No title

VOLUME 6 NUMBER 1 1994 Manufacturing Technology for Apparel Automation – Layup Module Part II The Impact of Fabric Prop...

Author: Emerald

186 downloads 765 Views 2MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

VOLUME 6 NUMBER 1 1994

Manufacturing Technology for Apparel Automation – Layup Module Part II The Impact of Fabric Properties on the Gap Length between Two Slats A. Seyam and F. Sun Department of Textile and Apparel Management, College of Textiles, North Carolina State University, Raleigh, USA

the gap between two slats Lg assuming the cloth trailing end falls freely under its own weight (Figure 1). The approach of free fall offers the most inexpensive solution as opposed to developing a mechanism to push down the fabric trailing end between the slats. Lg must be long enough to prevent fabric from touching the top or the side surfaces of the slat during wiping off the fabric from the conveyor surface in order to obtain a straight and well aligned stack of plies, which in turn reduces fabric waste during pattern cutting. Considering the above approach and the requirements of Lg, it is clear then that slat

Received 26 August 1993 Accepted 12 December 1993

Introduction As has been mentioned in Part I[1], the overall goal of this research work is to develop inexpensive automatic cloth layup machine to eliminate the century-old technology currently in use by creating a machine to (at a rate of about 100 m/min) stack rapidly fabric plies as fast as the Gerber cutter can cut. The layup machine will enable a Gerber cutting room to operate without spreading tables, and without spreading machines. The recognized advantages in such development are quick response to customer’s orders, high production rate, elimination of unnecessary machines, less floor space, reduced labour cost, and less manufacturing cost. The principle of the layup module is given elsewhere[1,2]. Part I was devoted to designing the dimensions of the slat, the main element of the cloth-stacking conveyor. The focus of this part is on estimation of the minimum length of

We would like to acknowledge the Defense Logistic Agency, USA for the financial support under contract number DLA900-87-CO509. Additionally, the authors wish to extend their sincere appreciation to Mr T. McDevitt and Dr J.W. Eischen of NCSU for providing the computerized numerical solution, Drs T.G. Clapp and T.J. Little of NCSU for their useful suggestions, and Professor E.M. McPherson of NCSU, the project co-ordinator, for giving us the opportunity to join the project team.

International Journal of Clothing Science and Technology, Vol. 6 No. 1, 1994, pp. 5-13, © MCB University Press, 0955-6222

5

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

flexible beams under large overall deformation is considered. The computerized numerical solution developed by McDevitt[14], which is based on Simo and Vu-Quoc model, is used to determine the locus of the fabric trailing end at different time intervals during wiping off the fabric from the conveyor surface.

Lg Platen Slat

O

A

h

B C

Theoretical Consideration

Stacking table

Textile fabrics are very complex non-linear mechanical systems for which it is difficult theoretically to predict their bending behaviour in terms of their constituents with reasonable accuracy. A simpler theoretical approach is to consider the fabric as a flexible continuum that undergoes large displacement and rotation. Thus the approach considers fabric bending behaviour without accounting for fibre, yarn, and weave parameters. Two methods can be applied to our case. The first is theoretical coupled with experimental determination of the moment/curvature relationship to predict the deformed cantilever shape of the fabric.

Figure 1. A Ply Trailing End Pictured at Different Times during Descending through a Gap between Two Slats

height and more importantly the fabric bending behaviour are the key elements in deciding the length of the gap between two slats. Since early this century, the area of bending behaviour of textile fabrics has drawn the attention of many researchers for its significant impact on cloth hand, performance of fabrics in garment manufacturing, and design of equipment and machinery that handle fabrics in automated apparel manufacturing. Two theoretical approaches have been adopted by researchers to predict the bending behaviour of textile fabrics. The first is to develop analytical equations expressing the bending behaviour in terms of fabric constituents characteristics (such as fibre properties, yarn properties, etc.) and the fabric geometry. Recently, Ghosh et al.[3] published an extensive critical review of previous work. Additionally, they developed[4,5] computational models, following the first approach, to predict the bending behaviour of plain woven fabrics in case of linear and bilinear thread bending. The second approach is to predict the deformed fabric shape from a measured moment/curvature relationships using computerized numerical solutions[6-14]. The measured bending characteristics can be determined very accurately by Kawabata[15] or FAST (Fabric Assurance by Simple Test)[16] systems. Very recently, the latter approach has been adopted in the area of fabric handling in apparel automation. Examples of these are fabric folding [7], laying down of fabrics on a work surface [8], and picking up of fabrics from a work surface [9]. We will follow this technique to help in estimating the minimum gap length between two slats of the layup module. Specifically, the model developed by Simo and Vu-Quoc[12,13] of in-plane dynamics of

n

The deformed co-ordinates of a fabric sample are recorded n The second is a purely experimental method which was developed recently by Clapp et al.[17]. In this technique, the deformed coordinates of a fabric sample as it is cantilevered under its own weight are recorded using sample cantilever fixture and vision system. The method could be adopted here only to determine the co-ordinates of the trailing fabric end at different bending lengths (points A, B, C of Figure 1). The first technique is adopted here. In the remainder of this section, summaries of theoretical works of Simo and Vu-Quoc[12,13], Eischen[11], and McDevitt[14] are given. Basic Kinematic Model The basic kinematic general model was developed[12,13] for a beam capable of undergoing large in-plane deflections as a result of concentrated forces, moments, and uniform distributed load (e.g. self-weight). Figure 2

6

VOLUME 6 NUMBER 1 1994

equilibrium employing the penalty method[18] as:

l

y ω

F1

m(0)

F2

l

m(l) n(0)

n(l)

q(0)

x+∆x

x

∫ [δεEAε + (δθ )' EIθ ' ]dx +

x

q(1)

0

a. Undeformed flexible beam

F2

m(l)

ω

1

q(l) x

n(0)

n(l)

F1

m(0)

l   u '   u '  a ∫ δ ( tan θ ) − δ  2   tan θ − 2  dx = 1 + u1'   1 + u1'    0

− ∫ δ u2ω dx + (n δu1 + q δu2 + mδθ ) l0 ,

x+∆x

q(0) b.Deformed flexible beam

where

Source:[14]

a

ε

Figure 2. Large Deflection Planer Flexible Beam

ω

EA EI GA E A I

m(x+∆x) n(x+∆x) q(x+∆x)

m(x) u2(x+∆x)–u2(x) n(x) y

q(x)

(1)

0

∆x+u1(x+∆x)–u1(x)

GA 1.2 = the axial strain =

= beam axial stiffness = beam bending stiffness = beam shear stiffness = modulus of elasticity of the beam = beam cross - sectional area per unit width = beam cross - section moment of inertia

(= t 3 /12) t G

x

Figure 3. Free Body Diagram of a Differential Element

= fabric thickness, and = shear modulus of the beam.

Primes indicate differentiation with respect to x. A quantity’s first variation, denoted with the δ symbol, is a hypothetical increment in the quantity without a corresponding increment in its independent variables. Since Equation (1) is a highly non-linear variational equation in terms of the displacements and rotation, finite element analysis of non-linear problem is invariably performed with an iteration scheme of solving the problem’s linearized counterpart. The directional derivative is used to systematically linearize the variational equation and prepare it for a finite element discretization[19].

shows the beam with an arbitrary external loading (concentrated forces F1, F2, and uniform load of density ω). The beam’s reaction forces, n(0), n(0), q(0), q(l), m(0), and m(l), restrict rigid body translation and rotation. Figure 3 shows a free body diagram of a differential element. Displacements in x and y directions, and the rotation of the cross-section are labelled u1(x), u2(x), and θ(x); respectively. The internal bending moment m and the internal forces of the element are decomposed into components q and n directed along the x and y axes; respectively.

Finite Element Discretization The four-node one-dimensional element shown in Figure 4 was used[14] for the spatial discretization of the beam’s domain. The notation used to identify an element’s degrees of freedom are:

Non-linear Variational Equilibrium Equations and Linearization Eischen[11], and McDevitt[14] neglected the kinematic energy terms of Simo and VuQuoc[12,13] model and by using the beam’s system of non-linear differential equilibrium equations, they derived a variational equation of

uij = ui at the jth node, and

θj= θ at the jth node (for i = 1, 2 and j = 1,4).

7

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

u21

θ1 1

u22

θ2

u11

2

u23

θ3

u12

u24

θ4

u13

3

The shear rigidity (GA)eff is estimated to be equal to the axial rigidity (EA)eff . Eischen and Kim’s[11] numerical experiments have shown that (GA)eff = (EA)eff for fabrics loaded primarily by their own weight.

4

u14 ξ

ξ = –1

ξ = –1/3

ξ =1/3

ξ=1

Source: [14]

n

Figure 4. Four-node Twelve-degree of Freedom Beam Element

The platen and the slats act as a fixture holding the fabric n

McDevitt[14] derived a system of algebraic equations using piecewise cubic shape functions by interpolating the displacements u1, u2, and cross-section rotation θ over the generic element. The variational quantities δu1, δu1', δθ δu2' and δθ ' were approximated using the same shape functions.

Table I shows the fabric basis weight, fabric thickness, and fabric bending rigidity per unit width of seven selected woven fabrics. These three experimentally measured properties are the only parameters needed to solve the system of algebraic equations. The fabrics were selected from previous publications[8,11] because they cover a wide range of apparel fabrics from light-weight shirting to heavyweight denimjean.

Numerical Results and Discussion Material Properties To produce simulations of fabric cantilever shape (computerized solution of the system of algebraic equations referred to above), the five fabric material properties data (E, A, t, I, and G) must be known. The thickness of the fabric t can be measured by thickness tester and hence A and I can be calculated (A = t and I =t3/12). An effective bending rigidity value (EI)eff per unit width of fabric then can be generated from moment-curvature relationship determined experimentally by Kawabata bending tester[15]. (EI)eff is the average slope of the momentcurvature response at curvatures of 0.5 cm and 1.5 cm, for positive and negative curvatures. From A, I and (EI)eff , (EA)eff can be calculated.

Thickness (cm)

Bending rigidity/ unit width (gf-cm2/cm)

1

0.0165

0.0292

0.080

2

0.0149

0.0267

0.090

3

0.0283

0.0483

0.275

4

0.0255

0.0406

0.386

5

0.0130

0.0229

0.040

6

0.0324

0.0584

0.292

7

0.0472

0.0940

1.571

1

Y Vertical deflection of fabric trailing end

Fabric Weight/unit area Code (gf/cm2)

Locus of the Fabric Trailing End We have adopted x-y plane with origin of point O(0,0) (Figures 1, 5-11). The platen and the slats act as a fixture holding the fabric which is considered as a cantilever of varying length of time dependent. The numerical solution allows the calculation of the co-ordinates of all the points of the cantilever at a given fabric length. Each of Figures 5-11 shows a group of cantilevers of different lengths (broken lines) representing the fabric at different time interval during wiping off the fabric to the stacking table.

0

-1

-2

-3

-4

-5 -3

Sources:[8,11]

-2

-1

0

1

2

X Horizontal displacement of fabric trailing end

Table I. Properties of Fabrics

Figure 5. Locus of Fabric Trailing End for Fabric 1

8

3

VOLUME 6 NUMBER 1 1994

1

Y Vertical deflection of fabric trailing end

Y Vertical deflection of fabric trailing end

1

0

-1

-2

-3

-4

0

-1

-2

-3

-4

-5

-5 -3

-2

-1

0

1

2

-3

3

-2

-1

0

1

2

3

X Horizontal displacement of fabric trailing end

X Horizontal displacement of fabric trailing end

Figure 9. Locus of Fabric Trailing End for Fabric 5

Figure 6. Locus of Fabric Trailing End for Fabric 2

1

Y Vertical deflection of fabric trailing end

Y Vertical deflection of fabric trailing end

1

0

-1

-2

-3

-4

0

-1

-2

-3

-4

-5

-5 -3

-2

-1

0

1

2

-3

3

X Horizontal displacement of fabric trailing end

-1

0

1

2

3

Figure 10. Locus of Fabric Trailing End for Fabric 6

Figure 7. Locus of Fabric Trailing End for Fabric 3

1

Y Vertical deflection of fabric trailing end

1

Y Vertical deflection of fabric trailing end

-2

X Horizontal displacement of fabric trailing end

0

-1

-2

-3

-4

0

-1

-2

-3

-4

-5

-5 -3

-2

-1

0

1

2

-2

3

-1

0

1

2

3

X Horizontal displacement of fabric trailing end

X Horizontal displacement of fabric trailing end

Figure 11. Locus of Fabric Trailing End for Fabric 7

Figure 8. Locus of Fabric Trailing End for Fabric 4

9

4

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Model:

x = b1 y + b2 √y Parameters b1 b2

R2 = 0.9977 Prob > F = 0.0001 Estimate P-value –1.11380 0.0001 2.88514 0.0001

Table II. Polynomial Regression for Fabric 1

Model: x = b1 y + b2 √y Parameters b1 b2

R2 = 0.9973 Prob > F = 0.0001 Estimate P-value –1.16088 0.0001 3.98068 0.0001

Table VIII. Polynomial Regression for Fabric 7

The square markers of Figures 5-11 represent the fabric trailing end at different time intervals. The co-ordinates of these square markers are the key to determine the minimum Lg as it will be seen later. The solid lines of Figures 5-11 represent the best fitting curve to the fabric trailing end. The regression equations of the seven fabrics were obtained using the least square method to smooth the numerical results and obtain analytical equation of x in terms of y. The resulting regression relationships are shown in Tables IIVIII. It is clear from Figures 5-11 that the general shape of the locus of the trailing end is of a maximum value of x(xmax) which can be obtained by solving the equation d x/ d y = 0 for each of the regression equations. The high R2-value (R is the correlation coefficient) and the low P-value (the probability of the error of the estimate) of the

R2 = 0.9979 Prob > F = 0.0001 Estimate P-value –1.15300 0.0001 3.06375 0.0001

Table III. Polynomial Regression for Fabric 2

Model: x = b1 y + b2 √y Parameters b1 b2

x = b1 y + b2 √y Parameters b1 b2

Model:

R2 = 0.9980 Prob > F = 0.0001 Estimate P-value –1.11520 0.0001 3.20948 0.0001

Table IV. Polynomial Regression for Fabric 3

(a)

Model: x = b1 y + b2 √y Parameters b1 b2

R2 = 0.9980 Prob > F = 0.0001 Estimate P-value –1.18614 0.0001 3.58407 0.0001

Platen

x h = ya

Table V. Polynomial Regression for Fabric 4

Slat

y |x max

y

Model: x = b1 y + b2 √y Parameters b1 b2

Path of fabric trailing end

Lg

A(xa, ya)

xmax h < y |x max

R2 = 0.9963 Prob > F = 0.0001 Estimate P-value –0.98616 0.0001 2.44655 0.0001

(b) Platen

Table VI. Polynomial Regression for Fabric 5

Path of fabric trailing end

Lg

x y |x max

h = yb

Slat B(xb, yb)

Model:

x = b1 y + b2√y Parameters b1 b2

R2 = 0.9975 Prob > F = 0.0001 Estimate P-value –1.08768 0.0001 3.11015 0.0001

y

xmax h ≥ y |x max

Table VII. Polynomial Regression for Fabric 6

Figure 12. The Estimate of Lg

10

VOLUME 6 NUMBER 1 1994

regression equations indicate that the model x = b1 y + b2 √y is highly significant.

(y|x max). Figure 12b illustrates the case when the slat height is greater than or equal to the y coordinate of the trailing fabric end at xmax. Considering the two cases, the following steps are necessary to estimate Lg min:

The Estimate of the Minimum Gap between Two Slats (Lg min) Two cases may take place and need to be considered to estimate the minimum gap between two slats in terms of slat height. These cases are best described by Figure 12. Figure 12a represents the case when the slat height is less than the y co-ordinate of the trailing fabric end at xmax

Fabric Code

y|x=x max (cm)

The value of y|x max and hence xmax can be obtained by solving the equation dx/dy = 0. ● From regression equation, x|y = h can be determined. ● Finally, Lg min can be estimated from the following: ●

x max (cm)

h (cm)

x|y=h (cm)

Lg min (cm)

1

–1.678

1.868

–0.635

1.592

1.592

2

–1.765

2.035

–0.635

1.709

1.709

3

–2.071

2.309

–0.635

1.849

1.849

4

–2.295

2.725

–0.635

2.111

2.111

5

–1.539

1.517

–0.635

1.323

1.323

6

–2.044

2.223

–0.635

1.788

1.788

7

–2.940

3.412

–0.635

2.435

2.435

1

–1.678

1.868

–1.270

1.837

1.837

2

–1.765

2.035

–1.270

1.988

1.988

3

–2.071

2.309

–1.270

2.201

2.201

4

–2.295

2.725

–1.270

2.544

2.544

5

–1.539

1.517

–1.270

1.505

1.505

6

–2.044

2.223

–1.270

2.124

2.124

7

–2.940

3.412

–1.270

3.012

3.012

1

–1.678

1.868

–1.905

1.860

1.868

2

–1.765

2.035

–1.905

2.032

2.035

3

–2.071

2.309

–1.905

2.305

2.305

4

–2.295

2.725

–1.905

2.701

2.701

5

–1.539

1.517

–1.905

1.498

1.517

6

–2.044

2.223

–1.905

2.221

2.221

7

–2.940

3.412

–1.905

3.283

3.283

1

–1.678

1.868

–2.540

1.769

1.868

2

–1.765

2.035

–2.540

1.954

2.035

3

–2.071

2.309

–2.540

2.283

2.309

4

–2.295

2.725

–2.540

2.715

2.725

5

–1.539

1.517

–2.540

1.394

1.517

6

–2.044

2.223

–2.540

2.194

2.223

7

–2.940

3.412

–2.540

3.396

3.396

Table IX. Lg min as a Function of Slat Height h

11

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

before it reaches to the predetermined zone of wiping. Currently, a model to predict the slat width in terms of fabric properties and the gap between two slats is being developed and will be reported in near future.

