, they can be categorized into two types of laser beam waves: continuous wave (CW) and pulse wave. 2.6.2.1

Combined Parameters for Continuous Wave (CW) Laser Beam

Steen et al. [42] showed that the beginning of non-wetting (i.e., discontinuous clad track) is correlated with combined parameter 2rPlwU , and the melting through of the substrate is correlated with combined parameter PUw where, Pw is the absorbed energy to the substrate laser power [W], U is process speed [mm/s], and rl is the laser spot radius on the substrate surface [mm]. They also showed that combined parameter 2Pwm˙Url is correlated with the aspect ratio, where m ˙ is the powder feed rate [g/s] [444]. In addition, they showed that a term 2rPlwm˙ has a maximum value before dilution occurs (e.g., 2500 J/(gmm) for Colmonoy Wallex PC6) and also combined parameter 2rPlwU represents the minimum energy required for cladding before the track starts to be discontinuous (e.g., 22 J/mm2 for Colmonoy alloy). They have arrived at a general plot which shows the correlation of these combined parameters and feasibility of cladding process as shown in Figure 2.40. The figure shows the hatched area that the cladding process is feasible. For continuous wave, Wu et al. [442] introduced two combined parameters for laser cladding by powder injection, which can be fitted to the process to present the critical states. These two combined parameters result in a simpler interpretation of the cladding process. They are called specific energy © 2005 by CRC Press LLC

Aspect Ratio

Power Per Spot Diameter (W/mm)

(e.g., 5 for Colmonoy Alloy)

Dilution Problem 1000

pw 2mrl

Porosity Problem

(e.g., 2500 J/(gmm) for Colmonoy Alloy) 500

le ib as e F

on gi Re

Pw 2rlU

(e.g., 22 J/mm for Colmonoy Alloy)

Non-feasible Cladding Region 0

0.1

0.2

0.3

0.4

0.5

Powder Feed Rate (g/s)

FIGURE 2.10 Correlation of aspect ratio, combined parameters, power per spot diameter, and powder feed rate with feasibility of laser cladding by powder injection [42].

Especif ic [J/mm2 ] and powder density G [g/dm2 ], which are expressed by Especif ic =

Pw 2Url

(2.6)

m ˙ (2.7) 2U rl where Pw is the laser power on the substrate [W], U is the process speed ˙ [mm/s], rl is the radius of the laser beam on the substrate [mm], and m is the powder feed rate [g/min]. Figure 2.44 shows the correlation between Especif ic and G [g/m2 ] and the critical situation in the laser cladding process of Cobalt-based alloy on an A3 steel substrate [442]. G=

2.6.2.2

Combined Parameters for Pulsed Wave Laser Beam

Two combined parameters, eective energy density Eef f [J/mm2 ] and eective powder deposition density #eff [g/mm2 ], are defined for a pulsed wave laser beam [3, 4, 443] and are expressed by Eef f = © 2005 by CRC Press LLC

EF Aef f

(2.8)

2.5

Fine Condition

50

2.0

40

1.5

30

1.0 Critical Condition 0.5

20

10

Single Track Height (mm)

Specific Energy (J/ mm2)

60

0 10

30

50

70

90

110

Powder Density (g/ m2) Correlation between specific energy and powder density Correlation between powder density and single-track height

FIGURE 2.11 Correlation between specific energy Especif ic [J/mm2 ] and powder density G [g/m2 ] and their eect on single track height for Co-based alloy on an A3 steel substrate [442].

#ef f =

mF ˙ W Aeff

(2.9)

where Aeff is the eective area per second which is irradiated by the laser beam and powder stream [mm2 /s], E is the pulse laser energy [J], F is the laser pulse frequency [Hz], W is the laser pulse width [s] and m ˙ is the powder feed rate [g/s]. This is determined not only by the substrate velocity but also by the pulse characteristics of the laser beam. Referring to Figure 2.42 and performing the geometrical analysis, the following equation is obtained for the eective area per second as

Aeff =

; h 1F W 2 A r + 2U r  2F Url  A l F A ? l A A A =

rl2 F + 2Url W F

© 2005 by CRC Press LLC

1F W 2F yU

 rl2 ( 2  sin1 for rl > for

³ ´i

1F W 2F

rl 5

y rl

U

1F W 2F U

(2.40)

where

y=

s

rl2



µ

1  FW U 2F

¶2

(2.11)

U is the process speed [mm/s], and rl is the laser spot radius on the substrate [mm]. Equations (2.10) and (2.11) are derived based on subtracting Ac , which is the half uncovered area of a rectangle created by two successive laser pulses (See Figure 2.12), from the total area covered by the pulses per second.

WU

1 − FW U F

rl

Ac = area uncovered by two succesive laser pulses FIGURE 2.12 A schematic of the aective area of cladding created by succesive laser pulses.

Inherent in Equation (2.9) is the assumption that when the pulse is o, no powder is deposited on the substrate due to the absence of the energy provided by the laser pulses. This aspect of the process is introduced through the inclusion of the duty cycle C = F W in Equation (2.9). Using these two combined parameters, a general plot can be obtained for iron-aluminide coating on mild steel or H13 [3, 4, 113] that presents the corresponding clad quality based on eective energy density and eective powder deposition density as shown in Figure 2.13 © 2005 by CRC Press LLC

160

Effective Energy Density (J/ mm2)

140

Good Quality Clads

120

Roughness, Some Pores & Cracks

100

80

Brittle Clads 60

No Cladding

40

20

1

1.25

1.5

1.75

Effective Powder Deposition Density (g/

2

2.25 x 10-3

mm2)

FIGURE 2.13 Correlation between eective energy density Eef f and eective powder deposition density # ef f for iron-aluminide coating on mild steel.

2.7

Comparison Between Laser Cladding and Other Metallic Coating Techniques

The application of laser cladding should compete with several major coating techniques such as thermal spray, welding, chemical vapor deposition (CVD), and physical vapor deposition (PVD). The thermal spray can be categorized into three methods, which are combustion torch (e.g., flame-spray, high-velocity oxy fuel, and detonation gun), electric (wire) arc, and plasma arc. In addition, PVD can be categorized into ion plating and ion implantation. CVD is also categorized into: sputtering, ion plating, plasma-enhanced CVD, low-pressure CVD, laser-enhanced CVD, active reactive evaporation, ion beam and laser evaporation [114]. Table 2.1 compares several major features of these coating techniques to provide the advantages and disadvantages of these processes for metallic and non-metallic coating applications. As it is listed in Table 2.1, laser cladding creates a very strong bond with low dilution, where a very low heat-aected zone (HAZ) is produced in the substrate. However, the investment cost and © 2005 by CRC Press LLC

TABLE 2.1

Comparison between laser cladding and other coating techniques. Feature Laser Welding Thermal CVD PVD Cladding spray Bonding strength Dilution Coating materials Coating thickness

Repeatability Heat-aected zone (HAZ) Controllability Cost

High High Metals, ceramics 50 µm to 2 mm

High High Metals

Low Nil Metals, ceramics 0.05 µm to 20 µm

Low Nil Metals, ceramics 0.05 µm to 10 µm

Moderate

Moderate Nil Metals, ceramics 50 µm to several mm Moderate

Moderate to high Low

High

High

High

High

Very low

Very low

Moderate to high High

Low

Moderate

Moderate

Moderate

Moderate to high High

Moderate to high High

1 to several mm

maintenance cost of the laser cladding machine are currently high, which are disadvantages of this process. It is anticipated that because of the fast growing of the new generation of lasers such as high-power diode and fiber lasers, which oer higher e!ciency and lower maintenance cost, laser cladding technology will play an important role in the metallic coating market in the near future.

2.8

Comparison Between Laser Cladding and Other Prototyping Techniques

There are about 20 methods of rapid prototyping. Some of the major rapid prototyping techniques are: stereolithography apparatus (SLA), fused deposition manufacturing (FDM), selective laser sintering (SLS), 3D printing, and laser cladding-based prototyping (i.e, DMD, LENS, ALPD, laser consolidation, etc.). Table 2.2 compares the features of these techniques. In this table, one of the features is called “support structure”, which refers to parts to place in the unsupported geometries during fabrication, such as the supporting part for fabrication of the top portions of a part with the shape of the letter “T”. These supports are usually calculated and added to the part by the system’s software and may be formed of the same material as the part, or from an entirely © 2005 by CRC Press LLC

TABLE 2.2

Comparison between laser cladding-based and other prototyping techniques. Feature Laser SLA FDM SLS 3D cladding Printing (e.g., DM D, ALPD, etc.)

Dimensional accuracy Prototype’s material

Moderate

Moderate

Metals and ceramics

Prototype’s quality

High

Support structure Cost of machine

Not required High (still under R&D)

Polymers and photopolymers Low to moderate (e.g., fragile) Required Relatively Moderate

Moderate to high Filament ABS plastic High

Required Moderate

Low Polymers, metals and ceramics Low (e.g., Porosity and cracks) Not Required Moderate

Low to Moderate Hard plastic, runner Low (e.g., fragile)

Required Low

dierent material. Support structures are either mechanically removed or dissolved away in secondary operations before the part can be used. As Table 2.2 indicates, prototyping techniques, which are performed based on laser cladding technology, have superior features for metallic rapid prototyping. However, the cost of this technology is still high due to the need of highly qualified personnel and the cost of laser systems. The development of an autonomous system for performing the laser cladding process without any need to expert personnel is under research and development in several research groups and industry. It is anticipated that providing a fully intelligent laser cladding apparatus will overcome the shortcomings of this technology.

© 2005 by CRC Press LLC

3 Laser Cladding Equipment

The laser cladding process requires the following equipment: a laser, a powder feeder along with delivery nozzles, and a positioning device equipped with CAD/CAM software. It is essential to understand the construction of these devices and their performance under dierent working conditions for the laser cladding process to be successful. This chapter provides a comprehensive comparison of available lasers, powder feeders, and nozzles to demonstrate their potential and suitability for use in laser cladding technology. The chapter also includes a brief review of available positioning devices and CAD/CAM systems suitable for this process.

3.1

Lasers

In the early 1960s, an enormous contribution was made to technology with invention of the first working laser. The word “laser” stands for light amplification by the stimulated emission of radiation. Miaman [40] invented the first ruby laser, which was the result of considerable discovery of Einstein [115], who demonstrated that lasing action should be possible. In general, the light emitted by lasers is dierent from the ordinary light sources such as incandescent bulbs, fluorescent light, and high-intensity arc lamps. Laser light has the following characteristics: • Highly monochromic. All regular light sources emit light (e.g., incandescent and fluorescent light) of many dierent wavelengths. Ordinary colored light consists of a broad range of wavelengths covering a particular portion of the visible-light spectrum. The beam of a laser, on the other hand, consists of an extremely narrow range of wavelengths within one single color portion of the spectrum meaning that it consists of light of almost a single wavelength. This nearly “monochromatic” or “single-colored” property is unique to laser light. • Highly coherent. The light waves within a highly collimated laser beam may be defined as a source of coherent light, unlike other regular light sources. This characteristic leads to a constant phase dierence in two or more waves over time. Two waves are said to be in phase if © 2005 by CRC Press LLC

their crests and troughs meet at the same place and at the same time, whereas the waves are out of phase if the crests of one wave meet the troughs of another. • Highly directional. All conventional light sources emit light in all directions, and it always diverges more rapidly than a laser beam. Directionality is the characteristic of laser light that causes it to travel in a single direction within a narrow cone of divergence. However, perfectly parallel beams of directional light (i.e., collimated light) cannot be produced even by a laser. In some applications, optical systems are employed with lasers to improve the directionality of the output beam. • Sharply focused. For laser light, the focused spot can be very small; for example, an intensity of 1017 W/cm2 is readily obtained by a laser, which is incredibly higher than any energy source (e.g., an oxyacetylene flame has an intensity of only about 103 W/cm2 ). In order to explain how a laser works, it is necessary to explain the following three processes by which the atom can move from one energy state to another: 1. Absorption. If the atom is placed in an electromagnetic field that is resonating at frequency f , the atom can absorb an amount of energy hf as represented by (3.1) hf = Ex  E0 and move to the higher energy state. In the equation, Ex is the higher level of energy and E0 is the ground level of energy for an atom. Figure 3.1a shows the atom in its ground and then in a higher level of energy. 2. Spontaneous emission. After a time, the atom will move of its own accord to its ground state, emitting a photon of energy hf in the process. This process, shown in Figure 3.1b, is called spontaneous emission because the event is not triggered by any outside influence. Usually, the mean life of excited atoms before spontaneous emission occurs is about 108 s. However, for some excited states, this means the life could be as much as 105 times longer; this longer state is called metastable. The light produced by the spontaneous emission of an atom is neither monochromatic and directional, nor coherent. 3. Stimulated emission. In this step, the atom is again in its excited state, but this time radiation with a frequency of f is present. A photon of energy of hf can stimulate the atom to move to its ground state, during the process, the atom emits an additional photon, whose energy is also hf. This process, shown in Figure 3.1c, is called stimulated emission because the event is triggered by the external photon. The emitted photon is in every way identical to the stimulating photon. The waves associated with the photons have the same energy, phase, © 2005 by CRC Press LLC

polarization, and direction of travel. Therefore, stimulated emission produces light that is monochromatic, directional, and coherent; this light appears as the output beam of a laser-stimulated emission for a single atom.

Ex a) Absorption

b) Spontaneous emission

c) Stimulated emission

Ex

hf

None E0

E0

Ex

Ex hf

None E0

E0

Ex

Ex

E0

E0

hf

Radiation

Matter

Matter

hf hf

Radiation

FIGURE 3.1 Interaction of radiation and matter in a) absorption, b) spontaneous emission, and c) stimulated emission.

In practice, generation of laser is subject to the interaction of a large number of atoms in the excitation field. Ludwing Boltzmann’s theorem shows that the number of atoms in a state of higher energy Nx is a function of the number of atoms in their ground state N0 and their corresponding energies, as represented by Nx = N0 e(Ex E0 )/T (3.2) where  is Boltzmann’s constant, E0 is energy of ground state, Ex is energy of atoms in a higher state, and T is thermal equilibrium temperature. This equation indicates that Nx < N0 because Ex > E0 . As a result, there are fewer atoms in the excited state than in the ground state. If a flood of atoms with photons of energy Ex E0 is generated, as shown in Figure 3.2a, photons will disappear via absorption by ground state atoms. Einstein showed that the © 2005 by CRC Press LLC

probabilities per atom for these two processes are identical. Therefore, because there are more atoms in the ground state, the net eect will be the absorption of photons. However, to produce laser light, the number of emitted photons should be more than absorbed photons. To accomplish this, a situation in which stimulated emission dominates should be occurred. The direct way to cause this is to begin with more atoms in the excited state than in the ground state, as shown in Figure 3.2b. This phenomenon is called population inversion. However, such a population inversion is not consistent with thermal equilibrium. Therefore, it is necessary to consider appropriate ways to improve the population inversion phenomenon in any laser type.

a)

Ex

Ex

Eo

Eo b)

FIGURE 3.2 a) equilibiruim distribution of atoms between the ground state E0 and excited state Ex , b) inverted population.

3.1.1

Laser Types

The numerous laser types can be categorized based on physical and operating parameters, which are involved in the laser beam generation. There are several ways to classify laser types; however, the most common way is to classify them based on their physical state of the active material. According to this criterion, lasers can be categorized as follows: • Gas lasers • Excimer lasers • Solid-state lasers • Semiconductor lasers • Liquid dye lasers • Fiber lasers © 2005 by CRC Press LLC

These classes of lasers can provide dierent wavelengths from  1 mm to  1 nm. Output powers cover even greater range of values. For continuous wave (CW) lasers, typical powers range from a few mW, used for signal sources; to tens of kW, used for material processing; and to a few MW, used in some military applications. In pulsed lasers, peak power can be much greater than in CW lasers. It can reach values as high as 1PW (1015 W). The pulse duration can vary from a ms level, typical of lasers operating in the so-called free-running regime (i.e., without any Q-switching or mode-locking elements in the cavity), to about 10 fs for some mode-locked lasers. In the following sections, the construction of the above-mentioned classes of laser will be briefly explained and their potential applications to the laser cladding process will be addressed. 3.1.1.1

Gas Lasers

Gas lasers utilize a gas or gas mixture as the active medium and may be operated in either CW or pulsed modes. Gas lasers are grouped into four categories according to the type of gas used: neutral-atom gas, ionized gas, and molecular gas. Excitation is usually achieved by applying current through the gas. Neutral-atom gas lasers employ electrically-neutral gas atoms as the active medium. The HeNe laser is the most common neutral-atom gas laser. Ion lasers contain ionized gas molecules as their active medium. The most common lasers of this group are the argon and krypton gas lasers. Some lasers, such as helium-cadmium (HeCd), include metal ions in a gas. CO2 is, by now, the most common molecular laser, but several other molecular gases are employed as well, such as CO, HE, and OF. Figure 3.3 depicts the basic construction of a CO2 laser with dierent sources of excitation: RF and DC. As seen, dierent sources of excitation can be embedded in the gas tubes. The wavelength of a CO2 laser is 10.24 µm and output power of the commercial CO2 lasers can be even more than 45 kW. The optical e!ciency of this type of laser is about 40 percent and their wall plug e!ciency is about 20 percent. These e!ciencies are strong functions of temperature. Regardless of low e!ciency, CO2 lasers have a better beam quality and focusability than other types of lasers with the same power. CO2 lasers also have the advantages of being very well absorbed by organics, glass and ceramic materials and are relatively color independent. As a result, selecting a CO2 laser is a trade-o between economical issues and the performance of the laser in dierent industrial applications. Although the high maintenance cost and low wall plug e!ciency are two restrictions for applications of CO2 lasers, this laser has been widely adopted for usage in industrial applications such as welding, cutting, cladding, and processing of glass, ceramics, and organic (e.g., polymer textile, paper, tissue material, and food) materials.

© 2005 by CRC Press LLC

Cathode

Anode

Laser Beam

Discharge

Partially Mirror Mirror

Gas In

Gas Out a)

Mirror Electrode Laser Beam Uniform Discharge

Partially Mirror

Electrode

Gas Out

Gas In b)

FIGURE 3.3 A schematic of CO2 laser with a) DC excitation, b) RF excitation.

3.1.1.2

Excimer Lasers

Excimer stands for “excited dimer”. The principle of operation of an excimer laser is a chemical reaction. The excimer laser is very often dedicated to the generation of a single wavelength. Each molecule of the active medium of an excimer laser is composed of an inert gas atom and a halogen gas atom. Among others, these include krypton fluoride (KrF), xenon fluoride (XeF), argon chloride (ArCl), argon fluoride (ArF), krypton chloride (KrCl), and xenon chloride (XeCl). The rare-gas halide (compound made from a halogen) laser, which emits in the ultraviolet wavelength (126 to 558 nm), operates on electronic transitions of molecules with repulsive ground states, until a diatomic (having two atoms within one molecule) occurs. In general, the excimer laser is generated by combination of two identical atoms or molecules, one of which is excited and the other is at a ground state. For this laser, excitation can be accomplished by E-beam or electric discharge. Figure 3.4 shows the construction of an excimer laser, including electrodes, supplying and storing electrical lines, mirrors, lenses and © 2005 by CRC Press LLC

a chamber for chemical reaction. U0

Laser Beam

FIGURE 3.4 A schematic of an excimer laser.

Typical average output powers of excimer lasers range from less than 1 W up to approximately 700 W. This is two orders of magnitude less than traditional Nd:YAG or CO2 lasers, which operate in the infrared part of the spectrum. The high intensity beam of an excimer laser is the product of pulse energy with 10 to 1000 mJ and the pulse duration of approximately around 10ns. Excimer lasers are widely used in medical technology as well as micromachining, as they provide the ultimate method for skiving, ablation, and micromachining of flex circuits, plastics, and ceramics. With these lasers, the ability to control depth in microns provides an easy and cost eective method for removing excess material, exposing leads and pads, removing oxide coatings, and providing controlled depth cavities. 3.1.1.3

Solid-State Lasers

Solid-State (SS) lasers use a solid crystalline material as the lasing medium, and are optically pumped. These lasers have lasing material distributed in a solid matrix (e.g., the ruby). Solid-state lasers use a pumping source to excite the atoms and supply energy to the crystal rods; typical pump-sources can be flash lamps or diode lasers. Figure 3.5 shows a typical construction of a solid-state laser along with pumping source. © 2005 by CRC Press LLC

The first laser, invented in 1960, was a solid-state laser [40]. It used a synthetic ruby rod (chromium-doped aluminum oxide) with mirrors on both ends (one semitransparent) pumped with a helical xenon flash lamp surrounding the rod. The lamp was similar to those used for indoor and high speed photography. The intense flash of blue-white light raised some of the chromium atoms in the matrix (the aluminum oxide is just for structure and is inert as far as the laser process is concerned) to an upper energy state from which they could participate in stimulated emissions.

Pump Radiation

Pump Radiation

Pump Cavity

Laser Beam

Pump Radiation

Partially Reflecting Mirror

Pump Radiation

Solid State Laser Rod

Mirror

Pump Cavity

FIGURE 3.5 A typical construction of a solid-state laser along with pumping source.

Modern solid-state lasers are not too dierent from the original prototype. The majority of modern solid-state lasers use neodymium (Nd) doped materials such as Nd:YAG (yttrium aluminum garnet, which is Y3 A15 O12 ), Nd:YVO4 , Nd:Glass, and others. These materials have a much lower lasing threshold than ruby as well as other desirable physical and optical properties. The strongest output wavelength of neodymium-doped lasers is approximately 1064 nm which is close to IR (infrared), and it is totally invisible. The exact wavelength of the strongest lasing lines depends on the actual host material. In addition to Nd:YAG and Nd:YVO4 at 1064 nm, there © 2005 by CRC Press LLC

are types of solid-state lasers that lase at slightly shorter wavelengths such as Nd:LSB at 1062 nm, Nd:Glass at 1060 nm, Nd:YLF at 1053 nm, and Nd:NiNbO3 (neodymium-doped lithium niobate) at 1092 nm. Other materials include holmium-doped YAG (Ho:YAG) or Ho:YLF, which provide laser light at approximately 2060 and 2100 nm, respectively. Among the above-mentioned materials used as the main crystal in solidstate lasers, Nd:YAG and Nd:YVO4 are becoming increasingly important for high-power lasers (e.g., 4 kW at 1064 nm). Solid-state lasers can be pulsed, CW, or quasi-CW. In a pulsed solid-state laser, Q-switching (Q stands for quality) is used to stabilize and boost peak power output by preventing the laser cavity from resonating (e.g., one of the mirrors is blocked or forced to be misaligned by a mechanical mechanism) until the population inversion is built up fully. CW solid-state lasers may use xenon or krypton arc lamps or other sources of intense broad spectrum light. However, the trend today is toward the use of arrays of high-power laser diodes for pumping. These can be designed to have a wavelength that matches an absorption band in neodymium (around 800 nm), making for very e!cient excitation. The diode pumped technique is rapidly taking over due to their higher e!ciency than flash one. This results in lower power consumption and heat dissipation, reduction in size, as well as an increase in reliability and decrease in maintenance. This type of laser is further discussed in the next section. Quasi-CW solid-state lasers are actually pulsed lasers but operating with a pulse repetition rate (PRR) that is high enough to appear to be continuous. 3.1.1.4

Semiconductor Lasers

Semiconductor lasers, which are also called diode lasers, are not solid-state lasers. These electronic devices are generally very small and use a low amount of power. They may be built into larger arrays such as the writing head in some laser printers or compact disc players. Figure 3.6 shows the construction of a diode laser. Some of the properties of diode lasers include wide spectrum band (2-20 nm), large beam divergence (up to 40 half-angle), non-symmetrical beam distribution (2.5-6 times dierence in beam divergence in the two orthogonal axes), and lower energy intensity per area. Diode lasers use nearly microscopic chips of gallium-arsenide or other exotic semiconductors to generate coherent light in a very small package of laser. These materials are based on semiconductors of group III-V components. The energy level dierences between the conduction and valence band electrons in these semiconductors are what provide the mechanism for laser action. Population inversion, as a result of electron transitions from the valency band to the conduction band of a doped semiconductor, is achieved by forward biasing the p-n junctions. Spontaneous emission and stimulated emission occur when electrons in the conduction band recombine with the holes in © 2005 by CRC Press LLC

the valency band. The optical cavity in a diode laser is formed by splitting two opposite facets of the semiconductor wafer to form a Farby-Perot lasing cavity [116].

Metal Contact

Active Region Voltage Bias

P N Laser Beam Heat Sink Metal Contact

FIGURE 3.6 A schematic of a diode laser.

The active element in a semi-conductor laser is a solid-state device not all that dierent from an LED. The first type of this laser was developed quite early in the history of lasers but they became widely available and more economical in the early 1980s. Today, there are various diode lasers in terms of output power. The most common types, found in popular devices like CD players and laser pointers, have a maximum output in the 3 to 5 mW range. The new generation of high-power diode laser (HPDL) can produce 4 kW. The high-power terminology is used for CW diode lasers with output power in excess of 0.5 W. Diode lasers have several disadvantages such as poor beam coherence and symmetry. These disadvantages can be overcome by a diode-pumped solidstate laser such as a diode-pumped Nd:YAG laser as well as the use of optical fiber beam delivery. Diode-pumped Nd:YAG lasers are an integration of crystal and diode pumping unit, as it was mentioned in the last section. A schematic of a typical type © 2005 by CRC Press LLC

of this laser is shown in Figure 3.7. In this laser, the p-polarized diode light, which is transmitted into the rod with low loss on the surface of the rod, is used to pump the YAG rod. The diode light can be sent to the rod through three dierent orientations. The laser crystal is mounted inside a flow-tube whose outer surface has AR (anti reflective) and HR (high reflective) coatings for the diode wavelength [117]. The diode pumped Nd:YAG is an established tool for micro-cutting applications; however, there are several disadvantages of diode pumped solid-state lasers such as low wall plug e!ciency, high running costs, and poor thermal stability.

Diode Laser (Pumping Source)

Cylindrical Lens Flow Tube

YAG Rod

High Reflective Coating

Cooling Water LD Light

FIGURE 3.7 A schematic of a diode-pumped Nd:YAG laser, which has three source of diode pumping [117].

3.1.1.5

Liquid Dye Lasers

Dye lasers are unique due to the use of liquid as the lasing medium. Depending on the particular dye used, the output laser beam can be at a wide range of wavelengths spanning the visible spectrum and beyond. Commercial dye lasers are often pumped by other lasers. For example, rhodamine-B, a common dye used in dye lasers for the red region, is often © 2005 by CRC Press LLC

pumped with an argon ion laser at 514 nm for CW operation or with a doubled YAG laser at 532 nm when pulsed. An intensive flash lamp can also be used as a pump source. Figure 3.8 shows a schematic of a liquid dye laser. The most useful feature of dye lasers is their tunability. The lasing wavelength for a given liquid may be varied over a wide range. Taking advantage of the broad fluorescent linewidths (50-100 nm) available in organic dyes, a diraction grating can be used as a wavelength-dispersive optical element in the laser cavity to perform selective tuning. Such tuning can yield extremely narrow linewidths. The hazards of dye lasers are relatively moderate. Some of the organic dye materials used in this type of laser are toxic, and a high voltage power supply (low current but a large energy storage capacitor) is required to fire the flash lamp.

Lens Pump Beam

Laser Output

Grating Dye Cell

Output Coupler

Beam Expander

FIGURE 3.8 A schematic of liquid dye laser.

3.1.1.6

Fiber Lasers

For the past decade, rare-earth-doped fibers have received widespread attention for their applications as laser sources and amplifiers. With wall-plug e!ciencies greater than 20 percent, a huge increase in the output power of fiber lasers has been reported in recent years. The new development in fiber lasers is high-power output, which works at eye-safe wavelengths. In addition, advances in ultrashort pulsed fiber lasers, based on photonic crystal or holey fibers, have opened up an entirely new set of applications in sensing, materials processing and biomedical sciences. A fiber laser for producing very short pulses is formed by placing a laser fiber in a resonant cavity. The fiber laser is formed of two dierent types of fibers, which are joined in series. They are a gain fiber, which contains the laser gain medium, and a pulse shaping fiber, which uses the phenomenon of solution pulse shaping to shorten the pulses. An initially formed pulse is recirculated many times in the resonator. On each pass, the pulse is both © 2005 by CRC Press LLC

amplified and shortened until it reaches steady-state. The zero dispersion wavelength of the pulse shaping fiber is chosen to be slightly less than the laser wavelength. The fiber is pumped by a continuous source, particularly CW laser diodes [118]. Figure 3.9 shows a schematic construction of a typical fiber laser. Medium Lens

Resonant Cavity

Pump Source

Laser Beam Fiber Optic

FIGURE 3.9 A schematic construction of a typical fiber laser.

The first 2 kW continuous-wave fiber laser has been produced and immediately used in automotive applications in 2003 [119]. The spot size of this 2-kW laser is 50 µm giving a power density of 100 MW/cm2 . The size of the unit is only 110 × 60 × 118 cm, including the power supply and air-cooling system. This new high-power laser is seen as a replacement for solid-state Nd:YAG or CO2 lasers because of the scalable power and a beam quality that is up to ten times better. Investigations show that the single-mode fiber laser is an e!cient, reliable and compact solution for micro-machining. Fiber lasers are more easily integrated into industrial processes in comparison with conventional lasers for a number of reasons: standard wall plug operation and high electrical e!ciency, no water cooling required, single mode fiber delivery line, high quality focusable beam, high repetition rate, optimized pulse duration, exceptionally high reliability, and maintenance-free operation.

3.1.2

Laser Beam Characteristics

Laser beam characteristics play an important role in laser material processing including laser cladding. There are many parameters that indicate the quality of a laser beam. Several important laser beam parameters are beam parameter © 2005 by CRC Press LLC

r

z r (z )

θ

r0l

I (r )

z0

FIGURE 3.10 Laser beam geometry.

product (BPP), laser beam mode, energy distribution over the beam spot area, polarization, and focusability. The beam parameter product (BPP) is important because it provides an indication of the focused beam size and the focal depth. It is represented by BP P =

r0l  2

(3.3)

where r0l is the beam spot radius in the waist of the laser beam and  is the far-field full divergence angle, as shown in Figure 3.10. The argument is that reducing the divergence by using a beam expander would increase the beam spot size. Based on Figure 3.10 and Equation (3.3), it can be concluded that a low divergence angle produces a smaller focused spot and greater depth of focus. The laser energy can be distributed in a uniform or Gaussian form over the laser beam spot area. However, generation of Gaussian energy intensity is easier than the uniform energy intensity. In order to achieve a good beam quality, it is necessary to resonate the beam in a resonator. In the resonator, the distribution of the amplitude and phases of the electromagnetic field can be reproduced due to the repeated reflections between mirrors [116]. These specific field shapes produced in the resonator are known as transverse electromagnetic modes (TEM) of a passive resonator. Transverse electromagnetic modes in polar coordinates, which are also called Gaussian-Laguerre modes, are demonstrated by TEMpl . The subscript p indicates the number of nodes of zero intensity transverse to the beam axis in radial direction, and the subscript l indicates the number of nodes of zero intensity transverse to the beam axis in tangential direction. The intensity distribution Ipl (r, *) of a TEMpl mode can be represented by ¸2 · 2r2 M 2 l p 2r2 M 2 2r2 M 2 2 Ipl (r, *) = I0 ( ) ( ) cos (l*) exp( ) (3.4) L l rl2 rl2 rl2 where I0 is the intensity scale factor [W/m2 ], rl is the radius of the laser beam profile, M 2 is the beam quality factor (based on the ISO 11146), and Lpl is the © 2005 by CRC Press LLC

generalized Laguerre polynomial of order p and index l [120]. The intensity scale factor I0 is expressed based on node number and the average power Pl [W] as ( 2r2 M 2 Pl l=0 r2 (3.5) I0 = 4r2 Ml 2 p! P l = 1, 2, 3, 4, ... r2 (p+l) l l

Based on dierent TEM, various beam energy intensities are available. Figure 3.11 shows several TEMs with Gaussian energy intensities. The other important parameters for a laser beam are the beam propagation factor and the quality factor. The radius of a radially symmetric laser beam varies along the propagation axis, which can be expressed by 2 + 4 2 (z  z0 )2 rl (z)2 = r0l

(3.6)

where r0l is the beam radius of the waist [m], z0 is the waist location with respect to an arbitrary coordinate along the propagation axis [m], and  is the far-field divergence angle [rad]. Figure 3.10 shows the denoted parameters. The propagation can also be described by the beam propagation factor Q, or the quality factor M 2 , which are related as M2 =

nrol  1 = Q 2

(3.7)

where  is the laser wavelength in the used medium [m], and n is the index of reflection. The propagation factor k is then defined as k=

1 2  2 M nrol 

(3.8)

If k = M 2 = 1, the beam is Gaussian; if M 2 > 1, the beam is not Gaussian. However, all of the standard Gaussian propagation formulas may be used with appropriate correction factors (see ISO 11146). In most cases, a laser application requires a laser beam with low divergence emitted in fundamental Gaussian mode (TEM00 ). This is not guaranteed for every laser and is unlikely for especially high-power laser systems because the emission may be multimode or may be changed based on the life of laser systems. As a result, the beam quality should be measured by the available measurement devices such as laser beam analyzer (LBA), slit, knife-edge, and CCD-based instrumentation [121]. The other important parameters of laser beams are reflectivity and polarization. The values of absorptivity and reflectivity are related by the following equations: ½ 1A (for opaque materials) R= (3.9) 1AT (for transparent materials) where R is reflectivity, A is absorptivity, and T is transmissivity. The reflectivity R for normal angles of incidence from air to opaque materials with perfect flat and clean surface is derived by © 2005 by CRC Press LLC

TEM

cross-section

distribution I0

M2 = 2

TEM00

I0

M2 = 2

TEM10

I0

M2 = 2

I0

M2 = 2

TEM11

TEM01

I0

M2 = 2

TEM01*

I0

TEM02

FIGURE 3.11 Several TEM modes with Gaussian energy intensity. © 2005 by CRC Press LLC

M2 = 2

TABLE 3.1

Optical properties of several materials for 1.06 micron light wavelength in room temperature. Materials k n Al 8.50 1.75 Cu 6.93 0.15 Fe 4.44 3.81 Ni 5.26 2.62 Pb 5.40 1.41 Ti 4.00 3.80 Zn 3.48 2.88 Glass 0.10 0.50

R = [(1  n)2 + k2 ]/[(1 + n)2 + k2 ]

(3.10)

where n is the refraction coe!cient and k is the extinction coe!cient of material. The absorptivity, A, of an opaque metal surface can be obtained by A = 1  R = 4n/[(n + 1)2 + k2 ]

(3.11)

Table 3.1 lists the optical properties of several materials for the light radiation with 1.06 µm wavelength. It has to be considered that the optical properties will change with temperature and light wavelength. Photons with shorter wavelengths are easier to be absorbed by the materials than photons with longer wavelengths. Therefore, R normally decreases as wavelength becomes shorter. When temperature rises, there will be an increase in the photon population. Therefore, the probability of interaction between the electrons and material increases causing a decrease in the reflectivity and an increase in the absorptivity. Of interest is the fact that the reflectivity is a function of light polarization and angle of incidence. Light can be described as an electromagnetic wave that propagates through a sinusoidal oscillation of an electric field. The direction in which the electric field oscillates as it propagates is known as the polarization. A laser is defined as “polarized”, if 90% or more of its energy is in a given polarization state such as linear, circular, or elliptical. In general, a laser pulse injects polarized electrons, whose spins have a definite orientation determined by the laser’s polarization. The desired polarization state is generated by a combination of dierent optic systems. Figure 3.12 shows examples of two polarization states: linear and circular. In circular state, the electromagnetic wave propagates as a function of time and rotates around a reference line as shown in the figure. Drude [122] showed a variation in reflectivity with both angle of incidence and plane of polarization. If the plane of polarization is in the plane of incidence, the beam is called a p-ray. If the beam has a plane of polarization © 2005 by CRC Press LLC

y y

z x

E

z x

E

E

Ey E

Ey

a)

Ex

b) Ex

FIGURE 3.12 a) Linear polarization, b) circular polarization.

which is normal angles to the plane of incidence, it is called s-ray as shown in Figure 3.13. The corresponding reflectivities for these two types of polarized beams can be obtained by Rp =

[n  (1/ cos *i )]2 + k2 [n + (1/ cos *i )]2 + k2

(3.12)

(n  cos *i )2 + k2 (n + cos *i )2 + k2

(3.13)

Rs =

where Rp is the reflectivity of a p-ray beam, Rs the reflectivity of an s-ray beam, *i is the incident angle, n is the refraction coe!cient, k is the material extinction coe!cient. In general, p-rays are more easily absorbed by materials than s-rays.

3.1.3

Types of Lasers and Laser Beam Characteristics in Laser Cladding Process

In the laser cladding process, it is essential to provide appropriate power density and interaction time between the laser beam and the material. Figure 3.14 shows the range of the power density and interaction time for various laser material processing techniques. As it is seen, the laser cladding process requires a power density from 70 to 100 W/mm2 and an interaction time of 0.01 to 1 second; any laser intended for use in the laser cladding process should provide this level of power density. In addition, the beam quality is a key factor for a successful laser cladding as © 2005 by CRC Press LLC

ϕincident

ϕ reflection

ϕtransmit a)

ϕ reflection

ϕincident

ϕtransmit b)

FIGURE 3.13 a) p-ray b) s-ray.

will be explained in the next section. The selected laser should provide the appropriate beam quality. Another important issue for any laser material processing is the light reflection from the surface of metals. The reflection is strongly a function of laser wavelength and it varies from metal to metal. Figure 3.15 shows the wavelength dependency of several metals’ reflection factor. It is also important to consider the contribution of temperature in reflectivity. As the temperature of the process zone rises, an increase in absorptivity occurs, which indicates the potential of more energy absorption by the material [101]. 3.1.3.1

Types of Lasers Used in Laser Cladding Process

As described in Section 3.1.1, there are many laser systems in the market. However, CO2 lasers, lamp-pumped Nd:YAG lasers, diode-pumped Nd:YAG lasers, and high-power diode lasers (HPDL) are most commonly used in the laser cladding process. There is no report on the use of liquid dye lasers in the laser cladding process. This laser is not widely used in laser material industry due to its low power capacity. To the best knowledge of the authors, there is no report on the use of fiber lasers in the laser cladding process. However, fiber lasers can be adapted to the process due to its high beam quality, cost eectiveness, and e!ciency in near future. There seems to be only a few reports about the application of excimer lasers to the laser cladding process. Panagopoulos et al. [123] carried out a coating of copper on mild steel by a KrF excimer laser with a wavelength of 248 nm. The power density per pulse was varied between 150 and 430 MW/cm2 and © 2005 by CRC Press LLC

Power Density (W/ mm2)

10

8

Shock Hardening

VAPORIZATION

Drilling 6

10

Glazing Cutting MELTING Ablation

10

4

Welding

Magnetic Domain Control

Melting Alloying

Bending

2

10

Transformation Hardening HEATING

1

Cladding LCVD

Stereolithography 10-8

10

-6

-4

10

10-2

1

100

Interaction Time (s)

FIGURE 3.14 Power density and interaction time for various laser material processing [101].

the pulse frequency was 10 Hz. Except for this work, research groups and industry have not utilized the excimer laser as a source of energy for the laser cladding process due to its low average power. Although the peak power of the excimer laser per pulse is high, this high power per pulse can vaporize the powder particles. Excimer lasers, on the other hand, have the potential for use in coating of micro-devices (e.g., MEMS). Table 3.2 summarizes characteristics of these four types of lasers which have been widely used in laser cladding. Both pulsed and continuous wave lasers have been used in laser cladding; however, with pulsed lasers, it is necessary to maintain the peak power of each pulse in a limited range. Pulses with high peak power energy (even those with low average power) can vaporize the powder particles prior to reaching the process zone. There are major performance dierences between Nd:YAG, HPDL, and CO2 lasers. Nd:YAG and HPDL light are emitted at wavelengths of 1.024 and 0.85 µm, respectively, which are in the near infrared, while CO2 light is emitted at 10.6 µm. The material interactions at these wavelengths dier. © 2005 by CRC Press LLC

1

Cu

Al

0.9

Reflectivity

0.8 0.7

Ni

0.6 0.5

W

0.4 0.3

Si

0.2 0.1 0

200

300

400

500

600

700

800

900

1000

7000

8000

9000

Wavelength (nm) a) Cu

1

Reflectivity

Al 0.8

Au Steel

0.6

0 1000

2000

3000

4000

5000

6000

Wavelength (nm) b)

FIGURE 3.15 Correlation of reflectivity and beam wavelength for dierent materials in two dierent wavelength ranges, a) from 200 to 1000 nm, b) from 1000 to 9000 nm. TABLE 3.2

Characteristics of common lasers used in laser cladding. Characteristics Wavelength [ µm] E!ciency [%] Maximum power [kW ] Average power density

CO2

Nd:YAG

Nd:YAG

lamp-pumped

diode-pump ed

HPDL

10.64

1.06

1.06

0.65-0.94

5-10

1-4

10-12

30-50

10

45 6...8

10

4 5...7

10

5 6...9

10

6 3...5

2 [W /cm ]

Service period [hour] Beam parameter product (BPP)[mm × mrad] Fiber coupling © 2005 by CRC Press LLC

1000-2000

200

5000-10000

5000-10000

12

25-45

12

100-1000

No

Yes

Yes

Yes

Metals are more reflective at 10 µm than at 1 µm as shown in Figure 3.15; as a result, Nd:YAG and HPDL are more e!cient than a CO2 laser for metallic processing. Aluminum is relatively highly reflective compared to the CO2 beam, whereas a beam from a Nd:YAG or HPDL laser is almost perfectly absorbed. On the other hand, most carbon steels and stainless steels absorb CO2 and Nd:YAG beams very much the same. HPDL in comparison to CO2 and Nd:YAG lasers has the shorter wavelength and thus higher absorption of the direct diode laser. Figure 3.15 shows the reflection factor as a factor of wavelength for several metals. A CO2 laser can provide a very high power such as 45 kW. Commercial Nd:YAG lasers are available with powers up to 4 kW (continuous) and pulsed Nd:YAG lasers with lower average powers (e.g., 1.5 kW) but have much higher pulse peak power. That is due to the cooling of the solid rod of solid-state lasers, which is a di!cult task [116]. As a result, the solid-state lasers have problems with high average powers. In contrast, CO2 lasers do not have a serious problem with thermal lensing; therefore, they can be fabricated in high power capacity with a good beam quality [116]. CO2 laser beams are focused to smaller spots and they are more symmetrical, which improves the clad width. A 1-kW CO2 laser can be focused to a 100 µm spot, whereas a 1-kW Nd:YAG is generally used with fiber optics for beam delivery and cannot be focused smaller than 400 µm. A HPDL laser provides a wide beam distribution and has a low beam quality. As a result, HPDL can be used only for coating in which a lower energy per area is required. HPDL lasers in today’s market cannot be used for high melting temperature materials. It is reported that a HPDL laser can be used for laser cladding; however, it is applicable to a limited number of materials and coating thicknesses [124]. Another important issue in selecting a laser is the beam delivery. It is impossible to transport the CO2 beam through a fiber optic cable due to its wavelength (i.e., 10.6 µm). As a result, the maneuverability of a motion system along with a CO2 laser is very limited. Although, a flying optic can be integrated into a CO2 laser to provide an extra degree of freedom, its usage in fabrication of complex parts with laser cladding is still limited. In addition, the flying lenses are very sensitive to powder intrusion into the moving lenses. Nd:YAG and HPDL lasers, on the other hand, can be run through a fiber optic cable and as a result, can be connected to the end eector of a robot with any degree of freedom. CO2 lasers generally produce either a dot-mode (TEM00 ) or a ring-mode (TEM01 ) beam, which can focus down to either a single point or a very tiny ring. Nd:YAG lasers can produce a multi-mode beam (i.e., TEM02 , TEM11 , TEM01 , TEM22 , etc.). The time constant for a CO2 laser is very high compared to Nd:YAG and HPDL. Therefore, a CO2 laser is not appropriate, when the power needs to be changed in a short period of time. This weakness can be overcome by the integration of a fast shutter system into the CO2 laser. Diode-pumped Nd:YAG lasers have a very impressive e!ciency. For small © 2005 by CRC Press LLC

(less than 10 W) lasers, total e!ciency is usually greater than 50 percent. Diode-pump Nd:YAG lasers with average power above 4 kW are being introduced to the market. High-power diode lasers HPDL are particularly compact and at the same time, highly e!cient. The development of lasers with an output power of over 1 kW opened a gate towards the use of diode lasers in laser cladding processes. HPDL lasers have been used for generating and repairing molds and motor parts. With HPDL, it is necessary to use a standard lens to achieve an appropriate working distance from the focus point. This distance provides enough space for the cladding modules (powder and inert gas nozzles). There is, however, a very high risk that the protective glass and the lens quickly become dirty or even damaged by the powder particles. It is also not possible to process surfaces with complex shapes (e.g., crankshafts). As a result of this shortcoming, researchers have undertaken projects to develop an appropriate lensing system. There are several claims about the higher dilution between the substrate and the clad layer when a HPDL is used [125, 126]. This is mostly due to the higher energy absorbed in the case of HPDL. However, using HPDL for laser cladding provides the user with a unique line source that produces clads with a controllable width without scanning many times over the surface. CO2 and Nd:YAG lasers have a smaller spot size such that the laser must be scanned over the coated area several times. The shorter wavelength of the HPDL allows for higher absorption into the material being coated so that a higher process speed can be achieved. Figure 3.16 shows typical cross sections of clad and the substrate region (Stellite 6 on steel) performed by HPDL and CO2 lasers [125]. The wavelength of HPDL laser beam was 0.94 µm. The experiments for both cases performed at process speed of 900 mm/min, and the maximum clad rate was 0.5 kg/hr. The diameter of CO2 laser beam was 4.7 mm on the substrate, and the diode beam cross section on the substrate was 4.5 × 4 mm×mm. The laser average power in CO2 experiment was 3900 W and in the HPDL was 1400 W. As seen in the figure, the cross section of the clad region for HPDL case is 1.9 mm2 , and for CO2 laser is 2.1 mm2 . The microstructures of both samples show a fine-dendritic structure, which are metallurgically bonded to the substrate. The power of CO2 laser should be set to 3.9 kW to produce the same clad as produced by the 1.4 kW power of HPDL. Table 3.3 lists the types of lasers currently used by researchers/organizations involved in the laser cladding process. 3.1.3.2

Laser Beam Characteristics in Laser Cladding

Since the processing zone in laser cladding is usually positioned below the focal point, a larger distance between the optical system and the workpiece is available, which facilitates the protection of the optical system. In general, in laser cladding, it is preferred to have a larger focal distance due to reduction of the sensitivity of the spot dimensions to the changes in beam characteristics © 2005 by CRC Press LLC

TABLE 3.3

Laser types of organizations/research groups involved in the laser cladding process. Organization Type of Laser Application Material Ref. used Fraunhofer-Institut fur Lasertechnik, Germany GE Aircraft Engines, USA

HPDL, CO2

coating, prototyping

Co and Febased alloy, SS 304

[125, 127]

CO2

Ni alloy

[128]

Laser X. Co., Japan

CO2

repairing of engine turbine blades coating

[110]

University of Waterloo, Canada

Nd:YAG (pulsed) lamp pumped, HPDL Nd:YAG(CW) lamp pumped

prototyping, coating

Cr-Ni based materials, Stellite 6 Fe-Al, H13

coating

SS 304

[130]

Nd:YAG(pulsed and CW) lamp pumped CO2

prototyping

SS 316L, IN625

[19, 21]

prototyping, blade repairing

SS, Stellite 6, superalloy, CMSX-4

[103, 36].