If h (= ya) < y|x max, then Lg min = xa (Figure 12a). If h (= ya) ≥ y|x max, then Lg min = xmax (Figure 12b).

n

Table IX shows the results of Lg min as a function of slat height by following the above procedure for the seven fabrics of Table I using the seven regression equations of Tables II-VIII. The results of Table IX are applicable to hollow-rectangular cross-section slats (shown in Figures 1 and 12) as well as C cross-section slats. The two crosssections were recommended in Part I[1]. The method can be applied to other cross-sections if so desired.

References 1. Seyam, A., Sun, F., McPherson, E., Clapp, T. and Little, T., “Manufacturing Technology for Apparel Automation – Layup Module, Part I: Slat Design”, International Journal of Clothing Science and Technology, Vol. 4 No. 5, 1993, pp. 44-59. 2. Sanborn, C.H., Apparatus for Stacking Pieces of Limp Material, US Patent No. 5,098,079, 24 March 1992. 3. Ghosh, T.K., Batra, S.K. and Barker, R.L., “The Bending Behavior of Plain-woven Fabrics, Part I: A Critical Review”, Journal of the Textile Institute, Vol. 81, 1990, pp. 24554. 4. Ghosh, T.K., Batra, S.K. and Barker, R.L., “The Bending Behavior of Plain-woven Fabrics, Part II: The Case of Linear Threadbending Behavior”, Journal of the Textile Institute, Vol. 81, 1990, pp. 255-71. 5. Ghosh, T.K., Batra, S.K. and Barker, R.L., “The Bending Behavior of Plain-woven Fabrics, Part III: The Case of Bilinear Threadbending Behavior and the Effect of Fabric Set”, Journal of the Textile Institute, Vol. 81, 1990, pp. 272-87. 6. Konopasek, M., “Classical Elastica Theory and Its Generalization”, Mechanics of Flexible Fibre Assemblies, NATO ASI Papers, 1980, pp. 255-74. 7. Lloyd, D.W., Shanahan, W.J. and Konopasek, M., “The Folding of Heavy Fabric Sheets”, International Journal of Mechanical Sciences, Vol. 20, 1978, pp. 521-7. 8. Brown, P.R., Buchanan, D.R. and Clapp, T.G., “Large-deflection Bending of Woven Fabric for Automated Material Handling”, Journal of the Textile Institute, Vol. 81, 1990, pp. 1-13. 9. Clapp, T.G. and Peng, H., “A Comparison of Linear and Non-linear Bending Models for Predicting Fabric Deformation in Automated Handling”, Journal of the Textile Institute, Vol. 82, 1991, pp. 341-52.

n

The gap length between two slats is affecting the slat width n As was expected, the results of Table IX prove that as the fabric bending rigidity gets higher, the minimum gap between two slats increases. Additionally, as the slat height increases Lg min also increases. The results suggest that a variable gap length between two slats should be considered to process wide range of fabrics. Preliminary experimentation using the prototype layup module verified some of the results reported above.

Conclusion The findings of this work have proved that the fabric characteristics significantly affect the design of the conveyor belt of the layup module. The minimum gap length between two slats (which is needed to prevent the fabric trailing end contact with the slat surface or side as it descends during wiping off) of the conveyor belt generally increases when the bending stiffness or the slat height increases. An attractive design of the layup module would be of variable gap length between two slats to accommodate the processing of wide range of textile fabrics. It is recognized that the gap length between two slats is affecting the slat width. The slat width is supporting the fabric as it is being transferred to the area under the platen. Too wide gap with respect to the slat width may result in insufficient support of the fabric by the conveyor belt and consequently the fabric may fall between the slats

12

VOLUME 6 NUMBER 1 1994

10. Postle, J.R. and Postle, R., “Fabric Bending and Drape Based on Objective Measurement”, International Journal of Clothing Science and Technology, Vol. 4 No. 5, 1992, pp. 7-15. 11. Eischen, J.W. and Kim, Y.G., “Optimization of Fabric Manipulation during Pick/Place Operation”, Proceedings of the Fourth Annual Academic Apparel Research Conference, Raleigh, 8-9 February, 1993. 12. Simo, J.C. and Vu-Quoc, L., “On the Dynamics of Flexible Beams under Large Overall Motions – The Plane Case: Part I”, Journal of Applied Mechanics, Vol. 53, 1986, pp. 849-54. 13. Simo, J.C. and Vu-Quoc, L., “On the Dynamics of Flexible Beams under Large Overall Motions – The Plane Case: Part II”, Journal of Applied Mechanics, Vol. 53, 1986, pp. 855-63. 14. McDevitt, T.W., “Flexible Fabric Mechanics Analysis Using Large Deflection Beam Theory”, Master of Science Thesis, North Carolina State University at Raleigh, 1993.

15. Kawabata, S., “The Standardization and Analysis of Hand Evaluation”, The Hand Evaluation and Standardization Committee, The Textile Machinery Society of Japan, Osaka, Japan, 1975. 16. Ly, N.G., Tester, D.H., Buckenham, P., Roczniok, A.F., Andriaasen, A.L., Scaysbrook, F. and De Jong, S., “Simple Instruments for Quality Control by Finishers and Tailors”, Textile Research Journal, Vol. 61, 1991, pp. 402-6. 17. Clapp, T.G., Peng, H. and Ghosh, T.K., “Indirect Measurement of the MomentCurvature Relationship for Fabrics”, Textile Research Journal, Vol. 60, 1990, pp. 525-33. 18. Zienkiewicz, O.C. and Taylor, R.L., The Finite Element Method, McGraw-Hill Book Company, London, 1977. 19. Hughes, T.J.R., The Finite Element Method: Linear Static and Dynamic Finite Element Analysis, Prentice-Hall, Englewood Cliffs, NJ, 1987.

13

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

A Study of Thread Tensions on a Lockstitch Sewing Machine (Part I) F.B.N. Ferreira Universidade do Minho, Azurem, Portugal S.C. Harlock Department of Textile Industries, the University of Leeds, Leeds, UK and P. Grosberg Shenkar College, Ramat Gan, Israel

Received 2 April 1993 Accepted 22 November 1993

tension in terms of tension variations in a stitch cycle and how different factors affect their behaviour.

Introduction The tensions generated in the needle and bobbin thread are an important factor in determining the quality of a lockstitch seam. Improper tensions can cause various problems, such as puckered seams, thread breakage and unbalanced seams. Several studies[1-7] have already been made on this subject. However, the bobbin thread tension has always been considered in a passive way, i.e. the setting of the bobbin thread tension was made at the beginning of the work and then the needle thread tensioner was adjusted to determine its effect on stitch and seam quality. To date, no systematic analysis of the bobbin thread tension during a stitch cycle of a lockstitch sewing machine has been found in the literature. Questions therefore remain as to how the setting of such factors as the bobbin thread tensioner, needle thread tensioner, different thread qualities, the number of plies and sewing speed affect the bobbin thread tension. Therefore this article investigates the simultaneous measurement of both the needle thread tension and bobbin thread

Needle and Bobbin Thread Tension Measurement Needle Thread Tension Measurement The tension on the needle thread during a stitch cycle was measured on the thread line between the take-up lever and the needle by sensing the deflection of a cantilever beam using bonded foil strain gauges. The natural frequency of the cantilever was about 4kHz. Bobbin Thread Tension Measurement In order to measure the tension of the bobbin thread, consideration had to be given to the design of the components in relation to the storage and feeding of the bobbin thread during the stitch cycle, namely the bobbin, bobbin case and rotary sewing hook. Therefore, a cantilever beam was made to support the strain gauges and convert the variation in tension into a deflection of the device in the area where the strain gauges were fixed. The cantilever beam using bonded foil strain gauges was fixed in front of the bobbin case

International Journal of Clothing Science and Technology, Vol. 6 No. 1, 1994, pp. 14-19, © MCB University Press, 0955-6222

14

VOLUME 6 NUMBER 1 1994

where there was a small gap bridged by the bobbin thread. The natural frequency of this cantilever beam was about 5.5 kHz. The analogue signals from both transducers were processed through an operational amplifier and an analogue to digital converter and finally stored in a computer memory for subsequent analysis. Details of the design of the transducers, electronic instrumentation and software are given in[8]. Plates 1 and 2 show the transducers in position on a Singer Centurion sewing-machine which was used in the experimental programme.

placed in contact with each transducer in turn in a path that simulated, as closely as possible, the path taken by the needle and bobbin threads respectively in situ.

General Pattern of Needle and Bobbin Thread Tension Variations Typical traces showing the variations on the needle and bobbin thread tension measured during a complete stitch cycle are illustrated in Figures 1 and 2, respectively. The needle thread tension traces are very similar to those obtained in previous studies [1-7]. Typically, the graphs show four significant peaks of tension on the needle thread tension trace and two significant peaks of tension on the bobbin thread tension trace. In the following text, these peaks will be referred to as Peak 1 to Peak 6, as marked in Figures 1 and 2.

Calibration The calibration of the transducers was performed by suspending weights on a free end of a thread

Tension (GRF) 450 400 350 1 300 250

4

200

3

150 2

100 50 0 0

45

90

135

180

225

270

315 360 Stitch cycle

180

225

270

315 360 Stitch cycle

Sewing speed – 3,000rpm Tensioner adjustment – 300G

Plate 1. Needle Thread Tension Transducer

Figure 1. Needle Thread Tension

Tension (GRF) 450 400 350 300 250 200

6

150 5

100 50 0 0

45

90

135

Sewing speed – 3,000rpm Tensioner adjustment – 300G

Plate 2. Bobbin Thread Tension Transducer

Figure 2. Bobbin Thread Tension

15

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Table I relates the timing of these peaks to the events occurring during a sewing cycle. An interesting observation is that the peak tension for the bobbin thread as it is pulled off the bobbin (Peak 5) is approximately one-third the value of the needle thread tension in its corresponding peak (Peak 1). This will be considered in more detail in Part III of this series of papers.

reference will be made to the alteration in Table II. Experimental Results Different experiments were made under the following conditions (as shown in Table II): (1) Changes to needle thread tensioner setting by varying the tension applied on the needle thread by the tensioner from 200gf to 400gf. (2) Changes to bobbin thread tensioner setting at a nominal constant needle thread tension of 300gf with the nominal bobbin thread tension varying from 10gf to 80gf. The nominal needle and bobbin thread tensions were based on the static calibration method. Adjustments to the needle thread tension were made using the spring/wire tensioner on the machine. Adjustments to the bobbin thread tension were made by changing the screw setting on the bobbin thread case. (3) Changes in sewing speed from 1,500 to 4,500 rpm but at a constant nominal needle thread tension of 300gf. (4) Number of plies (as experiment 1, but with 2, 3 and 4 plies; NP = 2, NP = 3, NP = 4). (5) Fabric quality (as experiment 1, but with different fabric qualities, A, B and C as shown in Table III). (6) Sewing thread qualities (as experiment 1, but with different thread qualities, A, B and C as shown in Table IV).

Effect of Some Factors on Needle and Bobbin Thread Tensions Sewing Conditions The general sewing conditions used during the experiments are shown in Table II. In the following section, for each different experiment,

Peak no.

Timing Event (°)

1

0-95

2 3

4

5

6

The thread take-up lever completes its upward stroke, drawing the slack needle thread through the fabric and pulling the bobbin thread off the bobbin and setting the stitch 105-150 Needle eye penetrates the fabric 280-325 Multipeak effect owing to the movement upwards of the needle thread and consequent tightening of the needle thread loop around the bobbin case 325-355 The needle thread slips out of the rotating sewing hook section jib and the case holder position bracket 0-40 The bobbin thread is pulled off the bobbin by the needle thread loop on its way upwards 80-90 Contact between the bobbin thread and the thread deflector on the sewing hook

Analysis Methodology The results obtained during the experiments on different sewing conditions were analysed considering two different aspects: the effect of the sewing factor on the tension traces and on the peak tension values. For each experiment, a quadratic equation for each peak tension was fitted to the data as follows:

Table I.

1. Sewing-machine – Singer 121D300BA Needle thread static tension – 70gf Bobbin thread static tension – 15gf Stitch density – 3 stitches/cm Sewing speed – 3,000rpm Needle – Singer 90 2. Fabric composition – 100 per cent cotton Fabric structure – Plain, 27 ends/cm, 24 picks/cm Fabric weight – 147g/m2 3. Sewing thread – Ticket number 120 Corespun – 61 per cent polyester – 39 per cent cotton

Peak i = a0 + a1 *T + a2 *T2 where Peak i – tension of peak i a0,…, a2 – coefficients defined by multiple regression technique. T – Principle variable: thread tension, bobbin tension, etc. The data collected in these experiments were processed using the multiple regression technique. With the regression equations obtained, curves

Table II. Sewing Conditions

16

VOLUME 6 NUMBER 1 1994

Fabric properties Composition Structure Ends/cm – Picks/cm Courses/cm – Wales/cm Weight (g/m2)

Fabric A 100 per cent cotton Plain 27-24 – 147

Fabric B

One of the important conclusions was the fact that, from the analysis of the data obtained, most of the multiple regression equations derived presented a very good correlation coefficient, in most cases greater than 0.90 with interval limits less than ± 15gf, for a confidence of 95 per cent. This shows that it is possible to predict, with a good degree of accuracy, the tensions generated on the different peaks, according to the sewing conditions. Furthermore, it was possible to conclude that:

Fabric C

100 per cent 100per cent wool wool Twill 2 × 2 Interlock 19-17 – 363

– 11-10 338

Distinctive tensions applied on the needle thread by the main tensioner significantly

●

Table III. Fabric Properties

450

Thread properties

1

400

Thread A

Thread B

350

Thread C

4 3

Cotton 100 per cent

2

6

400

390

380

370

360

350

340

330

320

310

300

290

0 280

7

50

270

17

5

260

27

100

250

40

150

240

36 49

200

230

120 25

250

220

Polyester 100 per cent

210

Polyester 61 per cent 39 per cent cotton Ticket number 120 Tex 25 Tenacity (cN/tex) 39 Elongation (%) at break 20

200

Composition

Tension (gf)

300

Needle tensioner adjustment

Figure 3.

Table IV. Sewing Thread Properties

350 1 300 3

250 Tension (gf)

were produced for each peak tension in each experiment, illustrating the influence of the parameter studied on the peak tensions under analysis. The curves are presented in Figures 3-8, for experiments 1-6 respectively.

200

4 6 5

150 100

2 50

78

74

70

66

62

58

54

50

46

42

38

34

30

26

22

18

14

10

0 Bobbin tensioner adjustment

Discussion of Results Analysing the tension variation traces obtained during a stitch cycle during the various experiments, it was found that no significant variations occurred as far as the timing and shape of the peak tensions referred to before were concerned. Variations on the shape of the peak tensions were noticed only in experiment 6, where thread C was analysed. It was found that the timing of the peak tensions remained unchanged, but the slope of both the increasing and decreasing tension curves was greater. Significantly, thread C had a much lower extensibility than the other threads, which clearly must be significant.

Figure 4.

450

1

400 350 4 3

Tension (gf)

300 250

2

200 150

5 6

100 50

Sewing speed

Figure 5.

17

4,500

4,350

4,200

4,050

3,900

3,750

3,600

3,450

3,300

3,150

3,000

2,850

2,700

2,550

2,400

2,250

2,100

1,950

1,800

1,650

1,500

0

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

450

1

400 350

Tension (gf)

300

4 3

250

●

200 150

56

100

2

50

400

390

380

370

360

350

340

330

320

310

300

290

280

270

260

250

240

230

220

210

200

0

Tensioner adjustment Number of plies =2 =3 =4

●

Figure 6. ● 1

450 400 350

4 3

Tension (gf)

300 250 200

5

6

150 2

100

6

50

400

390

380

370

360

350

340

330

320

310

300

290

280

270

260

250

240

230

220

210

200

0

Tensioner adjustment A B C

●

Figure 7.

500

1

400

3

Tension (gf)

4 300

200 562 6

100

400

390

380

370

360

350

340

330

320

310

300

290

280

270

260

250

240

230

220

210

200

0

Tensioner adjustment A B C

of the tension applied on the bobbin thread by the tensioner, no significant influence was noticeable on the tensions generated on needle thread. The peak tension variations obtained in the experiment where the effect of the sewing speed on the needle and bobbin thread tension was studied showed that the influence of the sewing speed on the peak tensions is not significant. The number of plies does not affect significantly the peak tensions generated on either the needle or the bobbin threads during the lockstitch formation. In the experiment made with different fabric qualities, no significant differences were found between the peak tensions generated for the different fabric qualities. However, although the magnitude of the peak tensions generated and the trends detected for each fabric quality were very similar, a more detailed analysis of the results obtained suggested that different equations should be used for each distinctive fabric quality, if peak tensions are to be accurately predicted. In the study of the effect of different sewing thread qualities on the tensions generated on the needle and bobbin thread, it was found that distinctive trends were obtained for each sewing thread quality. Also, it was found that the thread tension traces obtained with the 100 per cent cotton sewing thread presented a different timing, i.e. during the sewing cycle the different peak tensions started later and finished earlier but the maximum tension on each peak occurred at the same timing as the other sewing thread peak tensions. It was also noticeable that Peaks 1 and 4 for the experiment made with the 100 per cent polyester sewing thread exhibited very large variations.

It was concluded that different sewing thread qualities result in different trends and, for that reason, distinctive regression equations should be used for each sewing thread quality.