Nd:YAG (CW) lamp pumped CO2

prototyping

[28]

prototyping

Inconel alloy 690 Al, H13

CO2

coating

Co, Al, Ni

CO2

coating

[77]

CO2 Nd:YAG (CW) diode pumped

coating prototyping

CO2

low volume manufacturing prototyping, coating,

Ni-Al-Cr-Hf on Ni Stellite 6 SS 316, SS 304L, H13, IN718,IN 600 Ti, Ti-6Al-4V, Ti-5Al-2.5Sn H13, Cobased material H13, Ti-based alloy, Copper Ti- Ni, Ceramics

[82]

Ishikawajima-Harima heavy Industries Co., Japan National Research Council of Canada Swiss Federal Institute of Technology, Switzerland Los Alamos National laboratory, USA University of Illinois, USA DRL Institute, Germany Illinois University, USA Westinghouse, USA Sandia National Laboratories, USA AeroMet tion, USA

Corpora-

POM Inc., USA

CO2

University of Michigan, USA South Dakota school of Mines & Technology, USA

CO2 , Nd:YAG diode pumped Nd: YAG (CW) lamp pumped

© 2005 by CRC Press LLC

prototyping, coating prototyping, coating

[129, 68]

[64, 14] [131]

[55] [13]

[34]

[14] [132]

HPDL Laser

CO2 Laser

a)

b)

FIGURE 3.16 Cross section of a single track at process speed of 900 mm/min using dierent laser sources, a) HPDL laser at 1400 W, b) CO2 laser at 3900 W (Source: Courtesy of Fraunhofer Institute for Material and Beam Technology, Germany [125]).

and also reduction of the peak power intensity in the spot point, which can cause plasma formation [108]. In pre-placed laser cladding, a circular laser spot with uniform power distribution seems to be more suitable than a Gaussian beam. The main reason for this is the need to transfer a homogeneous energy on the pre-placed powder layer. If a Gaussian beam were used, it would cause non-homogeneous distribution of energy, which may cause plasma formation or even an unexpected clad width. In laser cladding by powder injection, a Gaussian beam may result in better bead quality, dilution, and homogeneity over the clad microstructure. In the case of lateral nozzle, the powder particles usually have a Gaussian distribution, which is compatible with laser power distribution as shown in Figure 3.17. A TEM00 has been used in many reports dealing with laser cladding [133, 36, 134]. However, Schneider [135] claimed that a laser spot perpendicular to the direction of cladding with a homogeneous distribution provides a uniform temperature distribution over the melt pool. Also, Weerasinghe et al. [44] used a TEM01 mode beam in their experiments and they arrived at a uniform heating eect. There may be cases in which the other mode shapes are preferred, especially when producing a thin wall clad. This line shape beam laser can be generated using two cylindrical mirrors or a segmented mirror [107]. Frenk et al. [136] showed that cladding using far-infrared radiation (e.g., CO2 ) should be done with linearly polarized beams at angles of some 70 to 80 degree. In this way, the transmitted energy can be improved by a factor of 3 to 4. © 2005 by CRC Press LLC

Laser Beam

Powder Stream

Substrate

FIGURE 3.17 A Gaussian laser energy distribution versus a Gaussian powder particles distribution.

In laser cladding, it is also possible to use a rectangular spot with a uniform power intensity, as generated by a diode laser [125]. Such a spot can also be generated using a two-dimensional beam integrator. An alternative to the use of integrating optics for achieving a uniform temperature profile over the width of the track is the use of scanning optics. High-power lasers can become instable when run for a long period in the cladding process. The laser’s properties can be influenced by the process itself, causing the process to fail or be unsatisfactory. Therefore, monitoring and control of laser beam parameters is an important task in the laser cladding processes [137, 138].

3.2

Powder Feeders and Powder Delivery Nozzles

Powder feeders are among the most important pieces of equipment in a number of industrial applications involving powder conveyance, such as thermal spraying, laser cladding and advanced materials processing. As dierent powders have very distinct sizes, shapes, and other physical and mechanical properties, it is nearly impossible to convey each type of them with a steady-state flow using a single feeder machine. With decreasing powder grain size (e.g., ultra-fine powder with size of less than 15 µm), the flowability of the powder is decreased, which causes problems in the powder transporting. Flowability also dramatically decreases with sticky and cohesive powders. Void factor (i.e., ratio of the space of air to that of solid) also plays an important role © 2005 by CRC Press LLC

in flowability of powder. For these reasons, dierent powder feeders are required for each type of powder. For example, the required powder feed rate for thermal spraying can be relatively large, whereas the required powder feed rate for prototyping by laser cladding is relatively small. Therefore, a powder feeder machine needs to be carefully controlled in order to ensure that a stable powder stream with a desired feed rate is generated. Selection of a suitable powder feeder is a vital factor for a successful laser cladding process. A powder feeder should provide a continuous and uniform powder stream with high accuracy in terms of flow rate at a desired feed rate. It is crucial to control the feed rate in real-time with minimum time constant. Also, in a laser cladding process, particular attention has been given to minimizing pulsations and agglomerations in the powder stream. Unfortunately, the current powder feeders in the market cannot provide a low time constant (e.g., 0.5 second) and low powder feed rate at high precision (e.g., 0.1 g/min), which are two important parameters in the laser cladding technology. For this reason, special powder feeders with dierent control strategies have been designed and introduced [139, 140, 141, 142]. Also, researchers are developing feeders for ultra-fine powders to arrive at a continuous stream with low feed rate. These powder feeders are vibrationbased or pressure-assisted feeders, which can even be used in direct-write deposition [143].

3.2.1

Powder Feeder Types

There are many types of powder feeders used in industry. In general, powder feeders can be categorized into the following classes based on dierent principles of operation: • Gravity-based • Mechanical wheel • Fluidized-bed • Vibrating In some powder feeders, a combination of the above methods is used to arrive at a better stability in the powder stream. In all types of powder feeders, a carrier gas should be supplied to transport the powder particles from the starting point to the desired location. A brief explanation of the above powder feeders is provided in the following sections. 3.2.1.1

Gravity-Based Powder Feeder

The principle of operation of gravity-based powder feeders is similar to a simple sand clock. The powder feeder machine essentially consists of a load cell based electronic weighing mechanism and an orifice. Due to the weight, © 2005 by CRC Press LLC

the material flows from hopper to the orifice if the powder particles have the required flowability. By reducing or increasing the area of the orifice, the amount of powder delivered to the nozzle decreases or increases. Figure 3.18 shows a schematic of a gravity-based powder feeder.

Powder Container

g

FIGURE 3.18 A schematic of gravity-based powder feeder.

In order to increase the controllability of gravity-based powder feeders, different devices such as a metering wheel can be integrated into the powder feeder. Also, a back pressure can be supplied on the powder funnel to increase the stability of the powder stream, which can be aected by the change in the height of powders in the funnel. Adding the external component for the measurement of powder is an essential device for obtaining a feed rate with high precision. One of these devices can be a rotating disk with holes around it as shown in Figure 3.19. The feeder machine consists of a powder container from which powder flows by gravity into a slot on a rotating disk. The powder is transported to a suction unit by a gas stream. The dimensions of the slot and the speed of the disk control the volumetric powder feed rate [139]. The other idea for integration of a metering wheel into a gravity-based powder feeder is shown in Figure 3.20. The size of holes around the rotating shaft and the angular velocity of the shaft determine the powder feed rate. The other design can be an integration of a lobe gear with the gravitybased powder feeder as shown in Figure 3.21. This design is not suitable for an application requiring the low powder feed rate.

© 2005 by CRC Press LLC

Powder Container

Gas

Rotating Powder Slot To Powder Inlet

To Powder Nozzle

Powder Pick-up

FIGURE 3.19 A typical gravity-based powder feeder with a rotating wheel for metering.

Back Pressure

Gas

Metering Wheel

FIGURE 3.20 A typical gravity-based powder feeder with a metering wheel. © 2005 by CRC Press LLC

Powder Container

Gas

Metering Wheel

FIGURE 3.21 A typical gravity-based powder feeder with a lobe gear.

3.2.1.2

Mechanical Wheel Powder Feeder

Mechanical wheel powder feeders are also known as screw powder feeders. Mechanical wheel feeders handle a wide range of powders with dierent mesh sizes. They do not seal against an uncontrolled flow of fine powders and normally operate with zero or low-pressure dierential between outlet and inlet. A typical mechanical wheel feeder has a pitch with dierent diameter ratio or a rotor which can grab powder particles from the storage area. There are many screw configurations that can be used to promote uniform flow with dierent feed rates. Figure 3.22 depicts two types of the configuration of mechanical wheel powder feeders. One disadvantage of this type of powder feeder is the interaction of moving parts and abrasive powder particles, which cause rapid wear in the wheel. This can result in variations in coating quality and also increase maintenance costs. 3.2.1.3

Fluidized-Bed Powder Feeder

A fluidized powder feeder operates based on fluidics principle, which eliminates the need for mechanically moving parts to deliver powder. The fluidics © 2005 by CRC Press LLC

Powder Container

Wheel

FIGURE 3.22 A schematic of mechanical wheel powder feeder.

powder feed delivery principle provides a continuous, non-pulsating feed of powder, thereby insuring the user optimum process control and improved coating quality. Another benefit is reduced maintenance and replacement part cost. The system is designed so that a predetermined quantity of gas is delivered to a closed hopper containing powder. The hopper is constructed so that the gas is passed through a filter located at the bottom of the unit, where it is diused through the powder, causing the powder to enter into the gas and therefore become fluidized. A powder pickup tube is positioned above the fluidizing gas inlet allowing the fluidized media to be delivered under a shed on the pickup tube through a number of controlled apertures and then to a carrier area where it is propelled by the carrier gas to the feed hose. Figure 3.23 shows the construction of a fluidized-bed powder feeder. 3.2.1.4

Vibratory-Based Powder Feeder

A vibratory feeder, which is also called a vibratory tray feeder or oscillating feeder, consists of a shallow flat-bottomed tray. As powder flows from the hopper outlet onto the tray, an external drive vibrates the tray, throwing the powder down to control the powder feed rate into the process. A vibratorybased powder feeder can feed most powders from at least 8 g/min to 2000 g/min with ±1% precision. In order to increase the precision, the vibratory powder feeder can consist of a vibrating tray with a number of plates set on a specified angle. Having these plates, the flowing of powder bulk can be controlled. Figure 3.24 shows the construction of a vibratory powder feeder. © 2005 by CRC Press LLC

Back Pressure

Gas In Carrier Gas

FIGURE 3.23 A typical fluidized-bed powder feeder.

Feeder Vibrating

FIGURE 3.24 A schematic of a vibratory-based powder feeder.

© 2005 by CRC Press LLC

TABLE 3.4

Powder feeder types of organizations/research groups involved in the laser cladding process. Organization Type of Type of Used Ref. applicapowder material tions feeder University of Waterloo, Canada University of Waterloo, Canada

Coating

Fluidized bed

Prototyping

NRC, Canada

Prototyping

Gravity-based powder feeder along with metering wheel Fluidized bed

University of Michigan, USA

Prototyping

Sandia, USA

Prototyping

University of Missouri at Rolla, USA

Prototyping

University of Liverpool, UK

Coating

3.2.2

Gravity-based powder feeder along with metering wheel Fluidized bed Mechanical wheel powder feeder Mechanical wheel powder feeder

H13, IronAluminide H13, Ni-based alloys

[3]

316 SS, IN625 H13, Ti-based alloys

[19]

Ti-based alloys, SS H13

—

Al, 316 SS

[144]

—

[14]

[141]

Applications of Powder Feeders to Laser Cladding

So far, dierent types of powder feeders have been used in the laser cladding process. However, it is hard to say which type of powder feeder is more suitable for this process. Due to the wide range of applications of laser cladding, dierent powders, with dierent mesh sizes at various powder feed rates are required for the process. Many research groups, which are developing the laser cladding apparatus, have designed and manufactured their own powder feeder which suits their applications. As it was mentioned, it is impossible to convey every powder with a steadystate flow using a single feeder machine. As a result, various types of powder feeders have been developed for laser cladding to provide the smooth and steady flow in the required flow rates. Table 3.4 lists several types of powder feeders, which are being used in laser cladding by dierent research groups and organizations. © 2005 by CRC Press LLC

3.2.3

Nozzles

In laser cladding by powder injection, the powder delivery nozzle can have dierent configurations as • Coaxial • Lateral Basic layouts of these two nozzles are shown in Figure 3.25. The coaxial supply of powder can be integrated with the optical system [145, 146]. One of the advantages of a coaxial nozzle is its independence from the direction of motion; however, experimental work has shown that its powder e!ciency, which is the ratio between the deposited powder on the substrate and the delivered powder by the powder feeder in a specified period, is significantly less than that of the lateral nozzle [145]. In both types, the powder can be preheated when it passes through the nozzle to increase e!ciency. Several forms of nozzles have been invented based on the above two mentioned nozzles. Islam et al. [20] invented a multiple nozzle processing head for manufacturing and repairing of turbine blades or compressor components. Jeantette et al. [139] invented a coaxial nozzle which is used for producing complex shapes. Their developed nozzle has been licensed to Sandia Corporation. Keicher et al. [147] invented a multiple beam and nozzle system to increase the deposition rate. Their developed nozzle and laser processing head has been currently licensed to Optomec Design Company. The interactions of powder particles, the laser beam and the inert gas with the melt pool are important parameters for arriving at a good quality clad. The interactions of powder particles with dierent surfaces in the process zone may result in dierent impact phenomena as [148] • Solid particles to solid surface impact causing a ricochet • Solid particles or liquid particles to liquid surface of melt pool causing catchment • Liquid particles to solid surfaces causing catchment The adhesion behavior of powder particles on solid or liquid surfaces surrounded by turbulent streams have been carried out by Zimon [149]. The type of nozzle, the angle of powder stream with respect to a reference line, the powder profile in the process zone, and powder stream diameter in the melt pool area will influence the interaction of powder particles with surfaces. An appropriate nozzle is the one that provides the minimum solid particles with solid surfaces. Minimizing impact between the solid particles and solid surfaces increases the powder catchment e!ciency. © 2005 by CRC Press LLC

Laser Beam

Shield Gas

Laser Beam

Shield Gas

Shield Gas

Powder Flow

Shield Gas

Lateral Nozzle

Shaping Gas

Clad Bead

Substrate

Clad Bead

a)

Substrate b)

FIGURE 3.25 a) Coaxial nozzle, b) lateral nozzle.

3.2.3.1

Lateral Nozzle and Powder Profile Quality

The powder delivery system plays an important role in the clad quality. Regardless of the type of the nozzle, knowledge of the intersection of powder stream and laser beam, diameter of powder stream on the workpiece, stability of powder feed rate, homogeneous shape of powder profile and velocity of powder particles are crucial to a successful process. In order to address the eect of nozzle diameters on the above-mentioned parameters, a simple measurement test rig was developed to take pictures of the powder stream. The images were then processed to find the profile characteristics of the powder stream in terms of the distance from the tip of the nozzle. It was found that the profile of the powder stream can be approximated by a parabolic equation as d = z 2 + d0

(3.14)

where d is the profile diameter at any z [mm],  is the powder profile quality coe!cient [1/mm], z is the distance of desired point from the nozzle tip [mm], and d0 is the nozzle diameter [mm]. In the study, several nozzles of PRAXAIR with serial number of TWEP2250 with diameters of 0.8, 1, 1.2, 1.4 and 1.8 mm were used. Figure 3.26 shows the identified parameter corresponding to each nozzle. As it is seen in the table,  is valid for a specific range, which represents the range of stable powder stream. © 2005 by CRC Press LLC

d0 d = profile diameter at any z (mm) d0 = nozzle diameter (mm) z = distance from the nozzle tip (mm)

z d

Powder Shield gas d 0 = 0.8 mm Feed rate Feed rate Valid λ 3 (1/mm) for (g/min) (m /s)

d 0 = 1 mm

λ

d 0 = 1.2 mm

λ

d 0 = 1.4 mm d 0 = 1.8 mm

λ

λ

(1/mm)

Valid for

(1/mm)

Valid for

(1/mm)

Valid for

(1/mm)

Valid for

1

1.56e-5

9.03e-3

z <21

1.27e-2

z <24

1.29e-2

z <24

1.62e-2

z <29

2.39e-2

z <32

1

2.34e-5

1.25e-2

z <19

1.45e-2

z <21

2.02e-2

z <21

2.05e-2

z <25

2.42e-2

z <28

1

3.12e-5

1.30e-2

z <18

1.98e-2

z <20

2.06e-2

z <20

2.17e-2

z <23

2.47e-2

z <26

1

3.9e-5

1.42e-2

z <15

2

1.56e-5

1.04e-2

z <21

1.80e-2

z <25

1.92e-2

z <25

1.93e-2

z <30

2.16e-2

z <30

2

2.34e-5

1.20e-2

z <18

2.17e-2

z <20

2.41e-2

z <20

2.51e-2

z <29

2.23e-2

z <29

2

3.12e-5

1.24e-2

z <17

2.27e-2

z <19

2.45e-2

z <19

2.57e-2

z <25

2.42e-2

z <25

2

3.9e-5

2.46e-2

z <18

2.59e-2

z <23

2.58e-2

z <23

FIGURE 3.26 Powder feed profile characteristics/dimensions.

Figure 3.27 shows a typical powder stream for 1 mm nozzle. As seen, the powder stream is expanded due to the interaction between the air resistance and powder flow and finally is spread out in the space. 3.2.3.2

Coaxial Nozzle

The invention of the coaxial nozzle had a great impact in laser cladding technology. The first important feature of this nozzle is the path independence. Also, better powder e!ciency can be obtained by this nozzle. In order to build the clad with accurate dimensions and high e!ciency of the powder deposition in a coaxial laser cladding, it is essential to analyze the powder flow structure. In the coaxial nozzle, three streams are involved, which impinge on the solid surface. These are shield gas, powder stream and shaping gas, as shown in Figure 3.25. © 2005 by CRC Press LLC

1 mm

FIGURE 3.27 A typical view of a powder stream when nozzle diameter is 1 mm, powder feed rate is 2 g/min, and shield gas feed rate is 2.34e  5 m3 /s.

A powder stream is formed around the laser beam, which has to be distributed homogeneously. A homogeneous distribution of powder particles is a key factor in the formation of a good quality clad. It is essential to hold the powder stream in a laminar flow, parallel to the laser beam profile, to arrive at a homogeneous powder stream in the outlet of the nozzle [85]. To provide good powder e!ciency and high quality clad, the powder stream’s focus point must be at the level of the melt pool [85]. The role of the tip of the nozzle in formation of laminar or quasi-laminar flow of powder stream is also important. As seen in Figure 3.28, several nozzles with dierent tips are available, which provide dierent powder stream profiles and powder focus points. Figure 3.29 shows the powder stream in exit of the coaxial nozzle connected to the end-eector of a robot. As seen, the powder profile has a homogeneous shape for a distance, and then spread out in the space.

3.3

Positioning Devices

Laser cladding can be used for a variety of applications, such as rapid prototyping, coating and repairing. For all of these applications, a solid model of the desired part is required. This model should be then sliced by a CAM soft© 2005 by CRC Press LLC

FIGURE 3.28 Dierent coaxial nozzles with dierent tips (Source: Courtesy of Fraunhofer Institute for Material and Beam Technology, Germany [85]).

FIGURE 3.29 Coaxial nozzle connected to a robot end-eector (Source: Courtesy of Fraunhofer Institute for Material and Beam Technology, Germany [85]). © 2005 by CRC Press LLC

ware based on the trajectory planning algorithm to identify the required paths for fabrication or coating of the desired part. A positioning device should not only provide enough workspace and maneuverability for the required paths, but also provide the appropriate velocity, which is a major process parameter. The positioning device can be a CNC table or a multi-DOF robot manipulator. In both cases, the dynamics constraints should be carefully considered to arrive at desired velocities and accelerations in dierent points of trajectory. Also, in the case of a robot, the singularity of the robot’s joints is another main constraint for developing the desired trajectory. The singularity will happen if one or more joints no longer represent independent controlling variables, causing a limitation on the workspace of the robot. As a result, fabrication of a complex shape using laser cladding, where a robot is selected as a positioning device, strongly depends on the DOF and kinematics of the selected robot. In general, the number degrees of freedom plays an important role in a positioning device used in laser cladding. In fabrication of complex shapes, it is necessary to deposit the layer in a non-planar fashion. Having said that, the higher degree of freedom provides the capability of non-planar motion. Process speed of a laser cladding process varies from 1 mm/s to about 20 mm/s based on the laser power and powder feed rate available. As a result, the dynamic of any positioning device selected for the process should meet the desired range of velocity. In fabrication of parts, the paths trajectories are extracted through a CAD/CAM software. This program provides a sequence of discrete points representing the desired paths, which are then fed into the motion controller. As a result, a positioning device should be controlled through a point-to-point (PTP) control mode, as long as the speed is also tunable at any time during the motion. This method of control can encounter an oscillatory motion. To overcome the oscillatory motion, many dierent motions’ algorithms have been developed. For instance, a trapezoidal velocity profile is used in a high e!cient algorithm. The slope of such a curve at the initial and final ramp are the maximum acceleration and deceleration. The top level of the trapezoid is the maximum velocity. This kind of profile gives a continuous acceleration, but there might exist jerks in the motion [150]. The algorithm, which is based on trapezoid velocity, is called PVT (position, velocity, and time). This algorithm is based on a second-degree polynomial of trajectory over the time. Figure 3.30 shows such an example with the desirable velocities at dierent positions which are depicted by the slope of the arrows. The dashed line provides a feasible trajectory that satisfies these position-velocity constraints. As seen in the figure, a trajectory can be simply described by expressing the position and velocity as a function of time. However, in the laser cladding process, velocity is an independent variable of time. This issue arises from the fact that the process speed should be changed during the laser cladding process to overcome the influence of disturbances in the process. The changes in the process speed should not, however, aect the pre-specified trajectory. © 2005 by CRC Press LLC

FIGURE 3.30 A typical position-velocity trajectory.

This type of motion results in infinite solutions. As a result, a method of approximation needs to be taken. A possible approximation or constraint on the motion system is to have a constant acceleration between consecutive points. Using piecewise interpolation, a feasible trajectory can be given in terms of series of second-degree polynomials such as p1 (t), p2 (t), p3 (t), etc. Assuming the terms are in a second degree of polynomial, the trajectory can be represented as (3.15) p(t) = at2 + bt + c Therefore, the velocity will turn into v(t) = 2at + b

(3.16)

where a, b, and c are constant coe!cients. Having the positions and velocities at initial time of t0 and t1 (which represents two successive points) provides four equations, which are enough to obtain four unknowns a, b, c,and t1 . Performing the above analysis iteratively for all position-velocity points results in a trajectory with a smooth velocity.

3.3.1

CAD/CAM System for Trajectory Generation

Like other rapid-prototyping processes, laser cladding is a way to fabricate a three-dimensional object designed with a computer aided design (CAD) software. Most existing prototyping processes are adapted into slicing technology, in which a CAD model is represented by a stack of flat and thin layers. © 2005 by CRC Press LLC

In many layered manufacturing methods, such as stereolithography, all points on each layer are deposited in a one-step growing procedure; however, in prototyping by laser cladding, each sliced layer should be resliced into paths, which represent the trajectory of the end-eector. This re-slicing is due to the nature of laser cladding for producing layers. In fact, by overlapping the clad’s tracks, a thin and flat layer of the final part can be produced. 3.3.1.1

CAD Formats

Several CAD model formats are available for developing 3D objects in a computer. The most popular formats are IGES, SAT, DXF, STP and STL. IGES, which stands for “initial exchange specification” is an ANSI standard that defines neutral file format for the exchange of CAD drawings or models in dierent CAD programs. Due to its comprehensive structure, it is not commonly used by low-end CAD software. Instead, it is most common among expensive high-end CAD software [151]. IGES does not support solid models. Although IGES files can be relatively large, they are very well compressible. SAT, which stands for “standard ACIS text” is also called ACIS. It has been supported by many solid model programs. SAT can store modeling information in external files. These files have an open format such that external applications, even those not based on ACIS, can have access to the ACIS geometric model, which is one of the most important advantages of this format. SAT files are ASCII text files that may be viewed with a simple text editor. A SAT file contains carriage returns, white space and other formatting that makes it readable to human eyes [151]. As with other solids translators, the SAT translator does not include the history tree, which is used to create the solid model. DXF, which stands for “drawing exchange format”, is one of the properties R ° of AutoCAD company. DXF is probably one of the most widely supported vector formats in the world today. DXF files are relatively easy to parse since they are tagged and text based, and therefore human readable. Tagged data means that each data element in the file is preceded by an integer number that is called a group code. A group code’s value indicates what type of data element follows. This value also indicates the meaning of a data element for a given object (or record) type. Virtually all user-specified information in a drawing file can be represented in the DXF format [151]. STP, which stands for “STEP”, is an emerging format. This format, not only includes the drawing points, facet and geometry, but also a large amount of machining information, including the model, material and tool information. Although this format is anticipated to become more widespread in the future, it is not currently used as widely as other formats, such as IGES. In addition, most packages that currently output STP files do not output all of the associated information that STP is intended to support [151]. STL, which stands for “stereolithography”, is an ASCII or binary file used in manufacturing. It is a list of triangular surfaces that describe a computer© 2005 by CRC Press LLC

FIGURE 3.31 A typical STL file, in which the density of triangle facets change according to the geometry (STL file is generated by MATLAB).

generated solid model. It is composed of triangular facets of data that represent 2D and 3D shapes. This is the standard input for most rapid prototyping machines. Although STL files represent 3D shapes, they do not actually contain the surface or the solid model data. In addition, STL files can be very large if a small tolerance is required [151]. Among these mentioned formats, STL format has been widely adapted with layered manufacturing. An STL file consists of a list of facet data. Each facet is uniquely identified by a unit normal (a line perpendicular to the triangle and with a length of 1.0) and by three vertices (corners). The normal and each vertex are specified by three coordinates each, so there is a total of 12 numbers stored for each facet. More technical details of the STL file format have been given in [152]. Figure 3.31 shows a typical STL file, which the density of triangle facets changes according to the geometry. 3.3.1.2

Slicing Technology

Slicing technology is formed to serve the needs of layered manufacturing technologies. Several trajectory planing algorithms have been developed to produce the required paths. In many of them, regardless of the part geometry or process used to produce the part, the slice thickness remains constant. Under this strategy, a given solid model is sliced horizontally into a set of planar layers, and these planar layers are then built one at a time from bottom up. In this case, planning the build sequence of a given model is no more than © 2005 by CRC Press LLC

a)

b)

c)

FIGURE 3.32 a) Desired solid model, b) layered part, c) layering error.

listing the sliced layers along the build direction, which is straightforward and can be automated easily. However, several issues limit the application of this technology. First, parts built by this technology generally show stair steps on the layer-to-layer boundaries [153]. Figures 3.32 shows the layering errors due to the regular slicing technology. This lack of continuity on the part surfaces is undesirable especially for parts made of ceramic materials, because each stair step can serve as a crack initiation point. In addition, the mechanical properties of parts will be aected by the bonding between layers. It has been reported that reducing the number of layers by increasing the slicing interval can improve the mechanical properties. However, the surface polishing of the fabricated parts may become worse as the interval between layers becomes larger. As a result, a trade-o has to be made case by case [32, 154]. Slicing the layers in a fixed thickness fashion does not usually conform to the part geometry, therefore, the slice thickness may be changed manually, which increases the production time and sacrifices the part quality. As a result, researchers have developed algorithms that allow the slice thickness to be adapted as a function of both the CAD model geometry and the process used to produce the part. This method is called adaptive slicing technology. By incorporating the geometry (the slope and curvature) of the part with the characteristics of the forming process, the optimum slice thickness is automatically selected. By merging these characteristics, the optimum slice thickness is automatically selected without human intervention. In using the algorithm, the surface texture can be controlled, production time decreases, the part quality increases, and the slicing process becomes more e!cient [155]. In laser cladding, the deposition path generation is dependent on the nature of the deposition process. Deposition of paths in each layer which have fine and homogeneous distributed clad are somewhat independent of geometric complexity. Also, the properties of the deposited material are influenced by the deposition path trajectory. Thus it is important to develop an appropriate path planning to reduce the path eect on the mechanical and metallurgical characteristics of the parts. Also, one of the factors that limits the quality of parts in laser cladding is the accumulation of residual thermal stresses. It © 2005 by CRC Press LLC

might be possible to reduce this eect by selecting an appropriate deposition path. In general, several factors will aect the type of deposition path, which are: • Thermal stress: Thermal stress accumulates if there is an unbalanced heating in the surface. Therefore, if the deposition path is not symmetrical, it will result in thermal stress. • Number of clad’s tracks: If there are many discontinuous deposition tracks, the laser shutter should be switched on and o between dierent tracks. It is desired to reduce the number of activation of shutter in the process. • Robustness: The deposition paths should be robust enough to be generated irrespective of the complexity of the geometry. • Non-planar material deposition: Redepositing the same points in the path should be avoided as this will result in extra material deposition causing the associated reheating. Therefore, a non-planar coating may overcome this shortcoming especially for fabrication of curved shape. • No gaps in the path: Having gaps will result in low quality parts. As a result, the deposition path should be continuous as much as possible. The above items make the deposition path generation a more constrained problem compared to regular cutting manufacturing and therefore alerts a dierent approach. The deposition paths are generated for the 2-D cross section obtained by the XY projection of the object. These cross sections have to be set by a certain distance before generating the deposition paths. The deposition paths are classified into two classes: • Zigzag paths: The path tracks correspond to back and forth motions in a fixed direction within the boundary of the 2-D cross section (see Figure 3.33a). • Spiral paths: The path comprises of a series of contours that are parallel to the boundary of the 2-D cross section (see Figure 3.33b). It is possible to arrive at a variety of deposition path patterns from these two classes. Some research groups have studied the influence of these two classes of paths in the microstructure of the final product. Hua et al. [16] showed that the pore/void level can be well suppressed by proper design of deposition tool-path. Figure 3.34 shows dierent tool trajectories which are used to fabricate a cubic components. They showed that the level of porosity/void may be increased especially in the case of zigzag-xy tool-path. That is due to irregular surfaces caused by clad track. © 2005 by CRC Press LLC

a)

b)

FIGURE 3.33 Dierent path patterns for fabrication of a logo (UW): a) zigzag paths, b) spiral paths.

FIGURE 3.34 Dierent tool trajectory patterns used for fabrication of a cubic component (Source: Courtesy of Professor J. Choi, Department of Mechanical and Aerospace Engineering and Engineering Mechanics, University of Missouri at Rolla).

© 2005 by CRC Press LLC

4 Laser Cladding Process Modeling

This chapter addresses the physics of the process along with several modeling techniques applied to the laser cladding by powder injection. Developed models assist us to improve and understand the underlying process and theory. These models can be used in the process prediction as well as designing a controller without performing any experiments. An accurate model is also important to reduce the cost of system development in an automated laser cladding process.

4.1

Physics of the Process

Figures 2.1 and 4.1 show the physical phenomena occurring in a laser cladding by powder injection process. The process can be sequentially listed as follows:

• The laser beam reaches the substrate and a significant part of its energy is directly absorbed by the substrate. A small part of laser energy is absorbed by powder particles. The energy absorbed by the substrate then develops a melt pool. The melted particles are simultaneously added into the melt pool (see Figure 2.1). This step of the process is expressed only by the heat conduction equation.

• Surface tension gradient drives the fluid flow within the melt pool. As far as the flow field penetrates in the substrate, the energy transfer mechanism changes to a mass convection mechanism. During this phenomenon, the melted powder particles are mixed rapidly in the melt pool (see Figure 4.1). This step of the process should be expressed by the momentum, the heat transfer, and continuity equations.

Based on these physical phenomena, three appropriate governing equations are heat conduction, continuity, and momentum. © 2005 by CRC Press LLC

Mushy Region

Solidification interface

Solidified Clad

Powder Stream

HAZ (Heat Affected Zone)

FIGURE 4.1 Schematic of convection influence during laser cladding process.

4.2

Governing Equations

For a laser cladding process, a moving laser beam with a general distribution intensity strikes on the substrate at t = 0 as shown in Figure 4.2. Due to additive material, the clad forms on the substrate as shown in the figure. The transient temperature distribution T (x, y, z, t) is obtained from the threedimensional heat conduction in the substrate as [156]: C(cp T ) + u · (cp UT )  u · (KuT ) = Q (4.1) Ct where Q is power generation per unit volume of the substrate [W/m3 ], K is thermal conductivity [W/m·K], cp is specific heat capacity [J/kg·K],  is density [kg/m3 ], t is time [s], and U is the travel velocity of the workpiece (process speed) [m/s]. In the laser cladding process, the conservation of momentum is one of the important governing laws. The equation of momentum is Newton’s second law applied to fluid flow, which yields a vector equation. The momentum equation is represented as C(U) + (Uu)U = g  up + µu · (uU) (4.2) Ct where g is gravity field [m/s2 ], µ is viscosity [kg/s·m], and p is pressure [N/m2 ]. The last equation is continuity and is represented by uU = 0

(4.3)

The above equations may be solved analytically (in special cases) or numerically along with required assumptions and/or simplifications. © 2005 by CRC Press LLC

FIGURE 4.2 Schematic of the associated physical domains of the laser cladding process.

4.2.1

Essential Boundary Conditions

For the laser cladding process, a set of complicated boundary conditions should be satisfied. However, a set of important boundary conditions are as follows: • The eect of the laser beam and the powder flux can be modeled as a surface heat source and heat flux, defined by the boundary condition as ½

I(x, y, z, t)  hc (T  T0 )  ²t (T 4  T04 ) if 5  if 5 / hc (T  T0 )  ²t (T 4  T04 ) (4.4) where n is the normal vector of the surface, I(x, y, z, t) is the laser energy distribution on the workpiece [W/m2 ],  is the absorption factor, hc is the heat convection coe!cient [W/m2 ·K], ²t is emissivity,  is the StefanBoltzman constant [5.67 × 408 W/m2 ·K4 ], is the workpiece surfaces [m2 ],  is the surface area irradiated by the laser beam [m2 ] and T0 is the ambient temperature [K] [156]. K(uT ·n)| =

• On the surface of the melt pool, if g is vertical, the surface tension © 2005 by CRC Press LLC

should be derived by p +  g z = (2µ

CU · n) + /R Cn

(4.5)

and U·n=0

(4.6)

where z is a vertical coordinate [m],  is surface tension [N/m], and R is the clad surface curvature [m] [157]. • At the solid /liquid interface f(x, y, z, t) = Const.

(4.7)

ux = uy = uz = 0

(4.8)

T = Tm

(4.9)

and and where f(x, y, z, t) is a function that presents the melt pool interface with substrate, and ux , uy , and uz is fluid velocity in x, y, and z directions, respectively [m/s] [158]. This condition is valid for pure elements. For the alloys, the freezing range should be considered. • At initial time and infinite time, the following conditions should be satisfied (4.10) T (x, y, z, 0) = T0 and T (x, y, z, 4) = T0

4.3

(4.11)

Laser Cladding Models in Literature

Models, which are based on physical laws, contribute to better understanding of the laser cladding process. A precise model can support the required experimental research to develop laser cladding. Several steady-state models for laser cladding have been proposed, whereas few papers have dealt with the dynamic nature of the process. Steady-state models are those models that are independent of time, whereas dynamic models refer to those models that take into account the transient response of the process. Table 4.1 lists the papers that deal with modeling of laser cladding. In the following, several developed models are briefly explained. © 2005 by CRC Press LLC

TABLE 4.1

List of modeling techniques and modeled paramaters for laser cladding. Reference Description Chande et al., 1985, [159] Kar et al., 1987, [160] Hoadley et al., 1992, [157] Lemoine et al., 1993, [161] Picasso et al., 1994, [162] Picasso et al., 1994, [163] Jouvard et al., 1997, [164] Colaco et al., 1996, [165] Romer et al., 1997, [166, 167] Kaplan et. al , 1997, [168]

Frenk et al., 1997, [136] Lin et al., 1998, [169] Bambeger et al., 1998 [170] Romer et al., 1999, [171] Kim et al., 2000, [104] Toyserkani et al., 2003, [172, 173] Zhao et al., 2003 [174] Toyserkani et al., 2002, [67, 134] Labudovic et al., 2003, [175]

© 2005 by CRC Press LLC

numerical model to obtain convection diffusion of matter in the melt pool analytical model to obtain a diusion model for extended solid solution numerical model to obtain temperature field and longitudinal section of a clad track analytical model to obtain powder e!ciency numerical model to obtain clad geometry and melt pool temperature numerical model to obtain fluid motion, melt pool shape analytical model to obtain critical energies for Nd: YAG cladding geometrical analysis to obtain clad height and powder e!ciency analytical model to obtain temperature of the melt pool numerical model to obtain powder particles temperature, melting limits, melt pool cross section analytical model to obtain the total power absorbed by melt pool numerical model to obtain powder catchment e!ciency analytical model to obtain the depth of clad and melt pool temperature stochastic model to obtain melt pool temperature numerical model to melt pool shape and dilution numerical model to obtain clad bead geometry numerical model to obtain dilution, melt pool temperature neural network and stochastic models to obtain the clad height numerical model to obtain dimensions of fusion, residual stress and transient temperature profiles

4.3.1

Steady-State Models

Several analytical and numerical models have been developed to show the process dependencies on the process parameters. They address several important physical phenomena such as thermal conduction, thermocapillary (Marangoni) flow, powder and shield gas forces on melt pool, mass transport, diusion, laser/powder interaction, melt pool/powder interaction, and laser/substrate interaction in the process zone. Kar et al. [160] studied one-dimensional diusion for extended solid solution in pre-placed laser cladding. They solved the energy transport and diusion equations. They also obtained equations for the dimensions and composition of the clad layer restricted by many assumptions. These assumptions are: no convection in the melt pool, semi-infinite plane, cylindrical clad shape, thermal independent coe!cients, and two-stage laser cladding. They improved their model by considering the heat convection in the surface and finding a model for changing the partition coe!cient [176]. Hoadley et al. [157] developed a two-dimensional finite element model for powder injection laser cladding. The model simulated the quasi-steady temperature field for the longitudinal section of a clad track. They took into account the melting of the powder particles in the liquid pool and liquid/gas free surface shape and position. Their results are for an idealized problem, where there is almost no melting of the substrate material in the clad. The results also demonstrate the linear relationship between laser power, the processing velocity, and the thickness of the deposited layer. One of the most simple but realistic models was obtained by Picasso et al. [162]. They considered the interaction between the powder particles and laser beam in the melt pool. They assumed the particles were melted by the laser beam before they arrived in the melt pool. The model predicted some of the processing parameters of laser cladding, including the beam velocity, the powder feed rate in the given laser power, beam width, and geometry of the powder injection jet. Picasso et al. [163] also developed a two-dimensional, stationary, finite element model for laser cladding by considering heat transfer, fluid motion, and deformation of the liquid-gas interface. They solved a stationary Stephan equation and found the shape of the melt pool in a known clad height Lin et al. [145], [177], and [169] developed a simple model for laser cladding with a coaxial nozzle. This model was proposed to characterize the particle bonding under heating in the cladding process. The results showed that particle sticking on the clad surface was increased when the particle size, velocity and the bonding temperature were decreased. Frenk et al. [136] developed a quantitative analytical model of the process based on the overall mass and energy balance. This model allowed them to calculate the mass e!ciency and the global absorptivity for laser cladding of Stellite 6 powder on mild steel, taking into account the incorporation of the powder into the melt pool as well as the energy absorbed by the powder jet © 2005 by CRC Press LLC

and the substrate. Colaco et al. [165] developed a simple lumped model for correlation between the geometry of laser cladding tracks and the process parameters. They assumed a circular shape for the cross sections of the laser clad tracks. Based on their results, a correlation between width and height of clads with the operating parameters such as powder feed rate, the process speed and e!ciency of the powder was obtained. Lemoine et al. [161] developed a model to predict laser cladding parameters such as the laser power and powder feed rate for a desired temperature. They considered a homogeneous temperature distribution inside each powder particle during the interaction time. Romer et al. [166] obtained an analytical process model, which relates the depth of the melt pool to the laser power and relative velocity of the laser beam to the sample. This model accounted for the latent heat of fusion and energy produced in the melt pool by exothermic reactions within the melt pool. The model showed a linear dependence of the melt pool depth on laser power and an inverse dependence on the square root of the relative beam velocity. Jouvard et al. [164] developed a model for an Nd:YAG laser operated at low powers, typically less than 800 W. Their theoretical study relied on a calculation of the powder feed rate fed into the melt pool and on a model of heat transfer in the substrate. They realized that the first power threshold is the power required for substrate melting, the second power threshold is the power that melts the powder particles directly and therefore they are in liquid phase when contacting the substrate.

4.3.2

Dynamic Models

An important step in control of laser cladding is to find a precise dynamic model for it. Bamberger et al. [170] developed a simplified theoretical model for estimating the operating parameters of laser surface alloying and cladding by the direct injection of powder into the melt pool. They developed an analytical dynamic expression for laser cladding when the table velocity was related to the melt pool temperature. They also achieved a steady-state expression, which related the height of the clad to the velocity of the table. The work done by Kim et al. [104] has modeled the melt pool formed during laser cladding by wire feeding using a two-dimensional, transient finite element technique. Due to the complexity of laser material processing, several authors have tried to identify a dynamic model for dierent methods of laser material processing using system identification techniques; however, there seems to be no report regarding the laser cladding dynamic model prediction using such techniques. Battaille et al. [178] tried to identify a dynamic model for laser hardening by system identification methods. They used a preliminary identification © 2005 by CRC Press LLC

experiment with a pseudo random binary sequence (PRBS). Romer et al. [171] found a stocastic-based dynamic model for the laser alloying process, when the table velocity or laser power was selected as the input and the melt pool surface area as the output. They used the auto regressive exogenous (ARX) system identification technique to obtain a dynamic model for laser alloying. The authors recognized nonlinearity in the process and, as a result, they used a linearized model for an operating point. They have reported that their model has performed poorly in many dierent cases due to its operating point dependency.

4.4

Lumped Models

In a lumped model, the dependency of the process (equations) and spatial variables is ignored and time becomes the only independent variable. This simplification will render ordinary dierential equations as opposed to partial dierential equations. In laser cladding by powder injection, a lumped model can be proposed by a balance of energy in the process. The balance of energy in the process is shown in Figure 4.3. In the figure, the total laser energy absorbed by the substrate and powder particles as well as dierent source of losses (e.g., reflection, radiation and convection) are shown. The balance of energy can be expressed by Qc = Ql  Qrs  QL + (  4)Qp  Qrp  Qradiation  Qconvection

(4.12)

where Qc is the total energy absorbed by the substrate [J], Ql is laser energy [J], Qrs is reflected energy from substrate, QL is latent energy of fusion [J],  is powder catchment e!ciency, Qp is energy absorbed by powder particles [J], Qrp is reflected energy from powder particles [J], Qradiation is energy loss due to radiation [J], and Qconvection is energy loss due to convection [J]. The laser energy is presented by Ql = Al Pl ti = rl2 Pl ti

(4.13)

where Al is the laser beam area on the substrate [m2 ], rl is the beam spot radius on the substrate [m], Pl is laser average power [W], and t is interaction time between the material and laser [s] which is presented by ti =

2rl U

(4.14)

The process speed is shown by U [m/s]. The reflected energy from the substrate is Qrs = (4   w )(Ql  Qp ) © 2005 by CRC Press LLC

(4.15)

T∞

T∞ Q radiation

Q radiation

Ql Qrp

Qr

Qr

Qrp

Qp

Qconvection

Qconvection

Clad

T0

Substrate

Qc

FIGURE 4.3 Balance of energy in laser cladding by powder injection.

where  w is workpiece absorbed coe!cient. The latent heat energy can be expressed by QL = Lf V

(4.16)

where Lf is latent heat of fusion [J/kg],  is average density in clad area [kg/m3 ] and V is the volume of melt pool, including the clad region [m3 ]. In order to find an expression for V in a lumped fashion, we can consider a portion of a cylinder laying on the substrate as shown in Figure 4.4, where the width of the clad is equal to the laser beam diameter on the substrate. Also, we assume that the length of the melt pool is equal to the laser beam diameter. The cross section area Ac can be found by Ac =

m ˙  p U

(4.17)

where p is particles density [kg/m3 ]. If dilution is ignorable, which is an appropriate assumption for laser cladding, and the width of the melt pool is equal to laser diameter, the volume of melted area V can be expressed by V = 2rl Ac © 2005 by CRC Press LLC

(4.18)

Ac 2rl Clad Clad Substrate Substrate

2rl FIGURE 4.4 A lumped cross section of the clad and substrate.

In a lumped model, the powder e!ciency can be assumed as the ratio between the area of laser beam and powder stream on the substrate. Therefore =

rl2 rs2

(4.19)

where rs is the powder stream diameter on the substrate [m]. In order to derive an equation for the energy absorbed by powder particles in a lumped model, consider a homogeneous distribution of powder particles over the laser beam cross section as shown in Figure 4.5. If powder particle radius rp is known, the number of particles n in the laser beam area over a time period of ti is given by n=

3mt ˙ i 4p rp3

(4.20)

where m ˙ is powder feed rate [kg/s], and p is powder density [kg/m3 ]. The overall area of the powder particles in the laser beam indicates the attenuated area Aat [m2 ] by the powder particles as Aat = nrp2 =

3mt ˙ i 4p rp

(4.21)

As a result, the absorbed energy by the particles can be obtained by Qp = Ql

3Ql mt Aat ˙ i = Al 4p rp rl2

(4.22)

The reflected energy from the powder particles can be derived from Qrp = (4   p )Qp © 2005 by CRC Press LLC

(4.23)

2rl

2r p

Laser beam diameter

FIGURE 4.5 Attenuated area by powder particles.

where  p is powder particles’ absorbed coe!cient. The radiative loss can be presented by Qradiation = Al ²t (T 4  T04 )

(4.24)

where ²t is emissivity,  is the Stefan-Boltzman constant [5.67×408 W/m2 ·K4 ], T is melt pool temperature [K], and T0 is ambient temperature [K]. The convective loss in a lumped model, assuming a concentrated heating zone in the laser beam area, can be presented by Qconvection = Al hc (T  T0 )

(4.25)

where hc is the heat convection coe!cient [W/m2 ·K]. Calculating hc is di!cult and Goldak [179] and Yang [180] suggested an experimental expression, which is (4.26) hc = 24.4 × 404 ²t T 4.64 Equation (4.26) introduces a nonlinear term in the final energy balance equation. This term can, however, be ignored for simplification of final dierential equation. The energy Qc can be presented in an integral form as Z Qc = cp T (x, y, z, ti )dVs (4.27) Vs

where Vs is the heat-aected volume in Cartesian coordinates (x, y, z) [m3 ] and  is the average density in the clad region [kg/m3 ]. Substitution of the parameters in Equation (4.12) leads to an equation for Qc . Plugging the derived equation for Qc into Equation (4.27) leads to a lumped dierential equation that presents the lumped model of the process. This dierential equation can be solved by dierent numerical approaches. © 2005 by CRC Press LLC

4.5

Analytical Modeling

In general, there is no closed form solution for an analytical model with both heat conduction and momentum equations for the laser cladding process. However, when only heat conduction is considered for a moving heat source, the solution can be found. A well-known approach to solve the simplified heat conduction Equation (4.1) with given boundary and initial conditions is Green’s function [156]. Based on this function the temperature distribution at time t and point (x, y, z) is represented as +4 Z +4 Zt Z G(x, y, z, t, x´, y´, 0, t´, u)I(x´, y´, t´)dx´dy´dz´ (4.28) T (x, y, z, t) = T0 + 0 4 4

where 4 G(x, y, z, t, x´, y´, 0, t´, u) = s (4.29) 4 k[(t  t´)]3/2 K ½ ¾ [(x  x´) + u(t  t´)]2 + (y  y´)2 + z 2 exp 4k(t  t´) and k=

K cp

(4.30)

T (x, y, z, t) [K] represent the temperature at (x, y, z) at time t [s] due to a point source of laser generated at (x´, y´, z´) at time t0 [s] that is moving with velocity of u [m/s], K is thermal conductivity [W/m·K], cp is specific heat capacity [J/kg·K],  is density [kg/m3 ], and T0 is ambient temperature [K]. In order to evaluate the steady-state and transient part, Green’s function can be rewritten as the product of a steady-state term W and a timedependent term V as G(x, y, z, t, x´, y´, 0, t´, u) = W (x, y, z, x´, y, u)V (x, y, z, t, x´, y´, t´, u) where W (x, y, z, x´, y0 , u) =

u 4 exp[ (x  x´+ ")] 2K" 2k

(4.31)

and ½ ¾ " ["  u(t  t´)]2 V (x, y, z, t, x´, y´, t´, u) = s exp  4k(t  t´) 2 k(t  t´)3/2 where "= © 2005 by CRC Press LLC

p (x  x´)2 + (y  y´)2 + z 2

(4.32)

(4.33)

After reversing the order of integration, the temperature distribution Equation (4.1) can be written as +4 Z +4 Z I(x´, y´, t´)W (x, y, z, x´, y, u)V´(", t, u)dx´dy´ (4.34) T (x, y, z, t) = T0 + 4 4

where V´(", t, v) =

Zt

U (x, y, z, t, x´, y´, t´, u)dt´

(4.35)

0

4 k(tt´)

which can be rewritten, using  = s " V (", t, v) = s  0

4 = 2

Z4

s 4/ kt

exp[

as

(" 2  u/k)2 ]d 4 2

(4.36)

¾ ½ " + ut "  ut 4  erf( s ) + exp["v/k)(4  erf( s )] 2 kt 2 kt

As an example, when the substrate is Fe, the laser beam is at (0, 0) position, and process speed is 0.005 m/s; W and V 0 for dierent " are shown in Figures 4.6a and 4.6b, respectively. The W plot represents the steady-state nature of the thermal domain, whereas the V plot shows the time dependency of the thermal domain. In the following section, we present a case study based on numerical solution of the laser cladding process.

4.6

Numerical Modeling — A Case Study

As a case study, we develop a numerical model for the laser cladding by powder injection. The main objective of developing a 3-D transient finite element model of laser cladding by powder injection is to investigate the eects of laser pulse shaping, traveling speed and powder feed rate on the clad geometry as a function of time. To improve and understand the underlying process and theory, several models have been developed in the literature as addressed before. These models show the dependence of the process on the important parameters involved. These models can also be used in predicting the process for dierent parameters as well as controller design. An accurate model can significantly reduce the development cost of automated laser cladding systems. Although the literature indicates several laser cladding models, there is a significant lack of more accurate models that take into account the eects of © 2005 by CRC Press LLC

a) 1

χ = 0.1

0.8 0.6

χ =1 χ =2 χ =3

0.4 0.2 0 0

0.05

0.1

0.15

0.2

Time (s) b)

FIGURE 4.6 a) Illustration of a typical value for function W , in which the laser beam is in the center of the plane (0, 0), the velocity of laser beam is 0.005 m/s, and the material is Fe under a CO2 beam, b) illustration for V 0 vs. time for dierent ".

laser pulse characteristics, melt pool geometry, power attenuation due to the powder particles, absorption factor deviation during the process (Brewster eect), and temperature dependencies of material properties. The literature also shows the absence of a model for the prediction of the clad geometry in the transient and dynamic period of the process. In order to develop a more precise model, a solution strategy is proposed. In this strategy, the interaction between the powder and the melt pool is assumed to be decoupled and as a result, the melt pool boundary is first obtained in the absence of powder spray. Once the melt pool boundary is calculated, it is assumed that a layer of coating material based on powder feed rate and elapsed time is deposited on the intersection of the melt pool and powder stream in the absence of laser beam. The new melt pool boundary is then calculated by thermal analysis of the deposited powder layer, substrate, and © 2005 by CRC Press LLC

laser heat flux. For implementation of the proposed solution strategy, a finite element technique is used to develop a novel 3-D transient model for laser cladding by powder injection. The model is then used to study the correlation between the clad geometry and the process parameters. In the first set of simulations, the eects of laser pulse shaping parameters (laser pulse frequency and energy) on the clad geometry are investigated when the other process parameters such as travel speed, laser pulse width, powder jet geometry and powder feed rate are constant. In the second set of simulations, the eects of the process speed and powder feed rate on the clad geometry are investigated when laser pulse shaping including energy, frequency and width of the pulse and powder jet geometry are constant. The quality of cladding of Fe on mild steel for dierent parameter sets is experimentally evaluated and shown as a function of eective powder deposition density and eective energy density. The comparisons between the numerical and experimental results are also presented. In the following section, a thermal mathematical model is developed and required assumptions are addressed.

4.6.1

Thermal Mathematical Model

For a laser cladding process, a moving laser beam with a Gaussian distribution intensity strikes the substrate at t = 0 as is shown in Figure 4.2. Due to material added, the clads form on the substrate as shown in the figure. The transient temperature distribution T (x, y, z, t) is obtained from the three-dimensional heat conduction in the substrate as expressed by Equation (4.1) [156]. As discussed earlier, the boundary conditions for the heat transfer process are Equations (4.4), (4.10) and (4.11). Equation (4.1) along with boundary conditions (4.4), (4.10) and (4.11) cannot comprehensively express the physics of the process. Therefore, to incorporate the eects of the laser beam shaping, latent heat of fusion, Marangoni phenomena, geometry growing(changing the geometry), and Brewster eect, the following adjustments are considered: • A pulsed Gaussian laser beam with a circular mode (TEM00 ) [101] is considered for the beam distribution. The laser power distribution profile I [W/m2 ] is [181] 5 Ãs ! 6 2 2 I(r) =  I0 exp 7 r2 8 (4.37) rl where r=

p x2 + y 2 ,

© 2005 by CRC Press LLC

I0 =

2 Pl , rl2

and

Pl = EF

(4.38)

and rl is the beam radius [m], I0 is intensity scale factor [W/m2 ], Pl is the laser average power [W], E is the energy per pulse [J], and F is the laser pulse frequency [Hz]. When the laser beam is on  = 4 and when it is o  = 0. The parameter  is changed based on the laser pulse shaping parameters such as frequency F and width W that is the time that the laser beam is on in one period. • The eect of latent heat of fusion on the temperature distribution can be approximated by increasing the specific heat capacity [182], as cp =

Lf + cp Tm  T0

(4.39)

where cp is modified heat capacity [J/kg·K], cp is the original heat capacity [J/kg·K], Lf is latent heat of fusion [J/kg], Tm is melting temperature [K], and T0 is ambient temperature [K]. • The eect of fluid motion due to the thermocapillary phenomena can be taken into account using a modified thermal conductivity for calculating the melt pool boundaries. Experimental work and estimations in the literature [183] suggest that the eective thermal conductivity in the presence of thermocapillary flow is at least twice the stationary melt conductivity. This increase can be generally presented by K  (T ) = aK(Tm )

if T > Tm

(4.40)

where a is the correction factor and K  is modified thermal conductivity [W/m·K]. • Power attenuation is considered using the method developed by Picasso et al. [162] with some minor modifications. Figure 4.7 shows the proposed geometrical characteristics in the process zone which is used in the development of the following equations. Based on their work µ ¶ Pat (4.41) P4 = Pl  w () 4  Pl µ · ¶¸ Pat Pat 4 + (4   w ()) 4  (4.42) P2 = Pl  p  p Pl Pl where P4 is total power directly absorbed by the substrate [W], P2 is power that is carried into the melt pool by powder particles [W], Pat is attenuated laser power by the powder particles [W],  w () is workpiece absorption factor,  p is particle absorption factor, and  is the angle of the top surface of the melt pool with respect to the horizontal line as shown in Figure 4.7 [deg]. Consequently, the total power absorbed by the workpiece Pw [W] is Pw = P4 + P2 = Pl © 2005 by CRC Press LLC

(4.43)

FIGURE 4.7 The proposed geometrical characteristics of process zone.

where  is the modified absorption factor. The ratio between the attenuated and average laser power can be obtained by [162] ; · m if rjet < rl Pat ? 2c rl rp vp cos(jet ) = (4.44) · m = Pl if r  r jet l 2 rjet rp vp cos(jet ) c

·

In these equations, m is powder feed rate [kg/s], c is powder density [kg/m3 ], rl is radius of the laser beam on the substrate [m], rp is radius of powder particles [m], vp is powder particles velocity [m/s], jet is the angle between powder jet and substrate [deg], and rjet is radius of powder spray jet [m]. The powder catchment e!ciency p can be considered as the ratio between the melt pool surface and the area of powder stream (Figure 4.7) as p =

Aliq jet Ajet

(4.45)

where Aliq jet is the intersection between the melt pool area on the workpiece and powder stream, and Ajet is the cross-section area of the powder stream on the workpiece. © 2005 by CRC Press LLC

If we assume the absorption of a flat plane inclined to a circular laser beam depends linearly on the angle of inclination  as shown in Figure 4.7, and  w (0) is the workpiece absorption of a flat surface,  w () can be calculated from (4.46)  w () =  w (0)(4 + w ) where  is the angle shown in Figure 4.7 and w is a constant coe!cient obtained experimentally for each material [101], [162]. • The temperature dependency of material properties and absorption coe!cients on the temperature are taken into account in the model. • In order to reduce the computational time, a combined heat transfer coe!cient for the radiative and convective boundary conditions is calculated based on the relationship given by Goldak [179] and Yang [180]: hc = 24.4 × 404 ²t T 4.64

(4.47)

Using (4.47), (4.37), and (4.43), the boundary condition in (2.2) is simplified to ; · ³s ´ ¸ 2 ? 2 2 2 Pl exp  rl r  hc (T  T0 ) if 5  rl2 K(uT ·n)| = = hc (T  T0 ) if 5 / (4.48)

4.6.2

Solution Algorithm

A method can be proposed to obtain the clad geometry in a 3-D and timedependent laser cladding process using the model discussed in the previous section. This proposed numerical solution has two steps as follows: 1. Obtaining the melt pool boundary in the absence of the powder spray. In this step, the interaction between the powder and melt pool is assumed to be decoupled, and, as a result, the melt pool boundary can be obtained by solving Equation (4.1). 2. Adding a layer of the powder to the workpiece in the absence of the laser beam. In this step, once the melt pool boundary is calculated, it is assumed that a layer of coating material based on the powder feed rate, elapsed time, and intersection of melt pool/powder jet is deposited on the workpiece. The new deposited layer creates a new tiny object on the previous domain which is limited to the intersection of the powder stream and the melt pool. For each increment in time, t, its height is given by mt ˙ (4.49) h = 2 rjet c © 2005 by CRC Press LLC

where h is the thickness of the deposited layer [m] and t is the elapsed time [s]. For numerical convergence, the temperature profile of the added layer is assumed to be the same as the temperature of the underneath layer, which will be discussed in the end of this section. The new temperature profile of the combined workpiece and the layer of powder is then obtained by repeating Step 1. Figure 4.8 shows the sequence of the proposed numerical modeling. On the left side of the figure, a moving laser beam is shown while the deposition of coating material (Step 2) is presented on the right side. The numerical solution is carried out in two dierent time periods. The first one is the time between two deposition steps, and the second one is the time period for calculating the melt pool area. After performing Step 2 and before repeating the first step, the following corrections are applied: • All thermo-physical properties and absorption factor  w (0) are updated based on the new temperature distribution. • The new  w () is calculated based on the updated  and Equation (4.46). • The new p is obtained based on the new melt pool geometry using Equation (4.45). • The new Pw is calculated using Equation (4.43). Many numerical methods for solving Equation (4.1) have been reported since 1940. Finite element method (FEM) is one of the most reliable and e!cient numerical techniques, which has been used for many years. FEM can solve dierent forms of partial dierential equations with dierent boundary conditions. In this work, the governing PDE Equation (4.1) is highly nonlinear due to material properties with dependency on temperature and a moving heat source with a Gaussian distribution. To implement the numerical solution strategy, code was developed using the MATLAB (www.mathworks.com)/ FEMLAB (www.femlab.com) software. The code discretizes the heat conduction equation and generates the initial mesh in the substrate using the available options in FEMLAB. By solving Equation (4.1) and calculating the melt pool boundary, the geometry of the domain is modified to incorporate the clad into the substrate. For meshing, the domain is partitioned into tetrahedrons (mesh elements) as shown in Figure 4.10. Due to the deposited layer and changes in the substrate geometry, an adaptive meshing strategy is used. As it is seen, the mesh is finer for the portion of the domain in which the clad is generated. A time-dependent solver is used to solve the nonlinear time-dependent heat transfer equation. The solver is an implicit dierential-algebraic equation (DAE) solver with © 2005 by CRC Press LLC

STEP 1

STEP 2

∆x = U∆t Melt pool's large diameter

∆h

a)

b)

FIGURE 4.8 Sequence of calculation in the proposed numerical model: a) Step 1, b) Step 2.

automatic step size control which is patented as fldask [184]. The solver is suitable for solving equations with singular and nonlinear terms. To justify the assumption made in Step 2 of the solution strategy regarding the temperature of the added layer, we note that the power transferred to the workpiece by the powder particles is considered by Equation (4.42). As a result, the temperature of the deposited layer in Step 2 should be assumed to be the same as the ambient temperature, which we consider contrary to it to be the same as that of the melt pool of the underneath layer. Numerical calculation indicates that this added energy is about 1% of the power, which does not have a considerable eect on the overall results. This assumption will help the convergence of the numerical solution and reduction of the computational time by eliminating the large temperature gradient between the added and underneath layers which cause instability in most of today’s numerical solvers. © 2005 by CRC Press LLC

TABLE 4.2

Process parameters. rl [m] rp [m] rjet [m] Laser pulse width [ms]  a

7.0e  4 [187] 22.5e  6 7.5e  4 3 4.67e  2 [162] 2.5 [183]

T0 Tv Tm vp jet p

[K] [K] [K] [m/s] [deg.] [%]for Nd:YAG

293 [185] 3343 [185] 4844 [185] 26 55 34 [186]

In the following section, the numerical parameters that are considered for the numerical model are addressed.

4.6.3

Numerical Parameters

A 50×40×5 mm block is selected for the initial substrate in a Cartesian coordinate system as shown in Figure 4.10. The thermo-physical properties of Fe are considered for both substrate and powder. All thermo-physical properties such as thermal conductivity, specific thermal heat, emissivity and density are considered to be temperature dependent. The thermo-physical properties for temperatures higher than vapor temperature, Tv , are fixed to the amount of thermo-physical properties in Tv . All the thermo-physical parameters have been obtained from Wong [185]. Also,  w (0) as a function of temperature is obtained from Xie and Kar [186] for Fe when a Nd:YAG laser is used. The other process parameters are listed in Table 4.2. In order to investigate the independence of the solutions on the number of nodes, simulations were performed in the dierent number of nodes. The computational results are shown in Figure 4.9. As seen, with increase of number of nodes, the curve shows the calculated temperatures at x = 0.030 m, y = 0.000 m, z = 0.005 m and at t = 20 s becomes flat such that the dierence between the calculated temperatures with 8, 400 and 40, 203 nodes is 8 K. As a result, the number of elements was initialized with 40, 203 nodes and 43, 577 elements which were mostly concentrated on the top of the surface as seen in Figure 4.10. The simulation was performed for 20 seconds. The time step between layers’ depositions was set to 20 ms and the other time step was controlled by the solver; however, it was not greater than 0.2 ms. The developed software was then used in studying the laser cladding process for dierent physical parameters.

4.6.4

Numerical Results

Simulations were carried out for dierent process parameters to study various aspects of laser cladding, which can be categorized into two sections as follows: 1. In the first study, the eects of laser pulse shaping on the clad geometry © 2005 by CRC Press LLC

Tempearture (K) at x=0.03 m, y=0.0 m, z=0.005 m

2400 2300 2200 2100 2000 1900 1800 1700 1600 1500 1000

2000

3000

4000

5000

6000

7000

8000

9000

10000 11000

Number of nodes

FIGURE 4.9 Comparison between the calculated temperature at a desired point in dierent number of nodes to investigate the independency of solutions on the number of grid.

were investigated when the other process parameters were constant. 2. In the second study, the eects of travel speed and powder feed rates on the clad geometry were examined when the other parameters were constant. In the following section, these two studies along with a selection of 3D numerical results will be addressed. 4.6.4.1

Eects of Laser Pulse Shaping on Clad Geometry

In order to evaluate the contribution of the laser pulse energy E and the laser pulse frequency F on the clad geometry, a multistep laser pulse energy and pulse frequency were selected as shown in Figures 4.11a and 4.11b, respectively. The laser pulse energy was changed from 2.5 J to 4 J in four steps. The laser pulse frequency was also changed from 70 to 100 Hz in four steps. The average laser power for both cases are shown in Figures 4.11a and 4.11b. In the numerical simulations of the first study, U = 0.004 m/s and m ˙ = 4.67e  5 kg/s. Also, in all numerical simulations, it was assumed that the laser was turned on at t = 0 s and the beam was at a position of x = 40, y = 0 and z = 5 mm. Figure 4.12 shows the temperature distribution of the workpiece at t = 20 s in dierent views for a multistep laser pulse energy. The figure illustrates the isothermal lines in the domain where the maximum temperature was 2, 388 © 2005 by CRC Press LLC

Finer mesh in the clad domain

m

Z X

Y m

FIGURE 4.10 A typical mesh in the proposed domain.

K. It also shows a rapid cooling in the domain due to the concentrated moving heat source. Along the object, the isothermal lines expand, whereas the condensed isothermal lines exist in the area close to the laser source. Figure 4.13 shows the generated clad after 20 s for a multistep laser pulse energy. In order to have a better view of the generated clad on the substrate, a virtual light source to illuminate the domain is considered. The ripples on the generated clad were discovered to be dependent on the size, shape and number of elements used to mesh the domain. Increasing the number of meshes can reduce the ripples; however, the average height remains the same. Although increases in the number of meshes and reducing their size result in elimination of ripples, the computational time dramatically increases. As seen in Figures 4.12 and 4.13, the clad height and width increase with increasing laser pulse energy as expected. 4.6.4.2

Eects of Process Speed and Powder Feed Rate on Clad Geometry

In order to evaluate the contribution of the travel speed, U, and the powder feed rate, m, ˙ on the clad geometry, a multistep travel speed is applied to the numerical procedure for five dierent powder feed rates. The selected multistep speed is shown in Figure 4.14. This multistep speed is applied for ˙ 2 = 2.09e5, m ˙ 3 = 2.54e5, m ˙4= five dierent feed rates: m ˙ 4 = 4.67e5, m 2.92e  5, and m ˙ 5 = 3.34e  5 kg/s. For all numerical simulations, laser pulse energy is E = 3.5 J, laser pulse frequency F = 400 Hz and laser pulse width is W = 3 ms.. Figure 4.15 shows the temperature distribution of the workpiece at © 2005 by CRC Press LLC

Laser pulse energy (J)

4 3.5 3 2.5 2

0

2

4

6

8

10

12

14

16

18

20

12

14

16

18

20

12

14

16

18

20

12

14

16

18

20

Average laser power (W)

Time (s) 400 300 200 100

0

2

4

6

8

10

Time (s)

Average laser power (W)

Laser pulse frequency (Hz)

a)

100 80 60 0

2

4

6

8

10

Time (s) 300

200

100

0

2

4

6

8

10

Time (s)

b)

FIGURE 4.11 a) Multistep laser pulse energy along with corresponding average laser power, b) Multistep laser pulse frequency along with corresponding average laser power.

t = 20 s for a multistep velocity at m ˙ 4 = 4.67e  5 kg/s in dierent views. The figure illustrates the isothermal lines in the whole domain where the maximum temperature is 2, 030 K. The figure also shows a rapid cooling in the domain due to the concentrated moving heat source. Along the object, the isothermal lines expand whereas the condensed isothermal lines exist in the area close to the laser source. Figure 4.16 shows the eect of powder feed rate on the maximum temperature in the object at t = 20 s. Increase in the powder feed rate (m) ˙ causes the maximum temperature to reduce as expected from Equation (4.44). The equation shows that the increase in m ˙ increases the power attenuation and consequently decreases the total absorbed energy. Thus, the eective energy absorbed by the substrate decreases. Further increase in the powder feed rate drops the maximum temperature below the melting temperature so that no clad can be produced. Of interest is the fact that some experiments conducted with the same process speeds but powder feed rate of 5.0e5 kg/s (3.0 g/min) © 2005 by CRC Press LLC

FIGURE 4.12 Temperature distribution (in Kelvin) at t = 20 s for a multistep laser pulse energy with W = 0.003 s, F = 400 Hz.

in which the model predicts a maximum temperature of approximately 4, 800 K showed that the cladding was impossible due to the unmelted and weak bond between the clad and substrate. Figure 4.17 shows the melt pool at t = 4 and t = 20 s when the process velocity is 0.5 and 2 mm/s, respectively, at a powder feed rate of 4.67e  5 kg/s. As it is seen, the shape of the melt pool depends on the process velocity and deposited clad. The isothermal lines are also illustrated in the figure. Figure 4.18 shows the generated clad after 20 s for a multistep velocity at m ˙ 4 = 4.67e  5 kg/s. In order to have a better view of the generated clad on the substrate, a light source to illuminate the domain is considered. As it was discussed, the ripples on the generated clad was discovered to be dependent on the size, shape and number of the elements used to mesh the domain. Figure 4.19 shows the clad heights for dierent powder feed rates and different process velocities. As seen in Figure 4.19, the clad height decreases © 2005 by CRC Press LLC

FIGURE 4.13 Generated clad after 20 s for a multistep laser pulse energy (domain is illuminated by a virtual light).

with increasing process speed, while it increases by increasing the powder feed rate as expected from Equation (4.49). The clad height is decreased gradually when the process speed is suddenly stepped up. The main reason for this occurrence is the contribution of transient temperature in the melt pool shape. The results also show that when the velocity increases, the clad height decreases. This is an indication of the nonlinearities in the process. In order to compare the numerical and experiment results, we will next explore an experimental analysis for the evaluation of the clad quality. Then, we will use the experimental analysis along with numerical results to interpret the modeling results.

4.6.5

Experimental and Numerical Analysis

In order to validate the numerical results, an experimental analysis is performed not only to investigate the experimental dependency of laser cladding of Fe on mild steel, but also to obtain a criteria for verification of numerical results. The bases for the experimental analysis are similar to the method which will be developed in detail in Chapter 6 and is also discussed in [3, 4]. The experimental analysis relates all process parameters in Equations (2.8), © 2005 by CRC Press LLC

Travel speed (m/s)

2.5

x 10

-3

2 1.5 1

Laser Power is on at t=0 when beam centerline is emitted on point ( x=0.001 , y=0.000 , z=0.005 m)

0.5 0

0

2

4

6

8

10

12

14

16

18

20

Time (s)

FIGURE 4.14 Multistep process speed.

z(m)

y(m)

x(m)

z(m)

x(m)

y(m)

FIGURE 4.15 Temperature (in Kelvin) distribution at t = 20 s for a multistep travel speed (m ˙ = 4.67e  5 kg/s). © 2005 by CRC Press LLC

FIGURE 4.16 Maximal temperatures at t = 20 s for dierent powder feed rates (U = 2 mm/s).

(2.9) and (2.10), which represent the eective energy density Eef f [J/mm2 ] and eective powder deposition density #ef f [g/mm2 ] as a function of eective area Aef f [mm2 /s]. Calculated values for Eef f and #ef f of the processing conditions listed in Table 4.3 are plotted in Figure 4.20. By observation, and mechanical and metallurgical tests, four regions are distinguishable for the generated clads as shown in Figure 4.20. The region called “good quality clad” provides a good bond between the substrate and clad where the clad has a relatively smooth surface and good profile without cracks and pores. The region called “roughness, some bonding” indicates that the clad has some bonding with the substrate; however, the clad has many cracks and pores and may be easily removed from the substrate after the process. The region called “brittle” indicates that the clad has been generated without any bonds to the substrate and even the clad itself may be brittle (i.e., poorly consolidated). The region called “no cladding” indicates that no clad can be created in this region. The evidence is shown in Figures 4.21 through 4.24, which are explained in the following sections. 4.6.5.1

Experimental Setup

The experiments were performed using a 1000 W LASAG FLS 1042N Nd:YAG pulsed laser, a 9MP-CL Sulzer Metco powder feeder unit, and a 4-axis CNC table. The spot point diameter on the workpiece was set to 1.4 mm where the © 2005 by CRC Press LLC

FIGURE 4.17 Temperature distribution and clad shape in dierent views at a) t= 4 s, U = 0.5 mm/s and m ˙ = 4.67e  5 kg/s, b) t = 20 s, U = 2 mm/s and m ˙ = 4.67e  5 kg/s.

© 2005 by CRC Press LLC

FIGURE 4.18 Generated clad at t = 20 s for a multistep travel speed (domain is illuminated by a virtual light).

Clad height (mm)

4

U = 0.5 mm / s

U = 2 mm / s

U = 1 mm /s U = 1.5 mm / s

3

m 5 = 3.34e − 5 kg / s

m 4 = 2.92e − 5 kg / s

2 1 0

m 3 = 2.51e − 5 kg / s m 2 = 2.09e − 5 kg / s m 1 = 1.67e − 5 kg / s

Laser on 0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Position (cm)

FIGURE 4.19 Numerical results for the clad heights at dierent powder feed rates (m) ˙ and process velocities (U ).

© 2005 by CRC Press LLC

180 9 160

10 8 3

140

13

Good quality clad

4

14

2

Effective energy density (J/mm )

7 120

2

11 12

100

15

6

5

16

1

80

Roughness some bonding

Brittle

60 Conditions 1 through 16

40 20

The number of the condition is shown beside the marker

No cladding

1

1.5

2

2.5

3

3.5

Effective powder depostion density (g/mm2 )

4

4.5

5 x 10

-3

FIGURE 4.20 Eective energy versus eective powder deposition densities for conditions 1 to 16 of cladding of pure iron on the mild steel.

laser intensity was Gaussian. The laser beam was shrouded by argon shield gas. In the experiments, Argon as the shield gas and inert gas were set to 2.34e  5 m3 /s (3 SCFH). The angle of the nozzle spray was set to 55 from the horizontal line and the size of powder stream profile was approximately 1.5 mm on the workpiece. The powder used in the experiments was Fe with a purity of 98% and mesh size of 45 µm (-325). Sandblasted mild steel plates (0.25 to 0.28 C; 0.6 to 1.2 Mn) with dimensions of 50×40×5 mm were selected as the substrate. The laser was aimed at 10 mm away from the edge. The height of the clad was measured in real-time by the device discussed in [68]. Two sets of experiments were performed to mimic the numerical simulations as follows: 1. The laser pulse energy and laser pulse frequency were changed based on those which are shown in Figures 4.11a and 4.11b when the pulse width was fixed to 3 ms. The process speed was set to 1 mm/s for this set of experiments, similar to the numerical simulation. The powder feed rate was also set to 1 g/min (4.67e  5 kg/s) (Conditions 1 to 8). 2. The travel speed was changed as shown in Figure 4.14. This multistep speed is applied for two dierent feed rates: 4.67e5, and 3.34e5 kg/s. © 2005 by CRC Press LLC

TABLE 4.3

Conditions of experiments. Condition E [J] F [Hz] U [mm/s] m ˙ [kg/s] 1 2.5 100 1 4.67e  5 2 3 100 1 4.67e  5 3 3.5 100 1 4.67e  5 4 4 100 1 4.67e  5 5 3.5 70 1 4.67e  5 6 3.5 80 1 4.67e  5 7 3.5 90 1 4.67e  5 8 3.5 100 1 4.67e  5 9 3.5 100 0.5 4.67e  5 10 3.5 100 1 4.67e  5 11 3.5 100 1.5 4.67e  5 12 3.5 100 2 4.67e  5 13 3.5 100 0.5 3.34e  5 14 3.5 100 1 3.34e  5 15 3.5 100 1.5 3.34e  5 16 3.5 100 2 3.34e  5

For these experiments laser pulse energy, laser pulse frequency and laser pulse width were set to 3.5 J, 400 Hz and W = 3 ms (Conditions 9 to 16), respectively.

4.6.6

Comparison Between Numerical and Experimental Results

Figure 4.21 shows the deviation between the numerical and experimental results for the change of laser pulse energy. As seen, the numerical modeling predicts a clad height which does not agree well with the experiment for Condition 1, while for Conditions 2, 3 and 4 there is a good agreement between the model and experimental results as listed in Table 4.4. Based on the quality analysis shown in Figure 4.20, it can be concluded that the quality of Condition 1 listed in Table 4.3 is not acceptable due to weak bonding between the clad and substrate. The reason for this is the lack of su!cient energy to melt the powder and substrate. As a result, the clad is easily removed from the substrate following the process as shown in Figure 4.21. Recalling the numerical modeling, the layer can be deposited only if a melt pool area on the substrate is expanded for any given time. The same justification can be mentioned for the case when the laser pulse frequency is changed as seen in Figure 4.22. For this case, Condition 5 does not provide a deposit that is bonded to the substrate and Condition 6 provides a low quality clad with high roughness. For Conditions 7 and 8 the quality of clads are desired. The average errors between the experimental and numerical © 2005 by CRC Press LLC

TABLE 4.4

Numerical and experimental average clad height. Condition

# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Numerical average clad height (mm) 0.25 0.77 1.45 1.73 0 0.38 1.20 1.44 2.02 1.51 1.11 0.82 2.18 2.61 1.91 1.39

Experimental average clad height (mm) 0.76 (broken) 0.89 1.42 1.57 0.84 (broken) 1.20 (low quality) 1.10 1.23 2.51 2.01 0.91 0.73 3.80 (broken) 3.50 2.11 1.7

Error (%) – 38 2 9 – – 8 14 24 33 18 10 – 34 10 22

results for Conditions 7 and 8 are also listed in Table 4.4. Figures 4.23 and 4.24 show the comparison between the clad heights ob˙ 5 = 3.34e5 tained from the model and experiments for m ˙ 4 = 4.67e5 and m kg/s. As seen in Figure 4.23, there is excellent agreement between the numerical and experimental results for Conditions 10, 11, 12, 15, and 16. The average error between the two results are listed in Table 4.4. In Figure 4.24, where m ˙ 5 = 3.34e5 kg/s, the experimental and numerical results are matched with an average error listed in Table 4.4, except for the starting point. Regardless of the relatively large errors between the numerical and experimental results in Conditions 9 and 13, the transient nature of the clad generation is correctly predicted by the model. At the starting point which represents Condition 13, the model shows a delay in the clad generation which is missing in the experimental results. After analyzing the quality of the clad, it was observed that the initial part of the clad on the substrate had very poor quality and was easily removed from the substrate as shown in Figure 4.24. This shows that the model has correctly predicted the melt pool temperature at the start and the delay was due to the time required for developing the melt pool after applying the laser onto the substrate. To further investigate the clad/substrate geometrical profile and comparison between the numerical and experimental results, sections through the clad/substrate couples were made for selected samples. These sections were then mounted and polished to disclose their profiles. Figures 4.25a and 4.25b show the clad/substrate macrostructure for Con© 2005 by CRC Press LLC

2.5

Condition 1 E=2.5 J

clad height (mm)

2

Condition 3 E=3.5 J

Condition 2 E=3 J

Condition 4 E=4 J

Numerical model Experimental result

1.5 1

Laser on

Laser off

0.5 0

1

1.5

2

position (cm)

2.5

3

3.5

Clad is broken

FIGURE 4.21 Comparison between the experimental and numerical results for Conditions 1 to 4.

Clad height (mm)

2

Condition 5 F=70 Hz

Condition 6 F=80 Hz

Condition 8 F=100 Hz

Condition 7 F=90 Hz

Numerical model

1.5 Experiment 1

0.5

0

Laser on

Laser off

1

1.5

2

position (cm)

2.5

3

3.5

Clad is broken

FIGURE 4.22 Comparison between the experimental and numerical results for Conditions 5 to 8.

dition 4 with a E = 4 J, W = 3.0 ms, F = 400 Hz and U = 4 mm/s and Condition 8 with a E = 3.5 J, W = 3.0 ms, F = 400 Hz and U = 4 mm/s, respectively. The clad deposit is clearly visible and the clad has a good profile. The comparison between the numerical and experimental profiles shows that the model has predicted the clad profile very well. © 2005 by CRC Press LLC

Clad height (mm)

4

Condition 9 U=0.5 mm/s

3

Condition 12 U=2 mm/s

Condition 11 U=1.5 mm/s

Laser off

Experiment

2 1 0

Condition 10 U=1 mm/s

Numerical Model 0

0.5

1

1.5

2

2.5

3

3.5

4

Position (cm)

Laser on

FIGURE 4.23 Comparison between experimental and theoretical data for Conditions 9 to 12 (m ˙ 4 = 4.67e  5 kg/s).

Clad height (mm)

5

Condition 13 U=0.5 mm/s

4 3

Condition 14 U=1 mm/s

Condition 16 U=2 mm/s

Condition 15 U=1.5 mm/s

Laser off

Experiment

2

Numerical Model

1 0

0

0.5

1

Laser on

1.5

2

2.5

3

3.5

4

Position (cm)

FIGURE 4.24 Comparison between experimental and theoretical data for Conditions 13 to 16 (m ˙ 5 = 3.34e  5 kg/s).

© 2005 by CRC Press LLC

Condition 8

Condition 4

Numerical clad profile

Numerical clad profile

Comparison Comparison

Experimental clad profile

a)

Experimental clad profile

b)

FIGURE 4.25 Comparison between numerical and experimental clads’ profile for a) Condition 4, b) Condition 8.