Figure 8.

affect the peak tensions on the needle thread, namely Peak 1, Peak 3 and Peak 4, and that Peak 1 follows very closely the variation on the tension applied on the needle thread by the tensioner. This merely confirms what has been reported earlier[1-7] and is already well known. ● While both the bobbin thread peak tensions increased proportionally with the increment

Conclusion These experiments and their results showed that it is possible to measure and predict with good accuracy the influence of different factors on the tensions generated on both the needle and the bobbin thread. In this way, the significant

18

VOLUME 6 NUMBER 1 1994

parameters that influence the peak tensions can be selected and general regression equations relating these significant parameters with the tension generated on the peak tensions can be derived. This is important in the development of algorithms to define the limits of tension for the production of high quality seams. This aspect will be considered in more detail in Part II.

4. Kamata, Y., Kinoshita, R., Ishikawa, S. and Fujisaki, K., “Disengagement of Needle Thread from Rotating Hook: Effects of Its Timing on Tightening Tension on an Industrial Single Needle Lockstitch Machine”, Journal of the Textile Machinery Society of Japan, Vol. 30 No. 2, 1984, pp. T40-9. 5. Kamata, Y., Sakai, T., Onoue, M. and Chatani, Y., “Analysis of Tightening Tension Waves in Single Needle Lockstitch Machine”, Journal of the Textile Machinery Society of Japan, Vol. 39 No. 1, 1986, pp. T7-15 and Vol. 39 No. 6, 1986, pp. T86-96. 6. Matsubara, T. and Jinbo, Y., “Analysis Approach for Stitch Construction and Stitch Tightening of Lockstitch Sewing Machine”, Journal of the Society of Fibre Science and Technology, Japan, Vol. 10, 1984, pp. T387-94. 7. Onoue, M., “Influences of the Sewing Conditions of the Lockstitch Sewing Machine for Industrial Use on the Needle Thread Tension”, Journal of the Society of Fibre Science and Technology, Japan, Vol. 10, 1984, pp. T395-401. 8. Ferreira, F.B.N., “A Study of Thread Tensions on a Lockstitch Sewing Machine”, PhD Thesis, University of Leeds, 1991.

n

References 1. Deery, W.A. and Chamberlain, N.H., “A Study of Thread Tension Variation during the Work Cycle in a Lockstitch Sewing Machine”, Technical Report No. 15, The Clothing Institute, 1964. 2. Greenberg, N.G., “An Instrument for Measurement of Thread Dynamic Tension Characteristics during the Sewing Operation”, Clothing Research Journal, Vol. 3 No. 2, 1975, pp. 77-84. 3. Horino, T., Miura, Y., Ando, Y. and Sakamoto, K., “Simultaneous Measurements of Needle Thread Tension and Check Spring Motion of Lockstitch Sewing Machine for Industrial Use”, Journal of the Textile Machinery Society of Japan, Vol. 28 No. 2, 1982, pp. 30-7.

19

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Productivity and Production in the Apparel Industry Shu-Hwa Lin Consumer Affairs Department, Auburn University, Alabama, USA, Doris H. Kincade Clothing and Textiles Department, Virginia Polytechnic Institute and State University, Blacksburg, Virginia, USA and Carol Warfield Consumer Affairs Department, Auburn University, Alabama, USA Received 6 June 1993 Accepted 10 December 1993

The outward signs of a mature domestic industry have been highly published and include the issues of reduced productivity, failure to reinvest profits for capital improvements, an alleged lack of dedication of workers, and investor’s focus on short-term profits[3]. For the apparel industry and many other industries, these issues are compounded by the pressures of competition from overseas manufacturers[4,5] and changes in the purchasing patterns of the domestic consumers.

Apparel marketing trends dictate shorter lead times, smaller quantities, and many style changes[6]. Apparel manufacturers are searching for ways to solve the production and costing problems encountered when addressing these marketing changes. For the apparel industry, one approach to increase productivity is the selection of an appropriate sewing system[7]. For many manufacturing firms, the use of appropriate technology can improve productivity and lower the associated production costs[5]. For a producer with high labour costs, the choice can be the use of low-cost labour or the selection of a sewing system which can make production more cost effective[7,8]. Apparel firms should evaluate alternative sewing systems as one component of the effort to change inefficient and high-cost operations into efficient, lower-cost operations. Apparel producers should use: (1) marketing skills to identify products which will sell well; (2) speed of production to deliver goods on time with a minimum of products in process; and (3) sewing systems to produce quality products at lower cost[9]. Effective sewing systems, combined with new plant layout and new sewing-machines (e.g. computer-controlled or robot-set) should give some apparel manufacturers a significant

International Journal of Clothing Science and Technology, Vol. 6 No. 1, 1994, pp. 20-27, © MCB University Press, 0955-6222

This research was supported in part by the Alabama Department of Economic and Community Affairs.

Introduction Until the mid-nineteenth century, clothing was either home-made or tailor-made. Tailor-made clothing was affordable only by a few people. With the advent of the sewing-machine in 1851, the apparel industry began mass-producing clothing, making manufactured clothes available to more people[1]. Even today apparel production remains one of the most difficult and labourintensive of all factory operations. Sewing has been difficult to mechanize, and productivity improvement has been slow[2]. For this reason, the measurement of productivity is a major concern for apparel manufacturers.

Background

20

VOLUME 6 NUMBER 1 1994

advantage in many market areas now dominated by producers with lower-cost labour[2]. Although many definitions of sewing systems have been provided in the past 40 years, the important point is that the best system is that which will enable the apparel producer to meet consumer demands. Apparel producers should use the sewing system which will provide the style changes, lead times, and quality and quantity levels required most efficiently and effectively to meet the requirements of their customers[8].

productivity. The introduction of dedicated equipment in an attempt to improve output can complicate the situation. A change in the length of a collar point or the shape of a pocket can mean several hours of downtime to reset a specialized machine. Style variations produce changes in the degree of operator skill required. The AAMA reported that style variations, without changes in production methods, were inversely related to volume of production[14]. To illustrate this problem, the following example is given. Production line A produces 10,000 dozen of one style compared with production line B which produces 10,000 dozen of four styles. Production line A is mass production and line B is more fashion-oriented. Production line A will generally have higher productivity than production line B. Sewingmachine operators on production line A will spend less time, on average, to produce the 10,000 dozen than will the production line B sewing-machine operators. The more product produced per style, the more time is saved. If the two production lines use the same methods and equipment, the line with the higher number of style changes will have reduced productivity.

Productivity To evaluate the sewing system, one can use the measure of productivity. Productivity is a measure of work with origins tied to the development of scientific management[3]. Definitions of productivity are numerous, but in the most simple terms, productivity is measurement of output relative to an input[3,5]. The output for a production facility is often measured in units produced, and the input is some variation of numbers of workers or hours worked[10,11]. This definition of productivity is applicable for apparel production, because the apparel industry has been described as the most labour-intensive of all industries[12].

Improvement of Productivity The challenge for an apparel production manager is to minimize input for an improved output while remaining flexible and meeting customer demands. A reduced ratio of number of workers to number of machines can achieve labour savings and increased productivity. A survey by the Institute of Industrial Engineers reported that 38 per cent of the respondents observed increased productivity as a result of the development of indirect labour standards and controls[15]. Systems innovations resulted in a 37 per cent improvement in productivity, while use of robotics accounted for a 29 per cent improvement in productivity. One can achieve increased productivity for human activity through multitask operations (i.e. cross-training operators) and other methods of doing the same amount of work in less time[16]. New sewing systems (i.e. modular manufacturing systems) should meet these requirements[17]. In modular manufacturing, the need for labour (i.e. number of workers) is reduced with a corresponding rise in productivity and savings[2,10].

n

Changes in production line characteristics can reduce productivity and increase costs n In a production facility, the variables in the productivity ratio and its outcome can be affected by a number of factors (e.g. management, raw materials, equipment, and worker dedication)[5]. Traditionally, changes in product line characteristics (i.e., increased style variation and reduced production volume) can reduce productivity and increase costs. A 1979 survey indicated that lower productivity was the biggest problem apparel manufacturers had when producing a wide variety of types of products in small quantities[13]. The American Apparel Manufacturers Association (AAMA) reported that constant style change has become a reality for apparel manufacturers[14]. Even small changes (e.g. a different fabric, additional trim items, or a colour variation) force the production line to face a new set of factors. Each change affects costs and

Research Study This exploratory study of the apparel industry was designed to examine the relationship of product

21

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

product line) were derived from the questionnaire data. Production volume was defined as dozens of units per season per style. It was calculated as follows:

line characteristics and sewing systems to calculated productivity. Alabama apparel producers were chosen as the sample for the study for the following reasons: ● Alabama was one of only two states to increase in apparel industry employment between 1978 and 1988[18]; ● Alabama has textile or apparel plants in every county in the state[19]; and ● the textile and apparel industry is the largest employer in the state accounting for approximately one-fifth of the manufacturing employment[20]. A mail questionnaire was sent to the 447 Alabama apparel producers listed in the Directory of Mining and Manufacturing, 19891990[19]. Questions requested information about these areas: machines and employees; products and production volume; and sewing systems/factory layout. The questionnaire was pilot-tested with a small sample of apparel manufacturers and was refined before mailing. A modified Dillman Total Design System[21] was used to increase the response rate. A follow-up questionnaire and phone calls were used to encourage additional responses and to collect information about non-respondents. Ninety-six apparel plant managers completed the questionnaire giving a 39 per cent adjusted response rate. The most common reason for nonresponse was lack of time. The non-respondents were similar in size of operations and type of product to the respondents.

Production volume = Monthly production × 3(months/season) (No. style changes + 1) / season Productivity was defined as production volume (dozens of units per season per style) per operator. It was calculated as follows: Productivity = Monthly production × 3 (No. style changes + 1/season) × No. operators Although these measures concentrate on specific factors and take “licence” in assuming other factors stay equal, they are appropriate and typical measures for use in comparisons[5]. Monthly production, production volume and productivity were reported in dozens of units per season per style. Rosner[22] pointed out that using percentiles has the advantage of limited influence from the sample size. Product line types, shown in Table I, were developed from the categories described by Johnson-Hill[23] and the frequency of style change dimension derived from the AAMA[14] special report on manufacturing. The four seasons per year (i.e. 13 weeks per season) was selected as the standard definition of season[14]. Comparisons of style changes must always be tempered by the fact that the definition of season and of a style change may vary among apparel producers[14,24]. Frequency of style change was obtained from the question, “How many times per season do you change style in the production line?” The responses, which ranged from “never” to “daily”, were grouped into the categories 0-1, 2-3, 4-6, and over six style changes per season.

Data Analysis Data were analysed with descriptive statistics, and phi-coefficients were used for testing comparisons. The variable (i.e. sewing system used) was gathered directly from the questionnaire. The three other variables (i.e. production volume, productivity, and type of

Type of product line

Type of style change (frequency)

Volume

Staple Semi-staple Fashion High-fashion

Basic Variable Variable Very variable

Mass production Mid-volume production Low production Special production

0-1 changes per season 2-3 changes per season 4-6 changes per season Over 6 changes per season

Table I. Type of Product Line

22

VOLUME 6 NUMBER 1 1994

Findings and Discussion

size of operation as described by the number of operators was 76-100 with 13 responses. The second most frequent size of operation was 101-150, and 201-300 operators, with 12 firms reporting each of these sizes. The minimum size was two sewing-machine operators and the maximum size was 3,600 sewing-machine operators. In this study production volume was defined by dozens of units per season per style. Production volume is summarized in Table II by reporting the minimum and maximum monthly production volume as well as the maximum volume for each dozen of units. The reported minimum monthly production volume was 33 dozen and the maximum monthly production volume was 350,000 dozen. Productivity was defined as production volume (i.e. number of units) per operator. The specific measure of productivity used in this study was the output (i.e. the number of apparel units in dozens produced) divided by the input (i.e. the number of styles per season and the number of operators). Data are summarized in the same manner as for monthly production (see Table III). For the company in the study with the highest productivity, each operator produced 94,286 dozen garments per style per season.

Size of the Operation The size of an apparel production operation can be described in several ways including the number of sewing-machines, the number of operators, and the volume of production. The most frequent size of apparel operations among respondents was 201-300 sewing-machines, represented by 19 firms (20.2 per cent of the sample). The second most frequent size of apparel operation was 101-150 sewing-machines with 18 firms (19.1 per cent) reporting this number. The smallest number of machines reported was one, while the largest was 6,750. Nationally, in 1982, 40 per cent of the apparel plants had fewer than ten employees and 87 per cent of the plants employed less than 100 workers[25]. In this study of Alabama apparel firms, 48 per cent employed fewer than 100 workers and 90 per cent employed fewer than 500 workers. These data indicate that the responding Alabama firms were somewhat larger than the 1982 national average[20]. The most frequent

Dozens of unita 0-1,000 1,000-5,000 5,000-10,000 over 10,000

Number of plants

Per cent

14 20 13 39

16.3 23.3 15.1 45.3

Type of Product Type of product can be described by merchandise classifications, type of garments, and type of product line. The products produced by the responding companies included men’s, women’s, children’s, and other types of garments. For the majority of respondents, the companies produced a mixed product line. In fact, many plants produced two or more of the merchandise categories. Type of garment included shirts, jeans,

Frequency missing = 8 volume is defined by dozens of units per season per style

aProduction

Table II. Production Volume

Productivityb (doz/units)

Cumulative frequency

Cumulative per cent

Frequencya

Per cent

0-50

42

48.8

42

48.8

51-100

13

15.1

55

64.0

101-150

5

5.8

60

69.8

151-200

4

4.7

64

74.4

Over 200

22

25.6

86

100.0

aFrequency

missing = 8 was defined as dozens units per style per season per operator

bProductivity

Table III. Productivity

23

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Number of changes 0-1 2-3 4-6 Over 6

Frequency(n)

Per cent

22 20 22 27

24.2 22.0 24.2 29.7

extreme low and high numbers of quality inspectors could be reflections of the use of the modular and TSS sewing systems. The following sewing systems were being used: bundle, progressive bundle, modular, TSS, and others (i.e. multiple systems). Approximately 77 per cent of the respondents used either the conventional bundle or the progressive bundle sewing system. Some apparel producers (5.2 per cent) used two or more sewing systems. Firms using more than one sewing system reported the following combinations: bundle-modular and progressive bundle, modular and TSS. Overall, 85.4 per cent of the respondents indicated at least moderate satisfaction with the sewing system or systems they were using. Only one-quarter of the respondents had plans to change sewing systems. Those managers who were planning to change sewing systems indicated that they were going to adopt modular, UPS, or TSS. The most common reason to change sewing systems was for increased efficiency.

Table IV. Frequency of Style Changes per Season

pants, skirts, socks, underwear, uniforms, T-shirts and sweats. The type of product line reported was evenly distributed among staple, semi-staple, fashion, and high-fashion production. In this study, 30.3 per cent of the respondents reported more than six style changes per season (Table IV).

Sewing Systems Sewing systems are a blend of workers, machines, and handling devices[17,26]. Characteristics of such systems involve the number of sewingmachine operators, number of sewing-machines, number of assistant operators, ratio of sewingmachines to sewing-machine operators and number of quality inspectors. The characteristics may also be examined in total to include the overall type of system used for production. Thirty-five apparel producers (40.7 per cent) used 26-50 sewing-machine operators in the production line. The second most frequent size of production line (15 plants or 17.4 per cent of the respondents) was one to ten sewing-machine operators in the production line. The reported range was one to 952 operators in a production line. The following sizes of production lines were reported: one to ten machines (15.3 per cent); 1125 (16.5 per cent); 26-50 (25.9 per cent); and 5175 (16.5 per cent). Overall, 74 per cent of the respondents reported from one to 75 sewingmachines in the production line. In this study, one was the smallest number of machines in a production line and 2,000 the largest number. All but two of the plants had at least one quality inspector. Over 70 per cent of the respondents (65 firms), reported one to ten quality inspectors in their plants. These numbers may be misleading, however, for in some sewing systems (i.e. the modular and Toyota Sewing System (TSS) production lines), every operator is responsible for quality. For this reason, plants using TSS and modular sewing systems may not need specially designated quality inspectors or may report all operators as quality inspectors. The

Production Volume and Productivity Production volume and productivity are measures of efficiency for apparel manufacturers and are concerns when selecting new sewing systems. The relationship between production volume and productivity, found from this study, is shown in Table V. This relationship was statistically significant when tested with phi-coefficients (phicoefficients = 0.819, p = 0.000). The data in Table V indicate that firms with a higher production volume tend to have a higher level of productivity. Conversely, firms which fell within the lower categories of production volume tended to fall in the lower categories of productivity as well. This finding supports the theory of AAMA[14] that lower production volume tends to correspond with lower productivity.

Style Change and Productivity The ability to handle frequent style changes in the production line is an indication of flexibility, a condition necessary in an increasingly competitive environment. Traditionally, style variations cost more to make because of the increased labour component[14]. Table VI illustrates the relationship between the frequency of style change and productivity in a two-way frequency distribution. This relationship is statistically significant (phi-coefficient = 0.752, p = 0.000).