4.7

Flow Field Modeling at the Exit of Coaxial Nozzle

This section addresses models of flow field in the exit of a coaxial nozzle. Laser cladding is a complex process involving interaction between the laser beam, the powder particles and the melted region of substrate. In order to build the clad with accurate dimensions and high e!ciency of the powder deposition in a coaxial laser cladding, it is essential to analyze the powder flow structure [148]. In coaxial laser cladding, the powder is carried by a flow stream impinging on the substrate. Some designs also include a shaping gas flow helping the powder flow stream to concentrate on the melt region of the substrate. Impinging jet flow on a solid surface as used in coaxial laser cladding has applications in many industrial processes such as water jet cutting and rocket exhaust during the take o, and, therefore, it has been studied extensively [188, 189, 190]. However, for the problem of compound jets, including three dierent coaxial jets with the middle flow containing the powder, not much information is available. In coaxial laser cladding, three dierent flows are encountered. At the center, there is an air (or argon) flow for protecting the lenses from the hot powder particles that may bounce o the substrate. Next is the flow with powder particles aiming at irradiated region, and finally the shaping gas as shown in Figure 3.25. All these flows and their interactions aect the catchments of the cladding powder at the laser irradiated region, and, therefore, aect the e!ciency and the quality of the clad. Lin [191] is among the first researchers who numerically studied the powder flow structure of a coaxial nozzle for laser cladding with various arrangements © 2005 by CRC Press LLC

of the nozzle exit. He used the commercially available FLUENT software to study the powder concentration in the air-powder flow. The flow at the exit of the nozzle can be laminar or turbulent depending on the nozzle exit Re number. It has been shown that turbulence-free jet cannot be sustained for Re < 4000 [192]. Typical flow parameter values for flow at the exit of the coaxial nozzle indicate that both laminar and turbulent jet can exist depending on the size and the exit velocity of the powder stream. Therefore both flow patterns are discussed in the following.

4.7.1

Laminar Model

The governing equations for the laminar flow are Navior-Stocks and continuity equations as 

CU  u2 U+(U · u)U + up= F Ct u·U= 0

(4.50) (4.51)

where U is velocity field [m/s],  is density of the gas [kg/m3 ],  is dynamic viscosity [m2 /s], p is pressure field [N/m2 ], F is external force [N]. The boundary conditions depend on the physical domain of interest. As a case study, a domain with the boundaries shown in Figure 4.26 is considered. These boundary conditions are • On the solid surface (i.e., substrate, and solid parts of the nozzle), the conditions are set to a no-slip boundary condition, in which v = 0 and u = 0, where v and u are the velocity components [m/s]. • On the top free surface, a neutral condition is considered, in which n · ((uU) = 0, where n is a normal vector on the free surfaces. • On the side free surface, a straight out flow is considered, in which t · U =0. • On the axisymmetric axis, the condition of slip can be considered, in which n · U = 0. • The shield gas, the powder stream, and shaping gas velocities are presented by Ul , Up , and Us , respectively. The physical domain can be solved by a numerical method. We developed a code using MATLAB/FEMLAB to obtain the flow field around the coaxial nozzle. The code discretized the momentum equation and generated the initial mesh in the substrate using the available options in FEMLAB. The flow domain was considered as shown in Figure 4.26. The shield gas velocity components Ul were set to (0, 0.5), the powder stream velocity components Up were set to (0.72, 2), and shaping gas velocity components were set to © 2005 by CRC Press LLC

Exit of coaxial nozzle

Axial direction (m)

Ul

Up

Us

Substrate surface Axisymmetric axis

Radial direction (m)

FIGURE 4.26 Geometry and boundary conditions for a typical coaxial nozzle exit (Ul is shield gas velocity, Up is powder stream velocity, and Us is shaping gas velocity).

(0.72, 2). Air was selected as the carrying gas with properties of  = 0.7 kg/m3 and  = 0.000037 m2 /s, which were the corresponding values in an average temperature of the domain under high temperature of the melt pool. A typical flow field based on the above-mentioned boundary conditions is shown in Figure 4.27, in which the flow field is shown by arrows and surface plot of the velocity field. As seen, the dark region indicates the low or even zero velocity, whereas the brighter color illustrates the higher velocity regions. The interaction between the powder stream and shaping gas results in a complex flow pattern including the formation of a vortex close to the shaping gas as seen in the figure. In order to investigate the trajectory of particles in the above flow field, the following equations were employed mp

d(Up ) = (mp  mf )g  6rp (Up  U) dt

(4.52)

dxp = Up (4.53) dt where mp is particle mass [kg], mf is fluid mass that the particle has displaced [kg], Up is particle velocity vector[m/s], U is the fluid velocity vector [m/s], © 2005 by CRC Press LLC

FIGURE 4.27 A typical laminar flow field at the exit of the coaxial nozzle.

rp is particle radius [m], g is gravity [m/s2 ],  is dynamic viscosity of the fluid [m2 /s], and xp is the tracing particle position in the flow field [m]. Inherent in Equation (4.52) is the assumption that the acceleration of a particle is influenced by gravity force and drag force [193]. As a case study, Fe particles were considered with rp = 22.5 µm in the above velocity field. A typical trajectory of a particle is shown in Figure 4.28. As seen, the particles follow the powder stream direction at the exit of the coaxial nozzle. However, near the melt pool, it tends to spread out due to loss of initial momentum and the existence of shield gas along the axis of symmetry. 4.7.1.1

Turbulent Flow

In this section, we investigate the flow field at the exit of the coaxial nozzle, when the flow is turbulent. In a turbulent flow, in addition to Navior-Stocks and continuity equations, the kinetic energy of turbulence k and dissipation of kinetic energy of turbulence % should be solved. This type of turbulence modeling is referred to as k  % turbulence modeling. The general form of governing equation is presented by U · u! + u · (D! u!) = P! + S!

(4.54)

! = (u, v, k, %)

(4.55)

where © 2005 by CRC Press LLC

Shield gas

Powder stream

Shaping gas

Substrate Axisymmetric axis

FIGURE 4.28 A typical trajectory of Fe particles in the flow field.

and D! is diusion coe!cient and S! is the source term corresponding to each ! components. Details of these terms can be found in the FEMLAB documentation [193]. As a case study, the above flow field was solved with the k  % turbulent model. For this model, in addition to the above laminar boundary conditions, a set of pre-defined boundary conditions in FEMLAB was used [193]. The k  % turbulent modeling capability of FEMLAB was employed to simulate the flow field. A typical result of the numerical modeling is shown in Figure 4.29. Comparison between the laminar flow field and turbulent patterns indicates a major change in the flow field. The vortex strength is weaker and the flow streams exit the domain mainly from the side free boundary condition.

4.8

Experimental-Based Modeling Techniques

This section addresses the application of experimental-based modeling techniques including stochastic and artificial neural networks to the laser cladding process. In many physical processes, it is very di!cult or even impossible to develop an analytical model due to process complexities. In laser material processing, © 2005 by CRC Press LLC

FIGURE 4.29 A typical turbulent flow field at the exit of the coaxial nozzle.

the complexity arises from the nature of the governing equations which are partial dierential and also the interaction of thermal, fluid, and mass transfer phenomena in the process as addressed in Section 4.2. There are several models for steady-state analysis of laser material processing and particularly for laser cladding [145, 160, 162, 166, 164], which provide many insights into the process. However, these models cannot be used directly in real-time control because of their limitations and intensive numerical calculations [194]. As a result, several authors have used stochastic techniques and neural network analysis to identify a dynamic model for the laser material processing. Bataille et al. [178] identified a dynamic model for laser hardening by stochastic methods. Romer et al. [171] found a dynamic model for the laser alloying process, where the table velocity or laser power was selected as the input and the melt pool surface area as the output. They used the auto regressive exogenous (ARX) system identification technique to obtain a dynamic model for the process. The authors recognized nonlinearity in the process, and, as a result, they used a linearized model around the operating point. They reported that their model performed poorly in many dierent cases due to its operating point dependency. There are several papers that deal with neural network analysis for laser material processing such as laser sheet bending and laser marking. Dragos et al. [195] used an artificial neural network to predict the future shape obtained by laser bending. They used the laser power and process speed as the variable parameters and the thickness of the material as the output of the model. Peligrad et al. [196] developed a model using an artificial neural network to predict the dynamics and parameter interactions of laser marking. Their © 2005 by CRC Press LLC

model considered the laser power and traverse speed as the inputs and the melt pool temperature as the output. However, to the knowledge of the author, there is no article on the application of neural networks for laser cladding to be used in the development of an intelligent system. All experimental-based modeling techniques such as the stochastic, artificial neural network, and neuro-fuzzy approaches are essentially based on optimizing the parameters of a given model to result in the minimum error between the measured and model prediction data. There are basically three general model structures that are used for nonlinear model prediction, based on prior and physical knowledge [197]. These models are white-box, when the model is perfectly known; grey-box, when some physical insights are available; and black-box, when the system is completely unknown. A black-box model is much more complex compared to the other two cases due to the variety of possible model structures. One of the model structures for black box modeling is artificial neural networks. Furthermore, experimental-based modeling techniques are well developed for linear systems; however, for nonlinear systems, the techniques are very limited, and they require many considerations for the selection of the model structure, inputs/outputs, and optimization techniques used to find the system parameters. Selecting a proper set of inputs and outputs and collecting data are critical in any dynamic model. The collected data due to the excitation signals should be rich enough and allow for identifying necessary higher modes in order to present the dynamics of the system accurately. Independent of the chosen model architecture and structure, the characteristics of the data determines a maximum accuracy that can be achieved by the model. For linear systems, a pseudo-random binary signal (PRBS) is the best choice for the excitation signal. For nonlinear systems, however, the PRBS signal is inappropriate [198]. For a nonlinear system, the minimum and maximum of amplitude and length of the excitation signals are essential to the identification process. The maximum and minimum of the amplitude reflect the range of process parameters over which the model should accurately predict the process. The magnitude of amplitude should also be changed around the desired points of operation. The length of excitation signals (duration) chosen should not be too small nor too large. If it is too small, the process will have no time to settle down, and the identified model will not be able to describe the static process behavior properly. On the other hand, if it is too long, only a very few operating points can be covered for a given signal length. The other concern about the data collection is noise within the data. The noise can arise from sensors or from the side eects of the other process parameters that are not included in the model. It is essential to generate a rich excitation signal for the data collection in terms of amplitude and duration to compensate the eects of noisy signals. Laser cladding is a thermal process, and for a thermal process, the response to an excitation signal is essentially slow. As a result, the minimum length for excitation signals is set to 10 s. It is experimentally tested that the process © 2005 by CRC Press LLC

response is settled down after 10 s, which indicates the required time for obtaining a steady-state response. Laser pulse energy, width, frequency, and table velocity are important excitation signals in a laser cladding process. The clad geometry and microstructure are two geometrical and physical properties that can be selected as the output signals. In this study, dierent sets of inputs/outputs are selected for each experimental-based modeling technique, which will be discussed in corresponding sections. In the following two sections, the stochastic and neural network analyses are applied to the laser cladding process and the identified models are presented.

4.8.1

Stochastic Analysis

Stochastic analysis, which is also known as system identification in engineering, is a technique to identify accurate and simplified models of complex systems from noisy time-series data. It provides tools to create mathematical models of dynamic systems based on observed input/output data. Generally, the identification procedure can be itemized as follows: 1. Design an experiment and collect input-output data from the process to be identified. 2. Examine the data and select useful portions of the original data. 3. Select and define a model structure. 4. Compute the best parameters associated with the model structure according to the input-output data and a given cost function. 5. Verify the identified model using unseen data which are not used in the identification step. If the model verification is acceptable, the desired model is identified; otherwise, Steps 3 to 5 should be repeated by another model structure or with more data. In order to explain the applications of stochastic analysis to laser cladding, two model structures are addressed in two separate case studies. In the first part, a model that relates the process speed to the clad height will be disclosed, and in the second part, a model that relates the laser pulse energy to the clad height will be identified. 4.8.1.1

Case Study 1: Correlation of Process Speed to Clad Height

In this section, it is intended to identify a model to relate the process speed to the clad height. The selection of these parameters as input and output is due to the focus of our research on the application of laser cladding to free forming and prototyping. A structure is selected and some knowledge © 2005 by CRC Press LLC

about laser cladding is incorporated into the grey-box Hammerstein-Wiener model structure. In the next section, the method for data collection and experimental setup are addressed. 4.8.1.1.1 Experimental Setup and Data Collection The experiments were performed with a 350 W Lumonics JK702 Nd:YAG pulsed laser, a 9MPCL Sulzer Metco powder feeder unit, and a CNC table. The laser power was set to 343 watts with a pulse energy of 6.86 J, width of 5 ms and frequency of 50 Hz in the experiments. The spot point was set to 5.08 mm under the focal length where the laser intensity was Gaussian. As a result, the beam diameter on the workpiece was 1.21mm. The laser beam was shrouded by Argon shield gas with a rate of 2.34e  5 m3 /s (3 SCFH). The powder feeder has a fluidized-bed powder regulating system with a consistent feed control of the materials. In the experiments, the powder feed rate was set to 2 g/min with Argon as the shield gas at a rate of 3.93e  5 m3 /s (5 SCFH). The angle of the nozzle spray was set to 55 from the horizontal line for experiments and the size of the powder stream was approximately 2 mm on the workpiece. The powders used in the experiments were pure Fe and Al powders, both with a purity of 98% on a metal basis and a mesh size of 45 µm (-325). These powders were mixed to a bulk composition of 20 w% Al before being placed in the powder feeder. Sandblasted mild steel plates (0.25 to 0.28 C; 0.6 to 1.2 Mn) with dimensions of 400 × 40 × 5 mm were selected as the substrate. Using this experimental setup, several experiments were performed to obtain data for the proposed model identification. Figures 4.30 and 4.31 depict two sets of data, which are obtained by two table velocities (sinusoidal and multistep) as shown in Figures 4.30a and 4.31a, respectively. 4.8.1.2

Model Prediction Using the Hammerstein-Wiener Structure

The Hammerstein-Wiener model is one of the structures used in nonlinear system identification. Several authors have studied the Hammerstein-Wiener nonlinear system for dierent industrial applications such as PH neutralization and distillation column [199, 200, 201]. In this case study, the Hammerstein-Wiener model structure with a more e!cient algorithm is examined for the laser cladding process. Figure 4.32 shows the model structure where f and g are the Hammerstein and Wiener memoryless nonlinear elements, respectively. The addition of the nonlinear memoryless elements allows us to incorporate our physical knowledge of the process into the model while keeping the overall model as simple as possible. In order to find the nonlinear elements of the model (f and g), we use the results reported in Romer et al. [166] and Bamberger et al. [170]. In [166, 170], the authors have shown an inverse dependency of the clad height on the square root of the relative beam velocity. They have also shown the dependency of temperature and clad height on a sigmoid function of the beam © 2005 by CRC Press LLC

1.5

Process velocity (mm/s)

1.4 1.3 1.2 1.1 1 0.9 0.8 0.7

Laser on

0.6 0.5

0

100

200

300

400

500

600

sample number (a)

700

800

900

2.5

Clad height (mm)

2

1.5

1

0.5

0

0

100

200

300

400

500

600

700

800

900

1000

sample number (b)

FIGURE 4.30 Experimental data, a) sinusoidal process speed, b) clad height.

velocity. As a result, it can be inferred that the clad height depends on at least two nonlinear functions in the form µ ¶ 4 4 h=f s , (4.56) v 4 + exp(v) This reciprocal relationship between the laser velocity and height is also evident from experimental results. Therefore, the Hammerstein-Wiener nonlinear parts of the model can be defined as: 4 f=s v and g= © 2005 by CRC Press LLC

c4 c2 + c3 exp(c4 z)

(4.57)

(4.58)

Process velocity (mm/s)

2.5 Multi Steps 2

1.5

1

0.5

0

Laser on 0

100

200

300

3

400

500

600

700

800

900

1000

500

600

700

800

900

1000

sample number (a)

Clad height (mm)

2.5

2

1.5

1

0.5

0

0

100

200

300

400

sample number (b)

FIGURE 4.31 Experimental data, a) multi-step process speed, b) clad height.

In the Hammerstein-Wiener model structure, disturbances are modeled as additive terms in the linear part and the output signals as shown by w(t) and w (t) in Figure 4.32, respectively [200, 199]. Based on these assumptions and the notation of Figure 4.32, the output of the linear block is z(t) = G(q, )u + H(q, )e(t)

(4.59)

where G(q, ) and H(q, ) are the linear models (rational functions) of the shift operator q for the system and disturbances, respectively, and e(t) is assumed to be white noise. The other parameters are shown in Figure 4.32. Removing the bias problem from Equation (4.59) (see [202] for details), it can be written as A(q)z(t) = B(q)u (t) + e(t) © 2005 by CRC Press LLC

(4.60)

Linear Model

e(t ) Hammerstein Block

u (t )

u * = f (u (t ))

f

w* (t ) w(t )

H (q,θ )

Wiener Model

z (t )

G(q,θ )

yˆ (t ) = g ( z (t ))

g

y (t )

FIGURE 4.32 Hammerstein-Wiener nonlinear structure.

where A(q) = 4 +

nf X

ai q i

(4.61)

i=4

and

(nk +nb 4)

B(q) = 4 +

X

bi q i

(4.62)

i=4

and nk , nf and nb are orders of the delay, denominator and numerator, respectively. Using Equation (4.60) and applying the operator q, z(t), Equation (4.59) can be written as z(t) = 

nf X

(nk +nb 4)

ai z(t  i) +

i=4

X

bi u (t  i) + e(t)

(4.63)

i=4

Assuming the Wiener nonlinear part in Figure 4.32 is invertible, then z(t) = g4 (b y(t))

(4.64)

Substituting Equation (4.64) into Equation (4.63) results in z(t) = 

nf X i=4

(nk +nb 4)

ai g 4 (b y (t  i)) +

X

bi u (t  i) + e(t)

(4.65)

i=4

and y(t) becomes y(t) = g(z(t)) + w (t)

(4.66)

In general, all linear and nonlinear parameters are included in the optimization procedure to minimize the output error [202]. However, implementation of this algorithm usually suers from numerical divergence. In the following, an improved algorithm is proposed to predict the model parameters. Since the Hammerstein part of the system is assumed to be known for the laser cladding process, the algorithm only identifies the linear and Wiener nonlinear parts. The steps of the algorithm are: 1. Remove the mean value from the output data y(t). © 2005 by CRC Press LLC

b 2. Ignore the Wiener block g; guess the order of linear part G.

3. Use u (t) and y(t) as the input and output; find the primary linear b model G.

4. Repeat steps P 2 and 3 by changing the order of the linear model to minimize (y(t)  z(t))2 where z(t) is the output of the linear system.

5. Find the nonlinear parameters of g(z) based on z(t) using Gauss-Newton minimization method. 6. Find z  (t) = g4 (y(t)). 7. Re-identify the linear model based on u (t) and z  (t). + %k4 )|   where k and  8. Repeat from Step 5 until |%k4 + %k2  (%k4 4 2 are the iteration index and a small positive number, respectively, and: %k4 = kk k , %k2 = kyk (t)k .

2.5

Clad height(mm)

2

1.5

1

0.5

Actual data Model prediction

0

0

100

200

300

400 500 sample

600

700

800

FIGURE 4.33 Comparison of actual data and Hammerstein-Wiener model prediction.

Code was written in MATLAB using the System Identification Toolbox to implement the above algorithm for the structure shown in Figure 4.32. This code was applied to the collected data of the laser cladding process. During the parameter estimation, it was observed that there was an optimum order for the linear subsystem such that increasing or decreasing the order resulted in higher prediction error. The overall structure of the system was relatively simple; however, the optimization algorithm was very sensitive © 2005 by CRC Press LLC

TABLE 4.5

Hammerstein-Wiener model a4 -1.0764 a6 0.0235 a2 -0.0672 a7 0.0924 a3 0.0610 b4 0.0035 a4 -0.0069 b2 0.0700 a5 0.0024 b3 -0.0265

parameters. b4 -0.0196 c4 0.1570 c2 0.0020 c3 1.0125 c4 -1.6722

to the order of linear part as well as to initial parameters of the nonlinear block. The optimum value for the order of the linear subsystem was 7 for the denominator and 4 for the numerator with a delay of 1. A sample time of 0.08 s was used in the identification process. Table 4.5 lists the estimated parameters according to Equations (4.58) and (4.65).

3

2.5 Actual data

Height(mm)

2

Model prediction

1.5

1

0.5

0

0

100

200

300

400

500 sample

600

700

800

900

1000

FIGURE 4.34 Verification of Hammerstein-Wiener model.

Figure 4.33 compares the experimental data with the model predictions when the sinusoidal data shown in Figure 4.30 are used. As seen in Figure 4.33, good agreement between the model and experimental results is achieved. Because of the eects of other involved parameters such as instability of powder feeder spray, dependency of the beam reflectivity and focal point to the clad height, there are some discrepancies between the predicted and actual data. To verify the identified model, the multistep response shown in Figure 4.31a was applied to the estimated model. The simulation and actual data are compared in Figure 4.34. To evaluate the eect of the Wiener nonlinear block © 2005 by CRC Press LLC

on the overall system output, the results of the model without the Wiener structure are compared with the experimental data as shown in Figure 4.35. As seen in the figure, the nonlinear Wiener block has significantly improved the identified model.

3 Actaul data 2.5

Model output without Wiener function

Height(mm)

2

1.5

1

0.5

0

0

100

200

300

400

500 sample

600

700

800

900

1000

FIGURE 4.35 Eect of elimination of Wiener function on model prediction.

The identified model using the Hammerstein-Wiener model structure oers a simple and relatively accurate model for the laser cladding process. The sluggish nature as well as large settling time associated with laser cladding can be predicted very well by the model. This model will be used for designing a controller in Chapter 5. 4.8.1.3

Case Study 2: Correlation of Laser Pulse Energy to Clad Height

In the second part of the stochastic analysis, it is intended to identify a model to relate the laser pulse energy to the clad height. Experimental analysis shows that the energy has a linear relationship with the clad height about a desired operating point. Since there are many non-linear uncertainties in the process, it is essential to select an operating point and consider only small variations of the input signal around this operating point. The choice of operating point is determined by the quality of clad as a constraint in the process identification. This issue will be addressed in Chapter 6. Providing the above-mentioned requirements, the identification of a model that reflects the dynamics of the process due to the changes in the laser pulse energy can be carried out using a classic ARX method [202], which will © 2005 by CRC Press LLC

be discussed later. In the following section, the experimental setup and the selected data for the identification process will be addressed.

Laser pulse energy (J)

4.8.1.3.1 Experimental Setup and Data Collection The experiments were performed with a LASAG FLS 1042N Nd:YAG pulsed laser with a maximum of 1000 W power, a 9MP-CL Sulzer Metco powder feeder unit, and a 4 axis CNC table. The spot point on the workpiece was 5.08 mm under the focal point with a diameter of 1.4 mm where the laser intensity was Gaussian. The laser beam was shrouded by Argon shield gas with a rate of 2.34e  5 m3 /s (3 SCFH). In the experiments, the powder feed rate was set to 1 g/min with Argon as the shield gas at a rate of 2.34e  5 m3 /s (3 SCFH). The angle of the nozzle spray was set to 55 from the horizontal line for experiments and the size of powder stream profile was approximately 1.4 mm on the workpiece. The powders used in the experiments were pure Fe and Al powders both with a purity of 98% on a metal basis and a mesh size of 45 µm (-325). These powders were mixed to a bulk composition of 20 w% Al before being placed in the powder feeder. Sandblasted mild steel plates (0.25 to 0.28 C; 0.6 to 1.2 Mn) with dimensions of 100 × 10 × 5 mm were selected as the substrate. The clad height was measured by the device discussed in Chapter 3. Several sets of PRBS pulse energy signals around 3.5 J were applied to the apparatus as shown in Figures 4.36 and 4.37 when U = 1.5 mm/s, m ˙ = 1 g/min, F = 96 Hz and W = 3 ms.

4 3

Laser pulse frequency= 96 Hz Laser pulse width =3 ms Process speed=1.5 mm/s

2 1 0

0

10

20

30

40

50

60

70

80

90

Time (s)

a)

Clad height (mm)

1

0.5

0

0

10

20

30

40

50

60

70

Time (s)

b)

FIGURE 4.36 Experimental data, a) random laser pulse energy, b) clad height. © 2005 by CRC Press LLC

80

90

Laser pulse energy (J)

4 3

Laser pulse frequency= 96 Hz Laser pulse width =3 ms Process speed=1.5 mm/s

2 1 0

0

10

20

30

40

50

60

70

80

90

100

60

70

80

90

100

Time (s)

a)

Clad height (mm)

1

0.5

0

0

10

20

30

40

50

Time (s)

b)

FIGURE 4.37 Experimental results for random excitation signal, a) laser pulse energy, b) clad height.

4.8.1.3.2 ARX Model The auto regressive exogenous (ARX) system identification is one of the structures used in linear system identification. It is the most popular model structure, which describes the error by means of white noise [202]. A simple input-output model that can be considered for a linear, time-variant, discrete-time and single-input single-output (SISO) system (see Figure 4.38) is D(q)y(t) = C(q)u(t) + e(t) (4.67) where y(t) and u(t) are output parameter and input, respectively, e(t) is the white noise, D(q) and C(q) are functions of shift operator q in the discrete space and are denominator and numerator, respectively. These functions can be presented by C(q) = c1 q 1 + .........cnb q nc D(q) = 1 + d1 q 1 + .........dna q nd

(4.68) (4.69)

where ci and di are the coe!cients of the polynomials, nc and nd are the order of C(q) and D(q), respectively. Code was developed in MATLAB using its System Identification Toolbox to implement the ARX model structure shown in Figure 4.38. This code was applied to the collected data of the laser cladding process. During the parameter estimation, it was observed that an optimum order for the linear © 2005 by CRC Press LLC

e(t )

u

y

1 D

∑

C

FIGURE 4.38 ARX model structure. TABLE 4.6

ARX model parameters. d1 -0.5807 d2 -0.2305 c1 0.03834

system was 2 for the denominator D(q) and 1 for the numerator C(q). Table 4.6 lists the estimated parameters according to Equations (4.68) and (4.69). Several dierent orders were checked to investigate their eects on the model. However, results showed that increasing the order of the linear model does not reduce the overall error. As a result, the above-mentioned orders were selected for the model identification.

Clad height (mm)

1 0.8 0.6 0.4

Model output

0.2

Actual data

0

0

10

20

30

40

50

60

70

80

90

Time (s)

FIGURE 4.39 Comparison between the experimental and modeling results.

Figure 4.39 compares the experimental data with the model predictions when the PRBS shown in Figure 4.36a was used. As seen in Figure 4.39, good agreement between the model and experimental results is achieved. To verify the identified model, a set of unused data in the identification procedure, which was generated by a dierent PRBS shown in Figure 4.37a, was fed to the identified model. The simulation and actual data are compared © 2005 by CRC Press LLC

Clad height (mm)

1 0.8 0.6 0.4

Model output

0.2

Actual data

0

0

10

20

30

40

50

60

70

80

90

100

Time (s)

FIGURE 4.40 Verification of the ARX model structure.

in Figure 4.40. As shown, good agreement has been achieved. Several dierent models with the same orders and delay are identified around dierent operating points using the same strategy. These models are then used for controller design and also as a base for tuning the controller gains in the virtual atmosphere as will be discussed in Chapter 5.

4.8.2

Artificial Neural Network Modeling

Artificial neural networks have gained prominence recently among researchers of manufacturing technologies. As the name implies, these networks are computer models of the process that constitute biological nervous systems. In recent years, there have been continuous publication of studies focusing on the improvement of the artificial neural network techniques. Many researchers have applied the artificial neural network to dierent processes to obtain their precise models. However, to the best knowledge of the author, there is no article on the application of neural networks to laser cladding to be used in development of an expert system. The most important characteristics of neural networks are: made from a large number of perceptron units, strongly connected units, robustness against the failure of single units, and learning from data. Although neural networks can be trained to solve problems that are di!cult for conventional computers or humans, selecting an appropriate model structure for a particular problem is still a big challenge. This challenge is more noticeable in dynamic systems where memory elements are required for accurate system modeling. Neural networks as shown in Figure 4.41 are basically composed of simple elements called perceptrons. Each perceptron has multiple inputs with processing elements including a summation and an activation unit. The network’s function is determined largely by the connections between the processing elements. The signals flowing between the perceptrons are scaled by adjustable parameters called weights. A neural network can be trained to perform a particular function by adjusting the weights. Usually, neural networks are adjusted or trained so that a particular set of inputs leads to a specific set of target outputs. Typically, © 2005 by CRC Press LLC

Hidden layer

Output layer

x1

1

1

y1

x2

2

2

y2

. .

. .

K

M

. .

xp

x2

ym

x1

xp

∑ ∫ Perceptron

FIGURE 4.41 A multilayer perceptron with one hidden layer.

many input/output pairs are used to train a network. One of the methods used to train a neural network is backpropagation given by Principe et al. [203]. 4.8.2.1

Neural Network Analysis for Laser Cladding — Case Study

In this study, a recurrent neural network structure is applied to laser cladding by powder injection to predict the clad height and rate of solidification. The prediction, simulation, optimization, and verification are performed for the process identification. For identification, the process velocity and laser pulse shaping including laser pulse energy, laser pulse frequency and laser pulse width as the inputs, and the clad height and rate of solidification as the outputs are used, which are shown in Figure 4.42. Of interest is the fact that the rate of solidification has a direct relationship with clad microstructure [204] and can be considered a measure for microstructure evaluation. In the next section, the experimental setup and data collection will be addressed. © 2005 by CRC Press LLC

Laser Pulse Energy Laser Pulse Width Laser Pulse Frequency

Clad Height

Recurrent Neural Network

Rate of Solidification

Process Velocity

FIGURE 4.42 Inputs and outputs to the recurrent neural network.

4.8.2.1.1 Experimental Setup and Data Collection The experiments were performed with the same experimental setup discussed in Section 4.8.1.3.1. However, in order to develop a comprehensive model to reflect the influence of essential process parameters, in addition to those experiments shown in Figures 4.36 and 4.37, numerous other experiments were performed to create a robust model. A large set of data with dierent specifications was obtained in dierent working conditions. For instance, the laser pulse energy was increased in several steps as shown in Figure 4.43a, while the other process parameters were held constant as shown in the figures. Figures 4.43b, and 4.43c also show the output results, which are clad height and rate of solidification versus time for any of the multistep laser pulse energy excitation signals. In the next section, the selected structure for the model identification using the artificial neural network will be discussed, and the results of identification will be presented. 4.8.2.1.2 Elman Recurrent Neural Network Model Architecture A recurrent neural network as shown in Figure 4.44 is a particular form of neural network model that has a feedback signal in the network architecture. This feedback enables the network to be used in problems and applications that require state representation such as speech processing, plant control, adaptive signal processing, time series prediction, and so on. The universal approximation capabilities of the recurrent multilayer perceptron make it a popular choice for modeling nonlinear dynamic systems and implementing general-purpose nonlinear controllers [205]. There are dierent forms of recurrent neural networks such as Elman and Hopfield networks. The Elman network is a two-layer network with feedback from the first output layer to the first input layer [206] as shown in Figure 4.44. In the figure, u and y are the input and output matrices, respectively, IW are the network’s weights, LW is the feedback weight, D is the delay and b is the bias matrix. This recurrent connection allows the Elman network to both detect and generate time-varying patterns for approximating any dynamic systems with an arbitrary accuracy. The only requirement is that the hidden © 2005 by CRC Press LLC

Laser pulse energy(J)

5 4 3

Laser pulse frequency =100 Hz Laser pulse width = 3 ms Process velocity= 1.5 mm/s

2 1 0

0

20

40

60

Time (s)

80

100

120

80

100

120

80

100

120

a)

Clad height (mm)

1.5 1

0.5 0

0

20

40

60

Time (s)

Rate of Solidification (mm/s)

b) 1.5 1

0.5 0

0

20

40

60 Time (s)

c)

FIGURE 4.43 Step excitation signal showing: a) laser energy vs. time, b) clad height vs. time, c) rate of solidification vs. time.

layer must have a suitable number of neurons. Although more hidden neurons are needed as the complexity of the function’s being fit increases, increasing the number of hidden neurons causes noise identification instead of process identification in a noisy environment [207]. For the laser cladding process model prediction, an Elman recurrent neural network as shown in Figure 4.44 was implemented. The delay D was selected as 2 samples due to the observed delay between the excitation signal and the real process response. The number of neurons were 12 and 2 for hidden and output layers, respectively. The significance of the number of neurons has a direct relationship with training performance for noisy data. Selecting less than 12 neurons causes the network to be inaccurate in prediction, while higher number of neurons causes the network to model noise instead of the process dynamics. When more neurons are used, the training method tries to fit the model into the noisy data by adjusting the weights and biases, and, as a result, the identified model cannot predict the process very well. © 2005 by CRC Press LLC

Output Layer

Hidden Layer D

Input Layer LW11

u

IW21

IW11

1

b1

tansig

1

b2

y

linear

FIGURE 4.44 Elman recurrent neural network.

To train the network, backpropagation through the time technique [203] with 4, 309 sets of input/output data were used. Figure 4.45 shows the training performance. As seen, the network converged to 0.0422 after 14 epochs, which is the sequence of training. However, due to the noise within the data, a better performance cannot be achieved. It seems that there is a minimum point in the selected data which causes the training to be stuck at that local minimum. The training results in the weight matrices are listed in Table 4.7. After training, the network was simulated. The same inputs as shown in Figures 4.43a and 4.36a were initially fed into the trained network. The simulation results for the clad heights are shown in Figures 4.46 and 4.47. Also, Figure 4.48 illustrates the comparison between the rate of solidification obtained from the model and process. As seen, the trained network is able to predict the outputs relatively well. In order to validate the identified model, several sets of unseen and new data in the training procedure were fed into the network. As seen in Figure 4.49a, a multistep laser pulse frequency is fed into the network when the laser pulse energy, laser pulse width and process velocity are fixed. Figures 4.49b and 4.49c show comparison of the model output with the actual data. The figures indicate good agreement between the model and actual data. Verification confirms that the model has a good potential to predict the transient behavior of the process due to the suitable neural network structure and a large variety within the data fed into the model.

© 2005 by CRC Press LLC

TABLE 4.7

The RNN matrices of layers. LW 11 0 .5 7 8

0 .1 7 3 1

-0 .2 4 7

-0 .7 5 7

0 .5 1 3

0 .5 5 0

0 .3 3 2

-0 .5 7 5

-0 .2 0 0

0 .1 5 2

-0 .0 0 2

-0 .7 5 3

0 .4 0 3

-0 .6 4 2

-0 .1 4 7

0 .2 4 6

-0 .5 0 0

0 .3 8 9

- 0 .4 0 7

0 .6 5 9

-0 .6 3 8

0 .6 7 2

-0 .3 1 3

0 .4 7 8

-0 .1 2 2

-0 .4 0 6

-0 .4 2 0

0 .4 9 0

-0 .2 3 1

0 .9 8 3

- 0 .1 1 0

0 .4 4 1

-0 .6 5 6

-0 .4 6 1

0 .3 7 7

0 .2 6 4

0 .2 0 0

0 .1 8 0

-0 .1 4 7

-0 .3 6 1

0 .0 3 4

0 .0 9 6

- 0 .5 4 2

0 .5 7 5

-0 .7 4 6

0 .6 4 4

0 .7 5 8

0 .0 9 9

0 .6 7 0

-0 .4 8 2

-0 .3 1 8

-0 .3 5 3

0 .3 2 3

-0 .7 2 1

0 .6 9 5

-0 .4 3 1

0 .0 9 9 0

-0 .5 9 1

0 .3 9 5

-0 .3 7 9

0 .3 0 5

0 .2 6 3

-0 .2 3 1

0 .4 5 4

0 .1 5 1

0 .2 1 0

- 0 .3 1 4

-0 .0 9 7

-0 .8 2 4

-0 .9 4 8

0 .3 6 2

0 .7 0 2

-0 .4 1 5

-0 .4 1 8

0 .0 0 5

0 .4 6 7

-0 .0 6 4

0 .5 2 0

- 0 .7 4 0

-0 .6 8 8

0 .0 3 6 3

-0 .4 3 7

-0 .6 0 6

-0 .3 8 9

-0 .2 1 0

0 .1 2 5

0 .3 1 7

0 .6 9 4

-0 .0 7 8

0 .6 8 1

0 .3 6 4

0 .5 0 1 6

-0 .5 4 7

0 .6 1 9

-0 .5 7 8

0 .3 6 3

-0 .4 8 7

0 .0 1 0

-0 .3 0 2

-0 .0 4 1

-0 .6 4 2

-0 .4 3 3

0 .6 1 6

0 .0 9 7

0 .4 2 0

-0 .6 8 1

-0 .7 5 6

0 .2 4 8

-0 .0 2 0

-0 .0 7 4

-0 .8 2 3

0 .8 5 5

-0 .1 2 0

0 .4 3 2

- 0 .4 5 4

-0 .3 8 3

-0 .2 6 5

-0 .5 0 0

-0 .4 4 9

-0 .6 0 7

-0 .1 7 3

0 .0 8 0

-0 .1 0 9

-0 .0 9 3

-0 .2 5 5

0 .0 4 2

- 0 .4 7 6

-0 .2 4 0

0 .6 1 9

0 .9 5 6

0 .6 0 6

0 .1 9 5

-0 .0 5 0

0 .6 1 2

-0 .0 4 6

0 .4 2 1

-0 .2 7 3

0 .5 9 9

0 .5 9 9

0 .5 0 9

-0 .6 2 7

-0 .3 9 9

0 .6 7 1

0 .1 4 5

0 .2 3 3

-0 .0 0 2

-0 .0 0 4

-0 .1 4 6

- 2 .1 8

-0 .6 0 9

0 .0 7 0

-0 .0 0 3

-0 .0 0 6

-0 .1 9 0

- 1 .5 5 3

-0 .6 2 8

0 .1 6 8

-0 .0 0 9

-0 .0 1 8

0 .5 9 2

- 1 .5 2 1

-0 .5 2 1

-0 .0 0 3

-0 .0 0 5

-0 .1 7 8

1 .4 4 3

-0 .1 4 4

.0 0 3 1

0 .0 6 2

0 .2 1 0

0 .5 7 5

-0 .0 7 9

0 .0 0 6

0 .0 1 2

0 .3 9 0

0 .4 3 0

0 .7 3 7

-0 .0 0 3

-0 .0 0 7

-0 .2 4 1

0 .3 9 2

0 .1 6 6

-0 .0 0 3

-0 .0 0 6

-0 .1 9 1

- 0 .0 1 6

0 .1 4 6

0 .0 0 3

0 .0 0 5

0 .1 6 6

0 .2 4 4

-0 .0 4 6

-0 .0 1 6

-0 .0 2 1

-0 .7 0 4

- 0 .9 8 2

0 .5 1 2

0 .0 0 7

0 .0 1 4

0 .4 6 9

0 .5 1 8

-0 .2 4 7

0 .0 0 3

0 .0 0 5

0 .1 8 5

IW 11

b1

b2

- 1 .2 0 4

IW 21 0 .0 7 9

0 .0 3 2

0 .7 9 1

0 .2 0 0

0 .9 8 2

-0 .1 5 3

- 1 .0 7 7

0 .1 5 0 1

-0 .6 3 4

-0 .7 8 0

0 .1 3 3

0 .3 5 7

-0 .6 1 6

-0 .6 4 3

0 .3 2 0

0 .4 1 0

-0 .9 9 9

0 .5 2 3

- 0 .6 5 3

-0 .7 9 3

-0 .9 6 5

0 .0 4 6

0 .4 3 3

-0 .0 2 9

© 2005 by CRC Press LLC

FIGURE 4.45 Training performance.

Clad height (mm)

1.5 1 0.5 0

RNN output Experiment 0

20

40

60

80

100

120

Time (s)

FIGURE 4.46

Clad height (mm)

Comparison between model and experiments for the clad height for the multistep laser pulse energy excitation.

1

0.5

0

RNN output Experiment 0

10

20

30

40

50

60

70

80

90

Time (s)

FIGURE 4.47 Comparison between model and experiments for the clad height for random laser pulse energy excitation. © 2005 by CRC Press LLC

Rate of Solidification (mm/s)

1 0.8 0.6 0.4

RNN output Experiment

0.2 0

20

40

60

80

100

120

Time (s)

FIGURE 4.48

Laser Pulse Frequency (Hz)

Comparison between model and experiments for the rate of solidification for the multistep laser pulse energy excitation

100 80 60

Laser pulse energy=5 J Laser pulse width= 3 ms Process velocity= 1.25 mm/s

40 20 0

0

20

40

60

80

100

120

Time (s)

Clad Height (mm)

a) 1.5 1

RNN output Experiment

0.5 0

0

20

40

60

80

100

120

Rate of solidification (mm/s)

Time (s)

b) 1.5 1

RNN output Experiment

0.5 0

0

20

40

60

80

100

120

Time (s)

c)

FIGURE 4.49 Data for verification: a) Multistep laser pulse frequency, b) Verification of model for the clad height for multistep laser pulse frequency, c) Verification of model for the rate of solidification for multistep laser pulse frequency. © 2005 by CRC Press LLC

5 Control of Laser Cladding Process

In spite of the many advantages of laser cladding, its industrial applications are still limited. This is not only because of the relatively high cost of laser systems, but also the high sensitivity of this process to disturbances. A small change in absorbed laser power, for example, can cause a large change in the melt pool size. Also, changes in the mass flow of the powder can produce significant variations in the overall geometry and microstructure of the clad. To make the process more stable and less susceptible to disturbances, it is important to understand the eects of involved parameters in the process, and clad quality and dimensions. In Figure 2.7, the important parameters that determine the clad quality are listed under inputs and process columns. These parameters can be categorized into two groups: • Intrinsic • Extrinsic Intrinsic parameters are those related to the substrate and powder properties. Some of these parameters include absorptivity, thermal conductivity, heat capacity, thermal diusivity, and workpiece geometry. Extrinsic parameters are those related to the hardware used in the process such as laser, powder feeder, and positioning system. Some of these parameters include laser average energy, laser focal point, powder feeder mass flow, nozzle position and orientation, and relative position and velocity of the laser and substrate. In general, there is no direct control on intrinsic parameters; however, the eects of changes in intrinsic parameters can be compensated by controlling the extrinsic parameters. For example, change in the absorptivity can be adjusted by controlling the laser average power. The control of extrinsic parameters, on the other hand, is relatively easy for most of the parameters mentioned above. Commercial lasers, powder feeders, and positioning systems are equipped with built-in controllers that allow the user to set the desired values. This is in fact the approach that is used in many applications. After much trial and error, the parameters that result in a good quality clad or coating for a specific application are obtained. These parameters such as laser average power, focal point, speed, and powder feed rate are then used in an open-loop process control. © 2005 by CRC Press LLC

In Figure 5.1, the block diagram of the laser cladding process in both open and closed-loop is shown. The box indicating the open-loop process shows that the controllable parameters are set to predefined values and there is no feedback from the process output (clad) to adjust the parameters. In this method, any disturbance and error in the equipment outputs such as laser power, focal point, and velocity will aect the process output. Similarly, there is no mechanism to adjust the machines to compensate for the changes in the clad quality. In a closed-loop control of the process, it is important to evaluate the actual clad in real time in order to compare it with the desired clad. In Figure 5.1, the measurable parameters that directly or indirectly can be used to determine the clad quality are shown. Direct parameters include the clad dimensions and roughness that can be measured using a camera. Indirect parameters include melt pool temperature and size that indirectly indicate the clad quality. The clad microstructure and rate of solidification are other indirect parameters that can be used to determine the state of the clad. Having a measure of the clad condition and comparing this condition with the desired one, a controller can be used to change the controllable parameters on the fly to compensate for any disturbance or change in the process. It is important to note that all the controllable parameters are not independent and only some of them are needed to achieve a closed-loop control. For example, laser power and velocity are correlated, and only one of them is needed to change the eective power injected to a volume of substrate. In general, open-loop control of laser cladding is useful when the application is fixed and is repeated over and over. Finding the right parameters is very time consuming; however, as long as there is no change in the application this method of control is eective and e!cient. The advantages of this method are: Cost Eectiveness: In this method, there is no need for additional costs for feedback sensors and a central controller to adjust the parameters on the fly to improve the clad quality. Ease of Implementation: Since the devices used in the process have their own controllers, the parameters can be directly set on the machines. The disadvantages of the open-loop control are: Sensitivity: As mentioned earlier, laser cladding is very susceptible to changes in parameters. As a result, any variation in both intrinsic or extrinsic parameters results in a large deviation in the final product. Tooling Time: For any new application and material, a significant amount of time is needed to arrive at a suitable set of parameters. This may not be feasible in many situations such as prototyping and low rate part fabrication. © 2005 by CRC Press LLC

FIGURE 5.1

Open and closed-loop control of laser cladding.