24

VOLUME 6 NUMBER 1 1994

Productivity Production volume

0-50

51-100

101-150

151-200

Over 200

Total

0-1,000

14

0

0

0

0

14

1,000-5,000

17

1

1

1

0

20

5,000-10,000

5

2

3

1

2

13

Over 10,000

6

10

1

2

20

39

42

13

5

4

22

86

Total Frequency missing = 8

Table V. Frequency of Production Volume by Productivity Rates

Productivitya Production volume

0-50

51-100

101-150

151-200

Over 200

Total

Staple

1

2

2

2

14

21

Semi-staple

5

6

1

2

5

19

Fashion

15

4

1

0

2

22

High fashion

22

1

1

0

2

26

Total

43

13

5

4

23

88

Frequency missing = 8 aProductivity was defined by production volume per operator

Table VI. Style Changes by Productivity (per Season)

a way to increase productivity, especially when faced with the demands of a fast changing marketplace. For this study, the relationship between sewing system and productivity is statistically significant (phi-coefficient = 0.93, p = 0.04) (Table VII). The number of plants operating with the “new systems” such as modular and TSS, are very limited, and many of these also reported continued use of traditional bundle and progressive bundle systems. The bundle system represents 50-60 per cent of the least, as well as the most, productive plants in the study. A qualitative investigation of these manufacturers indicates that those with bundle systems and high productivity are producing a variety of product lines. Similarly, the “other” sewing systems in use are concentrated in both the most and the least productive of the firms in this study and contain a mix of product lines. Further study is needed to clarify these findings.

Plants with basic product lines (i.e. less frequent style changes) tend to have higher productivity, and the converse is also true. Plants producing high fashion items (i.e. over six style changes per season) were concentrated in the lowest two levels of productivity, while plants with staple production (less than one style change per season) were concentrated in the highest condyle of productivity. The less productive plants, as measured in dozens of units per season per operator, tended to be plants that produced more fashion-oriented products. These findings, which are consistent with previous studies[13], show the need to find ways to increase productivity while responding to consumer demands for frequent style changes. Sewing System and Productivity The introduction of a new sewing system is reported by trade literature and research studies as

25

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Summary

using the system and the training which may have been provided for workers. Of the managers who indicated that they were considering change, many indicated that they would select modular or TSS. With the preliminary findings from this study about the productivity rates of traditional systems, a manager should evaluate carefully the factors which may accompany a change in systems. Without more detailed analysis, managers should not change just to get a new system. The measures of productivity and attempts for improvement have been investigated in textiles, but rarely considered in apparel[2]. The findings of this study suggest the need for more exploration of sewing systems, reasons for using a certain system, and possible advantages of changing to another system. Further study could be done by a combination of case studies and experimentation with active sewing production lines and different sewing systems. These sewing production lines could be tested and analysed for the different characteristics of sewing systems and for their application to specific apparel production products and plants. Consumers’ demands have been increasingly diversified and individualized, creating the need for apparel producers to be responsive to the rapidly growing individualization of consumers’ needs. These new demands for consumer responsiveness call for a shortened product life cycle and increased diversification of fashion[27]. This increased responsiveness requires that successful apparel producers have the capability to produce many different types of products in small quantities in a shorter lead time.

Apparel manufacturers must be able to address rapidly changing consumer needs. The consumer focus for apparel manufacturers requires a shortened product life cycle and increased diversification of fashion. The apparel industry in the changing market must obtain the capability to produce many different types of products in small quantities in shorter and shorter lead times[27]. Previous authors have emphasized the importance of the right sewing system[7,8,14] for achievement of high productivity rates; however, the findings of this research agree with Losman and Liang[5] that the actual technology and systems of production are only one factor in developing high productivity. Product type was highly related to productivity. Sewing systems in this research are not clearly related to productivity level. Some manufacturers have managed to achieve high productivity with bundle systems. The sample type and size of this research limits the generalizability of the findings, but the indications of significance are important. The interaction between sewing system type, production level, style changes and specific product mix should be further investigated. Other factors should be tested for their impact on productivity. Beyond the number of style changes, one should also investigate merchandise classifications and product mixes. The human factor of the workforce should be considered. Education and motivation of the workforce as well as the worker’s orientation to the specific sewing system may have an effect. New systems have start-up time. Additional research on system type should consider how long firms have been

Productivity System

0-50

51-100

101-150

151-200

Over 200

Total

Bundle

22

8

2

3

14

49

Progressive bundle

16

3

2

0

4

25

Modulara

1

0

0

0

0

1

Bund, mod

2

2

0

0

0

4

Progressive bundle, mod, TSS

0

0

1

0

0

1

Other

2

0

0

1

4

7

Total

43

13

5

4

22

87

Frequency missing = 9 = all modular type sewing systems and all multiple systems (e.g. modular with bundle)

aModular

Table VII. System by Productivity

26

VOLUME 6 NUMBER 1 1994

14. American Apparel Manufacturers Association, Fashion Apparel Manufacturing, Arlington, VA, 1982. 15. Institute of Industrial Engineers, “Productivity Improvements – How to Make Them”, Bobbin, Vol. 25 No. 5, 1984, p. 58. 16. Groover, M.P. and Hughes, J.E., “Job Shop Automation Strategy Can Add Efficiency to Small Operation Flexibility”, Industrial Engineering, Vol. 13 No. 11, November 1991, pp. 67-76. 17. Cooklin, G., Introduction to Clothing Manufacture, BSP Professional Books, London, 1991, pp. 137-47. 18. American Textile Manufacturers Institute, Textile Hi-Lights, Washington, DC, March 1990. 19. Alabama Development Office, Directory of Mining and Manufacturing, Montgomery, AL, 1989-1990. 20. US Department of Commerce, Bureau of the Census, 1987 Census of Manufactures. Geographic Area Series. Alabama, US Government Printing Office, Washington, DC, 1987. 21. Dillman, D.A., Mail and Telephone Surveys: The Total Design Method, Wiley, New York, NY, 1978. 22. Rosner, B., Fundamentals of Biostatistics, PWS-Kent Publishing Company, Boston, MA, 1990. 23. Johnson-Hill, B., Fashion Your Future, Kurt Salmon Associates, The Clothing Institute, London, 1978. 24. Lin, S.H., Technique of Apparel Making, Ministry of Economic Affairs of ROC, Taipei, Taiwan, 1984, pp. 3-4. 25. US Department of Commerce, Bureau of Labor Statistics, 1982 Census of Manufactures. General Summary, Part 2, Industry Statistics by Employment Size of Establishment, Subject Series, MC82-S-1 (part 2), US Government Printing Office, Washington, DC, December 1985. 26. Schonberger, R.J., Japanese Manufacturing Techniques: Nine Hidden Lessons in Simplicity, The Free Press, New York, NY, 1982. 27. Ebihara, K., “Developing New Technology from an Apparel Equipment Supplier Point of View”, JSN International, December 1985, pp. 22-9.

Producers must be able to see the advantages of using various sewing systems for their particular plant(s) to achieve this increased responsiveness.

n References 1. “The First Hundred Years: 1776-1876”, American Fabrics and Fashions, Vol. 106, Winter, 1976, pp. 61-8. 2. Cline, W., The Future of World Trade in Textiles and Apparel, Institute for International Economics, Washington, DC, 1987. 3. Siegel, I.H., Productivity Measurement in Organizations: Private Firms and Public Agencies, Pergamon Press, New York, NY, 1986, pp. 4-6. 4. Black, J. and Friedman, A., “New Growth Spots Emerge from Far East Fall-out”, Daily News Record, 10 December 1991, pp. 8-9. 5. Losman, D.L. and Liang, S.J., The Promise of American Industry: An Alternative Assessment of Problems and Prospects, Quorum Books, New York, NY, 1990, pp. 117-38. 6. “Stores, SA Say Price Beats Pride”, Women’s Wear Daily, 3 December 1991, pp. 1,4-5. 7. US Office of Technology Assessment, The United States Textile and Apparel Industry: A Revolution in Progress, Congress of the United States, Washington, DC, 1987. 8. Wise, W., “Flexibility: The Key to Competitiveness for United States Swimwear and Intimate Apparel Producers”, Apparel Manufacturer, Vol. 2 No. 2, 1990, pp. 46-52. 9. Adams, M. and Ziemke, M.C., “Strategic Management for the Small Apparel Manufacturer and Control”, Apparel Manufacturer, Vol. 2 No. 4, 1990, pp. 70-4. 10. Gilbert, C.S., “Modular Manufacturing: Sizzle or Steak?”, Apparel Manufacturer, Vol. 2 No. 3, 1990, pp. 44-52. 11. Thurow, L.C., Packer, A. and Samuels, H.J., Strengthening the Economy: Studies in Productivity, Center for Democratic Policy, Washington, DC, 1991, pp. 21-32. 12. Vought, K.D., “Total Productivity”, Apparel Manufacturer, Vol. 2, 1990, pp. 74-81. 13. Juki Industrial Co. Ltd, Juki Unit Synchro System, Tokyo, 1982.

27

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

COMMUNICATIONS

Fractal Geometry A New Tool for Textile Design Development Applications in Printing Jorge Neves, Manuela Neves and Katja Janssens Department of Textile Engineering, University of Minho, Portugal

means that the parts and the whole are similar and, therefore, if the scale is changed, the geometrical appearance remains constant. This “scale invariance” is a fractal characteristic and the measure of its irregularity is the fractional or fractal dimension of the structure. But, more than a mathematical curiosity, fractals, as already mentioned, can help us to improve the uniqueness of science and art. In fact, anyone who chances to turn the pages of a popular fractal book, cannot be indifferent to the beauty and fascination of its structures and images: flowers, masks, trees, cascades, landscapes, clouds, new worlds, islands, waves, explosions and a great variety of special effects like roughness or depth. Fractal geometry, as an approach to the formation of a combination of technological ways of thinking and aesthetic considerations, will thus be of particular significance to the textile industry, in particular for textile printing and weaving (Jacquard) applications. But these applications can be made only with the technical support support of the most advanced computer-aided design technologies. The generation of fractals can be done using adequate softwares. However, the knowledge of the way in which a fractal is made can enhance the appreciation of its physical beauty and help in understanding problems which occur when attempts to apply it to CAD programs are made.

Introduction Fashion requirements impose to the textile industry a quick response to new tendencies. Because traditional ways do not agree with the needs of modern commercial politics, the European industry can compete with undeveloped countries’ labour costs with quality, new designs and quick response. The development of CAD/CAM systems follows this strategy because the aim of these systems is the improvement of versatility, design, short production lead times and the avoidance of a large quantity of samples production. At the same time they allow textile designers extra time for creative work instead of repetitive actions. On the other hand, the application of fractals to the CAD printing systems not only is a consequence of some fractal images’ beauty but also has resulted because fractals are mathematical objects which describe nature well, no matter how you look at it. They are one of the best ways to combine beauty, maths, microprocessors and design. We must mention that fractal geometry is a source of inspiration only in design creation and it must not be forgotten that, when images are adjusted to the technical parameters of the printing mill, they lose their fractal characteristics, in the scientific sense of the word.

Fractal Geometry There are many kinds of objects and other structures with “self-similar” characteristics. This

This article was delivered as a paper at the 5th International Congress of Graphical-Industrial Design, 2-4 June 1993 at the University of Oviedo, Spain. The authors wish to acknowledge the help of COELIMA and ART in the practical development of this work.

International Journal of Clothing Science and Technology, Vol. 6 No. 1, 1994, pp. 28-36, © MCB University Press, 0955-6222

28

VOLUME 6 NUMBER 1 1994

Fractal category

Examples

Formulae

Notes

Escape (to infinite) time algorithm

Julia sets

Zn+1 = Z 2n + C

C is a constant (complex number) and Z0 the starting-point of iteration, is the point to be tested Points are coloured, if iteration is not attracted to infinite

Mandelbrot set

Zn+1 = Z 2n + C

Escape to finite attractor

Newton’s method

Zn+1 = Zn –

Chaotic orbits

Lorenz attractor

Iterated function systems

Sierpinski gasket

C is the complex number corresponding to the point to be tested and Z0 = 0

F(Zn) F' (Zn)

dx dt = – 10 (x – y) dy dt = – xz + rx – y dz dt = xy – 8 z 3 (2x,2y – 1) if y > 0.5 (2x – 1,2y) if x > 0.5 Zn+1 = and y < 0.5 (2x,2y) otherwise

{

Fractal ferns Fractal dragon Fractal trees

The points are coloured according to the iteration when the ”orbit” is captured by a root of the polynomial F(Z) = Z p – 1 When the parameter r lies in the interval 24.7 < r < 145, the solution does not converge to a fixed point in the limit t–>∞ nor is there a limit cycle. The solution keeps moving around in a finite region Julia set producing a Swiss cheese triangle

These fractals are defined by exactly specifying the relationship between itself and its self-similar parts

Table I. Some Fractal Formulae Applied to CAD Programs

Table I shows formulae and additional notes of some of the most important fractal categories used in this work. The application of CAD-systems in the textile industry started only at the beginning of the 1980s, mainly in the area of textile printing. Its importance became stronger owing to the textile industry’s need to integrate new technologies in its design process. A CAD system will not replace the designer, but its advantages are very important: reduction of the designer’s work, reduction of time to change and control designs, reduction of costs, perfection of the designs, availability of alternatives and usage of common parts in various designs, and libraries of patterns enabling fast copy and repetition. However, they have also some negative aspects: downtime owing to failure of the equipment, waiting time owing to use of the computer for another application, optical weariness (fatigue), high cost of the equipment and additional training. The use of CAD/CAM systems will be involved in all the steps of the printing design process as shown below:

● ● ● ● ●

Creation of a design. Adaptation to the technical printing parameters. Colour separation. Engraving of the printing cylinder. Printing.

This process starts with the creation of a design, but this creation needs to be adapted to the technical textile printing parameters. A printing mill has an economical production limitation of 28 different colours per pattern and each pattern has to be adjusted to the repetitive sequence, i.e. when the design is placed on a roll, the beginning and the end should flow over into each other. The size of the pattern should be equal or a submultiple of the printing cylinder diameter (for rotational printing). Using a CAD system, the accomplishment of this task will be much easier, faster and more accurate. All the colours of a pattern will have to be separated to make a film, because, for each colour of the pattern, a printing cylinder will be engraved. Without a CAD system each film had

29

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Working with the half-tone shades module, the number of colours can be increased to 128 colourshades. On the other hand, it is necessary to pay attention to the different ways of colour storage. Usually, as CAD systems have true colour files (each pixel is represented by one colour of the palette) and fractals-generating softwares use colour-mapped files (each design colour results from a mathematical calculation), a colour alteration results when fractals are introduced by file transformation.

to be copied by hand. Using a CAD system, colour separation can be done automatically and with higher accuracy. The separates can be printed out on a plotter or sent to a laser engraving unit (CAM). The practical elaboration of the design and engraving process is, however, far more complicated. A lot of people collaborate in the preparation of one pattern: a designer for the creation, a colourist for the adaptation of the design, a “misonetist” to make the separates, an engraver to make the printing cylinders and finally a printer to print the pattern on the fabric. Using a CAD system, this intensive work can be done by one person, in a remarkably short time.

Practical Developments Material (1) Hardware: ● Computer: IBM-compatible 486/33 MHz; 16 MB RAM; 3,000 Disk. ● Digital board: 8514/A-compatible, Super VGA Paradise; VGA; Prodesigner II; GPIB (scanner). ● Peripherals: Scanner: Canon CLC 10. ● Colour printer: Canon FP 510; Canon CLC 10.

Application of Fractal Geometry into a CAD-System The introduction of fractals into a CAD system can be done by: ●

creation of the design directly in the system;

●

introduction with a scanner;

●

reading fractals created in another software in compatible format;

●

(2) Software: ● Fractal generator: Freeware FRACTINT developed by the Wait Group. ● CAD system: Info Design Vision. Modules: Textile Printing – Plain Colour Shade, Textile Printing – Half-tone Shade; Jaquard; Knitwear. ● File transformation: PL convertion programme (Wait Stone Group) – Imagein conversion modular programme: Image-in Scan & Paint, Image-in Plus and Image-in Colour.

reading images captured by a video cameras.

Scanning is the most common way. The zone to be scanned will have to be defined as well as the resolution of the scan (limited by the scanner resolution). The scanned image will be saved as an SCN.SCA file. Introducing designs created in other softwares is possible, if these designs have a compatible file format. The following formats can be converted into the system’s format without any problems: ICB, LBM, TGA, TIFF, VCR, Mayer, Aloha, Scitex Floppy, Scitex Tape, Scitex Ct, Scitex 3D, Postscript, Postscript film, Pic, Arcom and a lot of others. The files to be converted have to be stored in the program file’s system, to be read immediately by the system. It should be noted that printing machines do not work with more than 16 printing cylinders. So, when reading a fractal with a high number of colours, although all the colours appear on the screen, the computer will activate only 28 colours, selected out of 16.8 million by default.

Fractal Adaptation to the Printing Parameters Colour Reduction and Introduction into the System The colours reduction (usually to values less than 16) will depend on the way in which the image is introduced into the system. When the fractal images are introduced in CAD systems and converted into SCAN.SCA files, they can be treated as if they were scanned. The designer has the liberty to choose which and how many colours will remain in the design. These

30

VOLUME 6 NUMBER 1 1994

resulting design will show a great number of planes of the same colour. After this first general reduction, a pattern can be submitted for further colour reductions, independent of the manner of design introduction: ● Superimposition: The final colour will be obtained by mixing different colours. The separating of this superimposed colour results from a computer addition of different areas of the constituent colours. ● Addition: During the adaptation of the design, some colours can be substituted with others already existing to improve the aesthetic appeal of the final pattern. ● Elimination: One colour can be eliminated by protecting all the other colours with a mask, defining the substituting colour as ground colour and making a virtual clear. The eliminated colour will then be substituted by the ground colour.