© 2005 by CRC Press LLC

Desired Clad Properties

∑

- Type of Shield Gas - Powder Preheating

- Shield Gas Flow

- Other Parameters

- Inert Gas Flow Rate

- Powder Feed Rate

- Powder Feeder

- Velocity - Nozzle Location and Orientation

- Laser/substrate Position

- Positioning System

- Average Power - Focal Point - Laser Pulse Shaping - Shutter Speed

- Laser

Controllable Parameters (Inputs)

Open-Loop Process

Laser Cladding Process

- Surface Roughness - Microstructure (indirectly)

- Clad Quality

- Rate of Solidification

- Temperature - Size

- Melt Pool

- Height - Width

- Clad Geometry

Measurable Parameters (Feedback Signals)

Actual Clad

To overcome the shortcomings of the open-loop systems, many studies have been directed at developing closed-loop control systems. In the following sections, dierent sensors for measuring the above-mentioned parameters are first studied. Current research in closed-loop control of laser cladding is then reviewed. In the end, a successful closed-loop control of laser cladding is discussed in detail.

5.1

Sensors

The parameters that can be used as feedback signals in a closed-loop control of the laser cladding process are shown in Figure 5.1 as the measurable parameters. The measurement of these parameters can directly or indirectly indicate the clad quality. When the correlation of these parameters with the input parameters are found, a controller can be designed to close the control loop as shown in Figure 5.1. In general, measurable parameters related to the melt pool are temperature, size, rate of solidification. Measurable parameters associated with the actual clad size are height and width. The roughness of clad surface and associated microstructure can be indirectly estimated through the measurement of the clad geometry and rate of solidification. The sensors that are used for these measurements can be categorized into temperature and dimension sensors as follows. 5.1.0.2

Temperature Measurement

For temperature measurement of the melt pool, a non-contact sensor is needed. Most non-contact temperature sensors measure the thermal radiant power of the infrared or optical radiation of the object and are categorized as the following: • Radiation Thermometers (Pyrometers): These sets of sensors are based on Planck’s law of the thermal emission of electromagnetic radiation [208]. This group of non-contact temperature sensors includes pyrometers, infrared thermal imaging cameras (with temperature measurement capability), line-measuring thermometers (so-called line scanners), and infrared radiation thermometers. These sensors estimate the temperature from the amount of thermal electromagnetic radiation received from the object. This group of sensors includes both spot or point measuring devices in addition to line measuring radiation thermometers, which produce 1-D and, with known relative motion, can produce 2-D temperature distributions. In thermal imaging, measurement is done over an area from which the resulting image can be displayed as a 2-D © 2005 by CRC Press LLC

temperature map of the scanned area. In laser material processing, Smurov et al. [209] use a radiation pyrometer to measure the melt pool temperature. Several other authors apply a thermal camera to measure the temperature distribution of the process interaction zone [210, 211, 171]. The data from a thermal camera can be evaluated in terms of maximum temperature and its location, and temperature gradient calculations. The camera can also be used to determine the melt pool shape and area. Since there is a correlation between the melt pool area and the melt pool depth, the thickness of the resulting layer may be calculated indirectly. • Optical Pyrometers: These sensors are also based on Planck’s law of radiation. The main part of the sensor is a lamp filament, which is operated at a constant current. Brightness variation is caused by a variable density filter that changes the apparent filament brightness in the operator’s view when compared to the unchanging brightness of the object being measured [212]. Several authors [101, 213] applied optical pyrometers to measure laser power during the process. • Acoustic and Ultrasonic Pyrometers: The concept of measuring temperature of the melt pool by measuring the speed of sound in the process zone has been exploited by several groups [214, 101]. In this technique, the sound’s flight time, which is a function of temperature, is measured in the process zone. The sound’s flight time results in the temperature of the process zone. Acoustic pyrometers suer from setbacks such as low sensitivity and noise rejection. • Thermocouples: Thermocouples are used extensively in measuring temperatures in dierent industrial applications. Tungsten-rhenium and platinum-rhodium thermocouples are utilized for measuring temperature distributions during the laser cladding process [215]. The upper limits for the temperature measurement of tungsten-rhenium and platinum-rhodium thermocouples are 2400 and 1480 o C, respectively. Since thermocouples are contact-based sensors, they are not very suitable for the process because of the disturbances they develop in the melt pool. 5.1.0.3

Dimension Measurement

For the clad geometry measurement, cameras and phototransistors are the most common devices. The challenges encountered using these devices in the laser cladding process are contamination of the images with plasma, powder particles, etc., and also processing time of the images in real time to extract the clad geometry. Meriaudeau et al. [210] used a CCD camera to measure the height of a clad and optimize the process by averaging the height of the clad. However, their © 2005 by CRC Press LLC

work does not indicate the accuracy and the rate of measurements and also its application to a closed-loop control. Kinsman et al. [216] and Duley et al. [217] showed how a vision system can be used to monitor the process zone for laser material processing. In their technique, the size of the melt pool is determined by measuring the number of bright pixels in the images. Mazumder et al. [14, 218, 219] and Koch et al. [93] disclosed the usage of a phototransistor for process monitoring of the laser cladding process. The phototransistor is used to signal clad height deviation for a desired threshold. Another attempt at monitoring the process has been done by Stegemann et al. [220]. They combined microfocus-radioscopy with a high-speed video camera in order to observe the mass flow within the melt pool. In general, there is a little work in direct measurement of the clad dimensions and the quality in real time, although it has a great potential in automated laser cladding. In Section 5.3, a successful vision-based technique for measuring both clad geometry and the clad quality are introduced.

5.2

Closed-Loop Control of Laser Cladding

As seen in Figure 5.1, there are several parameters that can be used in controlling the laser cladding. The choice of a parameter for controlling the process largely depends on the type of equipment used. For example, a Nd:YAG laser compared to CO2 laser provides more flexibility in terms of controlling the input power, or a positioning system has a better bandwidth compared to a laser for controlling the average power delivered to a substrate. Developing a closed-loop control of the laser cladding involves the selection of a control input, the measurement of clad properties for the feedback signal, and finally a controller that relates the error between the desired and measured clad property to the control input. In practice, the parameters most commonly used as the control input and the feedback signal are the laser average power and melt pool temperature, respectively. In Nd:YAG lasers, the average power can be changed relatively quickly (usually on the order of some 1/10 of a second); however, in CO2 lasers, due to their large time constants the average power is changed by opening and closing a shutter. The shutter can be operated at a maximum of 10-20 Hz. Another approach in controlling the power is to change the laser focal point with respect to the substrate in which the input power density to the melt pool is changed. This technique, however, may result in a variable clad width, especially in cladding. According to Figure 5.1, a multi-input multi-output (MIMO) closed-loop control system may seem more appropriate for the laser cladding process. Due © 2005 by CRC Press LLC

Laser Beam Powder Container

Optic Lens CCD Camera or Pyrometer

Central Controller Positioning Device

FIGURE 5.2 Closed-loop control of laser cladding.

to the complexity of the process and also the fact that the input parameters are not independent (similarly the feedback signals), there is no report for a MIMO control strategy for the laser cladding process. For other laser material processing, few authors attempted the MIMO approach. For example, Bataille et al. [178] used both the laser power and process speed during laser surface hardening. Figure 5.2 shows the overall structure of a closed-loop control system of the laser cladding process. In a single input-single output (SISO) control system, the powder feeder and positioning system are not in the main process control loop and they set independently to predefined powder rate and specified path motion. Instead, the laser power is adjusted using direct control of the laser average power (in Nd:YAG lasers) or pulse modulation of the laser beam via a shutter (in CO2 lasers) through the feedback from the melt pool temperature or geometry using a pyrometer or a camera. Using the laser focal point technique, Morgan et al. [221] controlled the temperature of the melt pool via positioning the laser spot point relative to the workpiece. They performed experiments to demonstrate the eectiveness of closed—loop over open—loop control. Li et al. [222, 140] developed a real-time laser cladding control system to determine the optimal operating conditions for a given requirement and for online fault diagnosis and correction. Development of closed-loop control systems for laser material processing © 2005 by CRC Press LLC

has been performed by several researchers. Kinsman et al. [216] and Duley et al. [217] disclosed a fuzzy logic controller for manipulating the laser processing variables such as laser power, laser intensity and laser beam velocity to control the penetration depth of welding. Mazumder et al. [14, 218, 219] and Koch et al. [93] disclosed a feedback controller for adjusting the laser power based on the presence or absence of the laser beam from the process zone. This controller trims the control analog voltage, which is sent to the laser based on the TTL signal received from phototransistor. The modified analog signal sent to the laser causes the laser beam to be on and o for specific durations.

5.3

Closed-Loop Control of Laser Cladding, An Example

In the following, the details of a successful closed-loop control of the laser cladding process is discussed. Also, experimental results are presented to compare the open-loop with the closed-loop control approach in laser cladding.

5.3.1

Equipment and Configuration

Figure 5.3 shows the schematic of the integrated system. As seen in the figure, the system comprises: 1. Pulsed Nd:YAG laser, 2. Powder feeder machine, 3. Positioning device, 4. Central process controller consisting of frame grabber, image processing and pattern recognition software, CAD/CAM software, and interfaces, 5. Optical CCD-based detector, 6. Nozzles for powder injection, 7. Processing head, including optic system and shield gas delivery system, 8. Frame grabber, 9. Powder delivery tube, 10. Fiber optic, 11. Halogen light, 12. Substrate. © 2005 by CRC Press LLC

FIGURE 5.3 Schematic of the developed system.

Referring to Figure 5.3, the system includes a laser power source ° 1 , which 10 for transferring is connected to a fiber optic cable with 1.6 mm diameter ° the beam to a processing head °. 7 The laser has an RS232 serial port for sending/receiving the information to/from the main controller °. 4 12 move by a positioning The processing head ° 7 and/or the substrate ° device °. 3 The processing head ° 7 is connected to a shield gas such as argon and is integrated with the laser optical system and protective ceramic head. The spray nozzle ° 6 delivers a continuous powder stream through a flexible tube with 3 mm internal diameter °. 9 The powder stream with the desired powder feed rate and shield gas rate is provided by the powder feeder machine °. 2 The powder spray from the nozzle is conducted toward the intersection of 12 and, as a result, the powder particles and the laser beam and the substrate °, a thin layer of substrate are melted. Due to the metallurgical fusion between the deposited layer and the substrate a strong and uniform layer is built up on the substrate. The relative position and orientation of the processing head and workpiece are commanded through the main process controller ° 4 via an RS232 serial port to control the clad location and also the laser focal point. The motion controller is able to adjust the relative velocity between the process spot and © 2005 by CRC Press LLC

workpiece. Also, it traces the desired clad layer based on the CAD model of the workpiece available in the main controller °. 4 The CAD solid model of the desired object is sliced into many layers by the adaptive slicing technology, and the location, height and width of each layer are used by the main process controller to place and orient the workpiece and process head. The optical CCD-based detectors ° 5 monitor the processing zone illumi11 The number of optical detectors can be two or nated by a halogen light °. three based on the desired object and the layer pattern. Figures 5.4 and 5.5 show the two pictures of the system.

FIGURE 5.4 Integrated system.

5.3.2

Optical CCD-based Detector

14 is used to Referring to Figure 5.6, a charge coupled device (CCD) camera ° provide images from the process zone. The images are then grabbed by a frame grabber ° 8 as shown in Figure 5.3. In the process zone, the high temperature of the melt pool emits light with dierent intensities and wavelengths. In addition, plasma and vaporized metal illuminate the melt pool. In order to filter undesired wavelengths, an interference filter with bandwidth of 500—700 15 plus a neutral filter ° 16 are used. Magnification lenses ° 17 are used to nm ° magnify the process zone. Referring to Figure 5.3, the frame grabber ° 8 receives the images from the © 2005 by CRC Press LLC

FIGURE 5.5 A close-up of the positioning device, processing head, nozzle, machine bed and optical CCD-based detector.

FIGURE 5.6 Optical CCD-based detector.

© 2005 by CRC Press LLC

optical CCD-based detector °. 5 The size of original images is 640×480 pixels. To lessen the computational time, a preprocessing procedure is performed inside the frame grabber. The preprocessing reduces the images and limits them to the location of the center line of the laser beam in the right hand side of images and to the surface of the substrate in the lower part of images. Also, in the preprocessing procedure the images’ brightness and contrast are decreased and increased by 10%, respectively. Figure 5.7a shows a typical image after preprocessing. The preprocessed images are then fed into an algorithm developed for image processing and pattern recognition in the main controller. The software processes the images and finds the dimensions, roughness and the rate of solidification in real-time. The developed algorithm comprises: 1. Change an RGB image to a gray level 2. Create a threshold binary image using a threshold level which is computed by a global image threshold using Otsu’s method applied on every tenth frame [223] 3. Extract the dimensions and the angle of the solid/liquid interface by finding the border of the white object in the image matrix as shown in Figure 5.7b. The clad’s height can be calculated based on the number of bright pixels in the desired column of the matrix Also, the angle is obtained from the angle between the border of the bright area and a reference line. Laser beam centerline

1 mm

α Substrate surface

FIGURE 5.7 a) Typical image from a normal view to process zone, b) processed image. © 2005 by CRC Press LLC

h

The above-mentioned preprocessing and pattern recognition algorithm are also shown in Figure 5.8 using block diagrams. Experiments show that the height of the melt pool somewhat away from the laser centerline gives the most precise height. Extracting the height and the angle using the algorithm provides valuable information and insight into the process. The clad’s roughness and surface quality can be obtained by analyzing the fluctuation of the clad’s height obtained in Step 3. Large fluctuations in the clad’s height indicate a rough surface clad while lower fluctuation is an indication of a high quality clad with smooth surface finish. As reported by Gilgien et al. [204] and Pie et al. [37], the measured angle can be used to determine the rate of solidification. Using the rate of solidification, the microstructure of the clad can be estimated. The metallurgical properties can then be extracted from the microstructure of the clad. Referring to Figure 5.9, if the process speed U and the angle  are known, the local rate of solidification, Us , can be obtained from ´ ³  (5.1) Us = U cos 2 The microstructure of the solidified clad is directly dependent on the rate of solidification and the temperature gradient in liquid and can be experimentally obtained for any material. In general, a higher rate of solidification results in finer microstructure, which improves the mechanical properties of the clad [80]. 5.3.2.1

Performance of the Optical CCD-based Detector

The performance of a device that provides the feedback signal (sensor) in a closed-loop control system is usually defined based on range, accuracy, sensitivity, stability, and response time. The range of a sensor defines the limits of the parameters that can be detected by the sensor. The accuracy is defined based on the error between the actual and measured values. The sensitivity is the minimum possible tolerance that may exist between two successive values. The response time is defined as the time that the sensor requires to produce the output signal. The performance of the sensor was examined experimentally and listed in Table 5.1 for the dimension measurement. The response time of the detector directly depends on the speed of the computer and the software platform used to implement the image processing algorithm. Using a Pentium II-450 MHz computer and MATLAB/SIMULINK platform, a maximum rate of 16 Hz was obtained. This rate can be improved by converting the image processing algorithm into an e!cient high level computer language and use of a real time operating system. In order to examine the accuracy of the real-time measurement, an o"ine measurement was utilized by the image processing technique. Pictures of the © 2005 by CRC Press LLC

Frame Grabber 480x 640 pixels

Chop the frame to 116x167 pixels

Change the brightness (-10%) and contrast (+10%)

Laser centerline Substrate line Interface between frame grabber and MATLAB/SIMULINK

Pattern recognition algorithm RGB to GRAY level Compute the global image threshold using Otsu's method Select every tenth frame

Create a threshold binary image using the threshold level

Extract the height and angle using the matrix analysis

FIGURE 5.8 Optical CCD-based detector, frame grabber and pattern recognition algorithm.

© 2005 by CRC Press LLC

Us α

Melt pool

U

Substrate

FIGURE 5.9 Solid/liquid interface and corresponding angle.

TABLE 5.1

Optical CCD-based detector performance Specification Range Accuracy Sensitivity

© 2005 by CRC Press LLC

Geometry Measurement 0.25 to 2.2 mm of clad dimension for laser power of 80 W to 600 W and process speed of 0.5 to 3 mm/s 0.1 mm 0.02 mm

Clad height (mm)

clads were taken by a digital camera and using an edge detection approach, the height of the clads were obtained with a resolution better than ±50 µm. The results were then compared with real-time measurements. Figure 5.10 shows the comparison. The results were obtained when the laser pulse frequency was changed from 70 to 100 Hz, and the laser energy = 5 J, powder mass rate = 1 g/min, laser pulse width = 3 ms, substrate speed = 1.5 mm/s.

2 1.5

Detector fails

1

a)

0.5 0 0

10

20

30

40

Close up view

Time (s)

50

60

70

80

90

Real-time measurement Offline measurement

b)

c)

FIGURE 5.10 Comparison between the real-time and o"ine measurements of clad height, a) real time and o"ine measurements, b) comparison, c) actual clad.

5.3.3

Control Strategy

In this section, several classical and fuzzy logic controllers implemented to the laser cladding process are discussed. These controllers are able to adjust the laser pulse energy and the process speed. As mentioned earlier in Section 4.8.1, the clad height becomes a linear function of the laser pulse energy around the operating point. As a result, simple models can be developed around dierent operating points to predict the clad height as a function of the laser pulse energy. One of the models discussed in Section 4.8.1 is used in designing the controller in this section. For the control, a PID controller [224] is considered with the identified model and a threshold limiting the upper and lower ranges of the laser pulse energy as shown in Figure 5.11. In this figure, hd is the desired and hr is the actual generated clad height. The main reason for using the lower and upper thresholds is to address the clad bead quality and the laser system limitations, respectively. © 2005 by CRC Press LLC

Threshold

hd

+

E

PID Controller

−

Plant Model

hr

FIGURE 5.11 PID controller integrated with a threshold and the identified plant model.

For optimizing the PID gains, an optimization algorithm is used based on the minimization of the quadratic Hessian function [225]. The algorithm optimizes the PID gains based on the desired overshoot, settling time, and rising time. For the laser cladding process, the desired situation is one for which the overshoot, settling time and the oscillations around the set point become minimum, so that a clad with the desired height can be produced rapidly and precisely. The optimization results in the proportional, integral and derivative gains as Kp = 5.5, Ki = 0.8 and Kd = 0, respectively, for an operating point in which U = 1.5 mm/s, m ˙ = 1 g/min, F = 96 Hz, W = 3 ms, and desired clad height is 0.8 mm. The other process parameters were similar to those used in Section 4.8.1.3.1. Since Kd = 0, it can be concluded that a PI controller gives a better process performance.

tanh(20(e − 0.1)) + 1 2

hd +

e −

×

ek

PID Controller

Threshold E

Experimental Plant

hr

FIGURE 5.12 Application of a PID controller to the laser cladding process.

To evaluate the controller, experiments were conducted using the system shown in Figure 5.3. Initial results showed that the model in Figure 5.11 was © 2005 by CRC Press LLC

not eective because of the instability of the laser due to large deviations in the input signal to the laser. To overcome this setback, a filter in the form of tanh(20(e  0.1)) + 1 (5.2) 2 was added to the controller. In this filter e is the error and ek is the filtered error in the control loop as shown in Figure 5.12. This function smooths the error and eliminates large fluctuations form the input signal to the laser. Implementation of this filter in the apparatus significantly improves the controller and in turn the clad quality. Figure 5.13 compares the experimental and simulation results. ek = e

tanh(20(e − 0.1)) + 1 2

hd +

e −

×

ek

PID Controller

Threshold E Experimental Plant

hr

FIGURE 5.13 Comparison between the simulation and experimental results where Kp = 5.5, Ki = 0.8 and Kd = 0.

5.3.4

Closed-Loop vs. Open-Loop

In a layer-by-layer laser part manufacturing process, it is important to fabricate the desired slice with high precision. Initiation of a layer using an openloop laser cladding system with a constant energy causes a gradual growth in the clad height in the initial stage of the first layer. Also, the same energy used in the first layer is more than su!cient for the other layers, which in turn damages the underlying layers through remelting. A closed-loop control system can overcome these issues by applying higher energy in the start up of the clad and reducing the energy afterward. In order to investigate and compare a closed-loop versus an open-loop controller during process startup, an experiment with U = 2.25 mm/s, m ˙ =2 g/min, F = 96 Hz, and W = 3 ms is conducted. The gains of the controller are set to Kp = 10, Ki = 0.4 and Kd = 0. Figure 5.14 compares the results. For this highly repeatable experiment, the average error of the generated clad heights in 5 seconds with respect to the © 2005 by CRC Press LLC

Clad height (mm)

1.4

Set point 1

Closed-loop

0.5

Open-loop 00 00

1

2

3

4

5

6

Time (s) Laser pulse energy (J)

a)

4 3 2

Upper threshold=4.5 J Lower threshold=2.5 J

1 0

0

1

2

3

Closed-loop Open-loop 4

5

6

Time (s)

b) Closed-loop

c)

Open-loop

d)

FIGURE 5.14 Performance of the closed-loop PID control and the open-loop systems during clad startup: a) clad height, b) corresponding laser pulse energy, c) a side view of the generated clad by the closed-loop PID control system, d) side view of generated clad by open-loop PID control system.

© 2005 by CRC Press LLC

desired value (1 mm) was 31% for the open-loop and 7% for the closed-loop control systems. Another fluctuation in the process that a closed-loop control method can attenuate is the absorptivity factor, . Absorptivity factor is a function of surface condition and temperature [101]. Changes in the absorptivity creates disturbances during laser cladding, which causes the total energy absorbed by the substrate to change. Consequently, these changes aect the quality of the clad in terms of geometrical, mechanical and metallurgical qualities.

Open-loop control of laser cladding Clad height (mm)

Sandblasted area

Shiny area

1

0.5

0

0

5

10

15

20

25

Time (s)

Laser pulse energy (J)

a) 3 2 1 0

0

5

10

15

20

25

Time (s)

b)

HAZ

Top view

Sandblasted area

Shiny area c)

Side view

d)

FIGURE 5.15 Open-loop laser cladding under the absorption variation: a) generated clad, b) laser pulse energy, c) top view of the generated clad, d) side view of generated clad.

© 2005 by CRC Press LLC

Closed-loop control of laser cladding Sandblasted area

Shiny area

Clad height (mm)

1

0.5

Set point =0.8 mm 0

0

5

10

15

20

25

Time (s)

Laser pulse energy (J)

a) 4 3 2

Upper threshold = 4.5 J Lower threshold= 2 J

1 0

0

5

10

15

20

25

Time (s)

b)

HAZ

Top view

Sandblasted area

Shiny area c)

Sideview

d)

FIGURE 5.16 Closed-loop laser cladding under the absorption variation: a) generated clad, b) laser pulse energy, c) top view of the generated clad, d) side view of the generated clad.

© 2005 by CRC Press LLC

In order to evaluate the performance of the controller, an experiment is conducted. Mild steel samples, half shiny and half sandblasted, are prepared. Numerous tests with dierent laser pulse energy and U = 1.5 mm/s, m ˙ =1 g/min, F = 96 Hz, and W = 3 ms are performed when the other process parameters are the same as those described in Section 4.8.1.3.1. The main reason for performing these experiments is to find the critical energy on the substrate in the presence of the absorption disturbance. It is concluded that E = 2.6 J is the critical energy for this circumstance resulted in a poor clad in the shiny area as shown in Figure 5.15. As seen, the generated clad in the shiny area has a lower height and poor quality (it could be easily removed after the process) due to the lack of energy as shown in Figures 5.15c and 5.15d. Of interest is the fact that the HAZ (heat-aected zone) line is visible in Figure 5.15c. It also provides evidence of the lack of su!cient energy absorbed by the substrate when the surface is shiny. In contrast, the closed-loop system can sense and overcome the lack of energy in the system, when the same conditions of the open-loop experiment are applied to the apparatus. The result of the closed-loop control for the absorptivity disturbance is shown in Figure 5.16. As seen, the closed-loop control system compensates for the disturbance. Although due to the low control cycle rate (typically 5 to 7 Hz) and large response time of both the laser and the process, a portion of the clad had low quality. A geometrical disturbance is one of the most likely disturbances in the laser cladding process. It may arise from the desired part geometry or even from the deformation of the underneath layer due to thermal distortion. The process speed is the most important parameter that can significantly compensate the geometrical disturbances. However, the laser energy can also be selected as the command signal to compensate the eects of small geometrical fluctuation around the desired set point. In order to examine the performance of the controller under geometrical disturbances, a simple test is considered. A step is machined on some mild steel plates with a depth of 0.25 to 0.3 mm. The fabricated step on the flat plate not only changes the underneath layer positions but also changes the eective position of the laser spot point and the intersection of the laser and powder stream on the workpiece. Several experiments are performed for both open- and close-loop conditions, where U = 2.5 mm/s, m ˙ = 2 g/min, F = 93.5 Hz, W = 2.92 ms, and the other process parameters are set as those used in Section 4.8.1.3.1. Figure 5.17 shows the result of an open-loop laser cladding process where the laser energy is set to 2.8 J. As seen in the figure, the clad height is significantly changed. Figure 5.17c also shows a side view of the step zone. Figure 5.18 shows the performance of the closed-loop control system for a step down in the underneath layer. Figure 5.18c also shows a longitudinal view of the sample in the step zone. As seen, the controller compensates the eect of the geometrical disturbances very eectively. © 2005 by CRC Press LLC

Open-loop laser cladding Step down

Clad height (mm)

1.5 1 0.5 0

0

5

10

Time (s)

15

20

25

15

20

25

Laser pulse energy (J)

a) 3 2 1 0

0

5

10

Time (s)

0.25 to 0.3 mm

b)

Reference line

c)

FIGURE 5.17 Open-loop control of laser cladding experiment for a geometrical disturbance: a) clad height, b) fixed laser pulse energy, c) side view of the generated clad in the step zone.

5.3.5

Application of the Developed Controller to Fabrication of Two Simple Components

To investigate the role of a closed-loop controller on a layer-by-layer part manufacturing, the controller is used in fabricating a thin wall and a thin curved shape. Several layers need to be deposited on each other to form the desired object. For both shapes, an open-loop and closed-loop process is performed. In the experiments U = 2.25 mm/s, m ˙ = 2 g/min, F = 96 Hz, and W = 3 ms whereas the laser pulse energy for the open-loop experiment is 3.5 J. Other process parameters are similar to those used in Section 4.8.1.3.1. When a layer is finished the laser is moved up by 0.38 mm before the next layer is deposited. Figure 5.19 shows the results of the open-loop and closed-loop controller in the fabrication of the thin-wall prototypes. As seen, the closed-loop control system has enhanced the geometry of the prototypes significantly. © 2005 by CRC Press LLC

Closed-loop laser cladding Step down Set point Clad height (mm)

1.5 1 0.5 0

0

5

15

10

20

25

20

25

Time (s)

a) Laser pulse energy (J)

4 3 2

Upper threshold=4.5 J Lower threshold=2.5 J

1 0 0

5

10

Time (s)

15

0.25 to 0.3 mm

b)

Reference line

c)

FIGURE 5.18 Closed-loop control of laser cladding experiment for a geometrical disturbance: a) clad height, b) laser pulse energy, c) side view of the generated clad in the step zone.

Figure 5.20 shows the power changes in layers. It can also be concluded that the main reason for a poor quality in the open-loop system is the high amount of delivered energy to the process causing remelting of the underneath layer. Figure 5.21 illustrates the results of a closed- and open-loop prototyping of a thin curved shape. As seen in the figure, the controller has a significant eect in improving the process for prototyping. © 2005 by CRC Press LLC

1 cm

b)

a)

1 cm

d)

c)

FIGURE 5.19 Comparison between a closed-loop and open-loop controlled productions of the thin wall: a) closed-loop, b) closed-loop side view, c) open-loop, d) open-loop side view.

5.4

Application of Knowledge-Based Control to Laser Cladding

As noticed and discussed so far, laser cladding is a complex and nonlinear process with many uncertainties. For such systems, application of classical controllers in which a good model or well-predicted behavior is needed may not be appropriate. Knowledge-based control techniques such as fuzzy logic or neural network, on the other hand, can handle these systems more conveniently. In the following, we investigate numerically the application of fuzzy logic to the control of laser cladding. For the laser cladding process, we use the Hammerstein-Wiener model developed in Section 4.8.1.2. This nonlinear model relates the clad height to the process speed. In the control development process, we apply the standard and modified version of a fuzzy logic controller and compare the results. © 2005 by CRC Press LLC

5

Layer 1

Laser pulse energy (J)

4

Layer 2 Layer 3 Layer 4 Layer 5 Layer 6 Layer 7

3

2

1

0

0

10

5

15

Time (s)

FIGURE 5.20 Comparison between the laser pulse energy for dierent layers in a closed loop control system.

5.4.1

Fuzzy Logic Controller

Fuzzy logic was invented by Zadeh [226] and developed in the past three decades. With fuzzy logic, it is possible to analyze complex systems without having their mathematical models [227]. In this fashion, the fuzzy logic controller was introduced [228]. The basic paradigm for fuzzy logic control is a linguistic or rule-based control strategy, which maps the observable inputs of the given physical system into its controllable outputs with applying a set of implication linguistic rules. Figure 5.22 shows a fuzzy control strategy, where a fuzzy logic controller and fuzzy logic gain scheduler are incorporated in the controller decision making part. This strategy maps the error, e = yd  y, into the control action u. At the heart of this control scheme is a fuzzy logic control algorithm that maps the normalized error ek and rate of change of error dk to the change in the control output or uk . A typical rule in fuzzy logic controller is if ek is LP or SP and dk is LN or SN, then uk is Z. where LP, SP, LN, SN, and Z stand for large positive, small positive, large negative, small negative, and zero, respectively [228]. Figure 5.23 shows fuzzy membership functions over the normalized domains of definition of the relevant variables and the mathematical meanings of LP, SP, LN, SN, and Z in a normalized range. For a system with a large settling time such as the laser cladding, the process command can also be amplified at the beginning of the process. As a result, a controller with a gain scheduler is developed to address this issue using a fuzzy approach. The basics of the developed fuzzy logic control paradigm is to choose the value of normalization factor in output of the standard fuzzy controller by another set of fuzzy rules, which is a so-called fuzzy logic © 2005 by CRC Press LLC

a)

1 cm

b) c)

d)

e)

f)

FIGURE 5.21 Comparison between a closed-loop and open-loop fabrication of a curved shape: a) closed-loop perspective view, b) closed-loop top view, c) closed-loop side view, d) open-loop perspective view, e) open-loop top view, f) open-loop side view.

© 2005 by CRC Press LLC

Controller

yd

e

ek

ke

-

FLC + -

kd

dk

1 z

δuk

Fuzzy δu logic gain + scheduler

u +

Plant

y

1 z

FIGURE 5.22 Proposed fuzzy logic controller.

gain scheduler. To keep control over the response, it is proposed that the area below the steady-state line be divided into several appropriate subregions of L, N, H, as shown in Figure 5.23. During the startup, the value of gain is switched to a higher value to accelerate the system response. Conversely, it is decreased when the response approaches the steady-state response. Implementation of the designed fuzzy logic controller to the laser cladding process is subjected to the compensation of the nonlinear terms identified in the Hammerstein-Wiener model. As a result, the inverse functions of f and g which were introduced in Section 4.8.1.2 are incorporated in the fuzzy controller as shown in Figure 5.24. In order to investigate the eect of the fuzzy controller on the system response, a standard fuzzy controller is implemented into the model as well. The standard fuzzy controller has the same structure except the fuzzy gain scheduler is eliminated. The classic fuzzy controller results in a sluggish response as seen in Figure 5.25. Conversely, the fuzzy controller with a gain scheduler can considerably reduce response time of the system and improve the system response (generated clad height) significantly, as shown in the figure.

© 2005 by CRC Press LLC

yn

δu L

N

H

1

L N

-1

0

1

Membership Function

H

ek

LP

dk

LP LN

δ uk

LP

LN

SN

SP

Z SN

Z

SN

Z

SP

Z

SN Z SN

SP LP SP Z Z LP Z

LP Membership Function

-1

FIGURE 5.23 Control rules and membership function.

© 2005 by CRC Press LLC

0

1

Z

t

SP

Z

Z

SP

SP

Z

Controller hd

g −1

e

+

ek

ke

−

+ -

kd

dk

FLC1

δ uk

Fuzzy logic δu gain Sche.

u +

+

1 z

f

−1

Plant

hr

1 z

g −1

FIGURE 5.24 Proposed fuzzy logic controller integrated with knowledge-based functions for the laser cladding process.

1.1 1 0.9

Clad height (mm)

0.8

Standard Fuzzy Control

0.7 0.6 0.5

Proposed Fuzzy Control

0.4 0.3 0.2 0.1

0

2

4

6

8

10

12

14

16

18

20

Time (s)

FIGURE 5.25 Comparison between the standard fuzzy logic (without gain scheduler) and proposed fuzzy logic (with gain scheduler).

© 2005 by CRC Press LLC

6 Physical Metallurgy and Material Systems of Laser Cladding

This chapter will describe the basic physical metallurgy applied to laser cladding and review the material systems that are being investigated for suitability in laser cladding processes. In the first section, the general cladability of substrate/clad materials will be described including both processing and metallurgical considerations. In the next section, a basic description of the relevant solidification conditions encountered during laser cladding will be completed. This will include the microstructural features that develop in the clad as a result of the solidification conditions. In the final section of the chapter, the material combinations that are most widely being investigated will be reviewed.

6.1

Cladability

There are many possible definitions of cladability. In this text, we consider cladability to include primarily the formation of a continuous, high density clad deposit with a uniform or homogeneous microstructure, possessing a strong metallurgical bond to the substrate but with low dilution. However, in particular cases, cladability could also include the requirement of a certain level of dilution by the substrate, a certain clad height or deposition e!ciency or even specific materials properties that are required from the clad. Using the above definition, cladability is primarily determined by the processing parameters used during the cladding operation as well as the metallurgical interactions or compatibility of the clad and substrate materials. Both these topics will be considered in detail in the following sections. The discussion in this section will limit itself to the powder injection method of laser cladding. However, the principles discussed can be equally applied to other laser cladding methods.

© 2005 by CRC Press LLC

6.1.1

Processing Parameter Considerations

In order to achieve a dense clad that is metallurgically bonded to the substrate, enough energy must be delivered by the laser beam to the process volume such that the powder stream and a small volume of substrate surface melts. Because high temperature gradients are present in the melt pool, strong convection forces develop due to the Marangoni eect [35]. As a consequence, the liquid phases rapidly mix and become homogeneous. Upon cooling, this well-mixed liquid phase solidifies and forms the basis of the dense well-bonded clad layer. The energy delivered to the process volume is not only determined by the laser beam energy but also by its spot size, the velocity of the substrate, and in the case of a pulsed laser, the frequency F and duration of the pulse W . The volume of material that this delivered energy must melt is determined by the powder feed rate and substrate velocity. As described in Section 2.6.2.2 of Chapter 2, these various process parameters can be combined into two effective parameters defined as the eective energy density Eef f [J/mm2 ] and the eective powder deposition density #ef f [g/mm2 ]. The role of these parameters in producing a dense clad with good bonding to the substrate can be demonstrated by an experimental case study as described below. 6.1.1.1

Case Study: FeAl Cladding on Mild Steel

In this case study, the cladding of FeAl on mild steel is investigated. The aim of this study was to obtain a correlation between combined process parameters with the clad bead quality (or cladability). However, it also demonstrates the potential of laser cladding a FeAl coating on mild steel. Literature shows that iron-aluminide alloys for high temperature applications have been used for several years primarily due to their superior high temperature oxidation and sulfidation resistance [229, 230, 231, 232]. These alloys have improved electrical resistance, cyclic fatigue resistance, high temperature oxidation resistance, and both low and high temperature strength [233]. Its application in fuel cell technology has shown that it is extremely resistant to corrosion and metal loss when exposed to oxidizing atmospheres at high temperatures [234]. Furthermore, these alloys are lower in cost and have better corrosion resistance compared to conventional Ni-based and stainless steel type alloys [235]. In spite of the mentioned properties of the iron-aluminides, production of this alloy in bulk form is limited due to hydrogen-induced embrittlement. The possibility of producing the iron-aluminide coating on less corrosion-resistant materials, such as low carbon and stainless steels, has recently been investigated [230, 232]. These coatings were produced by weld overlay cladding processes using gas tungsten arc (GTA) and gas metal arc (GMA) welding techniques. The results indicate that at welding contents above 10 wt% Al, cold cracking in the iron-aluminide cladding occurs in a similar manner to that observed in bulk samples. © 2005 by CRC Press LLC

The advantages of laser cladding, including chemical cleanliness, localized heating, low dilution of the cladding material by the substrate and rapid cooling rates [101, 49], may provide the capability of production of iron-aluminide on mild steel to overcome some of the problems encountered in arc-welding based iron-aluminide coatings. In the following sections, the results of experiments for production of ironaluminide coating on mild steel will be addressed.

6.1.1.1.1 Experimental Setup and Procedure The powders used in this study were commercially pure Fe and Al powders (i.e., 98% pure) both with a mesh size of 45 µm (-325). The powders were mixed to a bulk composition of Fe - 20wt% Al by a dry milling procedure for one hour. Sandblasted mild steel plates (0.25 to 0.28 C; 0.6 to 1.2 Mn) with 200 × 20 × 5 mm dimensions were used as the substrate. The laser was turned on 10 mm away from the edge of the plate. The experiments were performed with a LASAG FLS 1042N Nd: YAG laser with a maximum power of 1000-W, 9MP-CL Sulzer Metco powder feeder, and a 4-axis CNC table. The laser beam and powder feed nozzle were stationary while the CNC table moved the substrate at a controlled velocity. An argon shield gas was fed along the powder stream. It was also used to protect the optical lenses. During the various laser-cladding experiments, clad height was measured in real-time using the CCD-based detector as explained in Chapter 5. Initial experiments were performed to investigate the influence of laser processing characteristics on the cladding process. The conditions studied are given in Table 6.1, where F is the laser pulse frequency [Hz], W the laser pulse width (duration) [ms], E the pulse laser energy [J] and U is the process speed [mm/s]. The other process parameters were constant for all experiments as listed in Table 6.2. Each condition was performed along a track. For Conditions 1 to 14, E was changed incrementally as shown in Figure 6.1, while F, W , and U were constant. For Conditions 15 to 19, F was changed as shown in Figure 6.2, while E, U , and W were fixed. As indicated in Table 6.1, Conditions 1 to 6 were used to determine the influence of laser pulse energy and process speed on the Fe-Al cladding process. Also, Conditions 7 to 11 and Conditions 12 to 14 were performed to investigate the same eects with dierent laser pulse widths and laser pulse frequencies. Experiments 15 to 19 were used specifically to explore the influence of the laser pulse frequency. Representative samples from experiments with a given set of process parameters were sectioned, mounted and polished, and metallurgical analysis of these samples were performed using optical microscopy to determine the clad bead quality as a function of the processing parameters. © 2005 by CRC Press LLC

TABLE 6.1

Laser processing parameters. Cond. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Pulse Frequency [Hz] 100 100 100 100 100 100 80 80 80 80 80 50 50 50 40 to 100 40 to 100 35 to 95 35 to 95 35 to 95

Pulse [ms] 3.0 3.0 3.0 3.0 3.0 3.0 4.1 4.1 4.1 4.1 4.1 6.4 6.4 6.4 3.0 3.0 3.0 3.0 3.0

Width

Pulse Energy [J] 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 2 to 4 4 4 5 5 5

Process Speed [mm/s] 0.75 1.00 1.25 1.50 1.75 2.00 0.75 1.00 1.25 1.50 1.75 1.00 1.25 1.50 1.00 1.50 1.25 1.5 1.75

TABLE 6.2

Additional process parameters, which were fixed for all conditions. Laser shield gas (m3 /s) Laser spot radius on workpiece(mm) Powder feed rate (g/min) Powder feeder shield gas(m3 /s) Powder stream diameter on the workpiece (mm) Nozzle angle (deg)

2.34e-5 0.7 1 2.34e-5 1.6 55

Conditions 1 to 14

FIGURE 6.1 Multistep laser pulse energies for Conditions 1 to 14. © 2005 by CRC Press LLC

FIGURE 6.2 Multistep laser pulse frequencies for Conditions 15 to 19.

6.1.1.1.2 Results The optical CCD-based detector provides continuous height measurements of the clad along the substrate as described in Chapter 3. The continuous measurement of the clad height oers valuable information about the surface roughness of the clad deposit, which can be used in evaluation of the clad bead quality. Figure 6.3 shows the continuous clad height measurements for two typical results associated with Conditions 1 and 4.

Clad height (mm)

3.0

E=2 J

E=2.5 J

E=3 J

E=3.5 J

Condition 1

E=4 J

2.5 2.0 1.5 1.0 0.5

Condition 4

0.0 0

20

40

60

80

100

120

Position along the substrate (mm)

FIGURE 6.3 Clad heights along the substrate for Conditions 1 and 4.

The roughness of both samples can also be determined by analyzing the measured clad height. Figure 6.3 indicates that the roughness of the clad was high for Condition 1 along the entire track with an average deviation of 0.43 mm. For Condition 4, the clad roughness was fairly good with deviation of 0.1 mm for the last two steps, when E = 3.5 J and E = 4.0 J. The corresponding macrostructure of the track region for Condition 4 at E = 4.0 J is shown in Figure 6.4. Also, Figure 6.5 shows the corresponding track macrostructure for Condition 1 at E = 3.5 J. For Conditions 1 and 4 at E = 2.0 J and E = 2.5 J, the quality of clad product was poor in terms of strength of bonding between the clad and substrate. © 2005 by CRC Press LLC

1 mm

a)

1 mm

b)

FIGURE 6.4 a) Longitudinal view of the macro-quality of a clad with good surface condition before polishing, b) longitudinal section of the macro-quality of the clad indicating good qualities (Condition 4 at E = 4.0 J).

For these cases, the clads were easily removed from the substrate following the process. The bond strength of the other part of a track at E = 3.0 through 4 J was tested by a simple bending experiment [236]. During the bending test, no separation was observed. However, some cracks were created after the clad and substrate were bent about 40 relative to the substrate as seen in Figure 6.6. As the figure shows, the cracks continue through the clad and end in the substrate with no cracks running along the substrate interface providing further evidence of a strong bond between the clad and substrate. The bending test for other parts of a track with lower energy (E = 2.0 and E = 2.5 J) showed a weak bond between the clad and substrate such that the clad was separated from the substrate. Clearly the clad height measurements made in real-time with the optical CCD-based detector relate directly to the clad quality produced. A smooth clad deposit with little porosity and a good bond with the substrate corresponds with the detector readings, which indicate little height variations. Conversely, the detector readings, which indicate large height deviations correspond to rough clad deposits with a significant level of porosity. The above-mentioned procedure was carried out for other samples. The obtained information will be then used in the development of a strategy for evaluation of the clad bead quality later. In the following subsections, we use the average of the measured clad height © 2005 by CRC Press LLC

1 mm

a) 1 mm

b)

FIGURE 6.5 a) Longitudinal view of a clad with high roughness, b) longitudinal section of a clad with high roughness (Condition 1 at E = 3.5 J). Clad

Substrate

1 mm

Magnified views of cracks in the clad and substrate Clad

0.2 mm

Substrate

FIGURE 6.6 Bending section of a track at E = 3.0 through 4.0 J. © 2005 by CRC Press LLC

at any portion of the track with constant laser pulse energy or frequency to investigate the influence of process parameters, including process speed, laser pulse shaping, average power and duty cycle on the clad height.

6.1.1.1.3 Influence of Laser Pulse Shaping and Process Speed on Average Clad Height Due to the repetitive nature of the laser beam generated by the Nd:YAG pulsed laser, two parameters were defined as duty cycle and average laser power, which are expressed by Equations (2.4) and (2.5), respectively. These two parameters are usually used to study Nd:YAG pulse laser-based experiments. In this section, we use these two parameters to study their eects on the average clad height.

2.5

U=0.75 mm/s

Average clad height (mm)

2

U=1 mm/s U=1.25 mm/s U=1.5 mm/s

1.5

U=1.75 mm/s U=2 mm/s

1

0.5

F=100 Hz W=3 ms

0 1.5

2

2.5

3

3.5

4

4.5

Laser pulse energy (J)

FIGURE 6.7 Fe-Al average clad height as a function of laser pulse energy and process velocity (C is fixed, Pl is variable).

Figure 6.7 illustrates the eect of laser pulse energy and process velocity on the average clad height while the laser pulse frequency and width are both constant, and, as a result, duty cycle is also fixed. The results confirm that as U decreases or E increases the clad height increases. The figure confirms the consistency in the process when the laser pulse energy or process speed are changed. The other issue of interest is the influence of the laser pulse frequency on the © 2005 by CRC Press LLC

average clad height. Figure 6.8 shows the influence of laser pulse frequency and process speed on the average clad height while the laser pulse energy and width are set to 5 J and 3.00 ms, respectively. The figure confirms a consistent nature for the process for the experiments explored. As U increases the clad height decreases, whereas if F increases the clad height increases. 1.6 U=1.25 mm/s

1.4 U=1.5 mm/s

Average clad height (mm)

1.2 U=1.75 mm/s

1 0.8 0.6 W=3 ms E=5 J

0.4 0.2 0 30

40

50

60

70

80

90

100

Laser Pulse frequency (Hz)

FIGURE 6.8 Fe-Al average clad height as a function of laser pulse frequency and process speed (C and Pl are both variable).

Figure 6.9 shows the influence of laser pulse energy and frequency on the average clad height while the process speed and laser pulse width are both constant. As seen for the conditions explored, an increase in both E and F causes the average clad height to increase as well. From Figures 6.8 to 6.9, it can be concluded that the clad height increases with either an increase in duty cycle or average power as determined by a variation in E and F or both. Figure 6.10 reports average clad height data for the case in which C and U are both fixed, but W and F are both varied. The figure indicates that the nature of the process is not consistent. Despite a fixed duty cycle, dierent combinations of the laser pulse frequency and width do not result in the identical clad heights. However, as E decreases, the clad height also decreases. The same behavior occurs with respect to F except for the condition with F = 50 Hz and W = 6.4 ms. Although the clad height for this condition is relatively close to the condition with F = 100 Hz and W = 3.0 ms, the clad quality is poor for the entire track in terms of bonding between the clad and © 2005 by CRC Press LLC

1.4 E=5 J

1.2

Average clad height (mm)

E=4

1 0.8 W=3.00 ms U=1.5 mm/s

0.6 0.4 0.2 0 30

40

50

60

70

80

90

100

110

Laser pulse frequency (Hz)

FIGURE 6.9 Fe-Al average clad height as a function of laser pulse frequency and energy (Pl and C are variable).

1.4 F=50 Hz, W=6.4 ms

1.2

Average clad height (mm)

F=100 Hz, W=3 ms

1

F=80, W=4.1 ms

0.8 0.6 0.4

U=1.25 mm/s C=~0.31

0.2 0 1.5

2

2.5

3

3.5

4

4.5

Pulse energy (J)

FIGURE 6.10 Fe-Al average clad height as a function of laser pulse energy, frequency and width while the duty cycle is constant (C is fixed, Pl is variable). © 2005 by CRC Press LLC

F=100 Hz, E is variable 1.4

Average clad height (mm)

1.2 1

F=100 Hz, E is variable E=5 , F is variable

0.8

E=5 , F is variable

0.6 0.4

W=3.00 ms U=1.5 mm/s

0.2 0 170

220

270

320

370

420

Average power (W)

FIGURE 6.11 Dependency of the Fe-Al average clad height on the average power at dierent laser pulse energies and frequencies.

substrate. Figure 6.11 shows the influence of the average power on clad height when using dierent process parameters. The results show an inconsistency in the process as the average power is selected for the clad height evaluation. For instance, at condition E = 5.0 J and F = 40 Hz the average power is 200 W which is identical to the situation where E = 2.0 J and F = 100 Hz; however, the generated heights are significantly dierent. The main reason of this behavior is the interaction of powder stream and the laser pulses and its contribution to the clad generation. Of interest is the fact that at the lower frequencies (lower duty cycle), the interaction time between the laser beam and powder particles decreases. Consequently, the rate of eective powder deposition decreases. As is evident in the figure, the influence of this reduces significantly at higher frequencies and energies. All the previous figures indicate the dependencies of the process on the various parameters. Although the figures provide valuable insight into the process, the evaluation of the clad quality may be di!cult to be described using them due to the variety of parameters involved. As a result, we develop a strategy by introducing combined parameters and their influence on the bead quality of clad, which will be addressed in the last section of this chapter. © 2005 by CRC Press LLC

6.1.1.1.4 Clad Substrate Macro/Microstructure and Hardness In the previous section, two dierent qualities of clads were addressed: clads with good surface roughness and strong clad/substrate bonds, and clads with poor surface roughness and weak clad/substrate bonds. To further investigate the clad/substrate macrostructures for these two possible types of cladding, sections through the clad/substrate couples in transverse and longitudinal directions were made.

0.2 mm

FIGURE 6.12 Macrostructure of transverse section of clad made using E = 4.0 J, W = 3.0 ms, F = 100 Hz and U = 1.5 mm/s.

Figures 6.12 and 6.13 show the clad/substrate macrostructure for a sample with E = 4.0 J, W = 3.0 ms, F = 100 Hz, and U = 1.5 mm/s in transverse and longitudinal sections. The clad deposit and grain boundaries are clearly visible. The clad has a good profile. The figures indicate that this condition produces a very dense clad with homogeneous structure and negligible porosity. In addition, no cracks were observed in the clads or at the clad/substrate interface. Furthermore, the microstructure of the clad in longitudinal section can be explored through Figure 6.13. It is evident from the figure that two structures exist in the clad microstructure. The first structure is the lines of banding which are parallel to the solid/liquid interface. The banding structure is generated due to the repetitive pulsing nature of the heat source used in the process. In the rapid solidification, the absence and presence of the heat source in a short period of time cause an oscillatory growth of the solid area. This oscillatory growth results in the composition deviation [237]. The second structure is the grain boundaries in the clad region. As seen, nucleation starts from the substrate/clad interface, and the grains grow normal to the © 2005 by CRC Press LLC

0.2 mm

Grain boundary

Banding

FIGURE 6.13 Macro/microstructure of longitudinal section of clad made using E = 4.0 J, W = 3.0 ms, F = 100 Hz and U = 1.5 mm/s.

solid/liquid interface (the lines that are indicated by the banding structure). Figures 6.14 and 6.15 show the clad/substrate macrostructure for a sample with E = 2.0 J, W = 3.0 ms, F = 100 Hz, and U = 1.5 mm/s in transverse and longitudinal sections, respectively. Conversely, this sample exhibits a very porous deposit which is poorly bonded to the substrate. The hardness of the clad shown in Figure 6.12 was measured across the bond area using a Vickers microhardness instrument. The results are shown in Figure 6.16. Clearly, the hardness on the clad deposit is much higher than that of the mild steel substrate. The hardness steadily rises within the heataected zone (HAZ) of the substrate until it reaches a value two times higher than the substrate, in the clad deposit. The clad hardness is significantly higher than the hardness that is reported for the arc weld overlay techniques [230, 231]. 6.1.1.1.5 Combined Parameters Calculated values for Eef f and #ef f using Equations (2.8) and (2.9), respectively, for the processing Conditions 1 to 14 are plotted in Figure 6.17, and for Conditions 15 to 19 are plotted in Figure 6.18. Based on continuous clad height measurement, optical microscopy of selected sectioned clads, bending test, observations, and the evidence represented earlier, four regions are distinguishable. These regions are labeled as follows: 1. Good quality clad: The selected Eef f and #ef f in this region pro© 2005 by CRC Press LLC

0.2 mm

FIGURE 6.14 Macrostructure of transverse section of clad at E = 2.0 J, W = 3.0 ms, F = 100 Hz, and U = 1.5 mm/s.

0.2 mm

FIGURE 6.15 Macro/microstructure of longitudinal section of clad at E = 2.0 J, W = 3.0 ms, F = 100 Hz, and U = 1.5 mm/s.

© 2005 by CRC Press LLC

Vicker's microhardness

500

400

300

200

Fe-Al Clad Deposit

100

Mild Steel

HAZ

0 -600

-800

-400

0

-200

200

400

600

Distance from clad/substrate interface (microns)

FIGURE 6.16 Microhardness of clad specimen made using E = 4.0 J, W = 3.0 ms, F = 100 Hz, and U = 1.5 mm/s.

160

1 140

Good quality clads

3

2

2

Effective energy density (J/mm )

120

E=4 J 100

80

60

3

6

E=2.5 J 4 11 5

10

E=2 J 4 5 11

14 10 14

3

14 1.25

E=3.5 J 7

1.5

13 13 13 13

1 12

9 E=3.5 J 12 E=3 J

12

E=2.5 J

12

E=2 J

13

E=3 J

7

E=2.5 J

7

8

9 2

E=4 J

Roughness, some pores and cracks

7

1

8 2

9

10 3

14

1

9

10

14 20

8 2

E=3 J 11 5 11

6

1

3

6

6

40

4

5 4 E=3.5 J 5 4

6

1

2

Brittle clads

8 E=2 J 8

No cladding Conditions 1 through 14 The number of the condition is shown beside the marker

12

1.75

Effective powder deposition density (g/mm2 )

2

2.25

2.5 x 10

-3

FIGURE 6.17 Eective energy density versus eective powder deposition density for Conditions 1 to 14. © 2005 by CRC Press LLC

160

17 (F=95)

140

E=5 J 15 (F=100)

18 (F=95)

E=4 J

2

Effective energy density (J/mm )

Good quality clads

120

19 (F=95)

17 (F=80) 15 (F=85)

18 (F=80)

Roughness, some pores and cracks

19 (F=80)

100

17 (F=65) 19(F=65)

80

0.6

0.8

Brittle clads

16 (F=70) 15 (F=55)

19(F=50) 18 (F=50) 19(F=35) 16 (F=40) 15 (F=40) 17 (F=35) 18 (F=35)

40 0.4

15 (F=70) 16 (F=85)

18 (F=65)

17 (F=50)

60

16 (F=100)

Conditions 15 through 19 The number of the condition is shown beside the marker

No cladding

1

1.2

1.4

Effective powder deposition density (g/mm 2 )

1.6

1.8

2 x 10

-3

FIGURE 6.18 Eective energy density versus eective powder deposition density for Conditions 15 to 19.

vide clads with a good bond between the substrate and clad where the clads have a smooth surface and good profile without cracks and pores. Figures 6.4, 6.12, and 6.13 show an example of this type of clad. The figures indicate that the parameters in this region produce a very dense clad with homogeneous structure, no cracks, and negligible porosity. 2. Roughness, some pores and cracks: In this region, the clad has a good bond with the substrate; however, its surface is not smooth and many bumps are observed along the track. Figures 6.5, 6.14, and 6.15 show an example of this type of clad. The figures indicate that the parameters selected in this region produce a clad with high roughness, as well as porosity and cracks. 3. Brittle clads: In this region, a clad exists; however, there is not a good bond between the clad and substrate. The clad in this region is easily removed by hand after the process. 4. No cladding: In this region, no clad can be created due to the significant lack of energy. Generally, Figures 6.17 and 6.18 can be used for clad bead quality evaluation. However, the range of validity of Figures 6.17 and 6.18 is 0.4  U  3.0 © 2005 by CRC Press LLC

[mm/s], 35  F  100 [Hz], 2.0  E  5.0 [J], 3.0  W  5.0 [ms], m ˙ =1 g/min, and D = 1.4 mm. As seen in Figures 6.17 and 6.18, a critical state exists in the iron-aluminide laser cladding process which should be met during the real-time control of laser cladding to produce a good quality clad and high cladability.

6.1.2

Metallurgical Considerations

The formation of a strong bond between the clad and substrate material also depends critically on the metallurgical interactions that take place between the two materials during the cladding process. There will be certain cases where incompatibility between the substrate and clad materials will make it virtually impossible for a strong bond to be formed, in which case cladability will be nonexistent. Even in cases where strong bonding can be achieved, the metallurgical reactions between the clad and substrate materials will have an important contribution to the actual values of Eef f and #ef f that are required to achieve an optimum clad. Therefore, it is important to understand the details of these metallurgical interactions so that intelligent material selections can be made when choosing a clad and a substrate couple, and in predicting and interpreting the resulting clad microstructure. Laser cladding is essentially an extremely rapid and localized alloying process, which takes place between the clad material and substrate. Normally the objective in laser cladding is to minimize alloying or dilution of the clad by the substrate. However, the requirement of a strong metallurgical bond means that alloying or dilution cannot be avoided entirely and must be carefully controlled. Before discussing metallurgical interactions, it is first necessary to understand the physical circumstances under which alloying takes place during laser cladding. Figure 6.19 illustrates a schematic of the melt pool created during laser cladding of pure B powder onto a pure A substrate. In this figure, it is assumed that the cladding process has reached a steady-state; the laser beam and powder stream simultaneously come into contact with the melt pool, which has been created by the molten clad/substrate couple. A portion of the energy from the laser beam melts a small volume of the substrate at the leading edge of the clad/substrate couple region X. A portion of the laser beam energy is also used to heat up the powder entering the process volume. Depending on the melting point of the powder, when it comes into contact with the melt pool it may be in the form of molten droplets or solid particles. In many circumstances, there are likely both liquid and solid powder particles that come into contact with the melt pool surface. Therefore alloying (or mixing) will be between liquid A and liquid and/or solid B. It is fairly well accepted that the high temperature gradients present in the melt pool during laser cladding set up intense convection due to the Marangoni eect [35, 238]. This leads to rapid homogenization or alloying within the melt pool. Convection in the liquid pool can be characterized by the surface tension © 2005 by CRC Press LLC

Laser Beam

Solidified clad deposit with composition Co

M el tp

oo l

Stream of powder B

X Substrate, A

FIGURE 6.19 Schematic of the melt pool in laser cladding where alloying of the constituents occurs.

number S, which is equal to [35, 239] S=

(d/dT )qd µU K

(6.1)

where d/dT is the derivative of surface tension with respect to temperature [N/m·K], q is the total absorbed power per area of the substrate irradiated by the laser [W/m2 ], d is the diameter of the laser beam on the substrate [m], µ is the viscosity of the melt pool [kg/s·m], U is the scanning speed of the laser beam (or substrate speed) [m/s], and K is thermal conductivity [W/m·K]. When S  45, 000 [158], convection is negligible; otherwise, convection is important. Clearly, the importance of convection is determined in large part by the physical properties of the materials making up the clad melt pool. Therefore, its contribution to mixing and alloying should be assessed for every material system in use. For example, while convection is generally accepted to be important, at least one detailed calculation by Almeida et al. [239] indicates that for a Nb-Al cladding couple, S was less than 45, 000 indicating that convection did not play a major role in the mixing process in this situation. Regardless of the mechanisms of mixing or alloying, the results of the process can be initially understood by referring to a few simple metallurgical cases as depicted in Figure 6.20. In these cases, the substrate and/or powder stream could be considered to be pure A and B or pure B and A, respectively or pure C and D or pure D and C, respectively. The figure considers two cases: © 2005 by CRC Press LLC

1. Simple eutectic binary alloy system formed between A and B, 2. Simple isomorphous binary alloy system formed between C and D. In most laser cladding situations, the melting point (Tm ) of the clad and substrate materials will be dierent, and this is reflected in the hypothetical phase diagrams where the Tm of A < B and Tm of C < D. In the following, these cases are explained.

Liquid

TmB

Liquidus line Solidus line

(B)

(B) + L

(A) + L

TmA (A) TE

(A) + (B)

a) Pure B

Co

wt% A

Liquidus line

TmD

Pure A

Liquid

(S) + L

T mC

Solidus line

(S)

b) Pure D

Co2

wt% C

Co1

Pure C

FIGURE 6.20 Hypothetical phase diagrams for a) simple eutectic, b) simple isomorphous binary alloy mixture.

6.1.2.1

A Simple Binary Eutectic Cladding Process

First, a cladding process where the clad powder is made of pure B and the substrate of pure A is considered. By design, dilution by the substrate will © 2005 by CRC Press LLC

be low, and therefore the bulk composition of the melt pool Co would be on the B rich side of the phase diagram. As will be discussed in Section 6.2, it is well recognized that the melt pool in laser cladding is not at equilibrium, and therefore cannot be quantitatively characterized by an equilibrium phase diagram [204]. More specifically under non-equilibrium conditions, the liquidus and solidus lines usually shift toward each other with a subsequent reduction in the partition coe!cient. However, the equilibrium phase diagram is still a useful tool in interpreting the general metallurgical interactions qualitatively that will take place during alloying in the melt pool of a laser cladding process. Consider the solid/liquid interfacial region X, where the solid A substrate is in contact with the melt pool. Because solid-state diusion is very slow compared to the process time in laser cladding, diusion of solute B from the melt pool into the solid substrate is negligible. However, there will be a nonequilibrium partition coe!cient kv established between the solid and liquid phases. In a eutectic system, alloying reduces the melting point compared to the pure constituents. Therefore, melting of the substrate ahead of the melt pool will be assisted by compositional dissolution into the molten pool. This should have the eect of reducing the laser beam energy required to reach a condition of cladability as described in Section 6.1.1. A similar conclusion can be made when examining the solid/liquid interfacial region between the melt pool and any unmelted pure B powders colliding with its surface. A non-equilibrium partition coe!cient kv will be established between the solid A particles and liquid phase. Again, in this eutectic system, melting of the solid B powder at the melt pool surface will be assisted by compositional dissolution into the molten pool, further reducing the laser beam energy required to reach a condition of cladability. This consequence of alloying, together with the fact that A and B exhibit good solubility in the liquid state, will create a situation of high cladability between either a pure A substrate and pure B powders or a pure B substrate and pure A powders. 6.1.2.2

A Simple Binary Isomorphous Cladding Process

Solubility between pure elements in an isomorphous system is also very good both in the liquid and solid states. Therefore, it may be expected that a system of this type would exhibit good cladability. However, the consequence of alloying on the melting point is very dierent than in a eutectic system and this could cause cladability problems depending on dierent selection of elements for the substrate and the powder. The first case to consider is that where the substrate is made of the higher melting point pure D and the powder the lower melting point pure C. Again, with the assumption that low dilution of the clad by the substrate is desired, it is considered that the molten pool will have a C rich composition close to Co1 . Consider the solid/liquid interface at the leading edge of the clad track where the solid D substrate is in contact with the melt pool. In an isomorphous system, alloying of D with C reduces its melting point. Therefore, alloying © 2005 by CRC Press LLC

will assist the melting of the substrate layer by compositional dissolution into the molten pool and improved cladability. An opposite conclusion can be made when examining the solid/liquid interfacial region, where solid C particles come into contact with the melt pool. In this case, alloying of C with the melt pool will raise its melting point. However, in order to have a sustainable molten zone with a composition of Co1 the melt pool temperature would have to be above the non-equilibrium liquidus temperature for that composition (and therefore above the melting point of the pure C powder). Therefore, the pure C particles will undergo dissolution into the pool provided the local composition remains above Co1 . Consequently this type of clad couple should exhibit good cladability, although the nature of the alloying will not reduce the process conditions necessary to produce a good clad to the extent of that expected in a eutectic system. The second case to consider is that where the substrate is made of the lower melting point pure C and the powder, the higher melting point pure D. The requirement for low dilution is made such that the molten pool will have a D rich composition close to Co2 . Consider the solid/liquid interface region X of Figure 6.19, where the solid C substrate is in contact with the melt pool. Alloying with the melt pool will raise the melting point of the substrate, and therefore compositional dissolution will not assist in substrate melting. Alloying of unmelted D particles with the melt pool of composition Co2 would lead to compositional dissolution of the powders and therefore assist in their melting. However, if the desired bulk composition of the clad is Co2 , then the temperature of the process volume must be relatively high and close to the melting point of the pure D powder. This requirement for a high melt pool temperature will tend to create excessive melting of the lower melting point C substrate and cause undesirable levels of dilution. Therefore, while a cladding couple of this type may exhibit good cladability from the point of view of a dense clad and good bond formation, it may not be cladable if a very low dilution is required, particularly if the melting points of C and D are very far apart. 6.1.2.3

Complex Multi-Component Alloy Systems

The above discussion points out that metallurgical interactions occurring between the substrate and powder stream can assist or hinder the cladding process through an alteration of the melting temperature of the melt pool. In addition, large dierences between the melting point of the two starting constituents can lead to di!culties in controlling the degree of dilution and lead to high process temperatures and, therefore, high energies required to be delivered by the laser beam. However, in all of the above cases, a processing condition for good cladability, in terms of a continuous, dense clad and the establishment of a strong metallurgical bond between the clad and substrate, should be achievable. This will not always be the case for more complex alloy systems where intermediate phases can form as a result of alloying between © 2005 by CRC Press LLC

the substrate and powder. One of the major industrial applications of laser cladding is to create a hard, wear-resistant coating on a softer, more ductile substrate [35]. Unfortunately, this usually means that the melting point of the substrate and clad material are very dierent, and this tends to result in a complex phase diagram with several intermediate phases forming when the two elements are combined. The hypothetical phase diagram of Figure 6.21 is an example of such a phase diagram where intermediate, intermetallic compounds  and  can form through alloying between pure E and F . In addition, elements with such large dierences in melting point also exhibit low solubility with each other. Depicted in Figure 6.21 is a case where there is essentially no solubility of element F in E in the solid state as well as in the liquid state except at very high temperatures. Binary alloy systems with phase diagrams similar to this include Nb-Al, Ni-Al, Co-Al, Cr-Al and Fe-Al.

TmF Liquid

(F) α

β

TmE

β+Ε Pure F

Co4

C/o4

wt% E

Co3

Pure E

FIGURE 6.21 Hypothetical phase diagram of a binary system where alloying between the pure constituents (i.e., F and E ) leads to the formation of intermediate (or intermetallic) phases.

As discussed in the previous section, the success of the process will depend on the objectives of the cladding operation including the composition of the substrate, powder, and the desired composition of the as deposited clad. © 2005 by CRC Press LLC

Therefore, several cases must be treated to fully explore cladability in alloy systems of this kind. The first case that will be considered is one where the substrate is made from the low melting point element E and the powder made from the high melting point element F . The requirement for low dilution is made such that the molten pool will have an F rich composition close to Co4 . If the melting points of the two elements are very far apart it will be very di!cult to melt element F without causing excessive melting of the substrate material. Alloying of F particles, which come into contact with the melt pool, can result in compositional dissolution and this could help to melt the F particles. However, the melt pool temperature would have to be maintained at a very high level for a bulk composition of Co4 to be sustained. This would lead to excessive melting of the substrate and high dilution levels. The clad 0 and, upon solidification, the clad would composition could shift toward Co4 consist entirely of brittle  and  intermetallics. This would result in a poor bond with the substrate and poor cladability. The authors are aware of only one study that investigates the cladding situation described above. Li et al. [76] attempted to clad a Co-Cr rich alloy powder (i.e., Stellite 6) onto an Al-Si substrate. Over 90 wt% of the clad alloy is represented by Co and Cr while over 90 wt% of substrate is Al. Therefore, a qualitative understanding of the alloying expected between the substrate and powder can be interpreted by inspection of the Co-Al and/or Cr-Al binary phase diagrams (see Figures 6.22 and 6.23). Both binary alloys have a complex multi-component phase diagram similar in nature to the hypothetical phase diagram of Figure 6.21, including the formation of intermetallic compounds and very dierent melting points between Co (or Cr) and Al. The results of Li et al. [76] indicate that this system exhibits very poor cladability including di!culties in establishing process parameters, the formation of intermetallic compounds at the substrate/clad interface and excessive cracking in the bond region during cooling. 6.1.2.4

Case Study: Cladding of Nb on Al

Other conditions that have been reported with diering degrees of success are those where the composition of the clad deposited on the low melting point substrate has an E rich composition close to Co3 of Figure 6.21. Clad couples that have been investigated under these conditions include Nb-Al, Fe-Al, NiAl. The work on Nb-Al oers an excellent case study of what can occur under theses cladding conditions. In practical circumstances, some of the above problems can be somewhat overcome and a cladding of practical use can be produced. An example of this is the case of cladding Nb on Al, which has been studied by several researchers [239, 77, 240, 241]. The Nb-Al phase diagram is depicted in Figure 6.24. As indicated, the melting point of Nb is 2469 C whereas that of Al is 661 C. Clearly the production of a Nb rich melt pool would be excessively di!cult due to the high temperatures required. For example, even with alloying up to 10 wt% Al requires a melt pool temperature in excess of 2000 C. Some © 2005 by CRC Press LLC

FIGURE 6.22 Equilibrium phase diagrams for Al-Co (Source: Reprinted from [242], courtesy of ASM).

researchers have attempted to avoid this problem by using a powder mixture of both pure Nb and Al powders with bulk composition in the range of 75% wt Al. Providing these powders mix and alloy to form a completely liquid melt pool, the melt pool temperature can be considerably lower (e.g., in the range of 1400 to 1600 C depending on the exact composition) even without dilution by the molten Al substrate surface. However, work by Almeida et al. [239] indicates that even under these circumstances complications arise. Due to the high melting point of the Nb powder within the mixture, it is reasonable to assume that when these particles contact the melt pool surface they are still solid. Therefore, incorporation of these particles into the liquid state will require compositional dissolution of the Nb particles into the Al rich melt. Since there is essentially no solubility of Nb in liquid Al up to temperatures as high as 1400 C, this dissolution process is controlled by solid-state diusion within the Nb rich particles. For example, when the Nb particle comes into contact with the melt pool, a diusion couple develops between the two phases. Assuming that a pseudo-equilibrium develops, the intermetallic compounds depicted on the phase diagram can develop at the Nb/liquid interface. Thermodynamically, all of the phases can form between © 2005 by CRC Press LLC

FIGURE 6.23 Equilibrium phase diagrams for Al-Cr binary alloys (Source: Reprinted from [242], courtesy of ASM).

the Nb particle and melt pool. However, the rate at which each layer grows will determine if the phase is detectable using characterization tools. The work of Almeida et al. indicates that in the Nb-Al system, Al3 Nb is the dominate intermetallic phase that forms under the laser cladding conditions they used. Dissolution of the Nb particle into the melt requires the migration of the solid/liquid boundary (as well as the other solid-state boundaries) toward the center of the particle. This, in turn, requires solid-state diusion. This diusion process is relatively slow compared to the liquid diusion required to homogenize a full liquid melt pool. However the fine scale of the powders can minimize the time required for full dissolution. Calculations by Costa et al. [243] for the Nb-Al laser cladding conditions described above indicate that for 100 µm Nb particles, it would take 22 s for complete dissolution to occur. While this is a relatively short time frame for a solid-state process, it is much longer than the 0.24 s interaction time over which the melt pool is present at any position along the clad track [239]. In the case of Almedia et al.’s work, this results in a very nonhomogeneous clad microstructure with pores and undissolved Nb particles at the bottom of the clad zone suspended in a pure Al matrix (see Figure 6.25a) and a solidified structure at the top © 2005 by CRC Press LLC

FIGURE 6.24 Equilibrium phase diagrams for Al-Nb binary alloys (Source: Reprinted from [242], courtesy of ASM).

of the clad consisting of Al3 Nb dendrites, also in pure Al matrix (see Figure 6.25b). The explanation oered for this is that undissolved Nb particles, with a higher density than liquid Al, sink to the bottom of the Al melt pool where very little or no convection forces are present (recall the calculation of Equation (6.1) in the previous section). The top layer of the clad obtains a liquid composition consisting of Nb and Al. Upon cooling, the top layer solidifies with the formation of primary Al3 Nb dendrites and a pure interdentric Al phase. This indicates that the composition of the melt pool was to the right of the Al3 Nb phase boundary in the phase diagram of Figure 6.24. Because Al3 Nb is a brittle intermetallic compound [239], the bond strength of the Nb-Al clad system is not expected to be good. While bonding was not evaluated in detail, no cracking during processing or cooling was observed, and the intermetallic coating had an increased hardness compared to the Al substrate. The lack of cracking was attributed to the fine scale microstructure produced by the rapid solidification conditions of the laser cladding process. © 2005 by CRC Press LLC

a)

b)

FIGURE 6.25 Some Al-Nb clad microstructural features: a) an undissolved Nb (white) particle surrounded by Al3 Nb intermetallics phases (light grey) and suspended in a -Al matrix (black), b) Al3 Nb dendrites (light grey) with interdendritic (black) -Al (Source: Reprinted from [239], courtesy of Elsevier).

6.1.2.5

Case Study Cladding of Al on Mild Steel

In the above case, the general situation of a multi-component alloy system was considered, where the low melting point element was the substrate and the high melting point element (or a mixture made from it) was the powder. Another important general case to consider is the opposite situation (i.e., a high melting point substrate and low melting point powder). In this case, the initial requirement of low dilution is considered such that the bulk composition of the clad would be Co3 in Figure 6.21. Unlike the above case, melting of the powder should pose no problem due to its low melting point. However, the nature of alloying with the substrate during cladding can present problems which will be sensitive to the bulk composition of the formed clad. Some excellent examples of the problems that can occur [244] and some possible © 2005 by CRC Press LLC

solutions [245] are available for Al on mild steel. While mild steel does contain a small amount of C and Mn, over 98 wt% is Fe. Therefore, a qualitative understanding of the metallurgical interactions taking place can be achieved with the use of the Al-Fe binary phase diagram. Inspection of the Fe-Al binary phase diagram of Figure 6.26 reveals that it is a complex, multi-component binary system with similar characteristics to the hypothetical phase diagram of Figure 6.21, and those of the other binary Al systems described above. As indicated by Figure 6.26, when Al alloys with the iron substrate, there is an increased melting (or liquidus) temperature compared to pure Al. In fact, the liquidus of the alloy sharply rises with a relatively small increase in Fe content. Therefore, maintaining a melt pool in this alloy will require laser energies that will be a function of the level of dilution by the substrate. On the other hand, melting of the substrate can be assisted by compositional dissolution by the Al-rich melt. This coupled with the lower melting point of Fe compared to Nb and higher solubility in liquid Al can result in the formation of a homogeneous and easily sustainable alloyed melt pool.

C Fe

C Al

FIGURE 6.26 Equilibrium phase diagrams for an Al-Fe binary alloys (Source: Reprinted from [242], courtesy of ASM). © 2005 by CRC Press LLC

However, other problems are encountered in this clad couple as described by Ellis et al. [244]. With even a small amount of dilution of Al by the Fe from the mild steel substrate, the bulk composition of the melt pool will shift toward CAl as indicated in the Figure 6.26. The solidification of Alrich Fe melt pools under the rapid solidification conditions of laser processing has been studied in detail [204, 246]. While the phases that form during solidification depend on the melt pool temperature and composition as well as the solidification rate, primary dendrites of brittle Al3 Fe growing epitaxially from the substrate interface are di!cult to avoid at Fe contents above 3 or 4wt%. This structure was observed by Ellis et al [244] and resulted in cracking of the substrate/clad bond during cooling and thus, poor cladability. To attempt to overcome this problem, Ellis et al [244] placed several types of thin metallic coatings on the Fe substrate to prevent metallurgical interactions between Al and Fe. In the case of Cu coatings, the metallic layer completely alloyed with the melt pool leading to exposure of Al with the Fe below the coating. With the appropriate choice of laser processing conditions, a Ni coating remained in place while interacting with the melt pool. This effectively prevented interaction between Fe and Al. However, the metallurgical interactions that take place between Ni and Al, as the phase diagram of Figure 6.27 indicates, are similar to the Fe-Al system with the formation of an Al-rich Al-Ni melt pool. Nix Aly intermetallic compounds grew from the substrate/melt pool interface during solidification. In the Ni-Al case, cracking did not occur during cooling. However, during bend tests, cracking and spallation of the clad coating from the substrate surface occurred. It was determined that this fracture process took place within the Ni-Al intermetallic layer. The final attempt by the researchers [245] was to coat the Fe substrate with an Al coating followed by laser cladding of Al powder. In this case, laser conditions were chosen where the Al coating was not completely melted. It was successful due to the elimination of the intermetallic phase formation on the clad/substrate interface. Another possible solution to the cladding problem of Al on Fe is to alter the melt pool composition to avoid the formation of intermetallic compounds during solidification. This approach involves a powder mixture consisting of pre-blended pure Fe and pure Al powder, and a bulk composition of 18 wt% Al was clad onto a mild steel substrate. According to the Fe-Al phase diagram, the nominal melting temperature of this powder mixture is approximately 1450 C. Therefore, one of the advantages of this approach is that it brings the clad and substrate melting point temperature closer together, leading to better control over the laser cladding process particularly in the case of the level of dilution. Dilution levels near zero could be achieved with the appropriate laser processing conditions. The second advantage is that the diusion couple setup between the steel substrate and melt pool produces no intermediate compounds and solidification of the melt pool takes place along the dashed line of Figure 6.26, labelled CFe . The solidification path does not lead to the growth of intermetallic compound but does result in © 2005 by CRC Press LLC

FIGURE 6.27 Equilibrium phase diagram of the Al-Ni binary alloy mixture (Source: Reprinted from [242], courtesy of ASM).

the solidification of -Fe single-phase solid solution. Figure 6.28 illustrates a region of the solidified clad/substrate region. The clad microstructure consists of single phase columnar grains of FeAl solid solution. As discussed in section 6.1.1.1, this leads to a very strong bond between the clad and substrate.

6.2

Solidification Conditions Encounter in Laser Cladding

In Section 6.1, the metallurgical aspects of cladability were considered in detail. In particular, it was clearly demonstrated that alloying between the clad powder and substrate determines the melt pool composition. In addition, the phases that form during solidification, which are determined by substrate/clad alloying, play a major role in the development of the bond strength of the clad/substrate interface. In the Al binary alloy case studies discussed above, all three cases had an Al rich melt pool formed during cladding. Because the © 2005 by CRC Press LLC

50 µm

Clad deposit

HAZ

FIGURE 6.28 Microstructure of FeAl clad/substrate (steel) region for m ˙ = 2 g/minute and U = 2.0 mm/s (W = 10.5 ms, F = 30 Hz, rl = 0.6 mm, Pl = 343W).

nature of metallurgical interactions in the systems are similar, during solidification Al-Fe (or -Ni or -Nb) intermetallic compounds grow epitaxially from the substrate interface during solidification. However, despite these similarities, dierences in the mechanical strength of the bonds formed do exist. For example, cracking during cooling was observed in the Al-Fe system, but not in the cases of Al-Ni or Al-Nb. Some of these dierences are due to the different physical and mechanical properties of the intermetallics being formed (i.e., Al3 Fe, Al3 Ni or Al3 Nb) and the dierent thermal expansion mismatches that develop during cooling for the three clad/substrate couples. Another important influence is what types of microstructural features form during solidification. For example Ellis et al. [244] report that the Al3 Fe phase grows as a continuous layer at the clad substrate interface. This continuous brittle phase would be very susceptible to cracking during cooling. Work on the Al-Ni and Al-Nb couples indicate that the Al3 Ni or Al3 Nb grow as discrete primary dendrites perpendicular to the substrate interface. This type of structure would be less susceptible to cracking. In particular, Almeida et al. [239] attribute the lack of cracking in the Al3 Nb layer to be due to the formation of interdendritic Al during solidification. This phase is soft and ductile and therefore contributes to an overall increase in the toughness of the Al3 Nb rich microstructure, making it much more resistant to crack propagation. The above discussion points out that the nature of the microstructure that forms during solidification plays a very important role in determining the bond strength achievable during laser cladding. In addition, the microstructure that forms throughout the bulk of the clad deposition will also largely determine the physical and mechanical properties of the coating and therefore its suitability and performance in an application. For these reasons it © 2005 by CRC Press LLC

is extremely important to understand the conditions of solidification which can occur in laser cladding and the microstructures that result. This is the primary purpose of this section.

6.2.1

Process Conditions

Figure 6.29 illustrates a schematic of the process volume developed in laser cladding. Laser processing provides a localized heat source and a short interaction time (i.e., 0.5 s or less). This, coupled with the rapid heat removal by conduction into the substrate below the melt pool, leads to very rapid solidification rates (i.e., solid/liquid interface velocities as high as 1 to 2 m/s [247]). The heat removal is also very directional which sets up large positive temperature gradients (i.e., on the order of 106 K/m [247]) in the melt pool ahead of the solid/liquid interface.

Laser Beam

Powder Stream

Solidified clad

C

U VSs M Velt b po ol

U

A

θ

B Substrate

Melt pool

FIGURE 6.29 A schematic of the process zone during laser cladding. Melting takes place from A to B while solidification takes place from B to C . The lines within the solidified clad deposit indicate the direction of solidification.

© 2005 by CRC Press LLC

The velocity of the solid/liquid interface Us (i.e., rate of solidification) can be related to the scanning speed of the laser beam U (or in the case of a stationary beam, the velocity of the substrate), through the formula [247] Us = U cos 

(6.2)

where  is the angle between the vector representing the direction of the substrate motion and the vector normal to the solid/liquid surface at a particular point (i.e., the solidification rate). Inspection of Figure 6.29 reveals that the solidification rate Us varies throughout the process volume from zero at point B (where  approaches 90 ) to a maximum at the clad surface (where  approaches 0 ). Both the solidification rate and temperature gradient can be evaluated numerically, an example of which is given in Figure 6.30 for laser surface remelting [247]. The figure indicates the dependence of both solidification rate and temperature gradient G on the process speed. Also indicated is that the temperature gradient is highest at the bottom of the melt pool where Us is zero and decreases toward the surface. This indicates that the solidification conditions vary throughout the melt pool. In order to understand how these processing conditions influence the microstructure that forms during solidification, the concept of constitutional supercooling must be described.

6.2.2

Constitutional Supercooling

Consider a simple eutectic alloy with a bulk composition of Co and a binary phase diagram of Figure 6.31. Further, consider that solidification has achieved a steady state as described by Porter et al. [248], where the interface temperature has reached a constant temperature of T1 (the solidus temperature for the bulk alloy), the solid/liquid interface is flat or planar and moving at a constant velocity Us . In addition, it is assumed that solidification is proceeding at a fast enough rate that solid-state diusion is negligible, and mixing in the liquid is limited to diusional mass transport of solute in the liquid phase. Under these solidification conditions, the solute profile that develops across the solid/liquid interface is similar to that described in Figure 6.32a, where it is assumed that the interface compositions (and partition coe!cient ke ) are determined by the equilibrium phase diagram. The width of the concentration profile in the liquid phase ahead of the solid/liquid interface has a characteristic width D/Us which is dependent on the solidification rate and liquid phase diusivity [248]. As the binary phase diagram indicates, the liquidus temperature is determined by the composition of the liquid. Since the solute composition changes near the solid/liquid interface so also does its liquidus temperature. The liquidus temperature as determined by the phase diagram for the liquid ahead of the solid/liquid interface is plotted in Figure 6.32b (i.e., curve marked TL ). Since the liquid composition varies from a maximum of Co /k at the interface © 2005 by CRC Press LLC

U a) U

U

Us

U

b)

FIGURE 6.30 Calculated solidification conditions for laser surface remelting at two dierent process speeds: a) the temperature gradient, Gr , b) the solidification rate, Us (d is the distance into the melted zone from the surface) (Source: Reprinted from [247], courtesy of ASME).

to Co a short distance into the liquid, the liquidus temperature increases from T1 at the interface to a constant value of T2 a distance into the liquid [248]. In the case of laser cladding and many other solidification processes the “real” temperature gradient in the liquid ahead of the solid/liquid interface, Gr is positive as indicated in Figure 6.32b. The actual magnitude of Gr depends on the solidification conditions. If it is of the magnitude of that shown in Figure 6.32b, then a region of liquid ahead of the interface is known to be constitutionally supercooled. In other words, a certain portion of the liquid with elevated solute composition “experiences” a real temperature which is below its liquidus temperature determined by its composition. If a small protrusion develops at the interface its rate of growth will be larger than the bulk of the planar interface [248] in the constitutionally supercooled region. Therefore, its growth is promoted and the planar interface will become unstable. Depending on the degree of constitutional supercooling, the interface may develop a cellular structure or dendritic structure. A critical temperature gradient Gc can be defined (see Figure 6.32b) above, which no supercooling occurs and the solidification front will remain planar. © 2005 by CRC Press LLC

Liquid

TmB T2

T1

Solidus line

Liquidus line

(B)

Pure B

TmA

(A) TE

Co

Co/ke

wt% A

Pure A

FIGURE 6.31 A schematic of a hypothetical phase diagram indicating steady-state solidification conditions.