Scanning introduction (SCAN.SCA file)

Colour reduction

Unprintable details removal

Pattern repeat development

Scale adaptation

Coloration

Figure 1. Fractal Adaptation to Printed Patterns

Unprintable Details Removal The main characteristic of fractal images is the self-similarity of the pattern. This self-similarity results in patterns with a lot of detailed information, represented by differently coloured pixels, too small to be printed. They will have to be eliminated without losing the aesthetic appeal of the fractal creation. Depending on the pattern, different approaches are possible: ● Automatic cleaning replacing all the defined points by the largest percentile surrounding colour. This cleaning can be made from 1 to 49 points appearing in a line or area for all the colours of the pattern). ● Creating a mask defining the colours or the areas to be protected.

parameters have to be defined during the colour separation. When the fractals are introducted by file transformation, Figure 2 shows what has to be done. Conversion programs allow an automatic colours reduction to 8, 16 or 256 colours. When the fractal images are introduced with a number of colours lower than or equal to 16, no further colour reduction will be necessary and the design can immediately be adapted to printing parameters. But, if 256 colours are introduced, the computer will activate only 28 colours by default. The non-activated colours will have to be removed. This can be done, creating a mask of the activated colours and proceeding with a virtual clear. The non-activated colours will be substituted by the predefined ground colour. The

Adaptation of the Repeat Pattern This is another action which largely depends on the kind of pattern. Primarily, the kind of repeat will have to be defined: the step of the repeat, the direction – horizontal or vertical – and the size. Then it is necessary to obtain a repeat pattern. The following methods are mainly used to adapt the fractal creations to repeatable patterns: ● duplication with vertical or horizontal symmetry (this is the easier solution and sometimes the only one for changing a design in a repeatable pattern); ● copying parts of the design; ● drawing the new parts, to create continuity; ● cutting the pattern.

GIF (24 bits/pixel): colour-mapped file GIF (24 bits/pixel): RGB true colour file TGA (32 bits/pixel): true colour file SCA (32 bits/pixel): true colour file

Figure 2. File Transformation

31

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Plate 1. Complex Mark’s Julia – Escape Time Fractal. Formula: z2 c(p – 1) + c; Parameters: Real Part = 0.3; Imaginary Part = 0.6; Real Part of Degree = 1

Plate 4. Barnsley j2 – Escape Time Fractal. Formula: (z – 1)c, xzyc + xcyz > = 0; (z + 1)c,xzyc + xcyz < 0

Plate 2. Popcorn Julia – Escape Time Fractal. Formula: x(n + 1) = x(n) – 0.5 sin (n) + tan(3y(n)); y(n + 1) = y(n) + 0.5 sin x(n) + tan(3x(n)). Step Size = 0.5

Plate 5. Unity-Escape Time Fractal. Formula: One = x2 + y2; y(n + 1) = (2-One) x(n + 1) = 2-One y(n + 1)

Plate 3. Lambdasin – Escape Time Fractal. Formula: z(n + 1) = c fn z(n); first fn (z) = sin (z) Parameters: Real Part = 1; Imaginary Part = 0.4

Plate 6. Ducks – User-defined Fractal. Formula: z = Pixel; tst = p 1 + 4, t = 1 + pixel; z = sqr(2) + t;z < tst

32

VOLUME 6 NUMBER 1 1994

Plate 7. Complex Mark’s Julia – Duplication with Vertical and Horizontal Symmetry. Aquarelle Effect Repeated

Plate 8. Popcorn Julia (Step Size = 2)

Plate 10. Popcorn Julia (Step Size = 3)

Plate 9. Popcorn Julia (Step Size = 2) – Automatic Colours Reduction (16). Colours Addition. Repetition of the Pattern

Plate 11. Popcorn Julia (Step Size = 3) – Automatic Colours Reduction (16). Colours Addition

33

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Plate 12. Popcorn Julia (Step Size = 3) – Duplication with Vertical Symmetry Repetition of the Pattern

Plate 15. Barnsley j2 – Repetition of the Pattern with Drawing New Parts (Knitwear Simulation)

Plate 13. Popcorn Julia (Step Size = 0.5) – Repetition of the Pattern. Automatic Colours Reduction (16). Colours Addition. Cutting of Bands. Automatic Cleaning up to 20 Points

Plate 16. Barnsley j2 – Repetition of the Pattern with Drawing New Parts (Knitwear Simulation)

Plate 14. Barnsley j1 – Automated Colours Reduction (8)

Plate 17. Sierpinski Gasket. Colours Addition. Cutting of Bands. Duplication with Vertical Symmetry. Pattern Adjusted with Copy of Parts of the Design and Repeated

34

VOLUME 6 NUMBER 1 1994

Plate 18. Lambdesin – Colours Elimination (3). Pattern Band Repeated and Knitwear Simulation

Plate 21. Ducks – Automatic Colours Reduction (8). Repetition of the Pattern

Plate 19. Unity – Automatic Cleaning up to Ten Points. Coloration

Plate 22. Cesar (Peano Curve Variant). Duplication with Vertical and Horizontal Symmetry

Plate 20. Lambdasin – Zoom. Colours Elimination (12). Cutting of Bands. Arrangement and Repetition of the Pattern

Plate 23. Cesar. Repetition of the Pattern. Coloration

35

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Scale Adaptation, Coloration and Preparation of Films The size of the pattern has to be adapted to the size of the printing cylinder. Scale adaptation can be done automatically, with the option to reduce in the XX and XY parameters equal or different amounts. After the separation of the colours, the separated colour files will be sent to a plotter to print the films of each colour. The films will be used to engrave the printing cylinders.

instead of in a scientific way, can result in fashion trends inspired by fractal geometry.

n

Bibliography Bainsley, M., Fractals Everywhere, Academic Press, London, 1988. Bunde, A. and Havlin, S., Fractals and Disordered Systems, Springer-Verlag, Berlin, 1991. Internal information of INFO DESIGN on the CADsystem. Mandelbrot, B., The Fractal Geometry of Nature, W. H. Freeman, New York, NY, 1983. Mandelbrot, B., An Eye for Fractals, AddisonWesley, Reading, MA, 1991. Prusinkiewcz, P. and Lindenmayer, A., The Algorithmic Beauty of Plants, Springer-Verlag, Berlin, 1990. Takayasu, H., Fractals in Physical Sciences, Manchester University Press, Manchester, 1990. Wegner, T., Fractal Creations, Waite Group Press, 1991.

Final Considerations Having established all the necessary modifications for applying fractal images into CAD systems, a world of other interesting applications is opened. Fractal geometry can be applied as a new source of inspiration for printing, knitting and Jaquard pattern creation. Owing to the positive reaction of the textile design departments contacted, it seems and is predictable that new CAD systems will integrate fractal geometry into its softwares, as a new indispensable creation tool for CAD intelligent systems development. The classification of fractal images by topics like nature, emotions, abstractions, feelings …

36

VOLUME 6 NUMBER 2/3 1994

Basic Testing Method for Designing Excellent Fabrics for Men’s Suits Masukuni Mori SUMI Research Laboratory

Chart 14 shows a collected snake line of different yarn count fabrics (fabric structures were the same). ● Graph 1 shows THV of 24 kinds of fabrics.

Introduction

●

Objective of the Present Test Generally, “excellent fabrics for men’s suits” may be defined as following points 1 to 3 below. Hereafter they will be called “fantastic fabrics”.

Analysis of Results

(1) Good appearance in colour and design and excellent handle (soft and sufficient fukurami).

(1) Testing samples were classified into seven factors and were compared with the THV values. It was found that there were only a few differences (1.2, i.e. 25 per cent) among respective samples. This implies that, if yarn quality, yarn count and fabric structure are the same, THV is not influenced by reasonable changes of other factors. (2) No. 111 sample containing cashmere has the highest THV and the highest fukurami and numeri. These data prove that THV depends mainly on material. The second highest THV sample was No. 110. This sample is believed to have elasticity because of slight damage owing to piece dyeing as opposed to highpressure package dyeing. (3) From this result I conclude high THV fabrics also have high EMT. These fabrics exhibit high “hygral expansion shrinkage”. On the other hand, D.C. Snowden performed fine experiments in the 1950s about a close connection of shrinkage to fabric structure, especially a relationship between e/p (weft density/warp density ratio) and total shrinkage (dimension change in the course of the weaving process to the final sponging procedure). Graph 3 shows such a close relationship. That is it shows that there is a mutual relationship between shrinkage, extension of the fabrics and THV. I believe

(2) Easy to tailor and well tailored. (3) Comfortable to wear and good durability for wearing. The objective of the present test is to determine how to produce ideal fabrics intentionally. I used yarn of pure wool (2/60s), available in the market, and I selected 2/2 twill which has an ordinary fabric structure. I produced 24 kinds of fabric as samples and evaluated them mainly by KES (see Table I). At first I selected No. 501 as a standard fabric. (This fabric was made at Zegna S.p.A. Italy and is composed of 100s quality.) Twenty-four kinds of samples were compared with this sample (No. 501). The test results are shown as follows: ●

Charts 1 to 6 show snake charts of respective functional groups.

●

Charts 7 to 13 show comparisons of snake lines of Nos 102, 110, 111, 301, 401 with that of the standard (No. 501).

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 7-10, © MCB University Press, 0955-6222

7

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Factor controlling high THV

Level

Quality of raw wool Yarn count Twist Dyeing method of yarn

5 4 4 3

Fabric structure (10~45)

Fabric structure Density

5 3

Finishing (10~30)

Finishing process

4

Material (50~60)

Remarks Commercially available yarn (2/48,2/52,2/72s)

Note: Values in parentheses show the degree of contribution Table I. List of the Factors for Analysis of High Total Handle Value (THV) Element

that this is the key factor in producing socalled “excellent fabrics”. I think that we must look into the details of this subject.

Procedure for Evaluation of Trial Products (1) Twenty-four kinds of fabrics were experimentally produced based on the items listed in Table I. (2) Order of finishing processes and relevant processing conditions are shown in Table II. (3) Measurements were taken of 17 mechanical properties proposed by KES, based on the data: ● three primary hands koshi, numeri, fukurami; ● THV;

Reconsideration of Test Results and Future Test Schedule In the future I want to add to known characteristics of fabric samples with high THV, and to incorporate these samples’ data into the KES database. I hope to construct a system which enables us to select accurate data for producing excellent fabrics. Furthermore, I would add that it does not follow that good suits are tailored only from high THV fabrics.

Step Continuous crabbing (90˚ C × 60 m/min) ↓ Washing (40˚ C × 120 min) ↓ anion 0.5% nonion 0.5% Milling (20 min) ↓ Washing (40˚ C × 120 min) ↓ Continuous crabbing (90˚ C × 60m/min) ↓ Drying (120 ˚C × 50 m/min) ↓ Shearing (only Face, 3 times) ↓ Full dicator (115 ˚C × 6 min) ↓ Shrunk (90 ˚C × 20 m/min) ↓ Semi-decatizing (100˚ C × 4 min) Table II. Order of Finishing Processes

8

205

202B

202C

202D

" ↓ "

"

"

"

"

"

"

"

50

70

" "

"

"

↓ ↓ "

" "

"

"

" "

"

"

-

" "

"

"

-

" -

"

"

-

" "

"

"

"

" "

"

"

VOLUME 6 NUMBER 2/3 1994

(5) “Excellent fabrics” can be manufactured when raw wool, fabric structure and finishing method are matched well. The aim of this article is to show that high THV fabrics were made of common raw wool.

estimation of functions from the standpoint of mechanical properties; ● TAV; ● preparation of snake chart. (4) Calculation of tailoring and formability and plotting on a chart; (5) Judgement by hand; (6) Total evaluation. ●

Analysis of Results Aim of the Test We have already tested the following. (1) The effect of balance of densities of warp and weft on KES properties and dimension stability, in the case of fabrics which have higher density of warp than weft, such as gabardine and tuxedo of 100 per cent wool. (2) The effect of a percentage of mixed fibres and density on KES properties, especially warp elongation, in the case of fabrics blended with polyester and wool. (3) The effect of textile weave on KES properties and dimension stability. In this article, we examined the effects of twist, density and material of weft yarn or textile weave on appearance, physical properties, dimension stability, etc. We hope the results will prove to be useful for designing excellent fabrics.

Discussion (1) We collected excellent fabric samples for men’s suit which have THV of 4~5. Their mechanical properties were measured, and the values of tailoring and formability were calculated. The aim of this test was to search for design procedure of fabrics which have the properties comparable with those of excellent samples. It seems that there is a relationship between the fabric standard (such as characteristics of raw wool, fabric structure, finishing method, etc.) and high THV. (This THV is calculated from mechanical properties.) In Table III, “standard” means the degree of contribution attained when the conditions of three items are common and independent. “Range of modification” means the degree of contribution attained when the condition of each item is modified as much as possible. For example, when fabric structure is modified, the degree of contribution to THV becomes 10~45 per cent. Therefore THV can be changed in a considerably wide range. (2) The aim of this test is to investigate the change of THV when the structure and density are changed. (3) Snowden showed experimentally that fabric shrinkage is changed by the change of e/p. We examined the correlation of e/p with THV. (4) Fabric structure and THV (including existing data) were examined.

Analysis Based on Items (1) Effect of twist (101~104). It is difficult to judge by only one test, but we have established that, when the number of twist is 800, 670 or 520, there is little difference in RS, HE, EMT, G and 2HG5, but B, fukurami, THV and TAV of “520” are higher than those of the others. (2) Effect of warp density (105~107). When we wove fabrics whose warp density is 78, 72, 64, the weight was 358, 323, 300 g/m respectively. There was little difference in RS and HE but, when warp density was 64, EM2 and fukurami were the highest of all, and G, 2HG5, B and koshi were the lowest of all.

Degree of contribution (%) Range of modification Standard by altered condition

Item High THV

Characteristic of raw wool

40

30~40

Fabric structure

40

10~45

Finishing method

20

10~30

Table III. Degree of Contribution for Standard and Modified Items

9

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

(3) Difference in yarn-dyeing method (108~110). We compared cheese or hank dyeing with piece dyeing: ● RS, HE and EMT in piece dyeing were higher than those in the other dyeings. ● G and 2HG5 in piece dyeing were lower than those in the others. ● Koshi in piece dyeing was lower than that in the other dyeings. ● Fukurami, THV and TAV in piece dyeing were higher than those of the other dyeings. ● THV and TAV in hank dyeing were higher than those in cheese dyeing. (4) Effect of material (111~115). There was little difference in weight after finishing except for woollen fabrics. There was a remarkable difference between No. 111 and the others: ● Weft yarn of No. 111 was made of cashmere. ● HE and EMT in the warp direction were higher and B was lower than those of the others. ● Koshi was lower and fukurami and numeri were higher than those of the others. ● THV was 4.7, and TAV was 6.4. These were the best data in this test. (5) Effect of physical properties and handling with textile weave (201~205). The difference in textile weaves has the strongest effect on KES properties: ● 201 – 2/1 twill; ● 202 – 2/2 twill; ● 203 – 2/2 twill with cotton stripe; ● 204 – herring-bone twill; ● 205 – plain. (6) There was no remarkable difference in RS, HE, EMT, G and 2HG5 among five kinds of fabric: ● B of “205” is low; ● Fukurami, numeri and koshi are very low; ● THV is 2.3 and TAV is 3.0. The value of THV for plain weave was low according to the equation for winter suits. THV of “205” was calculated to be 3.5 by the equation for summer suits.

General Conclusion (1) The range of THV and TAV is as follows: ● THV 3.3 ~ 4.5 (for only plain weave, 1.2); ● TAV 4.0 ~ 5.7 (for only plain weave, 1.7). When weft yarn of fabrics is 2/60s, and weave is 2/2 twill, and the finishing process is the same, then the range of THV and TAV is very narrow, i.e. 1.2 and 1.7 respectively. (2) When we used cashmere as weft yarn, fukurami, numeri, THV and TAV all became maximum. (3) Weakly twisted, piece-dyed, 2/2 twill and herring-bone twill fabrics have high THV and TAV values. (4) Plain weave has low THV and TAV values for winter suits.

Test Plans for the Future We had fruitful results, namely that we could reconfirm a way of designing “excellent fabrics” through this experiment: (1) It is desirable, that when we clarify the properties of fabrics according to purpose, we can get information on material, weave, standard and finishing method. It is necessary to construct a system which can be used based on various data, e.g. 2/60s, 2/48s, 1/30s, etc. (2) It is necessary to construct a more definite system using KES data for commercially available samples, which is equivalent to CCS (Computer Colour Searching) in CCM (Computer Colour Matching). (3) In order to increase the reliability of the system, it is necessary to incorporate the results of Snowden and the proper standards of Brierley. (4) It is necessary to examine the conditions for not only 100 per cent woollen fabrics but also polyester-mixed spinning fabrics for working wear and cotton or linen for casual wear.

10

VOLUME 6 NUMBER 2/3 1994

A Study of Factors Affecting Fabric Cover-shelter Properties Shi Meiwu, Lai Kan, Yao Mu and Zhang Yan North-west Institute of Textile Science & Technology, Xian City, People’s Republic of China

which the object can be discriminated by eye without fabric are proposed as follows: (1) The field angle θ of the detail of the object to the eyes is not less than the least distinguishable angle θcr of the eyes, i.e. θ > θcr. (2) The luminance eo of the bright part of the object is not less than the critical (minimum) luminance ecr which can be sensed by the eyes, i.e. eo > ecr. (3) The contrast grade K between the bright part and the dim part of the object is not less than the critical (minimum) grade Kcr which can be resolved by eye, i.e. K > Kcr. The detail of the object will be distinguishable when the above three conditions are satisfied; and (4) The luminance eo of the light from the bright part of the object to the eye which is passed through the fabric is greater than the luminance ef of the light reflected by the yarn

Introduction Fabric has its own cover-shelter properties according to its uses, for example, the fabrics used for curtains which always possess good coverage, or the fabrics used for gauze that possess good visibility. One of the methods of obtaining the above requirements is to arrange the luminance and relative distance adequately. But how to enhance the process for this purpose using the structure and property of fabrics has not been reported. This article, based on [1], analyses the impact of structure and optical property of fabrics on cover-shelter properties.