Inspection of the figure reveals that the slope of this critical gradient can be defined as T2  T1 (6.3) Gc > D/Us where T2 T1 is known as the equilibrium freezing range for the alloy. Another form of Equation (6.3) is (Gc /Us ) >

T2  T1 D

(6.4)

The left side of Equation (6.4) describes the important processing conditions present during solidification while the right side describes the material parameters of importance. It should be further noted that the magnitude of the equilibrium freezing range is determined by the equilibrium partition coe!cient defined as Cs (6.5) ke = Cl where Cs and Cl are the equilibrium solidus and liquidus compositions at T1 (see Figure 6.31). When the value of ke approaches unity this corresponds to solidus and liquidus lines that are close together, resulting in small freezing ranges. From Equation (6.4), it can be stated that high temperature gradients and slow solidification rates will be more likely to avoid constitutional supercooling © 2005 by CRC Press LLC

Interface

Co / ke Solid

Wt% A

Liquid

Co D /Us

Position, x a)

Interface

D / Us

Gr

Gc

Temperature

T2 TL Solid

Liquid

T1

Position, x b)

FIGURE 6.32 a) Compositional gradient, b) temperature gradient at the solid/liquid interface during steady-state solidification.

and favor a planar growth condition for a given alloy system. Alternatively, alloys with small freezing ranges will be more likely to solidify with a planar front under moderate values of Gr and Us . Applying these concepts to laser cladding, it can be stated by inspection of Figures 6.29 and 6.30 that at the base of the clad layer, region B, (Gr /Us ) is at a maximum and therefore solidification by a planar front would be favored in this location. On the other hand, (Gr /Us ) are at their lowest at the surface, region C. Therefore, a transition from planar solidification to dendritic growth from the clad deposit near the substrate interface to the clad surface would be expected. Figure 6.30 also illustrates that the change in the ratio of (Gr /Us ) per unit length of the melt pool depth is more significant at higher laser processing speeds. © 2005 by CRC Press LLC

An excellent example of how these changing (Gr /Us ) values alter the microstructure in a clad deposit is given by Vilar [249] for a Stellite 6 clad deposit. At the base of the clad near the substrate surface, a microstructurally featureless zone indicates a region of plane front solidification. This type of zone has been observed by many researchers [247]. The microstructure then transforms to a cellular structure and then finally a fully developed dendritic structure with secondary dendrite arms.

6.2.3

Rapid Solidification

Many complex engineering alloys have a significant equilibrium freezing range and therefore dendritic growth is often the dominant solidification process and therefore most common microstructural feature observed in the clad deposit. This is due to the di!culty in meeting the criterion of Equation (6.4) for a planar form when [T2 T1 ] is relatively large. However, with increasing solidification rates the compositions in the solid and liquid at the tip of the dendrites no longer follow their equilibrium values. This is because the atoms do not have the time to diuse and maintain local equilibrium at the solid/liquid interface. One primary way in which to describe this eect is through the use of the solidification rate dependent partition coe!cient kv [250] (6.6) kv = [ke + ao Us /D]/[1 + ao Us /D] where ao is a characteristic value of the thickness of the interface and the other parameters are as defined above. As the solidification rate Us increases, kv approaches unity. Kurz et al. [247] have applied non-equilibrium solidification modelling to an Al-4%Cu alloy to determine the concentrations in the solid and liquid (i.e., CS and CL , respectively) at the dendrite tip as a function of temperature for high solidification rates. Figure 6.33 plots these results. Also included are the equilibrium solidus and liquidus lines. The fine dashed line shows the values of CS and CL as a function of temperature. At low undercoolings (and therefore low solidification rates), the non-equilibrium values follow closely the equilibrium lines. However, with increasing undercoolings, they deviate significantly. Also plotted in the figure are the eective solidus and liquidus lines at a solidification rate of Va , which Kurz and Trivadi [247] define as the absolute stability point. Clearly, the value of kv is significantly closer to unity than the equilibrium partition coe!cient and therefore the effective freezing range of the alloy is considerably reduced. According to the criterion of Equation (6.4), this could lead to the re-establishment of a stable, planar interface at high solidification rates. This phenomenon has been modelled by Kurz and Trivadi [247], the results of which, for Al-4%Cu, are presented in Figure 6.34. Clearly the type of microstructure that develops depends on both the solidification rate and temperature gradient. A region of planar growth (P ) exists at both low and high growth rates but does depend on the temperature gradient at low values © 2005 by CRC Press LLC

FIGURE 6.33  Calculated compositions in the solid and liquid (CS and CL , respectively) and eective solidus and liquidus lines for an Al-Cu alloy as a result of rapid solidification. Va is the solidification rate corresponding to the absolute stability point. Solid lines indicate the equilibrium solidus and liquidus lines (Source: Reprinted from [247], courtesy of ASME).

of Us . At intermediate growth rates, non-planar growth, in the form of cells or dendrites, exists (see Figure 6.34b). Included in Figure 6.34a are the laser surface remelting traces of Figure 6.30. Based on these traces, one would predict a planar structure near the substrate interface followed by a cellular region and then dendritic structure. At very high rates the planar solidification front can also be replaced by a banded structure. This type of transition has been observed in Al-Cu [251] and Al-Fe [246] binary systems.

6.2.4

Microstructure Maps

From the above discussion, it is clear that both the laser cladding processing conditions, which determine the values of Gr and Us with the melt pool, and the alloy composition, which determines the magnitude of the freezing range (or the values of ke and/or kv ), will determine the type of microstructure formed in the clad layer during solidification. The case studies of Al-Nb, © 2005 by CRC Press LLC

a)

b)

FIGURE 6.34 Mode of solidification and the resultant microstructure as a function of the temperature gradient Gr and solidification rate Us , a) including laser conditions, b) including type of non-planar growth. P-planar, C(F) and C-cellular, D-dendrites and B-banding (Source: Reprinted from [247], courtesy of ASME).

© 2005 by CRC Press LLC

Us

Al-Ni and Al-Fe given in Sections 6.2.2.1 and 6.2.2.2 illustrate that this microstructure can have a large impact on the substrate/clad bond strength, and thus the success of the cladding process. It would, therefore, be extremely useful to be able to predict what microstructures would develop under certain processing conditions so that the optimum microstructure could be selected by the appropriate choice of laser processing parameters. Several researchers [204, 246, 251, 252] have developed microstructural maps with the aim of providing this type of information. An example of such a map is given in Figure 6.35 for an Al rich binary Al-Fe alloy. Similar maps have also been developed for Al-Cu alloys [251] and Ag-Cu [252]. The solid lines in the figure indicate results of experimental observations, whereas the dashed lines indicate calculated values from solidification models.

FIGURE 6.35 The type of microstructure observed (solid lines) and predicted (dashed lines) in an Al-Fe binary alloy as a function of solidification rate and Fe content (Source: Reprinted from [246], courtesy of Elsevier).

At low Fe contents and slow solidification rates, primary dendrites of Al with interdendritic regions of equilibrium eutectic (Al-Al3 Fe) are the preferred microstructure. As the solidification rate increases, a similar microstructure develops except that the equilibrium (or stable) eutectic is replaced by a metastable Al-Al6 Fe eutectic. The formation of metastable phases during rapid solidification has been observed in other alloy systems [240, 253, 254, © 2005 by CRC Press LLC

255]. With a further increase in the solidification rate, the condition for absolute stability of the solid/liquid interface is approached with the formation of a banded and then plane front structure. With an increase in the Fe content, the sequence of microstructural development as a function of solidification rate changes. In the slow to moderate range, the preferred microstructure changes from primary dendrites of Al with interdendritic Al-Al3 Fe (or Al-Al6 Fe), to an all eutectic Al-Al3 Fe (or Al-Al6 Fe) structure and then finally to primary dendrites of Al3 Fe with interdendritic Al-Al3 Fe (or Al-Al6 Fe). At very high solidification rates, the microstructure that forms becomes insensitive to the Fe content where banding or a plane front becomes the preferred structure. What this diagram points out is that with a careful control over the process parameters, including dilution by the substrate (or clad composition), a range of microstructures for the clad can be selected. In the particular case of the problems encountered by Ellis et al. [244] for an Al cladding on Fe, the problems encountered with the formation of brittle Al3 Fe dendrites at the surface may be avoidable if the level of Fe in the cladding (and thus dilution) could be kept to a minimum. Microstructure maps similar to Figure 6.35 are not yet available for more complex ternary and higher order alloy systems. However, the work of Kurz and others points out the tremendous value of these maps and the importance of developing them for other systems.

6.2.5

Microstructural Scale

Another consequence of rapid solidification is that it causes a refinement in the microstructure of the as-cast material. In the case of cellular or dendritic growth, this can be understood by referring to Figure 6.32. As the solid grows it must reject solute into the liquid ahead of the solid/liquid interface. With the formation of a cellular or dendritic structure, solute can be more e!ciently rejected in lateral directions as well as in front of the tip. As the solidification rate increases, the need to reject this solute more eectively increases and the dendrite tip radius becomes smaller. This results in a refinement in the dendrite spacing and overall microstructure. An example of how the laser processing speed can be used to refine the clad microstructure (through an increase in solidification rate) is given in Figure 6.36 for the cobalt-based Stellite 6 alloy [80]. At a slow scanning speed of 1.67 mm/s, the secondary dendrite arm spacing s is in the range of 2.5 to 6 µm, whereas at a scanning speed of 167 mm/s, s is reduced to 0.5 to 0.8 µm. Because the solidification rate varies from near zero to a maximum at the clad surface, it is expected that the microstructure of the clad should also be refined across its height. An example of this refinement is given in Figure 6.37 for an Al-Si-Ni clad material placed onto an Al-Si substrate [256]. In this case, solidification of primary dendrites of Al3 Ni2 and/or Al3 Ni (the white phase in the figure) with interdendritic Al-Al3 Ni are the dominant microstructural features. A clear refinement in the scale of the dendrites is visible in going © 2005 by CRC Press LLC

FIGURE 6.36 Microstructure of a Co-based Stellite 6 clad deposit at a) a laser scanning speed of 1.67 mm/s, b) a laser scanning speed of 167 mm/s (Source: Reprinted from [80], courtesy of Elsevier).

from the bottom to the top of the clad. Gilgien et al. [246] have constructed a microstructure map for Al-Fe which includes the dendrite tip radius (or in the regions where eutectic forms, the interphase spacing). A clear correlation exists between solidification rate and microstructural scale (see Figure 6.38). This diagram also illustrates the great potential that laser cladding has in producing nanoscaled microstructures. These ultra-fine microstructures, the finest of which will be located at the clad surface, are in large part the reasons for the high hardness values or increased wear resistance so often reported in laser cladding research [35, 239, 247, 249, 253, 254, 256, 257]. An additional consequence of rapid solidification is the development of extended solid solubility. This can be understood with reference to Figure 6.33. © 2005 by CRC Press LLC

FIGURE 6.37 Microstructure of an Al-Si-Ni clad on an Al-Si substrate showing the scale of the primary Ni aluminide dendrites (white phase) as a function of the clad depth (Source: Reprinted from [256], courtesy of Elsevier).

© 2005 by CRC Press LLC

Us FIGURE 6.38 Scale of the solidified microstructure of laser remelted Al-Fe as a function of solidification rate and Fe content (Source: Reprinted from [246], courtesy of Elsevier).

As the solidus line shifts to the right of the equilibrium case, more solute remains in solid solution within the primary phase. This phenomenon has been investigated by a number of researchers in systems including Ag-Cu [252], NbAl [241], and Ni-based alloys [51, 49, 160]. Solubility increases can be as high as several percent. This increased solubility will lead to higher levels of solid solution strengthening and contributed to increase hardness and strength in the clad material. Accompanying the non-equilibrium solute content is the formation of metastable phases such as the Al6 Fe based eutectic illustrated in Figure 6.35. The formation of metastable phases have been observed in Ni-based, Cobased, and Fe-based alloys mostly in the form of metal carbides and/or borides [253, 254, 255]. Again, these metastable phases will have an impact of the mechanical properties of the clad deposit.

6.3

Material Systems Used in Laser Cladding

There has been a rapid increase in research activities related to laser cladding, growing from a handful of published journal papers in the 1980’s to over 60 in the year 2003. A wide range of materials have been studied, and it © 2005 by CRC Press LLC

can be safely said that the feasibility of laser cladding using every major class of metallic alloy as well as ceramics, glasses and intermetallics has been investigated. Alloy systems that have been laser clad include Ni, Co, Nb, Ti, Al, and Cu-based alloys as well a full range of steels including tool and stainless steels. However, over the past several years, two general material systems have enjoyed particular attention and will be the focus of this section. By its nature, the major objective of laser cladding is to create a high performance surface with enhanced properties including wear resistance, corrosion resistance, and/or oxidation resistance. Traditionally, materials with these high performance attributes are high temperature materials, which are expensive and di!cult to process in bulk form. Therefore, creating a coating of these materials on less expensive, easily processed base materials is where laser cladding oers the best advantages.

6.3.1

High Temperature Alloys

Over the span of the last 20 years there has been a major emphasis on the laser cladding of high temperature Ni, Co and/or Fe-based alloys. These alloys are generally expensive and di!cult to process by conventional means but oer superior properties. In the majority of cases, these alloys are clad onto relatively less expensive low carbon or low alloy steels [75, 112, 77, 81, 253, 254, 255, 109, 112, 79]. The main alloying elements in most of the clad alloys include Ni, Co, Cr and Fe with smaller amounts of W and Mo as well as C, which forms metal carbides with the alloy ingredient. Generally, these alloys exhibit good cladability with low carbon steel substrates. This can be qualitatively understood in light of the discussion of Section 6.2.2 by realizing that Fe is very soluble in Ni, Co, and Cr in the solid-state near the melt pool temperatures with no intermetallic phases forming between the constituents during solidification. However, this general statement can be complicated by complex alloy compositions and with the presence of C, B and W . Metal carbides can form during solidification and brittle, intermetallic compounds can precipitate in the solid-state during cooling. Therefore, hot and cold cracking can take place in the deposit/substrate during laser cladding [258]. Often pre- and/or post-heat treatments are required to avoid these problems. However, a detailed discussion of these problems is beyond the scope of this text. Another common application of laser processing using high temperature alloys is the cladding of Ni-based substrates with Ni-based cladding containing rare earth elements such as Hf [77, 259, 260, 261]. In these cases the high temperature oxidation resistance of Ni-based superalloys can be improved by the rare earth additions. In these cases, since the clad composition is only slightly dierent than the substrate, cladability is generally very good.

© 2005 by CRC Press LLC

6.3.2

Composites

Since a primary application of laser cladding is to create a hard wear-resistant surface, it is not surprising that a significant research eort is being made to laser clad metal/ceramic composite coatings onto metal substrates [262, 263, 264, 265, 266, 267, 268, 269, 270]. Metal/ceramic composite (MCC) materials are relatively di!cult to process by conventional means such as powder metallurgy or ingot metallurgy, particularly when high ceramic volume fractions are required [271]. These composites also tend to have low ductility and fracture toughness, making them unsuitable in bulk form for many applications [272]. However, by virtue of their ceramic content they exhibit high hardness and excellent wear resistance. Therefore, their use in laser cladding is an ideal application [262, 263, 264, 265, 266, 267, 268, 269, 270]. Generally, there have been two approaches to the cladding of MCCs. In the first, ceramic particles, (sometime with metallic binders) are mixed with metal particles and added to the melt pool either by pre-placement on the substrate or by powder injection [262, 263, 264]. In these investigations, the goal is to limit the reaction between the ceramic particles and the metallic melt pool. WC particles are the ceramic particles of choice in this approach because they seem to be resistant to dissolution into the melt pool. The problems encountered during this process are similar to those encountered in more conventional ingot metallurgy techniques for producing MCCs [272, 273]. The ceramic particles must wet the liquid metal in order to be incorporated into the melt pool. However, there must not be an extensive deleterious reaction taking place between the particle and alloying elements with the metallic liquid. In addition, during solidification particle pushing can lead to inhomogeneous distributions of the ceramic particles within the metal matrix. These problems are only just beginning to be evaluated in laser cladding [263]. In cases where TiC and SiC are used as the ceramic particle, reaction does take place with the metal powder [71, 265, 266]. For example, when SiC and Ni are combined, nickel-silicide compounds form [265] and for TiC and Al, titanium-aluminide compounds form [71]. The reactions that take place between the ceramic particles and melt pool are not necessarily negative for either cladability or wear resistance. In fact, in a second approach, ceramic reinforcement is developed by in-situ reactions between the powder ingredients within the molten zone. [265, 266, 267, 268, 269, 270]. In some cases, ceramic particles are added to the clad deposit and undergo partial or complete dissolution into the melt pool and re-precipitate during solidification [265, 274]. In other cases elemental powders are added to the melt pool and then react to form the reinforcing phase. Examples of this include the addition of Mo, Si and C powders which form MoSi2 and SiC particles [267], the addition of Ni-based alloy and Zr powder to form ZrC particles [269], and Ni-based alloy, graphite and Ti powder to form TiC particles [270]. The advantages of this approach include the formation of fine ceramic particles that are well bonded to the matrix. As Figure 6.39 indicates, © 2005 by CRC Press LLC

a large increase in hardness can result from the in-situ formation of the laser clad MCCs. This fairly new approach to laser cladding shows great potential for the creation of novel materials with unique properties.

a)

b)

FIGURE 6.39 Vickers microhardness readings across clad/substrate couples: a)in-situ formation of ZrC MCC clad on a mild steel substrate, b) in-situ formation of MoSi2 and SiC MCC clad on a commercially pure Al substrate (Source: preprinted from [269] and [267], respectively, courtesy of Elsevier).

© 2005 by CRC Press LLC

7 Safety

Laser material processing is hazardous if all the safety guidelines are not followed. Safety in working with high-power lasers should be paramount because of the serious damage they could impose. In the following, the safety issues related to the laser material processing are discussed .

7.1

Laser Classification

All lasers are classified by the manufacturer and labelled with the appropriate warning labels. Any modification of an existing laser or an unclassified laser must be classified by a laser safety o!cer prior to use. The following criteria are used to classify lasers: 1. Wavelength. If the laser is designed to emit multiple wavelengths the classification is based on the most hazardous wavelength. 2. For continuous wave (CW) or repetitively pulsed lasers, the average power output [W] and limiting exposure time inherent in the design are considered. 3. For pulsed lasers, the total energy per pulse [J], pulse duration, pulse repetition frequency and emergent beam radiant exposure are considered. Class 1 Lasers These are lasers that are not hazardous for continuous viewing or are designed in such a way that prevent human access to laser radiation. These consist of low power lasers or higher power embedded lasers (i.e., laser printers). Class 2 Visible Lasers (400 to 700 nm) W This

chapter is reprinted from the University of Waterloo safety guidelines with permission.

© 2005 by CRC Press LLC

Lasers emitting visible light that because of normal human aversion responses, do not normally present a hazard but would if viewed directly for extended periods of time (i.e., many conventional light sources). Class 3A Lasers that normally would not cause injury to the eye if viewed momentarily but would present a hazard if viewed using collecting optics (fiber optics loupe or telescope). Class 3B Lasers that present an eye and skin hazard if viewed directly. This includes both intrabeam viewing and specular reflections. Class 3B lasers do not produce a hazardous diuse reflection except when viewed at close proximity. Class 4 Lasers Lasers that present an eye hazard from direct and specular reflections. In addition, such lasers may be fire hazards and produce skin burns. In this class, CO2 beams are much safer. The wavelength of a CO2 beam is 10.6 µm, a wavelength that is strongly absorbed by water. Since more than 70% of the human body consists of water, an unfocused or reflected CO2 beam does not penetrate beyond the first few layers of skin. More importantly, this wavelength does not penetrate the eye if a scattered beam is observed. Processing can usually be directly observed with minimal precautions. Lasersafety eye wear is normally su!cient to meet the guidelines. YAG beams have a 1.06 µm wavelength — just below the visible deep-red. This wavelength deeply penetrates the body. If you get hit by a YAG beam, there is more damage than just a surface burn. Worse, the 1.06 µm wavelength is focused by the eye just like “normal” light. A beam scattered from a process will be focused to a point in the eye, probably destroying the spot where it was focused. As such, all YAG workstations must be sealed o and light-tight during processing.

7.2 7.2.1

Laser Hazards Eye Hazards

The potential for injury to the dierent structures of the eye (Figure 7.1) depends upon which structure absorbs the energy. Laser radiation may damage the cornea, lens or retina depending on the wavelength, intensity of the radiation and the absorption characteristics of dierent eye tissues. Ocular Image Wavelengths between 400 nm and 1400 nm are transmitted through the curved cornea and lens and focused on the retina. Intrabeam viewing of a © 2005 by CRC Press LLC

Retina Iris Cornea

V itreous Lens

Fovea

Macula

Optic Disc A queous

Optic Nerve

FIGURE 7.1 Structures of the eye.

point source of light (see Figure 7.2a) produces a very small spot on the retina resulting in a greatly increased power density and an increased chance of damage. A large source of light such as a diuse reflection of a laser beam produces light called extended source which enters the eye at a large angle. An extended source produces a relatively large image on the retina (see Figure 7.2b); energy is not concentrated on a small area of the retina as in a point source. Details of Irradiation Eects on Eyes Cornea absorbs all UV light which produces a photokeratitis (weld). Ultraviolet -B+C (100 — 315 nm) The surface of the flash by a photochemical process causes a denaturation of proteins in the cornea. This is a temporary condition because the corneal tissues regenerate very quickly. Ultraviolet -A ( 315 — 400 nm) The cornea, lens, and aqueous humour allow ultraviolet radiation of 315 — 400 nm wavelengths of which the principal absorber is the lens. Photochemical processes denature proteins in the lens resulting in the formation of cataracts. Visible light and Infrared-A (400 — 1400 nm) The cornea, lens, and vitreous fluid are transparent to electromagnetic radiation of these wavelengths. Damage to the retinal tissue occurs by absorption © 2005 by CRC Press LLC

Extended Source

Point Source a) (a)

b) (b)

FIGURE 7.2 a) Point source viewing, b) Extended source viewing.

of light and its conversion to heat by the melanin granules in the pigmented epithelium or by photochemical action to the photoreceptor. The focusing eects of the cornea and lens will increase the irradiance on the retina by up to 100,000 times. For visible light (400 to 700 nm), the aversion reflex which takes 0.25 s may reduce exposure causing the subject to turn away from a bright light source. However, this will not occur if the intensity of the laser is great enough to produce damage in less than 0.25 s or when light of 700 — 1400 nm (near infrared) is used since the human eye is insensitive to these wavelengths. Infrared-B and Infrared-C (1400 to 106 nm) Corneal tissue will absorb light with a wavelength longer than 1400 nm. Damage to the cornea results from the absorption of energy by tears and tissue water causing a temperature rise and subsequent denaturation of protein in the corneal surface. Wavelengths from 1400 to 3000 nm penetrate deeper and may lead to the development of cataracts resulting from the heating of proteins in the lens. The critical temperature for damage is not much above normal body temperature (w 370 ). Laser Radiation Eects on Skin Skin eects are generally considered of secondary importance except for highpower infrared lasers. However, with the increased use of lasers emitting in the ultraviolet spectral region, skin eects have assumed greater importance. Erythema (sunburn), skin cancer and accelerated skin aging are produced by emissions in the 200 to 280 nm range. Increased pigmentation results from exposure to light with wavelengths of 280 to 400 nm. Photosensitization has resulted from the skin being exposed to light from 310 to 700 nm. Lasers emitting radiation in the visible and infrared regions produce eects that vary from a mild reddening to blisters and charring. These conditions are usually repairable or reversible; however, depigmentation, ulceration, and scarring of © 2005 by CRC Press LLC

CIE Band

U V-C

100

UV-B

280

UV-A

315

400

Photokeratitis

Visible

700

Retinal Burns

Erythema

IR-C

3000 (nm)

1400

C orneal Burns

Cataracts

Cataracts Adverse Effects

IR-B

IR-A

Colour Vision Night Vision Degradation Thermal Skin Burns

FIGURE 7.3 Interaction of optical radiation and various tissues.

the skin, and damage to underlying organs may occur from extremely highpower lasers. Summary of Wavelengths of Light and their Eects on Tissues A summary of interaction of optical radiation and various tissues is shown in Figure 7.3. The wavelengths are divided into bands as defined by the International Commission on Illumination (CIE).

7.2.2

Collateral Radiation

Radiation other than that associated with the primary laser beam is called collateral radiation. For example, x-rays, UV, plasma, and radio frequency emissions are collateral radiation. Ionizing Radiation X-rays could be produced from two main sources in the laser laboratories. One is high-voltage vacuum tubes of laser power supplies, such as rectifiers, thyratrons and crowbars, and the other is electric-discharge lasers. Any power supplies that require more than 15 kilovolts (keV) may produce enough x-rays to cause a health hazard. Interaction between x-rays and human tissue may cause a serious disease such as leukemia or other cancers, or permanent genetic eects which may show up in future generations. UV and Visible UV and visible radiation may be generated by laser discharge tubes and pump lamps. The levels produced may exceed the maximum permissible exposure (MPE) and thus cause skin and eye damage. © 2005 by CRC Press LLC

Plasma Emissions Interactions between very high-power laser beams and target materials may in some instances produce plasmas. The plasma generated may contain hazardous UV emissions.

7.2.3

Electrical Hazards

The most lethal hazard associated with lasers is the high voltage electrical systems required to power lasers. Several deaths have occurred when commonly accepted safety practices were not followed by operators working with high voltage sections of laser systems. Safety Guidelines 1. Do not wear rings, watches or other metallic apparel when working with electrical equipment. 2. Do not handle electrical equipment when hands or feet are wet or when standing on a wet floor. 3. When working with high voltages, regard all floors as conductive and grounded. 4. Be familiar with electrocution rescue procedures and emergency first aid. 5. Prior to working on electrical equipment, de-energize the power source. Lock and tag the disconnect switch. 6. Check that each capacitor is discharged, shorted and grounded prior to working in the area of the capacitors. 7. Use shock preventing shields, power supply enclosures, and shielded leads in all experimental or temporary high-voltage circuits.

7.2.4

Chemical Hazards

Many dyes used as lasing media are toxic, carcinogenic, corrosive or cause of a fire hazard. All chemicals must be accompanied by a material safety data sheet (MSDS). The MSDS will supply appropriate information pertaining to the toxicity, personal protective equipment and storage of chemicals. Various gases are exhausted by lasers and produced by targets. Proper ventilation is required to reduce the exposure levels of the products or exhausts below standard exposure limits. Cryogenic fluids are used in cooling systems of certain lasers. As these materials evaporate, they replace the oxygen in the air. Adequate ventilation must be ensured. Cryogenic fluids are potentially explosive when ice collects in valves or connectors that are not specifically designed for use with cryogenic © 2005 by CRC Press LLC

fluids. Condensation of oxygen in liquid nitrogen presents a serious explosion hazard if the liquid oxygen comes in contact with any organic material. Although the quantities of liquid nitrogen that are used are small, protective clothing and face shields must be used to prevent freeze burns to the skin and eyes. Compressed gases used in lasers present serious health and safety hazards. Problems may arise when working with unsecured cylinders, cylinders of hazardous materials not maintained in ventilated enclosures, and gases of dierent categories (toxins, corrosives, flammable, oxidizers) stored together.

7.2.5

Fire Hazards

Class 4 lasers represent a fire hazard. Depending on construction material beam enclosures, barriers, stops and wiring are all potentially flammable if exposed to high-beam irradiance for more than a few seconds.

7.2.6

Explosion Hazards

High-pressure arc lamps, filament lamps, and capacitors may explode violently if they fail during operation. These components are to be enclosed in a housing that will withstand the maximum explosive force that may be produced. Laser targets and some optical components also may shatter if heat cannot be dissipated quickly enough. Consequently, care must be used to provide adequate mechanical shielding when exposing brittle materials to high intensity lasers.

7.2.7

Eye Protection

The following is an account written by a researcher who sustained permanent eye damage viewing the reflected light of a Class 4 neodymium YAG laser emitting a 10-ns pulse of 6 mJ radiation at 1064 nm. “When the beam struck my eye, I heard a distinct popping sound caused by a laser-induced explosion at the back of my eyeball. My vision was obscured almost immediately by streams of blood floating in the vitreous humour. It was like viewing the world through a round fish bowl full of glycerol into which a quart of blood and a handful of black pepper have been partially mixed.” Dr. C.D. Decker. The researcher had eye protection available but failed to wear it. Eye protection is required, and its use is enforced by the supervisor when engineering controls may fail to eliminate potential exposure in excess of the applicable MPE. Laser radiation is generated both by systems producing discrete wavelengths and by tunable laser systems producing a variety of wavelengths. For © 2005 by CRC Press LLC

this reason it is impractical to select a single eye protection filter that will provide su!cient protection from all hazardous laser radiation. Therefore, it is important to choose an eye protection specific for the wavelength and power of a particular laser. Laser Protective Eyewear Requirements 1. Laser protective eyewear is to be available and worn by all personnel within the nominal hazard zone (NHZ) of Class 3B and Class 4 lasers where the exposures above the maximum permissible exposure (MPE) can occur. 2. The attenuation factor optical density of the laser protective eyewear at each laser wavelength shall be specified by a laser safety o!cer. 3. All laser protective eyewear shall be clearly labelled with the optical density and the wavelength for which protection is aorded. 4. Laser protective eyewear shall be inspected for damage prior to use. The use of beam attenuators to align visible lasers will reduce laser beam intensities to a level that will allow the operator to align the beam without personal protective equipment. Laser alignment cards for ultraviolet and infrared radiation allow operators to locate the beam during alignment procedures.

7.3

Powder Hazards

It is important to consult material safety data sheets (supplied with materials), and applicable national and health regulations before using any powder or spray materials. Unusual sensitivity to powder materials is seen in some individuals that may result in health hazards. In general, a suitable exhaust system is essential in laser material processing to avoid the toxic eects of powder and fumes that are generated by the process. In addition to the above-mentioned hazards, some materials are inherently dangerous to our health. For example, beryllium and tellurium are very harmful to the respiratory system, and the fumes of cadmium and chromium alloys are extremely hazardous. Similarly, the fumes of nickel components are potentiality hazardous. Before using any material, proper measures need to be considered by consulting an industrial hygienist.

© 2005 by CRC Press LLC

References

[1] U. Ritter, W. Kahrmann, R. Kupfer, and R. Glardon, “Laser coating proven in practice,” Sulzer Technical Review, vol. 73, no. 3, pp. 18—20, 1991. [2] L. Jianglong, D. Peidao, and S. Gongqi, “Scanning electron microscopy study of laser coating microstructures,” Materials Characterization, vol. 33, no. 4, pp. 387—391, 1994. [3] S. Corbin, E. Toyserkani, and A. Khajepour, “Cladding of Fe-Al coating on mild steel using laser assisted powder deposition,” Journal of Materials Science and Engineering A, vol. 354, no. 1-2, pp. 48—57, 2003. [4] E. Toyserkani, S. Corbin, and A. Khajepour, “Iron aluminide coating of mild steel using laser assisted powder deposition,” in Proceeding of International Symposium on Processing and Fabrication of Advanced Materials Processing XI, T.S.Srivatsan and R.A.Varin, Eds., pp. 244— 257. ASM International, Oct. 8-10 2002. [5] K.F. Tam, F.T. Cheng, and H.C. Man, “Laser surfacing of brass with Ni-Cr-Al-Mo-Fe using various laser processing parameters,” Materials Science and Engineering A, vol. 325, no. 1-2, pp. 365—374, 2003. [6] W.M. Steen and C.G.H. Courtney, “Surface heat treatment of En8 steel using a 2kW continuous-wave CO2 laser,” Metals Technology, vol. 6, no. 12, pp. 456—462, 1979. [7] C.G.H. Courtney and W.M. Steen, “Surface coating using a 2kW CO2 laser,” in Proceedings of the Technical Program, National Noise and Vibration Control Conference, pp. 228—233. NOISEXPO, 1979. [8] S. Das, “Physical aspects of process control in selective laser sintering of metals,” Advanced Engineering Materials, vol. 5, no. 10, pp. 701—711, 2003. [9] F. Abe, K. Osakada, M. Shiomi, K. Uematsu, and M. Matsumoto, “The manufacturing of hard tools from metallic powders by selective laser melting,” Journal of Materials Processing Technology, vol. 111, pp. 210—213, 2001. [10] Y. Tang, H.T. Loh, Y.S. Wong, J.Y.H. Fuh, L. Lu, and X. Wang, “Direct laser sintering of a copper-based alloy for creating three-dimensional © 2005 by CRC Press LLC

metal parts,” Journal of Materials Processing Technology, vol. 140, no. 1-3, pp. 368—372, 2003. [11] M.L. Gri!th, D.L. Keicher, J.T. Romero, J.E. Smugeresky, C.L. Atwood, L.D. Harwell, and D.L. Greene, “Laser engineered net shaping for the fabrication of metallic components,” American Society of Mechanical Engineers, Materials Division, Advanced Materials: Development, Characterization Processing, and Mechanical Behavior, vol. 74, pp. 175—176, 1996. [12] W. Hofmeister, M. Wert, J. Smugeresky, J.A. Philliber, M. Gri!th, and M. Ensz, “Investing solidification with laser-engineered net shaping process,” JOM, vol. 51, no. 7, 1999. [13] M.L. Gri!t, M.T. Ensz, J.D. Puskar, C.V. Robino, J.A. Brooks, J.A. Philliber, J.E. Smugeresky, and W.H. Hofmeister, “Understanding the microstructure and properties of components fabricated by laser engiR ),” in Materials Research Society, Symponeered net shaping (LENS° sium Y Proceedings, vol. 625, pp. 117—124. 2000. [14] J. Mazumder, D. Dutta, N. Kikuchi, and A. Ghosh, “Closed-loop direct metal deposition: Art to part,” Optics and Lasers in Engineering, vol. 34, pp. 397—414, 2000. [15] J. Mazumder, J. Choi, K. Nagarathnam, J. Koch, and D. Hetzner, “Direct metal deposition of H13 tool steel for 3-D components,” JOM, vol. 49, no. 5, pp. 55—60, 1997. [16] Y. Hua and J. Choi, “Feedback control eects on dimensions and defects of H13 tool steel by DMD process,” in Proceeding of ICALEO’2003, pp. LMP Section C, 103—112. LIA, 2003. [17] J. Choi and Y. Hua, “Adaptive aided DMD process control,” in Proceeding of ICALEO’2001, pp. 720—729. LIA, 2001. [18] M.A. McLean, G.J. Shannon, and W.M. Steen, “Laser direct casting high nickel alloy components,” Advances in Powder Metallurgy and Particulate Materials, vol. 3, pp. 21—3—21—16, 1997. [19] L. Xue and M. Islam, “Free-form laser consolidation for production functional metallic components,” in Proceeding of ICALEO’1998, pp. 15—24. 1998. [20] M.U. Islam, L. Xue, and G. McGregor, “Process for manufacturing or repairing turbine engine or compressor components,” U.S. Patent Number 6269540, August 7 2001. [21] L. Xue, A. Therialuit, and M.U. Islam, “Laser consolidation for the manufacturing of complex flextensional transducer shells,” in Proceeding of ICALEO’2001, pp. 701—711. Laser Institute of America, 2001. © 2005 by CRC Press LLC

[22] L. Xue, A. Theriault, B. Rubinger, D. Parry, F. Ranjbaran, and M. Doyon, “Investigation of laser consolidation process for manufacturing structural components for advanced robotic mechatronics system (ARMS),” in Proceeding of ICALEO’2003, LMP Section, pp. 134 of 143. LIA, 2003. [23] D. Kaser, “Laser powder fusion welding,” in Human Corporation Technical Report. Human Corporation, 2001. [24] J. Laeng, J.G. Stewart, and F.W. Liou, “Laser metal forming processes for rapid prototyping — A review,” International Journal of Production Research, vol. 38, no. 16, pp. 3973—3996, 2000. [25] M. Gremaud, J.D. Wagniere, A. Zryd, and W. Kurz, “Laser metal forming: Process fundamentals,” Surface Engineering, vol. 12, no. 3, pp. 251—259, 1996. [26] R. Mah, “Directed light fabrication,” Processes, vol. 151, no. 3, pp. 31—33, 1997.

Advanced Materials and

[27] J.O. Milewski, G.K. Lewis, D.J. Thoma, G.I. Keel, R.B. Nemec, and R.A. Reinert, “Directed light fabrication of a solid metal hemisphere using 5-axis powder deposition,” Journal of Materials Processing Technology, vol. 75, no. 1-3, pp. 165—172, 1998. [28] J.O. Milewski, P.G. Dickerson, R.B. Nemec, G.K. Lewis, and J.C. Fonseca, “Application of a manufacturing model for the optimization of additive processing of inconel alloy 690,” Journal of Materials Processing Technology, vol. 91, no. 1, pp. 18—28, 1999. [29] A.J. Pinkerton and L. Li, “Eects of powder geometry and composition in coaxial laser deposition of 316L steel for rapid prototyping for rapid prototyping,” CIRP Annals — Manufacturing Technology, vol. 52, no. 1, pp. 181—184, 2003. [30] X.Y. Xu and M.L. Zhong H.Q. Sun, “Synthesis and fabrication of WC particulate reinforced Ni3 Al intermetallic matrix composite coating by laser powder deposition,” Journal of Materials Science Letters, vol. 22, no. 19, pp. 1369—1372, 2003. [31] E. Toyserkani, A. Khajepour, and S. Corbin, “Vision-based feedback control of laser powder deposition,” in M2VIP03 Mechatronics and Machine Vision 2003: Future Trends, J. Billingsley, Ed., in press. A Research Studies Press Book, 2003. [32] Y.C. Chang, J.H. Kao, J. Dong, K. Ramaswami, and F.B. Prinz, “Automated layer decomposition for additive subtractive solid freeform fabrication,” in Proceeding of Solid Freeform Fabrication Symposium, University of Texas at Austin, Ed., pp. 111—120. 1999. © 2005 by CRC Press LLC

[33] Y. Li, H. Yang, X. Lin, W. Huang, J. Li, and Y. Zhou, “The influences of processing parameters on forming characterizations during laser rapid forming,” Materials Science and Engineering A, vol. 360, no. 1-2, pp. 18—25, 2003. [34] F.G. Arcella, D.H. Abbott,and M.A. House, “Rapid laser forming of titanium structures,” in Proceeding of Powder Metallurgy World Conference and Exposition. Grenada, Spain, 1998. [35] R. Vilar, “Laser cladding,” Journal of Laser Applications, vol. 11, no. 2, pp. 64—79, April 1999. [36] M. Gaumann, H. Rusterholz, R. Baumann D.J. Wagniere, and W. Kurz, “Single crystal turbine components repaired by epitaxial laser metal forming,” Materials for Advanced Powder Engineering, vol. 1479, pp. 1—6, 1998. [37] Y.T. Pei and J.T.M.D. Hosson, “Producing functionally graded coatings by laser powder cladding,” JOM, vol. 52, no. 1, pp. 641—647, http://www.tms.org/pubs/journals/JOM/0001/Pei/Pei0001.html 2000. [38] M.T. Ensz and M.L. Gri!th, “Critical issues for functionally graded R ),” in material deposition by laser engineered net shaping (LENS° Proceedings of the 2002 MPIF Laser Metal Deposition Conference. 2002. [39] X.C. Li, A. Golnas, and F. Prinz, “Shape deposition manufacturing of smart metallic structures with embedded sensors,” in Proceedings of SPIE, v 3986, pp. 160—171. The International Society for Optical Engineering, 2002. [40] T.H. Maiman, “Stimulated optical radiation in ruby,” in Nature. 1960. [41] D.S. Gnanamuthu, “Laser surface treatment,” Optical Engineering, vol. 19, no. 5, pp. 783—792, 1980. [42] W.M. Steen and C.G.H. Courtney, “Hardfacing of nimonic 75 using 2kW continuous-wave CO2 laser,” Metals Technology, vol. 7, no. 6, pp. 232—237, 1980. [43] V.M. Weerasinghe and W.M. Steen, “Laser cladding by powder injection,” IFS (Publ) Ltd, pp. 125—132, 1983. [44] V.M. Weerasinghe and W.M. Steen, “Laser cladding by powder injection,” in Transport Phenomena in Materials Processing, J. Mazumder M.M. Chen and C. Tucker, Eds., pp. 15—23. ASME, 1983. [45] V.M. Weerasinghe and W.M. Steen, Laser Processing of Materials(Laser Cladding with Pneumatic Powder Delivery), 166-174, 1983. © 2005 by CRC Press LLC

[46] V.M. Weerasinghe and W.M. Steen, “Computer simulation model for laser cladding,” American Society of Mechanical Engineers, Production Engineering Division (Publication) PED, vol. 10, pp. 15—23, 1983. [47] L.J. Li and J. Mazumder, “Study of the mechanism of laser cladding processes,” Metallurgical Society of AIME, pp. 35—50, 1985. [48] J. Mazumder, “State-of-the-art laser materials processing,” ASME, vol. 2, no. 115, pp. 599—630, 1987. [49] A. Kar and J. Mazumder, “One-dimensional finite-medium diusion model for extended solid solution in laser cladding of Hf on nickel,” Acta Material, vol. 36, no. 3, pp. 701—712, 1988. [50] J. Mazumder and J. Singh, “Laser surface alloying and cladding for corrosion and wear,” NATO ASI Series, Series E: Applied Sciences, no. 115, pp. 297—307, 1986. [51] J. Mazumder and A. Kar, “Solid solubility in laser cladding,” Journal of Metals, vol. 39, no. 2, pp. 18—23, 1987. [52] J. Mazumder and J. Singh, “Laser surface alloying and cladding for corrosion and wear,” High Temperature Materials and Processes, vol. 7, no. 2-3, pp. 101—106, 1986. [53] A. Kar and J. Mazumder, “Eect of cooling rate on solid solubility in laser cladding,” ASME, vol. 3, pp. 237—249, 1987. [54] J. Singh and J. Mazumder, “In-situ formation of Ni-Cr-Al-R. E. alloy by laser cladding with mixed powder feed,” Springer-Verlag, pp. 169—179, 1986. [55] M. Eboo and A.E. Lindemanis, “Advances in laser cladding technology,” in Proceeding ICAIEO’83, pp. 31—39. SPIE 527, 1983. [56] P. Koshy, “Laser cladding techniques for application to wear and corrosion resistant coatings,” Proceedings of SPIE — The International Society for Optical Engineering, vol. 527, pp. 80—85, 1985. [57] T.R. Tucker, A.H. Clauer, I.G. Wright, and T. Stropki, “Laser processed composite metal cladding for slurry erosion resistance,” Thin Solid Films, vol. 118, no. 1, pp. 73—84, 1984. [58] A.G. Blake, A.A. Mangaly, M.A. Everett, and A.H. Hammeke, “Laser coating technolgy: A commercial reality,” in Laser Beam Surface Treating and Coating, G. Sepold, Ed., pp. 56—65. SPIE 957, 1988. [59] S.B. Doran, “The laser where flexibility ultimately means economy,” in Production Magazine. 1987. [60] K. Tanaka, T. Saito, Y. Shimura, K. Mori, M. Kawasaki, M. Koyama, and H. Murase, “New copper based composite for engine valve seat di© 2005 by CRC Press LLC

rectly deposited onto aluminium alloy by laser cladding process,” Japan Inst. Metals, vol. 57, no. 10, pp. 1114—1122, 1993. [61] R.M. Macintyre, “Laser hardsurfacing of gas turbine blade shroud interlocks,” in LIM-4. 1983. [62] L. Shepeleva, B. Medres, W.D. Kaplan, M. Bamberger, and A. Weisheit, “Laser cladding of turbine blades,” Surface and Coatings Technology, vol. 125, no. 1, pp. 45—48, 2000. [63] C. Theiler, H. Kohn, and B. Gruger, “Laser beam cladding of turbine bolts,” VGB PowerTech, vol. 83, no. 3, pp. 75—78 + 6—7, 2003. [64] J.L. Koch and J. Mazumder, “Rapid prototyping by laser cladding,” in Laser Material Processing, ICALEO’4993, T.D. Mccay, A. Matsunawa, and H. Hugel, Eds., pp. 556—557. SPIE 2306, 1993. [65] M. Murphy, C. Lee, and W.M. Steen, “Studies in rapid prototyping by laser surface cladding,” in Proceeding of ICALEO’4993, pp. 882—891. 1993. [66] M.L. Murphy, W.M. Steen, and C. Lee, “A novel prototyping technique for the manufacture of metallic components,” in Proceeding of ICALEO’4994, pp. 31—40. Laser Institute of America, 1994. [67] E. Toyserkani, A. Khajpour, and S.F. Corbin, “Recurrent neural network based analysis for laser cladding dynamic model prediction,” in Proceeding of the International Congress on Application of Lasers and Electro-optics, ICALEO’02, p. CD reference Code 1210. LIA, 2002. [68] E. Toyserkani, A. Khajepour, and S. Corbin, “System and method for closed-loop control of laser cladding by powder injection,” Provisional U.S. Patent Number 60/422606 , Oct. 2002. [69] X. Wu, “In situ formation by laser cladding of a tic composite coating with a gradient distribution,” Surface and Coatings Technology, vol. 115, no. 2, pp. 111—115, 1999. [70] S. Yang, W. Liu, and M. Zhong, “Microstructure characteristics and properties of in-situ formed TiC/Ni based alloy composite coating by laser cladding,” in Proceedings of SPIE, vol. 4831, pp. 481—486. The International Society for Optical Engineering, 2002. [71] L.R. Katipelli, A. Agarwal, and N.B. Dahotre, “Laser surface engineered TiC coating on 6061 Al alloy: microstructure and wear,” Applied Surface Science, vol. 153, no. 2, pp. 65—78, 2000. [72] C. Hu, H. Xin, L.M. Watson, and T.N. Baker, “Analysis of the phases developed by laser nitriding Ti-6Al-4V alloys,” Acta Metallurgical, vol. 45, no. 10, pp. 4311—4322, 1997. © 2005 by CRC Press LLC