Theoretical Analysis The clarity of an observed object will decrease when a fabric is placed between the observer and the object (Figure 1). That is to say the least distinguishable distance Y (mm) increases[2]. The reason for this is that the fabric attenuates the “quantity” of optical information in its propagation. On the other hand, the reflection of light from the fabric surface decreases the signal/ noise ratio of optical information and the “quality” of optical information is decreased; that is, the contrast, i.e. the ratio of bright part to dim part of the object, is decreased. In other words, the intervention of fabric makes the “entropy of information” increase in the propagation of optical information. When the distance between the object and the observer is Lo (m), the premiss conditions under

EO U AI DE C B GHI FK H

F SD ZW X U JYT R

Lf Lo

Figure 1. Relationship between the Observer, Fabric and Objects

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 11-13, © MCB University Press, 0955-6222

11

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

of the fabric surface to the eye, i.e. eo > ef when the fabric is placed at the distance Lo. The luminance of the object Eo and the reflectances ηo and ηb of the bright part and the dim part of the object respectively are constant. The fabric is then placed at a distance Lf (m) away from the observer. The cover factor of the fabric is ε (decimal), the reflectiveness of the side to the observer is ηf1 and the reflectiveness of the other side is ηf2. After the fabric is in place, the condition of equation (1) is still satisfied, but the attenuation is: eo =

(

)

Eoηo (1 − ε ) + ε − η f 1 exp( −κδ ) . Lo 2

(

)

where kL = Lo /Lf , KE = Ef /Eo. When K < Kcr or ef > eo, it is impossible to resolve the object because of the existence of the ef . It can be concluded that the principal reason causing the visibility to decrease is the attenuation of the absolute factor e and the reduction of the relative ratio of eo + ef /eb + ef to the observed system according to equations (2) and (4), which include the optical properties (ηo, ηb) of the object, the circumstantial conditions (Ea, Ef , Lo , Lf), and the features of the fabric such as structure (ε, δ) and optical properties (ηf1, ηf2, κ).  L 2 Ef  Assume A = K L2 K E =  o    , η f 1 = η f 2 = η f  L f   Eo 

(1)

Z1 = εη f , Z2 = ε (1 − η f )κδ ,

Equation (1) consists of two parts. The first is the luminance from the light passing through the crevice of the yarn, the second is the luminance from the light passing through the area covered by the fabric after the reflected light is deducted. In the equation, κ and δ are the thickness of fabric and the absorption coefficient of the area covered by the fabric respectively. If exp (- κδ) is spread and the first two items are taken, the equation is approximately:

(

then eo = eo = K=

(

)

((

)

) ))

Eoηo 1 − εη f 1 − ε 1 − η f 1 κδ . Lo 2

=

(

=

Efηf 2 Lf 2

.

(2)

=

Efηf 2 Lf 2 Efηf 2 Lf 2

ηb (1 − εη f 1 − ε (1 − η f 1 )κδ ) + K L K E εη f 2

ηo (1 − Z1 − Z2 ) + AZ1 . ηb (1 − Z1 − Z2 ) + AZ1

(6) ( 7)

(1) The elements to influence the fabric covershelter properties primarily include structural factors (cover factor ε and thickness δ) and characteristic factors (reflectance ηf and absorbance κ). (2) The reflection factor Z1= ε ηf can show effectively the fabric cover-shelter effect with the proper κ and δ values. If the absorption factor Z2 = ε (1-ηf) κ δ is applied, the Figure will have better effect. (3) To the fabric surface facing the observed object, increasing of the reflection factor makes a large amount of effective information reflect and then attenuate, which enhances the shielding effect. To the other surface, increasing of Z1 also makes interference to the effective information strengthened under condition Ef = 0, and improves the covershelter properties of fabrics even more.

(3)

ηo (1 − εη f 1 − ε (1 − η f 1 )κδ ) + K L K E εη f 2

Z1

Conclusion

Then the contrast grade between the brightness and dimness which is observed by the eye is: eo + e f K = = eb + e f Eoηo ((1 − εη f 1 − ε (1 − η f 1 )κδ ) + Lo 2 = Eoηb ((1 − εη f 1 − ε (1 − η f 1 )κδ ) + Lo 2

Lf 2

Partial typical data of experiment results are shown in Table I. The relationship of Y and Z1 with parameters Lo and Ef is shown in Figures 2 and 3.

After the fabric is in place, one part of the actual luminance is from the object (eo, eb), the other part from the light reflected from yarns of the fabric surface (ef) because of the surface reflected light (which should be scattered light in the strict sense, and it is equal to the following): ef

Ef

(5)

Results

Eoηo (1 − ε ) + ε − η f 1 (1 − κδ ) Lo 2

eo =

Eoηo (1 − Z1 − Z2 ) Lo 2

( 4)

12

VOLUME 6 NUMBER 2/3 1994

Eo (Lx)

Ef = 0)

Ef (Lx)

(Eo = 200 Lx)

Z1

32

200

400

10

50

100

400

1

0.1743

18.0

14.5

11.5

18.0

18.0

23.0

-

2

0.0604

18.0

14.5

14.5

14.5

14.5

14.5

18.0

3

0.1335

14.5

9.5

9.5

9.5

9.5

11.5

18.0

4

0.1431

18.0

11.5

9.5

7.0

11.5

14.5

23.0

5

0.1969

23.0

18.0

14.5

18.0

23.0

29.0

-

6

0.1080

11.5

9.8

7.0

7.0

7.0

14.5

18.0

7

0.1743

11.5

9.5

7.0

7.0

7.0

11.5

23.0

8

0.0478

3.5

2.1

2.0

2.0

2.0

2.0

2.1

9

0.0675

3.5

2.1

2.0

2.0

2.0

2.0

2.1

10

0.0492

3.5

2.1

2.0

2.0

2.0

2.0

2.1

11

0.1170

14.5

11.5

7.0

9.5

11.5

11.5

-

12

0.2522

29.0

23.0

18.0

23.0

29.0

-

-

13

0.0758

5.0

2.7

2.7

2.1

2.1

2.1

3.5

No.

Table I. Experiment Data (Lo = 5m Lf = 2.5 m)

Y (mm)

Y (mm)

E0 = 32(X)

20

X

+

+

50

E ƒ=

X

X

X

X

0

= Eƒ A+

X

X

+

20

E0 = 400(0)

A

10

+

X

10

E0 = 200(X)

E ƒ=

Eƒ= 1

00

X

B

X

B

X

10

10

X

X

X

X

X

XX

0

X

X

X

X XX

10

20

(x 10–2)

Z1

X

0 10

20

(x 10–2)

Z1

Figure 3. Relationship between Reflectance Factor and the Least Distinguishable Distance under Different Luminance of the Fabric Ef

Figure 2. Relationship between Reflectance Factor and the Least Distinguishable Distance under Different Luminance of the Object Eo

References

Protective Clothing and Its Material, Xian, 1992, pp. 80-3 (in Chinese). 2. Wu, K.G., Optical, Beijing People’s Education Press, 1978, pp. 536-8 (in Chinese).

1. Meiwu, S. and Kan, L., A Study on the Theory and Practice of Fabric Cover-shelter Properties, National Symposium on

13

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Some Difficult Issues in Recent Apparel Manufacturing and a Counter-measure for the Future K. Ito Osaka Apparel Technology Modernization Association, Japan and M. Nitta Oga Clothing Hirakata Company, Nagao-Tanimachi, Osaka, Japan also high. On the other hand, their TAV (the value of suit appearance prediction) is low. The high THV suiting has high TAV. The unbalanced combination of the higher THV and lower TAV resulted in a difficult decision regarding suit quality and, as a result, consumers began a gradual rejection of these “new generation wool” suits. Figure 1 shows the correlation between THV and TAV of conventional suiting and the lightweight suiting. There is no correlation between THV and TAV of the lightweight suiting. (2) A recent general trend of weave structure is towards a lower picking number of weft yarn. This is because of the high efficiency of weave production and to make the weave light. (3) As shown in Figure 2, the fabric extensibility becomes lower because of the lower number of weft yarns and the fabric is lightweight. Also the LT (Linearity of tensile property; the higher value of LT makes fabric stiff in the small extension region) is high, especially high in the case of SIRO-fil yarn weaves. And tensile resilience and shear stiffness are small.

Recent Trend of Men’s Suiting One of the recent trends in men’s suiting is excessive lightweight suiting. This trend is especially seen in mid-summer suiting. The range of the fabric weight of lightweight suiting is 120160 gr/m2. The range of traditional summer suiting was 170-210 gr/m2. This light weight has been brought about by the use of SIRO-spun yarn or various types of SIRO-fil type yarn. These suitings are called “new generation wool”. Although lightweight is attractive, a problem was that the lightweight suiting had been developed without careful consideration as to the quality of suiting from both the consumer’s and the tailoring point of view. In fact, many consumer complaints were made to retail shops, mostly because of wrinkle problems. The difficulty in the tailoring of the “new generation wool” suiting caused many technical troubles during the tailoring. These problems and troubles were caused by the excessive lightweight of the suiting.

The Problems of Excessive Lightweight (1) The hand value of SHARI (crispness) of the lightweight suiting is high and, accordingly, the THV (Total hand value) of these fabrics is

This article was written on the basis of a discussion with Professor Sueo Kawabata of Kyoto University and prepared with his assistance. The authors appreciate his help.

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 14-16, © MCB University Press, 0955-6222

14

VOLUME 6 NUMBER 2/3 1994

from the fabric fitting to human body motion, mainly the fabric extensibility in weft direction. Not only fabric weight but also low LT with high RT (resilience) are related to this true feeling.

5

4

▲

●

●

▲ ▲

▲

▲

▲

▲

▲

▲

▲

▲

THV

▲

▲

●

▲

●

▲

▲

▲

▲

▲

▲

●

▲

▲

▲

▲▲

Normal zone

▲

▲

3

▲

●

● ●

●

▲ ▲ ▲ ●

▲

Difficulty in Tailoring of Lightweight Suiting

●

▲ ▲ ▲

▲

▲

▲

▲

▲

▲ ▲▲

▲

▲

▲

▲▲

▲

▲ ▲

▲

▲

▲

▲

▲

▲

▲

▲

▲

▲

▲

▲

●

2

●

●

The excessive lightweight structure of fabric reduces fabric stiffness from its moderate level. Accordingly, various chemical treatments tend to be applied to the fabrics for increasing the stiffness in spite of the fact that the stiffness should be basically controlled by the fabric structure design. As a result, many difficulties in tailoring and suit wearing become apparent as mentioned above. In addition, another serious problem in tailoring is the unpredictable behaviour of fabric steam-press shrinkage and recovery. During tailoring, steam-press operations are applied to fabrics frequently. After pressing, the fabric recovers the dimensional shrinkage caused by the pressing. The next sewing must be operated after the complete recovery of fabric or based on the prediction of the recovery amount in advance. The shrinkage recovery behaviour of a normal worsted fabric is shown in Figure 3 and of an abnormal example in Figures 4 and 5. In the Figures, the fabric shrinkage is measured at various stages as follows: Stage 0 S0: Initial (before steam press) state. Stage 1 S1: The shrinkage just after steam press. Stage 2 S2: The shrinkage at equilibrium state in room conditions after steam press.

●

▲

▲

▲

1

Siro filament (Nylon) Siro filament (PET) Siro spun

●

▲ ▲

Siro filament New generation wool

▲

0 0

1

2

3

4

5

TAV

Figure 1. The Correlation between THV and TAV of the Lightweight Suiting. There Is No Correlation between Them

The mechanical parameters of several light weight, “new generation wool” weaves are plotted on the tailoring process-control chart. The circle is SIRO-spun and the triangle is SIRO-fil yarn weave. These (2 and 3) cause the wrinkle problem during wearing. The mechanical parameters are outside the comfort zone shown by the shadowed zone in Figure 2. (4) The difficulties in tailoring cause an appearance defect in the suit. It must be understood that the true feeling of softness and the lightweight feeling of suits come

Non-control zone

Control zone

Control zone

▲

LT

0.5

0.55

0.58

0.6

▲▲ 0.65

0.7

0.81

High LT Difficult overfeed sewing

▲▲▲

RT

50

55

60

65

68 70

Steam-press shrinkage and recovery test by HESC FT 103A method

75

Fabric shrinkage S (per cent) 3–

▲ 4

4.5

5.5

6 10.6

EM2

▲

4

2

6

8

10

Normal (before 1982)

G

▲ 2HG5 0.8

▲

▲ 0.4

–

–

–

F (per cent)

–25

–20

–15

–10

–5

S0 0

▲ 0.5

▲ 1

3

–

2

–

▲ 1

S2 1– S4

15

▲▲

EM2/EM1

0.7

0.6

0.8

0.9

1.0

1.2

1.8

l ma

rve

2

High THV and comfort zone

2.5

3

▲

Summer:

Source: Ito and Kawabata, 1985.

–1 – –2 –

S3

Siro-spun Siro-filament

1 3 5 7 Dehydration

cu

r

No

▲ 1.5

S1

2–

▲ ▲

–

3.5

–

▲ 3

–

▲

–

EM1

–3 – Warp direction

Weft direction

Figure 3. The Shrinkage Behaviour of Normal Worsted Fabrics

Figure 2. Tailoring Process-control Chart

15

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

-

Fabric shrinkage S (per cent)

Steam-press shrinkage and recovery test by HESC FT 103A method

-

5

–

-

-

-

-

3

–

2

-

-

1

2hrs

10 min

–3 – Warp direction

0

Steam press

–2 –

Summer:

-

-

–5

-

145–

Fabric weight (gr/400cm2)

–

–10

-

–

–

–15

–

–

–20

–

–

–25

S2 S0 0 1 3 5 7 Dehydration –1 – –

–

1–

Normal

-

-

S3

150–

–

rmal

F (per cent)

S1

2 – S4

Abno

Fabric dimension (mm)

3–

Weft direction

Normal

5– -

Stage 3

S3: The shrinkage just after humidification which is applied after stage 2.

Stage 4

S4: The shrinkage at the equilibrium state in a room after humidification.

5

–

-

-

3

-

2

-

-

1

–

-

0

-

-

0

–

-

Figure 4. The Shrinkage Behaviour of the Recent Abnormal Summer Suiting. The Behaviour Has No Trend Like That of Normal Fabrics, but Changes at Random for Each Fabric

2hrs

10 min

Steam press

Note: Spring and Summer temperature 36-40˚C; relative humidity 95-100 per cent

Figure 5. The Shrinkage-recovery Process after Steam Press. The Recovery Process of Recent Summer Suiting Is Varied and Difficult to Predict

The abnormal behaviour is probably caused by the chemical treatment applied to the fabric to increase stiffness. The lightweight fabrics, especially the fabrics treated with chemicals, have higher LT value of KES measurement parameter. These high LT fabrics are difficult fabrics for overfeed operation in tailoring and cause pucker problems. These problems have increased since the lightweight suiting appeared in the marketplace. The lightweight fabrics and SIRO-fil yarn fabrics are not necessarily bad fabrics. If they are good fabrics from the consumers’ point of view and comfortable to wear, we can use them positively. The problem is that these “new generation wool” fabrics are not produced with careful design as to fabric quality, but are produced only because of a new fashion.

General Problems in Tailoring In the apparel industry, especially the men’s suiting industry, sewing-machines which have an automatic overfeeding mechanism are not available yet. The steam-press machines which have a precise temperature- and pressure-control system are not available yet. The apparel industry must study more advanced technology and put investment into research. Fabric buyers in the apparel industry must say to textile makers explicitly what properties of fabric are important for both consumers and apparel manufacturers from a technical viewpoint, and not buy those fabrics having inferior properties as mentioned before. These efforts will be a step towards new developments for the coming age of high quality.

16

VOLUME 6 NUMBER 2/3 1994

Difficulty with Shingosen A View from an Analysis of Fabric Hand Sueo Kawabata Department of Polymer Chemistry, Kyoto University, Japan Development of Shingosen Weaves –4

Bend

–3

EM (%)

Tensile strength

The Shingosen weaves have been developed by the initiative of fibre producers as follows. When PET fibres appeared in the marketplace around 1960, the producers of PET fibres expected a brilliant future for them because of their high utility performance. The consumers’ reaction to this new fibre was, however, one of rejection and the term “gosen” (short of synthetic fibre) began to have a bad image. This is a result of the fibre producers not understanding the importance of fabric. Their main interest was in fibre evenness and strength. The first step towards improving the PET weaves by the fibre producers was to approximate the hand of silk weaves by means of caustic treatment. The silk-like weaves were successful to some extent, but, the “gosen” image still remained. The difficult times of the fibre producers continued. Around 1980, the Shingosen weaves appeared based on progress in micro-fibre technology. The micro-denier fibres brought a fresh, new type of fabric hand. The new hand of the Shingosen weaves attracted young female customers[1,2].

–2

2

LT

3

0.3

WT

2

RT (%) BN

4

2

10

0.5

20

0.6

30

0.9

40

50

50

0.1

0.2

4

40

0.8

20

40

0.05

3

30

0.7

10

30

0.03

1

100

60

0.3

70

0.5

1

2HBN

G

0.01

0.2

2HG

0.2

0.1

0.02

0.3

0.3

0.06

0.03

0.4

0.1

0.5

0.3

0.2

1

0.4 0.5

1

3

2

1

0.5

2

3

4

4

5

5

10

20

2HG5 0.3

Compressibility

Shear strength

6

5

20

0.02

0.005

LC WC

0.4 0.5

1

0.1

0

0.03

2

0.2

0.04 0.05

0.3

3

4

0.4

0.1

10

5

0.5

0.2

0.3

0.4

20

0.6

0.7

0.5

0.8

1

2

RC 20

MIU

Surface

4

0.4

3

(X-X)/σ 0

–1

30

40

0.1

50

60

70

0.2

80

90

0.3

0.4

MMD 0.004

0.005

0.01

0.02

0.03

0.04

0.05

0.1

SMD (microns)

T (mm)

W (mg/cm2)

0.2

0.2

0.3

0.4 0.5

0.3

1

0.4

10

2

3

15

5

4

10

20

1

0.5

20

25

30

40

2

30

40

50

3

50

4

60

70

Shingosen (new worsted, N = 33) for women's suit Good zone

Figure 1. Mechanical Parameters of the Shingosen Weaves “New Worsted” Type for Ladies’ Suiting

Features of the Shingosen Weaves in Fabric Property The features in fabric property are: (1) Smooth, soft and fine appearance coming from micro-denier fibres. (2) Bending stiffness of the fabric is lower in general than conventional fabrics having similar end-uses. Figure 1 shows the average mechanical parameters of the ladies’ Shingosen suiting. They are plotted on the mechanical parameter chart of conventional ladies’ suiting. The shadowed zone is the conventional high-quality suiting zone. The bending and shear parameters are much lower

than conventional weaves. This property makes the Shingosen weaves very flexible and drapable (see Table I).

Men’s suit

Ladies’ suit

Conventional 1.75 1.81 (worsted base) Shingosen 1.45 1.24 Bending length = (B/W)1/3cm Table I. Bending Length of Shingosen Fabrics

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 17-19, © MCB University Press, 0955-6222

17

Ladies’ dress 1.37 1.24

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

(3) In the low-bending curvature of which the range is 0~0.2 cm–1, the hysteresis in the bending moment is very low, that is the fabric is very elastic and has high resilience. In the higher curvature, the hysteresis is as high as it is for normal PET filament weaves. This high resilience in the lower bending curvature makes a beautiful ripple motion of garment with body motion.

(1) The micro-denier fibre and the condition of 100 per cent PET restrict the wider variation of fabric hand of the Shingosen weaves. (2) Fabric quality of Shingosen is not necessarily high from conventional criteria[3]. Examples are shown in Figure 2 for ladies suiting and Figure 3 for men’s suiting of Shingosen weaves. The fabric primary hand of the Shingosen weaves are almost outside the conventional high-quality zone which is indicated by a shadowed area. It must be noted that the conventional criteria of the quality is based on man’s long experience in clothing. (3) Difficulties in the sewing and tailoring processes are a problem. These difficulties raise the price of Shingosen garments. The appearance of the suit made of Shingosen weave is not predicted to be of good appearance from fabric property on the basis of conventional criteria. Figure 4 shows the predicted total appearance value (TAV) of the Shingosen suiting and mechanical components of fabric relating to the TAV[3]. The good zone is shown by the shadowed area. Most of the

(4) High surface frictional coefficient of the fabric. These features of the property of Shingosen weaves brought the weaves a soft feeling and a new and elegant feeling. This is the reason why the Shingosen weaves were accepted by young female customers.

Signs of Progress in Textile Engineering From a technological viewpoint, the development of the Shingosen has stimulated fibre engineering, especially PET fibre engineering, into activity and brought a new concept for the development of new textiles. The signs of progress in textile engineering are:

Hand value:

(1) Fibre producers and engineers are now seeing the importance of the fabric and apparel market, whereas in the past they were concerned only with fibre.

Stiffness (Koshi) Smoothness (Numeri) Fullness (Fukurami) THV –winter

(2) The wide variation of fibre shape increased freedom in fabric design of PET weaves.

3

0

4

1

0

1

–3

5

2

+

3

2

3

1

–4

(3) Co-operation between the fibre engineers and the weaving and finishing people was initiated.

+

2

4

7

5

4

7

6

7

3

–1

8

6

5

2

–2

+

6

8

9

8

9

4

0

1

10

9 10 10

5

2

3

4

Shingosen (new worsted, N = 33) for women's suit Good zone + σ (SD, chain line) Average (black circle) and _

The Japanese textile industry has a subdivided structure. The textile industries from fibre to apparel are subdivided into many sections such as weaving companies, finishing companies, etc. The interrelationship between these companies has not been close in terms of technical cooperation. The Shingosen technology has been developed on the basis of the co-operation of these sections. This was good for the Japanese textile industry from a technical viewpoint; however, the relationship is still not entirely satisfactory. This is causing a new problem for the future of Shingosen weaves.

Figure 2. Fabric Hand of the Shingosen Weaves (“New Worsted” Type) for Ladies’ Suiting. The Hand Values Are outside the Conventional High-quality Zone (Shadowed Zone) of Ladies’ Suiting

–3σ

Hand value: Stiffness (Koshi) Smoothness (Numeri) Fullness (Fukurami) THV –winter

–2σ

2

–σ

4

0

4

2 1

6 3

3σ

8

6

4 2

2σ

6

2

0

σ

0

4

10

8

10

8

10 5

Figure 3. Fabric Hand of the Shingosen Weaves (“New Worsted” Type) for Men’s Suiting. The Hand of the Shingosen Weaves Fall outside the Conventional High Quality Zone (Shadowed Zone) of Men’s Suiting

Problems with Shingosen The future development of the Shingosen weaves is encountering difficulty, and the future is not so hopeful. The problems are:

18

VOLUME 6 NUMBER 2/3 1994

–4

Z1 (Formability)

–3

Mechanical parameters X11 log EL2 2 X12 log BS2 X13 log SS

Z2 (Elastic potential) Z3 (Drape)

0.02 0.4

3

4

0.03 0.04 0.05

0.5

X21 log BP 0.05 X22 log SP

Normalized value –1 0 1

–2

5

6

0.07

7 8 9 10

0.07

0.1

3

0.2 2.0

0.2 5

4

6

0.3

3

20

0.1

0.6 0.7 0.8 0.9 1.0

2

2

3.0 0.4 0.5

7 8 9 10

30

0.3 0.4 0.5 4.0

0.7

Yarn, apparel manufacturing

4

40

(a) Archaic system

0.7 5.0

6.0 7.0

1.0

20

30

40

50

3

X31 BS/W

3 X32

SS/W

1.0

1.5

2.0

2.0

3.0

4.0

Consumer

Polymer fibre producer

2.5 5.0

6.0

7.0

Weaver

Finisher

Apparel Trading Consumer manufacturer companies

Three basic components of tailorability Formability Z1 Elastic Z2 potential Z3 Drape

KN(eq.10) TAV

0

1

2

0

2 0

1 0

2 1

3

4

3

4

3 2

3

4

5

6 5

6

5 4

(b) Modern divided system

6 5

6

Polymer scientists Fibre producer Weavers Finishers Apparel manufacturer

Shingosen (new worsted, N = 33) for women's suit Good zone

Figure 4. Suit Appearance Prediction of the Shingosen Suiting (“New Worsted” Type) for Lady’s Suit. The TAV and Related Components Are Shown

Fabric design

Consumer

(c) The future system

Figure 5. The Textile Manufacturing System

components of Shingosen suiting fall outside the good zone. This is mainly caused by their low bending and shear stiffness. This low stiffness of the weaves causes difficulty in the pattern-cutting process. (4) Some weak points of PET fibres such as hydrophobic property, high friction between fibres, etc. remain in the Shingosen weaves because of the 100 per cent PET weave.

was quite effective in controlling fabric property for the fibre producers and, in fact, the Shingosen weaves were developed by the initiative of fibre producers. The result, however, has many problems, as shown in the previous section. For the further development of Shingosen technology we need to consolidate the technology from fibre to apparel by the joining of these subdivided sections as shown in Figure 5(c). The technology level in these sections must be equally high. However, this is a difficult problem for the Japanese fibre and textile industries.

The Future of Shingosen Weaves For the further improvement of Shingosen weaves, the traditionally subdivided nature of the Japanese textile industry will be a problem. In the past, that is before the industrial revolution, textile and apparel manufacturing had been consolidated, the products were produced in the same place and the manufacturer was closely connected with consumers (see Figure 5(a)). The manufacturer knew what property of fabric was good on the basis of consumer preference. After the modernization of industry, the textile process was divided into many sections (see Figure 5(b)). Production efficiency became high under this system but the engineers working in each section lost the criteria for real fabric quality. In Japan, the fibre industry became a large-scale industry and its technology produced the Shingosen. The big problem is, however, that weaving, finishing and apparel manufacturing of Shingosen weaves are supported by smaller-scale companies, mostly as subcontractors of the fibre producers, and their technology depends mostly on the expertise based on experience. In the early stages of the Shingosen development, this subcontract system

n References 1. Kawabata, S. and Niwa, M., “Synchronized Innovation and Marketing: The Case of Shingosen”, Proceedings of the Textile Institute World Conference 1993, The Textile Institute, Hong Kong, May 1993. 2. Kawabata, S., “Shingosen Fabrics: A New Generation of Textile and Fibre Engineering”, Bill Aldridge Memorial Lecture, Advances in Fabric Technology, Proceedings of the 19th Annual Conference, The Textile Institute (New Zealand Section), c/o WRONZ, Christchurch, New Zealand, 1991. 3. Kawabata, S. and Niwa, M. “Fabric Performance in Clothing and Clothing Manufacture”, Japan Textile Institute, Vol. 80 No.1, 1989, pp. 19-50.

19

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Investigation of the Performance of Sewing Thread Miyuki Mori and Masako Niwa Department of Clothing Science, Nara Women’s University, Japan is stretched by the take-up lever and the bobbin thread is stretched by the rolling of the bobbin mechanism and the needle thread. The needle thread is bent at the needle hole or the thread eyelet and rubs against the needle or the thread eyelet placed at several points in the machine. To characterize the sewing thread, tensile, bending and frictional properties were measured.

Introduction The mechanical properties of sewing thread play an important role in sewing. There have been some studies of seam strength from the point of view of the strength of sewing thread[1-4]. The mechanical modification of the seam has been tried[5,6] and the fatigue of sewing thread has been studied[7], but there have been few studies to date about thread from the point of view of making seams of good appearance[8]. Limited by the types of machines used by sewing factories, thread is chosen by empirical intuition. It is the purpose of this article to identify the suitable sewing thread for making seams of good appearance, and to investigate the interrelation between thread property and seam-line quality by using a collection of 53 commercial threads.

Tensile Property The thread needs intensity to some extent in order not to be cut off because of tension or friction by the machine. As an index to express the strength of the thread, the tensile property until breaking-point was measured, and breaking load, breaking strain, yielding load and yielding strain were chosen and defined as Fb, eb, Fy and ey respectively, as shown in Figure 1. The tension of the needle thread is as high as 200 gf when being sewn. This region is shown in Figure 1 by diagonal lines. The tensile property within this region, up to 200 gf, was measured. The stitch length is a few millimetres, while the stroke of the needle is a few centimetres, so the

Samples From sewing factories, 43 kinds of thread for men’s suiting and ten kinds of thread for ladies’ dresses, in all 53 kinds of commercial thread, were chosen, as shown in Table I. Among these commercial samples of threads, several were recommended as good thread for sewing by apparel manufacturers.

Sample: No.01(Silk 100 per cent, 22tex)

Fb

F(gf)

1,000

Measurement Conditions of Machine Threads

500

Fy

200

The threads are deformed in various ways at several points in the machine. The needle thread

ey

0 0

5

10

Figure 1. Breaking Tensile Curve

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 20-27, MCB University Press, 0955-6222

20

15 e (per cent)

20

25

eb

VOLUME 6 NUMBER 2/3 1994

Main use

Number of samples

Men’s suit

43

Fibre

Filament (19)

Polyester (15) Silk (2) Nylon (2) Polyester (13) Cotton (4) PET/Cotton (1) Polyester (5) PET/Cotton (1)

15-27 22-25 17-27 24-44 29-38 33 20-31 38

Polyester (1) Polyester (9)

24 21-34

Spun (18)

Core spun (6) Lady’s thin dress

10

Thread count tex

Thread

Filament (1) Spun (9)

Note: Number of samples shown in parentheses Table I. Machine Threads Samples for Lock-stitch

thread was stretched five to eight times at the same place. Taking this into account, tensile property up to 200 gf was measured seven times repeatedly. Figure 2 shows this curve and it is evident from this figure that the difference between the first and second cycles is clear, but less distinct between the sixth and seventh cycles. For the increase of the tensile time, the residual strain is higher. On the first cycle, tensile strain, linearity of tensile, tensile energy and tensile resilience were defined as EM, LT,

200

Properties Tensile Break Repeated tension (maximum load Fmax = 200 gf)

Sample: No.01(Silk 100%, 22tex)

Conditions

Sample length: 10 cm Strain rate: 0.04%/sec. Maximum load (Fmax): 200 gf Repetitive time: 7 cycles Sample length: 10 cm Strain rate: 0.02%/sec.

Bending

Sample length: 1 cm Maximum curvature:±2.5 cm–1 Rate of bending curvature: 0.5 cm–1/sec.

Friction

Thread tension: 150 gf Moving rate of the needle: 1 cm/sec Moving stroke of the needle: 1 cm

F(gf)

First cycle

Table II. Measurement Conditions of Machine Threads

LT linearity WT energy RT resilience

100

WT and RT[9] respectively. On the seventh cycle, the residual strain was defined as R7. These five parameters were chosen.

EM

0 0

1

e (per cent)

2

F(gf)

200

Bending Property Figure 3 shows the bending moment-curvature curve. Bending stiffness and hysteresis per thread were chosen and defined as B and 2HB[9] respectively.

Seventh cycle

100

Frictional Property The frictional property between the needle and the thread was measured by the device as shown by Figure 4. On the handy compression tester[10], the needle and the thread eyelet were

R7 0 0

1

e (per cent)

2

Figure 2. Repeated Tensile Curve

21

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Sample: No.01(Silk 100 per cent, 22tex)

B

-

-

-

-

-

–3

0.0025 -

PF (gf)

M(g cm/thread)

PF1 400 –

0.0050 -

–2

–1 2HB 0

1

2

3

300 –

F= PF1-PF2 PF2

200 – 100 –

K(cm–1) –0.0025 -

0

–0.0050 -

8

10

set, and a weight of 150 grams was hung on it so as to produce tension equal to the tension given to machine thread. The stroke and speed of the needle were established at 1 cm and 1 cm/sec respectively. Figure 5 shows the frictional curve. For example, in the case of thread sample No. 1, by attaching the weight, the force points out about 300 gf. The needle motion starts, and at the point of penetration, maximum value, designated as PF1, is reached. At the point of withdrawal, minimum value, designated as PF2, is reached. The difference between PF1 and PF2 was designated as FF. One cycle consisted of five repetitions of penetration and withdrawal, as this is the friction given to actual sewing thread. The

Thread eyelet Yarn W

Stroke: 10 mm Speed: 10 mm/sec

Figure 4. The Device for Measurement of Friction

Mechanical parameter

Mean

Tensile (break) log Fy log ey log Fb log eb

2.293 0.388 3.035 1.252

yielding load yielding strain breaking load breaking strain

Tensile (Maximum load, Fmax=200 gf) log EM tensile strain at Fmax LT linearity log WT tensile energy RT tensile resilience log R7 residual strain at 7th cycle

Friction PF1

2 4 6 Deformation (cm)

Figure 5. Frictional Curve

Figure 3. Bending Curve

Bending log B log 2HB

Sample: No.1(PE 22tex)

bending stiffness bending hysteresis

friction force between thread and needle FF difference of descending and ascending friction forces Number of samples: n = 53

SD

0.1405 0.1097 0.09186 0.1443

0.507 0.923 0.464 62.05 -0.0244

0.1820 0.1748 0.2276 16.925 0.3186

-2.811 -2.838

0.1909 0.1931

Maximum

Minimum

UNIT

361 4.07 1542 35.1

58 1.59 706 6.6

gf % gf %

5.25 1.44 8.28 88.2 6.90

1.65 0.626 0.74 25.5 0.30

% gf cm % %

0.00925 0.00903

0.00066 0.00058

gf cm2 gf cm

426.9

16.06

466

391

gf

162.5

22.24

216

113

gf

Table III. Mechanical Parameters of Machine Threads

22

VOLUME 6 NUMBER 2/3 1994

G2-1 tex Tensile Break Fy ey Fb eb

23

P1

No. 39 No. 07 No. 33

24

30

30

200 –

28

138 3.10 872 17.10

180 2.50 1090 19.00

171 3.01 1035 20.68

260 1.83 1263 14.07

252 2.43 1439 17.77

6.8 0.978 6.70 41.6 3.00

3.6 0.939 3.45 67.6 0.75

5.5 1.077 5.86 43.3 1.83

2.0 0.782 1.55 85.3 0.53

2.4 0.722 1.72 77.7 0.64

0.80 1.20 0.80

1.81 2.80 1.97

1.19 1.97 1.56

1.60 2.29 1.38

1.97 3.02 2.09

F (gf)

First cycle

100 –

Seventh cycle

0– 0

1

2

3

e (per cent)

200 –

Sample: P1(PET 100 per cent spun 24tex)

First cycle

413 140

435 172

415 145

430 168

436 174

F (gf)

Fmax = 200 gf EM LT WT RT R7 Bending B BS 2HB Friction PF1 FF

Table IV. Mechanical and Frictional Properties of Thread Samples

100 –

0– 0

measurement conditions of these parameters are shown in Table II.

Seventh cycle

1

e (per cent)

2

3

Figure 7. Repeated Tensile Curve

continued to be used. The other is not suitable for sewing, and its use was stopped. The tensile curves of the two are shown in Figure 6. Sample G2-1 is the former thread and sample P1 is the latter. Breaking load and strain of sample P1 are larger than sample G2-1 as shown in Figure 6. The curve of repeated tensile properties for both threads is considerably different, as shown in Figure 7.