[73] J. Feng, M.G.S. Ferreira, and R. Vilar, “Laser cladding of Ni-Cr/Al2 O3 composite coatings on AISI 304 stainless steel,” Surface and Coatings Technology, vol. 88, no. 1-3, pp. 212—218, 1997. [74] J.D. Damborenea and A.J. Vazquez, “Laser cladding of high temperature coating,” Journal of Materials Science, vol. 28, no. 4, pp. 4775— 4780, 1993. [75] L.C. Lim, Q. Ming, and Z.D. Chen, “Microstructures of laser-clad nickel-based hardfacing alloys,” Surface and Coatings Technology, vol. 106, no. 2-3, pp. 183—192, 2000. [76] Y. Li and W.M. Steen, “Laser cladding of Stellite and silica on aluminum substrate,” in Proceeding of ICALOE’92, M.zhou, Ed., pp. 602—608. SPIE 1979, 1992. [77] S. Sircar, C. Ribaudo, and J. Mazumder, “Laser clad nickel based superalloy: Microstructure evolution and high temperature oxidation studies,” in Laser Beam Surface Treating and Coating, pp. 29—41. SPIE 957, 1988. [78] A. Hidouci, J.M. Pelletier, F. Ducoin, D. Dezert, and R.E. Guerjouma, “Microstructural and mechanical characteristics of laser coatings,” Surface and Coatings Technology, vol. 123, no. 1, pp. 17—23, 2000. [79] D. Peidao, L. Jianglong, S. Gongqi, Z. Shouze, and C. Pengjun, “Laser surface alloying of a low alloy steel with cobalt,” Journal of Materials Processing Technology, vol. 58, no. 1, pp. 131—135, 1996. [80] A. Frenk and W. Kruz, “High speed laser cladding: Solidification conditions and microstructure of a cobalt-based alloy,” Materials Science and Engineering A, vol. 173, no. 1-2, pp. 339—342, 1993. [81] J. Przybylowicz and J. Kusinski, “Laser cladding and erosive wear of Co-Mo-Cr-Si coatings,” Surface and Coatings Technology, vol. 125, no. 1, pp. 13—18, 2000. [82] “Technical reports,” in Internet: http://www.pom.net. [83] F. Lusquinos, J. Pou, J.L. Arias, M. Larosi, A. De Carlos, B.L.P. Amor, S. Best, and F.C.M. Driessens, “Laser surface cladding: A new promising technique to produce calcium phosphate coatings,” Key Engineering Materials, vol. 218-220, pp. 187—190, 2002. [84] F. Audebert, R. Colac, R. Vilar, and H. Sirkin, “Production of glassy metallic layers by laser surface treatment,” Scripta Materialia, vol. 48, pp. 281—286, 2003. [85] St. Nowotny, S. Scharek, F. Kempe, and E. Beyer, “COAXn: Modular system of powder nozzles for laser beam build-up welding,” in Proceeding of ICALEO’2003, pp. LMP Section pp. 190—193. LIA, 2003. © 2005 by CRC Press LLC

[86] P. Bergan, “Implementation of laser repair processes for navy aluminum components,” in Proceeding of Diminishing Manufacturing Sources and Material Shortages Conference (DMSMS), Published on the Web. The Department of Defense, 2000. [87] Y.P. Kathuria, “Some aspects of laser surface cladding in the turbine industry,” Surface and Coatings Technology, vol. 132, no. 2-3, pp. 262— 269, 2000. [88] S. Krause, “An advanced repair technique: laser powder build-up welding,” in Sulzer technical review 4. Sulzer Turbomachinary services, 2001. [89] Y. Tanaka, S. Nagai, M. Ushida, and T. Usui, “Large engine maintenance technique to support flight operation for commercial airlines,” in Technical Review, Mitsubishi Heavy Industries Ltd., no. 2, vol. 40, 2003. [90] A. Rosochowski and A. Matuszak, “Rapid tooling: the state of the art,” Journal of Materials Processing Technology, vol. 106, pp. 191—198, 2000. [91] Stratasys Inc., “Redchip Scottsdale investor conference 2003 webcast,” Internet http://www.firstcallevents.com, 2003. [92] B. Wiedemann and H.A. Jantzen, “Strategies and applications for rapid product and process development in Daimler-Benz AG,” Journal of Computers in Industry, vol. 39, pp. 11—25, 1999. [93] J. Koch and J. Mazumder, “Apparatus and methods for monitoring and controling multi-layer laser cladding,” U.S. Patent Number 6122564, 2000. [94] C. Jimin, C. Tao, L. Shibin, and Z. Tiechuan, “The investigation on laser microfabrication with metallic powder,” in Proceeding of ICALEO’2003, pp. LMF Section E, 61—71. LIA, 2003. [95] X. Li, H. Choi, and Y. Yang, “Micro rapid prototyping system for micro components,” Thin Solid Films, vol. 420-421, pp. 515—523, 2002. [96] F.C.J. Fellows and W.M. Steen, Laser Surface Treatment, Chapman and Hall, Chapter of Advanced surface coatings and p. 244-277, 1991. [97] S. Tondu, T. Schnick, L. Pawlowski, B. Wielage, S. Steinhauser, and K. Sabatier, “Laser glazing of FeCr-TiC composite coatings,” Surface and Coatings Technology, vol. 123, no. 2, pp. 247—251, 2000. [98] Argonne National Laboratory, “Working on railroad,” Tech. Transfer Highlights, vol. 9, no. 2, 1998. [99] W.M Steen, Laser Surface Treatment, NATO ASI series, Vol. 307, pp. 1-19, 1984. © 2005 by CRC Press LLC

[100] X.B. Zhou and T.T.M. Hosson, “Spinel/metal interfaces in laser coated steels: A transmission electron microscopy study,” Acta Metal, vol. 39, no. 10, pp. 2267—2273, 1991. [101] W.M. Steen, Laser Material Processing, Springer, 1998. [102] J. Powell, P.S. Henry, and Steen W.M., “Laser cladding with preplaced powder: Analysis of thermal cycling and dilution,” Surface Engineering, vol. 4, no. 2, pp. 141—149, 1988. [103] M. Gremaud, J.D. Wagniere, A. Zryd, and W. Kruz, “Laser cladding with preplaced powder: Analysis of thermal cycling and dilution,” Surface Engineering, vol. 12, no. 3, pp. 251—259, 1996. [104] J.D. Kim and Y. Peng, “Melt pool shape and dilution of laser cladding with wire feeding,” Journal Materials Processing Technology, vol. 104, pp. 284—293, 2000. [105] J.D. Kim and Y. Peng, “Plunging method for Nd: YAG laser cladding with wire feeding,” Optics and Lasers in Engineering, vol. 33, pp. 299— 309, 2000. [106] E. Lugscheider, H. Bolender, and H. Krappitz, “Laser cladding of paste bound hardfacing alloys,” Surface Engineering, vol. 7, no. 4, pp. 341— 344, 1991. [107] G.J. Bruck, “Fundamentals and industrial applications of high power laser beam cladding,” in Laser Beam Surface Treating and Coating, G. Sepold, Ed., pp. 14—28. SPIE 957, 1988. [108] P.A. Vetter, T. Engel, and J. Fontanine, “On-line surface treatment feasibility of industrial components by means of power laser,” in Laser Materials Processing: Industrial and Microelectronics Applications, E. Beyer, M. Cantello, A.V.L Rocca, L.D. Laude, F.O. Olsen, and G. Sepold, Eds. SPIE 2207, 1994. [109] Y. Yang, “Microstructure and properties of laser-clad high-temperature wear-resistant alloys,” Applied Surface Science, vol. 140, no. 1, pp. 19— 23, 1999. [110] Y.P. Kathuria, “Laser-cladding process: a study using stationary and scanning CO2 laser beams,” Surface Coatings Technology, vol. 97, no. 1-3, pp. 442—447, 1997. [111] W.M. Steen, “Laser surface cladding,” in Proceedings of Indo-US Workshop on Principles of Solidification and Materials Processing, AIBS ONR, Ed., pp. 168—178. SOLPROS, Hyderabad, India, 1988. [112] X. Wu, B. Zhu, X. Zeng, X. Hu, and K. Cui, “Critical state of laser cladding with powder auto-feeding,” Surface Coatings Technology, vol. 79, no. 1, pp. 200—204, 1996. © 2005 by CRC Press LLC

[113] E. Toyserkani, “Modeling and control of laser cladding,” in Ph.D Thesis. University of Waterloo, Waterloo, On, Canada, 2003. [114] EPA, “Waste minimization for metal finishing facilities,” Washington, DC: O!ce of Solid Waste, 1995. [115] A. Einstein, “Quantum theory of radiation and atomic processes,” Physikalische Zeitschrift, vol. 18, pp. 121—128, 1917. [116] O. Svelto, Principles of Lasers, Plenum Press, Fourth edition, New York, 1998. [117] Y. Akiyama, H. Takada, M. Sasaki, H. Yuasa, and N. Nishida, “Eecient 10 kW diode-pumped Nd:YAG rod laser,” in Proceeding of First International Laser Macroprocessing, Vol. 4834, I. Miyamoto, Ed., pp. 96—100. SPIE, 2002. [118] K. Ueda, “Optical fiber laser device,” U.S. Patent Number 6178187 B1, June 2001. [119] “Fiber lasers address new markets,” in Fibers.org. Internet http://fibers.org/ articles/ news/ 5/ 6/ 15/ 1. [120] V.S. Aizenshtadt, Tables of Laguerre Polynomials and Fuctions, Oxford, Pergamon Press, chapter 3, 1966. [121] “Laser beam characterestics,” in Internet: http://www.mellesgriot.com. Melles Griot. [122] P. Drude, Theory of Optics, Longmans, 1922. [123] C.N. Panagopoulos and A. Markaki, “Excimer laser treatment of copper-coated mild steel,” Journal of Materials Science, vol. 32, pp. 1425—1430, 1997. [124] “Technical report,” Marco.

in Internet http://www.rofin.com. Rofin Laser

[125] S. Nowotny, A. Richter, and E. Beyer, “Laser cladding using high-power diode lasers,” in Laser Material Processing, ICALEO’4998, pp. 68—74 Section G. LIA, 1998. [126] C.M. Cook, J.M. Haake, M.S. Zediker, and J.M.T. Banaskavich, “Diode laser cladding produces high quality coatings,” International SAMPE Technical Conference, vol. 32, pp. 910—921, 2000. [127] E.W. Kreutz, G. Backes, A. Gasser, and K. Wissenbach, “Rapid prototyping with CO2 laser adiation,” Applied Surface Science, vol. 86, no. 1, pp. 310—316, 1995. [128] H. Jack, “Laser cladding investigation,” in Internet: http://claymore.engineer.gvsu.edu/ remeltsm/laser4.htm. 1997. © 2005 by CRC Press LLC

[129] E. Toyserkani, A.Khajpour, and S.F. Corbin, “Application of Hammerstein-Wiener nonlinear model to laser cladding dynamic model prediction,” in Proceeding of the International Conference on Modeling and Simulation, pp. 94—98. IASTED, 2001. [130] H. Fujimagari, M. Hagiwara, and T. Kojima, “Laser cladding technology to small diameter pipes,” Nuclear and Engineering and Design, vol. 195, no. 3, pp. 289—298, 2000. [131] R. Volz and K. Battling, “Cladding and alloying with the laser by single-step processing,” in Laser Materials Processing ICALEO’4993, pp. 999—1003. SPIE 2306, 1993. [132] J.W. Sears, “Direct laser powder deposition ’state of the art’,” Minerals, Metals and Materials Society/AIME, Powder Materials: Current Research and Industrial Practices, pp. 213—229, 1999. [133] R.C. Gassmann and M.F. Modest, “Laser cladding of partially yttria stabilized zirconia on titanium, nickel and steel alloys,” in Laser Materials Processing ICALEO’4994, pp. 125—134. SPIE 2500, 1994. [134] E. Toyserkani, A. Khajepour, and S. Corbin, “Application of experimental-based modeling to laser cladding,” Journal of Laser Applications, vol. 14, no. 3, pp. 165—173, 2002. [135] M. Schneider, “Laser cladding with powder,” Ph.D Thesis. University of Twente, Enschede, The Netherlands, 1998. [136] A. Frenk, M. Vandyoussefi, J.D. Wagniere, A. Zryd, and W. Kurz, “Analysis of the laser-cladding process for Stellite on steel,” Metallurgical and Materials Transactions: B, vol. 28B, no. 3, pp. 501—508, 1997. [137] S. Kaierle, S. Mann, J. Ortmann, E.W. Kreutz, and R. Poprawe, “Intelligent beam monitoring and diagnosis for CO2 lasers,” in Proceedings of SPIE — The International Society for Optical Engineering, vol. 4270, pp. 80—87. 2001. [138] F. Lemoine, J.M. Jouvard, H. Andrzejewski, and D.F. Grevey, “Influence of the nature of optical fibers in materials treatment by Nd:YAG lasers,” Journal of Materials Processing Technology, vol. 59, pp. 337— 342, 1996. [139] P.F. Jeantette, D.M Keicher, J.A. Romero, and L.P Schanwald, “Mutiple and system for producing complex-shape objects,” U.S. Patent Number 6046426, April 2000. [140] L. Li, W.M. Steen, and V.M. Weerasinghe, “Sensing, modeling and closed-loop of powder feeder for laser surface modification,” in Proceedings of ICALEO’4993, pp. 965—974. LIA, 1993. © 2005 by CRC Press LLC

[141] R.G. Landers, M. Hilgers, F.W. Liou, and B. McMillin, “Object oriented modeling and fault detection of a powder feeder for a laser metal deposition system,” in 43th Annual Solid Freeform Fabrication Symposium. Austin, Texas, 2002. [142] P.A. Caravalho, N. Braz, M.M. Pontinha, M.G.S. Ferreira, W.M. Steen, R. Vilar, and K.G. Watkins, “Automated workstation for variable composition laser cladding-its use for rapid alloy scanning,” Surface Engineering, vol. 72, no. 1, pp. 62—70, 1995. [143] P. Kumar, K.J. Santosa, E. Beck, and S. Das, “Direct-write deposition of fine powders through miniature hopper-nozzles for multi-material solid freeform fabrication,” Rapid Prototyping Journal, in press, 2003. [144] L. Li and W.M. Steen, “Sensing, modeling and closed loop control of powder feeder for laser surface modification,” in Proceeding of ICALEO’4993, pp. 965—974. Laser Institute of America, 1993. [145] J. Lin, “Concentration mode of the powder stream in coaxial laser cladding,” Optics and Laser Technology, vol. 31, no. 3, pp. 251—257, 1999. [146] S. Carty, I. Owen, W.M. Steen, B. Bastow, and J.T. Spencer, “Catchment e!ciency for nozzle designs used in laser cladding and alloying,” in Laser Processing: Surface Treatment and Film Deposition, J. Mazumder, O. Conde, R. Vilar, and W. Steen, Eds., pp. 395—410. NATO 307, 1996. [147] D.M. Keicher and W.D. Miller, “Multiple beams and nozzles to increase deposition rate,” U.S. Patent Number 6268584, July 31 2001. [148] J. Lin, “A simple model of powder catchment in coaxial laser cladding,” Optics and Laser Technology, vol. 31, no. 1, pp. 233—238, 1999. [149] A.D. Zimon, Adhesion of dust and powder, Plenum Publishing, 1982. [150] H. Bruyninckx, J. Rutgeerts, and W. Meeussen, “Motion API,” in http://www.orocos.org/ documentation/ motion-api.html. 2003. [151] J.M. Kirkpatrick, The AutoCAD Book: Drawing, Modeling, and Applications, Including Release 43, Prentice Hall, 1996. [152] W.H. Wah, in http://rpdrc.ic.polyu.edu.hk. 1999. [153] K. Tata, G. Fadel, A. Bagchi, and N. Aziz, “E!cient slicing for layered manufacturing,” Rapid Prototyping Journal, vol. 4, no. 4, pp. 151—167, 1998. [154] K. Ramaswami, “Process planning for shape deposition manufacturing,” Ph.D Thesis. Stanford University, 1997. © 2005 by CRC Press LLC

[155] K. Tata, A. Bagchi, and N.M. Aziz, “Apparatus and method for layered modeling of intended objects represented in STL format and adaptive slicing thereof,” U.S. Patent Number 5596504, 1997. [156] H.S. Carslaw and J.C. Jaeger, Conduction of Heat in Solids, Oxford University Press, second edition, 1959. [157] A.F.A. Hoadley and M. Rappaz, “A thermal model of laser cladding by powder injection,” Metallurgical Transactions: B, vol. 23, no. 1, pp. 631—642, 1992. [158] C. Chan, J. Mazumder, and M. M.Chen, “A two-dimensional transient model for convection in laser melted pool,” Metallurgical Transaction: A, vol. 15, no. 12, pp. 2226—2232, 1984. [159] T. Chande and J. Mazumder, “Two-dimensional, transient model for mass transport in laser surface alloying,” Journal of Applied Physics, vol. 57, no. 6, pp. 2226—2232, 1985. [160] A. Kar and J. Mazumder, “One-dimensional diusion model for extended solid solution in laser cladding,” Journal of Applied Physics, vol. 61, no. 7, pp. 2645—2655, 1987. [161] F. Lemoine, D.F. Grevey, and A.B. Vannes, “Cross-section modelling laser cladding,” in Laser Material Processing, Proceeding of ICALEO’4993, P. Denney, I. Miyamoto, and B.L. Mordike, Eds., pp. 203— 212. SPIE 2306, 1993. [162] M. Picasso, C.F. Marsdan, J.D. Wangniere, A. Frenk, and M. Rappaz, “A simple but realistic model for laser cladding,” Metallurgical and Materials Transactions: B, vol. 25, no. 2, pp. 281—291, 1994. [163] M. Picasso and A.F.A. Hoadley, “Finite element simulation of laser surface treatments including convection in the melt pool,” International Journal of Numerical Methods in Heat Fluid Flow, vol. 4, pp. 61—83, 1994. [164] J.M. Jouvard, D.F. Grevey, F. Lemoine, and A.B. Vannes, “Continuous wave Nd:YAG laser cladding modeling: A physical study of track creation during low power processing,” Journal of Laser Applications, vol. 9, no. 1, pp. 43—50, 1997. [165] R. Colaco, L. Costa, R. Guerra, and R. Vilar, “A simple correlation between the geometry of laser cladding tracks and the process parameters,” in Laser Processing: Surface Treatment and Film Deposition, J. Mazumder, O. Conde, R. Vilar, and W. Steen, Eds., pp. 421—429. NATO 307, 1996. [166] G.W.R.B.E. Romer and J. Meijer, “Analytical model describing the relationship between laser power, beam velocity and melt pool depth in © 2005 by CRC Press LLC

the case of laser remelting, alloying and dispersing,” in Lasers in Material Processing, L.H.J.F. Beckmann, Ed., pp. 507—516. Europt 3097, 1997. [167] G.R.B.E. Romer, Modeling and Control of Laser Surface Treatment, Print Partners Ipskamp, The Netherlands, 1999. [168] A.F.H. Kaplan, B. Weinberger, and D. Schuocker, “Theoretical analysis of laser cladding and alloying,” in Proceeding of ICALEO’4997, Vol. 3097 SPIE, pp. 499—506, 1997. [169] J. Lin and W.M. Steen, “An in-process method for the inverse estimation of the powder catchment e!ciency during laser cladding,” Optics and Laser Technology, vol. 30, no. 2, pp. 77—84, 1998. [170] M. Bamberger, W.D. Kaplan, B. Medres, and L. Shrprlrva, “Calculation of process parameters for laser alloying and cladding,” Journal of Laser Applications, vol. 10, no. 1, pp. 29—33, 1998. [171] G.R.B.E. Romer, R.G.K.M. Aarts, and J. Meijer, “Dynamics models of laser surface alloying,” Lasers in Engineering, vol. 8, no. 4, pp. 251—266, 1998. [172] E. Toyserkani, A. Khajepour, and S.F. Corbin, “3-D finite element modeling of laser cladding by powder deposition: Eects of laser pulse shaping on the process,” Journal of Optics and Laser in Engineering, in press, 2003. [173] E. Toyserkani, A. Khajepour, and S.F. Corbin, “3-D finite element modeling of laser cladding by powder deposition: Eects of powder feedrate and travel speed on the process,” Journal of Laser Applications, vol. 15, no. 3, pp. 153—161, 2003. [174] G. Zhao, C. Cho, and J.D. Kim, “Application of 3-d finite element method using Lagrangian formulation to dilution control in laser cladding process,” International Journal of Mechanical Sciences, vol. 45, pp. 777—796, 2003. [175] M. Labudovic, D. Hu, and R. Kovacevic, “A three-dimensional model for direct laser metal powder deposition and rapid prototyping,” Journal of Materials Science, vol. 38, pp. 35—49, 2003. [176] A. Kar and J. Mazumder, Laser Processing: Surface Treatment and Film Deposition, NATO, Vol. 307, pp. 229-296, 1996. [177] J. Lin and B.C. Hwang, “Coaxial laser cladding on an inclined substrate,” Optics and Laser Technology, vol. 31, no. 4, pp. 571—578, 1999. [178] F. Bataille, J.M Cerez, and D. Kechemair, “A systematic method for the design of multivariable controller actuating power and speed during © 2005 by CRC Press LLC

a CO2 laser surface treatment,” Journal of Laser Applications, vol. 4, no. 1, pp. 43—47, 1992. [179] J. Goldak, “Computer modeling of heat flow in welds,” Metallurgical Transaction B, vol. 17, pp. 587—600, 1986. [180] L.X. Yang, X.F. Peng, and B.X. Wang, “Numerical modeling and experimental investigation on the characteristics of molten pool during laser processing,” International Journal of Heat and Mass transfer, vol. 44, pp. 4465—4473, 1986. [181] M.R. Frewin and D.A. Scott, “Finite element model of pulsed laser welding,” Welding Research Supplement, pp. 15—22, 1999. [182] S. Brown and H. Song, “Finite element simulation of welding of large structures,” Journal of Engineering for Industry, Transactions of the ASME, vol. 144, no. 4, pp. 441—451, 1992. [183] C. Lampa, A.F.H. Kaplan, J. Powell, and C. Magnusson, “An analytical thermodynamic model of laser welding,” Journal of Physics D: Applied Physics, vol. 30, pp. 1293—1299, 1997. [184] S. Li and L. Petzold, “Design of daspk for FEMLAB,” in FEMLAB Documentation. COMSOL company, 2001. [185] H.Y. Wong, Handbook of Essential Formulae and Data on Heat Transfer for Engineers, Longman Group United Kingdom, 1977. [186] J. Xie and A. Kar, “Laser welding of thin sheet steel with surface oxidation,” Welding Research Supplement, pp. 343—348, 1999. [187] LASAG, “Manual of 1000 W LASAG FLS 1042N Nd:YAG pulsed laser,” in Technical Manual. 2002. [188] G.N. Abramovich, The Theory of Turbulent Jets, MIT Press, Cambridge, MA, 1963. [189] H. Schlichting, Boundary Layer Theory, McGraw-Hill, New York, 1979. [190] J.O. Hinze, Turbulence, McGraw-Hill, New York, 1975. [191] J. Lin, “Numerical simulation of the focused powder streams in coaxial laser cladding,” Journal of Materials Processing Technology, vol. 105, pp. 17—23, 2000. [192] V.L. Streeter, Handbook of Fluid Dynamics, McGraw-Hill, New York, 1961. [193] FEMLAB, User Guide, COSMOL Inc., Switzerland, 2003. [194] E. Toyserkani, A. Khajepour, and S.F. Corbin, “A 3-D transient finite element model for laser cladding by powder injection,” in Proceeding of the International Congress on Application of Lasers and Electro-Optics, ICALEO’02. LIA, 2002. © 2005 by CRC Press LLC

[195] V. Dragos and R. Kovacevic, “Predicting the laser sheet bending using neural network,” in IEEE International Symposium on Circuits and Systems, pp. 686—689. May 28-31, Geneva, Switzerland. [196] A.A. Peligrad, E. Zhou, and L. Li, “System identification and predictive control of laser marking of ceramic materials using artificial neural networks,” Proceeding of the Institution of Mechanical Engineers. Part 4: Journal of Systems and Control Engineering, vol. 216, no. 2, pp. 181—190, 2002. [197] J. Sjoberg, O. Zhang, L. Ljung, A. Benveisite, B. Delyon, P. Glorennec, H. Hjalmarsson, and A. Juditsky, “Nonlinear black-box modeling in system identification: A unified overview,” Automatica, vol. 31, no. 12, pp. 1691—1724, 1995. [198] O. Nelles, Nonlinear System Identification, Springer, 2001. [199] E.W. Bai, “An optimal two-stage identification algorithm for Hammerstein-Wiener nonlinear system,” Automatica, vol. 34, no. 3, pp. 333—338, 1998. [200] S.A. Billings and S.Y. Fakhouri, “Identification of nonlinear systems using the Wiener model,” Electronics Letters, vol. 13, no. 17, pp. 501— 503, 1977. [201] A.D. Kalafatis, L. Wang, and W.R. Cluett, “Identification of Wienertype nonlinear systems in a noisy environment,” International Journal of Control, vol. 66, no. 6, pp. 923—941, 1997. [202] L. Ljung, System Identification, Theory for the User, Prentice Hall, 1999. [203] J.C. Principe, N.R. Euliano, and W.C. Lefebvre, Neural and Adaptive Systems, John Wiley and Sons, 2000. [204] P. Gilgien and W. Kurz, Microstructure and Phase Selection in Rapid Laser Processing, NATO: Laser processing surface treatment and film deposition, Vol. 307, pp. 77-91, 1989. [205] A.G. Parlos, O.T. Rais, and A.F. Atiya, “Multi-step-ahead prediction using dynamic recurrent neural networks,” Neural Networks, vol. 13, no. 7, pp. 756—786, 2000. [206] O. Nerrand, P. Roussel, D. Urrbani, L. Personnaz, and G. Drefus, “Training recurrent neural networks: Why and how? An illustration in dynamical process modeling,” IEEE Transactions on Neural Networks, vol. 5, no. 2, pp. 178—184, 1994. [207] S. Das and O. Olurotimi, “Noisy recurrent neural networks: the continuous-time case,” IEEE Transactions on Neural Network on Neural Networks, vol. 9, no. 5, pp. 913—936, 1998. © 2005 by CRC Press LLC

[208] W.J. Smith, Modern Optical Engineering, McGraw-Hill, New York, 1990. [209] I. Smurov and M. Ignatiev, “Real time pyrometry in laser surface treatment,” in Laser Processing: Surface Treatment and Film Deposition, J. Mazumder, O. Conde, R. Vilar, and W. Steen, Eds., pp. 529—564. NATO 307, 1996. [210] F. Meriaudeau and F. Truchetet, “Control and optimization of the laser cladding process using matrix cameras and image processing,” Journal of Laser Applications, vol. 36, pp. 317—324, 1996. [211] G.R.B.E. Romer, J. Meijer, and J.O. Benneker, “Process control of laser surface alloying,” Surface Engineering, vol. 14, no. 4, pp. 295—298, 1998. [212] “Pyrometer instrument Co. (pyro), Northvale, New Jersey, USA,” in Internet http://www.pyrometer.com. [213] F.O. Olsen, H. Jorgensen, C. Bagger, T. Kristensen, and O. Gregerson, “Recent investigations in sensorics for adaptive control of laser cutting and welding,” in Proceeding of the LAMP’92, pp. 405—413. 1992. [214] S. Bierman and M. Geiger, “Integration of diagnostics in high power laser systems for optimization of laser material processing,” in Proceeding of the Conference on Modeling and Simulation of Laser System II, pp. 330—341. SPIE 1415, 1991. [215] Y.P. Hu, C.W. Chen, and K. Mukherjee, “Measurement of temperature distributions during laser cladding process,” Journal of Laser Applications, vol. 12, no. 3, pp. 126—130, 2000. [216] G. Kinsman and W.W. Duley, “Fuzzy logic control of CO2 laser welding,” in Proceeding of the International Congress on Application of Lasers and Electro-Optics, ICALEO’4993, pp. 160—167. LIA, 1993. [217] W.W. Duley and G. Kinsman, “Method and apparatus for real-time control of laser processing of materials,” U.S. Patent Number 5659479, 1997. [218] J. Mazumder and J. Koch, “Apparatus and methods for laser cladding,” International Patent, Number WO 00/00921, 2000. [219] J. Mazumder, A. Schierer, and J. Choi, “Direct materials deposition: Designed macro and microstructure,” Journal of Material Research Innovation, vol. 3, no. 3, pp. 118—131, 1999. [220] D. Stegemann, H. Haferkamp, J. Gerken, and C. Reichert, “In-situ inderschung des harestotransportes beim laserstrahl-dispergieren mittels hochgeschwindigkeits-radioskopie,” Metal, vol. 50, no. 3, pp. 185—191, 1996. © 2005 by CRC Press LLC

[221] S.A. Morgan, M.D.T. Fox, M.A. McLean, D.P. Hand, F.M. Haran, D. Su, W.M. Steen, and J.C. Jones, “Real-time process control in CO2 laser welding and direct casting: Focus and temperature,” in Proceedings of ICALEO’4997, pp. 290—299. LIA, 1997. [222] L. Li, W.M. Steen, R.H. Hibberd, and V.M. Weerasinghe, “Real time expert systems for supervisory control of laser cladding,” in Proceedings of ICALEO’87, pp. 9—15. LIA, 1987. [223] R.H Tizhoosh and H. Hauecker, Fuzzy Image Processing: Overview, Academic Press, Boston, 1999.

An

[224] B. Kuo and F. Golnaraghi, Automatic Control, John Wiley, 8th edition, New York, 2002. [225] J.E. Marsden and A.J. Tromba, Vector Calculus, Freeman, 4th edition, UK, 2000. [226] M. Lotfi Zadeh, “Fuzzy sets,” Int. Control, vol. 8, pp. 338—353, 1965. [227] A. Nurenberger, A. Radetzky, and R. Kruse, “Using recurrent neurofuzzy techniques for the identification and simulation of dynamic system,” Neurocomputing, vol. 36, pp. 123—147, 2001. [228] J. Yen and R. Langarani, Fuzzy Logic Intelligence, Control, and Information, Prentice Hall, pp. 194-210, 1998. [229] S. W. Banovic, J.N. Dupont, and A.R. Marder, “Experimental evaluation of Fe-Al claddings in high-temperature sulfidizing environments,” Welding Journal, vol. 80, pp. 63—70, 2001. [230] S.W. Banovic, J.N. Dupont, P.F. Tortorelli, and A.R Marder, “The role of aluminum on the weldability and sulfidation behavior of ironaluminum claddings,” Welding Journal, vol. 78, pp. 23—30, 1999. [231] S.A. David, J.A. Horton, C.G. McKamey, T. Zacharia, and R.W. Reed, “Welding iron aluminide,” Welding Journal, vol. 68, pp. 372—381, 1989. [232] P.J. Maziasz, G.M. Goodwin, C.T. Liu, and S.A. David, “Eects of minor alloying elements on the welding behavior of FeAl alloys for structural and weld-overlay cladding applications,” Scripta Metallurgica et Materialia, vol. 27, pp. 1835—1840, 1992. [233] V.K. Sikka, S.C. Deevi, G.S. Fleischhauer, M.R. Hajaligol, and Lilly, “Iron aluminide useful as electrical resistance heating elements,” U.S. Patent Number 6280082, 2001. [234] R.R. Judkins, P. Singh, and V.K. Sikka, “Iron aluminide alloy for solid oxide fuel cell,” U.S. Patent Number 6114058, 2000. [235] S. Viswanathan S. Vyas and V.K. Sikka, “Eect of aluminum content on environmental embitterment in binary iron-aluminum alloys,” Scripta Metallurgica et Materialia, vol. 27, no. 2, pp. 185—190, 1992. © 2005 by CRC Press LLC

[236] Y.F. Tzeng and F.C. Chen, “Eects of operating parameters on the static properties of pulsed laser welded zinc-coated steel,” International Journal of Advanced Manufacturing Technology, vol. 18, pp. 641—647, 2001. [237] G. Duan, Y. Liu, G. Yang, and Y. Zhou, “Morphological evolution of banding structures at high solidification velocity,” Materials Letters, vol. 57, no. 5, pp. 1091—1095, 2003. [238] T.R. Anthony and H.E. Cline, “Surface rippling induced by surfacetension gradients during laser surface melting and alloying,” Journal of Applied Physics, vol. 48, no. 9, pp. 3888—3894, 1991. [239] A. Almeida, P. Petrov, I. Nogueira, and R. Vilar, “Structure and properties of Al-Nb alloys produced by laser surface alloying,” Materials Science and Engineering A, vol. 303, no. 1-2, pp. 273—280, 2001. [240] S. Sirca, K. Chattopadhyay, and J. Mazumder, “Nonequilibrioum synthesis of NbAl3 and Nb-Al-V alloys by laser cladding: Part I. Microstructure evolution,” Material Transaction, vol. 23A, pp. 2419—2429, 1992. [241] G. Agrawal, A. Kar, and J. Mazumder, “Theoretical-studies on extended solid solubility and nonequilibrium phase-diagram for Nb-Al alloy formed during laser cladding,” Scripta Metall Material, vol. 28, no. 11, pp. 1453—1458, 1993. [242] H. Baker et al., Alloy Phase Diagrams, ASM Handbook, Vol. 3, 1992. [243] L. Costa and R. Vilar, “Diusion-limited layer growth in spherical geometry: A numerical approach,” Scripta Metall Material, vol. 80, no. 8, pp. 4350—4353, 1996. [244] M. Ellis, D.C. Xiao, C. Lee, W.M. Steen, K.G. Watkins, and W.P.M. Brown, “Spinel/metal interfaces in laser coated steels: A transmission electron microscopy study,” Journal of Materials Processing Technology, vol. 52, no. 10, pp. 55—67, 1995. [245] Z. Liu, J.L. Sun, W.M. Steen, K.G. Watkins, C. Lee, and W.P. Broiwn, “Laser cladding of Al-Sn alloy on a mild steel,” Journal of Laser Applications, vol. 9, pp. 35—41, 1997. [246] P. Gilgien, A. Zryd, and W. Kurz, “Microsturcture selection maps for Al-Fe alloys,” Acta Metallurgica et Materialia, vol. 43, no. 9, pp. 3477— 3487, 1995. [247] W. Kurz and R. Trivedi, “Microstructure and phase selection in laser treatment of materials,” Transactions of the ASME, vol. 114, no. 9, pp. 450—458, 1992. © 2005 by CRC Press LLC

[248] D.A. Porter and K.E. Easterling, Phase Transformations in Metals and Alloys, 2nd ed., Chapman and Hall, London, UK, pp. 208-220, 1992. [249] R. Vilar, “Solid freeform fabrication: Laser cladding,” International Journal of Powder Metallurgy, vol. 37, no. 2, pp. 31—48, 2001. [250] M.J. Aziz, “Model for solute redistribution during rapid solidification,” Journal of Applied Physics, vol. 53, no. 2, pp. 1158—1168, 1982. [251] S.C. Gill and W. Kurz, “Laser rapid solidification of Al-Cu alloys:banded and plane front growth,” Material Science Engineering A, vol. 173, pp. 335—338, 1993. [252] W.J. Boettinger, D. Shechtman, J. Schaeferand, and P.S. Biancaniello, “The eect of rapid solidification velocity on the microstructure of AgCu alloys,” Material Transaction A, vol. 15, pp. 55—66, 1984. [253] J. Choi and J. Mazumder, “Non-equilibrium synthesis of Fe-Cr-C-W alloy by laser cladding,” Journal of Materials Science, vol. 29, no. 17, pp. 4460—4476, 1994. [254] Q. Li, D. Zhang, T. Lei, C. Chen, and W. Chen, “Comparison of laserclad and furnace-melted Ni-based alloy microstructures,” Surface and Coating Technology, vol. 137, pp. 122—135, 2001. [255] X. Wu, “Rapidly solidified non-equilirium microstructures and phase transformation of laser-synthesized iron-based alloy coating,” Surface and Coating Technology, vol. 115, pp. 153—162, 1999. [256] P. Sallamand and J.M. Pelletier, “Laser cladding on aluminium-base alloys:microstructural features,” Material Science Engineering A, vol. 171, pp. 263—270, 1993. [257] A. Tiziani, L. Giordano, P. Matteazzi, and B. Badan, “Laser Stellite coatings on austenitic steels,” Materials Science and Engineering, vol. 88, no. 1, pp. 171—175, 1987. [258] W.L. Song, J. Echigoya, B.D. Zhu, C.S. Xie, W. Huang, and K. Cui, “Vacuum laser cladding and eect of Hf on the cracking susceptibility and the microstructure of Fe-Cr-Ni laser-clad layer,” Surface and Coatings Technology, vol. 126, no. 1, pp. 76—80, 2000. [259] C.R. Ribanudo Nd J. Mazumder and D.W. Hetzner, “Laser-clad NiAlCrHf alloys with improved alumina scale retention,” Metallurgical Transactions B: Process Metallurgy, vol. 23, no. 4, pp. 513—522, 1992. [260] J. Singh, K. Nagarathnam, and J. Mazumder, “Laser cladding of NiCr-Al-Hf on inconel 718 for improved high-temperature oxidation resistance,” High Temperature Technology, vol. 5, no. 3, pp. 131—137, 1987. © 2005 by CRC Press LLC

[261] J. Singh, M. Zmuda, and J. Mazumder, “In-situ Ni-Cr-Al-Fe-Hf alloy by laser cladding with mixed powder feed,” Journal of Metals, vol. 37, no. 11, pp. A34—A34, 1985. [262] J. Przbylowicz and J. Kusinski, “Structure of laser cladded Tungsten carbide composite coatings,” Journal of Materials Processing Technology, vol. 109, no. 1-2, pp. 154—160, 2001. [263] X.Y. Zeng, Z.Y. Tao, B.D. Zhu, E.H. Zhou, and K. Cui, “Investigation of laser cladding ceramic-metal composite coatings: Processing modes and mechanisms,” Surface and Coatings Technology, vol. 79, no. 1-3, pp. 209—217, 1996. [264] S. Nowotny, A. Techel, A. Luft, and W. Reitzenstein, “Microstructure and wear properties of laser clad carbide coatings,” in Laser Materials Processing ICALEO’4993, pp. 985—993. SPIE 2306, 1993. [265] J.H. Ouyang, T.C. Lei, and Y. Zhou, “Microstructure of laser-clad SiC-(Ni alloy) composite coating,” Materials Science and Engineering A—Structural Materials Properties Microstructure and Processing, vol. 194, no. 2, pp. 219—224, 1995. [266] A. Hidouci, F. Simonin, and J.M. Pelletier, “Preparation by laser cladding and microstructural characterization of MoSi2 and (MoSi2 +ZrO2 ) coatings on steel and graphite substrates,” Lasers in Engineering, vol. 8, no. 1, pp. 17—, 1998. [267] S. Yang, N. Chen, W.J. Liu, and M.L. Zhong, “In situ formation of MoSi2 /SiC composite coating on pure al by laser cladding,” Materials Letters, vol. 57, no. 22-23, pp. 3412—3416, 2003. [268] B.J. Kooi, Y.T. Pei, and J.T.M. Hosson, “The evolution of microstructure in a laser clad TiB-Ti composite coating,” Acta Metallurgica et Materialia, vol. 51, no. 3, pp. 831—845, 2003. [269] Q.M. Zhang, J.J. He, and W.J. Liu M.L. Zhong, “Microstructure characteristics of ZrC-reinforced composite coating produced by laser cladding,” Surface and Coating Technology, vol. 162, no. 2-3, pp. 140— 146, 2003. [270] S. Yang, M.L. Zhong, and W.J. Liu, “TiC particulate composite coating produced in situ by laser cladding,” Materials Science and Engineering A—Structural Materials Properties Microstructure and Processing, vol. 343, no. 1-2, pp. 57—62, 2003. [271] S.F. Corbin and D.S. Wilkinson, “Influence of matrix strength and damage accumulation on the mechanical response of a particulate metalmatrix composite,” Acta Metallurgica et Mateialia, vol. 42, no. 4, pp. 1329—1335, 1994. © 2005 by CRC Press LLC

[272] S.F. Corbin, X. Zhao-jie, H. Henein, and P.S. Apte, “Functionally graded metal/ceramic composite by tape casting, lamination and infiltration,” Materials Science and Engineering A, vol. 262, pp. 192—203, 1999. [273] A. Mortensen and S. Suresh, “Functionally graded metals and metalceramic composites 1. processing,” International Materials Reviews, vol. 40, no. 6, pp. 239—265, 1995. [274] X. Wu and Y. Hing, “Microstructure and mechanical properties at TiC/Ni-alloy interfaces is laser-synthesized coatings,” Materials Science Engineering A, vol. 318, pp. 15—21, 2001.

© 2005 by CRC Press LLC