Mechanical and Frictional Properties of the Commercial Threads The value of the parameters mentioned above of 53 kinds of the commercial thread has a wide range, summarized in Table III. With mean and standard deviation of each parameter, the nomalized data chart was made. The sample threads from a maker in Japan were placed in two groups. One is suitable for sewing judged by an apparel engineer, and

n

The breaking strain of the suitable thread is smaller

1,500

n

Sample: P1 (PET 100 per cent spun, 24tex)

F (gf)

1,000

The value of tensile, bending and frictional properties of the two are shown in Table IV and plotted on the data chart as shown in Figure 8. The breaking load, yielding load and breaking strain of the suitable thread is smaller by one sigma than the mean, but yielding strain is greater by one sigma than the mean. EM, LT, WT and R7 are different for both and the suitable thread is considered softer than the unsuitable

500

Sample: G2-1 (PET 100 per cent spun, 23tex) 0

Sample: G2-1(PET 100 per cent spun 23tex)

0

5

10 e (per cent)

15

20

Figure 6. Breaking Tensile Curve

23

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Sample

Yarn count (text)

EM (%)

Tension of needle thread (g) MJ MB MS MP*

Tension of bobbin thread (g) MJ MB MS MP

G2-1 P1 No. 39 No. 07 No. 33

23 24 30 30 38

6.8 3.6 5.5 2.0 2.4

145 130 140 140 200

17 16 20 18 25

80 90 90 90 110

60 63 100 60 80

85 95 83 95 95

10 10 15 15 15

18 20 14 16 22

10 15 15 9 10

* Presumed value

Table V. The Tension of the Different Kinds of Thread

thread. The bending properties of these are not very different. The frictional force of the suitable thread is smaller than that of the unsuitable thread. In another example, the mechanical and frictional properties of three commercial threads which are suitable for sewing (sample No. 39 thread was judged by a thread maker in England, the others were judged by Japanese apparel makers) are shown in Table VI and plotted on the data chart as shown in Figure 9. Tensile property of No. 39 thread is similar to the suitable thread sample G2-1 as shown in Figure 8. Bending stiffness of this thread is higher than that of sample G2-1 because of its thickness. Frictional properties are similar to sample G2-1 as shown above. This thread is considered a soft thread. In contrast with this, the other sample, No. 33 thread made in Japan, was hard to stretch and

Tensile –4 (Break) Fy* Fy*

ey* Fb* eb*

–3 60

70

–2

80 90 100

1

(Fmax = 200gf)

5

600 6

7

EM*

RT

0

2

3

0.05

3

2

30 0.2

40

50

5

40

50

0.0003 0.0004 0.0005

10 1.5

5

70 1

10

80

90

2

3

100

20 110

4 5

0.001

0.002

0.003 0.004 0.005

0.001

0.002

0.003 0.004

0.005

PF1 FF

370

380

390

400

100

Thread sample: * log scale

410

420

440 450 460

430

G2-1,

470

200

150

Men’s summer suiting Wool/polyester blendc

A

B

C

D

log EL

8.147

5.569

3.56

2.37

log BS

0.0185

0.1135

0.0481

0.1163

log SS

0.332

0.615

0.822

2.745

EP

0.526

0.484

0.338

0.189

log BP

0.0467

0.2634

log SP

5.093

5.376

T

0.697

0.732

W

4.2

7.8

AIR

0.314

0.069

Wool 100% Wool 50%, polyester 50%

490 250

Figure 8. The Data Chart Plotting Mechanical and Frictional Parameter of the Thread for Lady’s Dress

SIRO spunb

c

480

P1

New worsted

Polyester 100%

120 10

New silky

b

60

Friction

Lady’s dress SHINGOSENa

a

600

2,000

5

4

60

0.3 0.4 0.5

0.0003 0.0004 0.0005

2HB*

500

1

20

0.1

4

30

4

4

3

350 400

1,500 20

1

10

2 300

3

9 10

1.5

0.4 0.5

1 250

800 900 1,000

0.5

WT*

B*

8

200

2

700

1

LT

R7* Bending

150

1.5

500

(X–m)/S 0

–1

0.1244

14.965

0.369

0.449

14.3 0.184

Table VI. Fabric Samples

24

0.2395

13.795

6.7 0.508

VOLUME 6 NUMBER 2/3 1994

(X–m)/S –3 70

Fy*

(Fmax = 200gf)

80 90 100

1 600 6

7

EM*

RT R7* Bending

B* 2HB*

200

2

0.05

40

5

50

3

2

30 0.2

40

50

60

0.3 0.4 0.5

60

1

1.5 4

5

70 1

10

80

90

2

3

100

0

2.0 110

4 5

MJ

120

MB MS Machine type

MP

10

Sample B (Shingosen new worsted) 0.0003 0.0004 0.0005 0.0003 0.0004 0.0005

3

0.001

0.002

0.003 0.004 0.005

0.001

0.002

0.003 0.004 0.005

Friction

PF1 FF

2

10

1

20

0.1

600

2,000

30

4

500 5

1,500

3

1

10

350 400

Sample A (Shingosen new silky)

4

3

4

20

0.5

0

250 300

800 900 1,000

1.5

0.4 0.5

2

3

9 10

1

LT WT*

1

2

700 8

3

0

150

1.5

500 5

–1

Puckering value

ey* Fb* eb*

–2

Puckering value

Tensile –4 (Break) Fy* 60

370

380

390

400

410

430

440 450 460

150

100

Thread sample: *log scale

420

No.39,

470

480

200

No.07,

490 250

No.33

2

1

0

Figure 9. The Data Chart Plotting Mechanical and Frictional Parameter of the Thread for Men’s Suiting

MJ

Thread sample:

had smaller strain than the thread made in England. Both of them were hard to bend and their frictional force was larger than the thread made in England. These threads are considered hard threads. Here, the definition of a soft thread is “a thread which has low elastic modulus and large residual strain, low bending rigidity, low frictional force”, and the definition of hard thread is “thread which has high elastic modulus and small residual strain, high bending rigidity, high frictional force”.

MB MS Machine type G2-1

MP

P1

Figure 10. Effect of Machine and Thread on Puckering Value Which Shows Mean Puckering Value and Standard Deviation

tension used in this experiment and machine types is shown. The thread tension was measured by two measurement tools, named “Somfy Tec”, made in France. The maximum measurement loads of the two are 100 gf and 500 gf, and accuracy of measurement are 2 gf and 10 gf respectively. The thread tension for the MJ type of machine was highest. Four fabric samples, SHINGOSEN “New Silky” type, “New Worsted” type, Wool/PET blend and SIRO spun, were chosen and their mechanical properties are shown in Table VI. Fabric length (warp direction) is 30 cm and width (weft direction) is 10 cm. The fabric sample is folded in two and sewn at a distance of 1 cm from its end along the warp direction. The distance between the two marks on the fabric is 20 cm. The sewing speed is 1,000-1,500 rpm. An Organ needle, size 11, was used. These four fabric samples were sewn by the threads mentioned above using four types of machine. After that the seam shrinkage was measured by auto measure tester[11] and the seam pucker was evaluated using a standard model which rates seam puckering as grade 0 (no pucker), grade 1

n

The tread tension was measured by two measurement tools n To clarify the effect of the mechanical and frictional properties of threads on the seam appearance, sewing experiments were carried out on each of the threads shown in Figures 8 and 9.

Sewing Experiments Because the best condition of thread tension varies with the machine type, in this article four types of machine, MJ, MB, MP and MS, were used. In Table V, the interrelation between thread

25

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

threads were expressed by closed symbols. Seam pucker was strongly influenced by the type of machine, and the effect of thread property on the seam pucker is different for each machine.

Sample B (Shingosen new worsted) Puckering value

3

2

1

Conclusions

0

(1) Commercial sewing threads have a wide range of mechanical properties.

MJ

MB MS Machine type

MP

(2) Every apparel factory has a different opinion about the best sewing thread. But there are mainly two types of opinion about good threads, a “soft type” of thread and a “hard type” thread. (3) Experiments were carried out by using different types of machines for lock-stitch. Seam pucker was strongly influenced by the type of sewing machine, and the effect of thread property on pucker is different for each machine. There may be an optimum thread for each type of machine. (4) Needle-thread tension of the sewingmachine must be adjusted to obtain the optimum seam. It has been found that serious seam pucker occurred with the sewing-machine where the optimum needlethread tension was high. For such sewingmachines, hard thread is more suitable than soft thread for reducing pucker. In the case of machines for which the lower thread tension is the best condition, soft thread is more suitable than hard thread.

Sample C (W/E blend) Puckering value

3

2

1

0

MJ

MB MS Machine type

MP

Sample D (SIRO spun) Puckering value

3

2

1

0

MJ

Thread sample:

MB MS Machine type

MP

No.39

No.33

No.07

n

Figure 11. Effect of the Machine and Thread on Puckering Value Which Shows Mean Puckering Value and Standard Deviation

References 1. Shimazaki, K., “Studies on Seam Strength – Tensile Strength of Seam Sewed by Hand”, Japanese Resource Association of Textile End-Uses, Vol. 20, 1979, p. 317. 2. Matsuo. M. and Aoki, I., “Study on the Seam Strength – Prediction of the Tensile Strength of a Curved Seam”, Japanese Resource Association of Textile End-Uses, Vol. 22, 1981, p. 191. 3. Utiyama, S., Mori, Y., Yamamoto, T. and Noshi, H., “An Experimental Study of Seam Strength”, Japanese Resource Association of Textile End-Uses, Vol. 20, 1979, p. 153. 4. Kawakami, K. and Masuda, Y., “Studies on Men’s Homewear (Part 6) – The Tensile Strength of Seat Seams”, Annual Reports of

(modest pucker), grade 2 (normal pucker), grade 3 (serious pucker), and grade 4 (very serious pucker)[12]. This experiment was carried out to achieve as good a seam as possible for each machine and thread.

Results Seam shrinkage has a close relation to the evaluated Puckering Value. So the results of this experiment were expressed by using Puckering Value as shown in Figures 10 and 11, which show the interrelation between machine type and Puckering Value for each thread; the soft threads were expressed by open symbols, the hard

26

VOLUME 6 NUMBER 2/3 1994

5.

6.

7.

8.

9.

Studies, Osaka Joshigakuen Junior College, No. 24, 1980. Amirbayat, J., “Profile of Lock-stitch Seams: A Theoretical Study”, Textile Resource Journal, Vol. 61, 1991, p. 119. Ajiki, I., “Analysis of Seam Structure by Mechanical Model – On the Good Stitch”, Japanese Resource Association of Textile End-Uses, Vol. 27, 1986, p. 208. Iwasaki, K. and Tanaka, M., “The Fatigue of Sewing Thread by Tensile and Bending Stress”, Annual Report of the Science of Living, Osaka City University, Osaka. Jojima, E., Kusakabe, A. and Mashima, T., “Effect of High-speed Sewing on the Thread”, Proceedings of the Jissen Women’s University Department of Home Economics, Vol. 23, 1986, p. 33. Kawabata, S., Niwa, M. and Matsudaira, M., “Characterization of Mechanical Properties of the Yarns Produced by New Spinning and the Effect of the Yarn Properties on Fabric

Handle”, Proceedings of the 13th Textile Research Symposium at Mt Fuji, Vol. 36, 1984. 10. Kawabata, S., “Analysis of Fabric Hand of High-quality Apparel Fabrics on the Basis of Objective Evaluation Technique and the Design and Development of the Highperformance Fabrics”, Research Project, Grant-in-Aid for Scientific Co-operation Research, The Textile Machinery Society of Japan, Vol. 156, 1987. 11. Shiomi, S., Niwa, M. and Kawabata, S., “Prediction of the Shrinkage Recovery of Fabrics after Steam Pressing – Part 2: Practical Use of the Prediction Theory”, Seni-Gakkaishi, Vol. 32, 1979, p. T1. 12. Yamada, Y. and Niwa, M., “ A Study on Seam Puckering Part 1: The Effect of the Basic Mechanical Properties on Seam Puckering in Fabrics for Men’s Suits”, Japanese Resource Association of Textile End-Uses, Vol. 34, 1992, p. 142.

27

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

A Study of the Objective Measurement of Fabric Lustre Yao Mu Northwest Institute of Textile Science and Technology, Xian City, People’s Republic of China, Pan Xunqi China Textile University, Shanghai, People’s Republic of China, and Lu Mingzhe Northwest Institute of Textile Science and Technology, Xian City, People’s Republic of China

Introduction

(2) The degree of total reflection brightness of a fabric surface, that is the characteristics of group target of brightness degree in an objective evaluation; (3) The traits of fabric reflection include the rate between outside surface and inside surface of reflection of the fabric surface’s fibres, that is “quality of lustrous degree” in an objective evaluation (above called “quantity of lustrous degree”). This is relative to the variety of fibre, the shape of fibre cross-section, the surface traits of fibre and the weave of fabrics, etc.; (4) The degree of chromatic dispersion in the various kinds of beams of different wavelengths of light in fabric reflection, that is “degree of colour” in an objective evaluation. At present many researchers or measurement instruments research or measure only one, or at most two, aspects of content[4-10], so the compatibility problems of the kinds of objective evaluation and subjective evaluation, considering the fine distinction in a large category or in similar category fabrics, cannot be solved. Meanwhile, fabric lustre still affects fabric colour (different evaluations exist for identical fabrics, with different dyeing, and the

The study on fabric lustre has been continuing for more than 50 years. Many research members have put forward many test methods and analysis methods[1,2]. Various kinds of special lustrous instruments were made by many companies. Some of the instruments have several parameters, which are rough – and sensitivity is not high – that can distinguish between several kinds of fabrics, but cannot distinguish micro-differences in the same kind of fabric. Other instruments can distinguish micro-differences in the same kind of fabric, but cannot describe characteristics of different kinds of fabrics. The authors comprehensively studied fabric lustre for more than 20 years with physics, physiology, psychology and engineering, and found the visual evaluation of fabric lustre includes four concepts[3]: (1) The difference in the degree of reflection brightness when the fabrics are observed from different directions, that is the characteristics of a group target of relative lustrous degrees in an objective evaluation; International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 28-31, © MCB University Press, 0955-6222

28

VOLUME 6 NUMBER 2/3 1994

differences also exist for fabrics with the same dyeing but different lustre in evaluation). The authors further improved the multidimensional and modifying-angle lustrous degree instrument which was manufactured by Northwest Institute of Textile Science and Technology, having improved its sensitivity, the overall measured brightness degree curve, and increased parallel and vertical polarization measurement, linked a micro-computer to gather and calculate data. The problem between the measurement of fabric lustre and its evaluation was thus further solved.

Objective Measurement Method The equator two-dimensional and angle conversion brightness curve was measured by the multi-dimensional conversion lustre instrument. The reflection curve of the reflection half-sphere (Figure 1) was cut out. Incidence direction is -45˚. Seven essential parameters, from the 27 obtained from the curve, were picked up, calculated according to the intensity of incidental light Ii. (1) 0 relative lustrous degree G0 = Im/I0 (2) -25 relative lustrous degree G-25 = Im/I-25 (3) -65 relative lustrous degree G-65 = Im/I-65

Research Method

(4) peak value reflection factor Gm = (Im/Ii)×100

Fabric lustre is the result of an objective physical quantity exerting physiological quantity through the eyesight of people and by the decision and determination of comprehensive brain psychology. So correct research methods should be relating objectivity to subjectivity (simulating subjectivity to a certain extent) and reaching the identity between them.

(5) equator reflection factor Gp = [( ∫-9090 Idθ)/180I2] ×100

(6) diffuse reflection factor Ga = ( ∫0θ m Udθ)/Im

(7) peak width rate Gw = [ (θ1 − θ2)/180] ×100 Two kinds of light source under three kinds of conditions were measured for all, that is: (1) natural light source;

(2) plane polarized light source (the direction of polarization of polarizer, that is the electric vector plane, is vertical to the fabric plane; the polarimeter is parallel to the direction of polarization of the polarizer);

Subjective Evaluation Method Fifty kinds of fabrics of different materials (cotton, wool, linen, silk, chemical fibre), different weaves (plain weave, twill weave, satin weave and various other types), different weight per square metre and different colour were chosen, and these samples’ lustre degree from the lowest to the highest, light reflection from the darkest to the brightest, the gentlest to the extreme, were also considered. Experts were organized to assess the subjective evaluation marks of fabric after thorough identical training. The evaluation had two aspects:

(3) plane polarized light source (the direction of polarization of the polarizer is vertical to the fabric plane; the polarimeter is vertical to the direction of the polarization of the polarizer). There are two other conditions of the plane polarized light source[9,10] but they are insignificant in the analysis of fabric lustre by theoretical and practical demonstration.

(1) Evaluate individual reflection light brightness in respect of the total, for ten general classes from 1 to 10, each class having ten smaller classes; mark from 0.0 to 10.0, the brighter, the higher mark, named M1.

I Im 0.9 Im

I–25

(2) Evaluate individual lustre degree according to brightness difference rate from different directions, for ten general classes from 1 to 10, each having ten smaller classes, mark from 0.0 to 10.0, the higher difference rate, the higher mark, named M2.

Io

I–65

–90

On the other side, comprehensive evaluation was also done according to types of cotton, wool and silk.

–65

–25

0

θ1 θm θ2

90

Figure 1. Reflection Curve of the Reflection Half-sphere

29

θ(°)

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

analysis method, new unrelated variables Z1p, Z2p, Z3p, Z4p, Z5p, Z6p, Z7p, are linearly composed, and selecting the two unrelated variables with the biggest contribution rates Z1p, Z2p, the calculated formula is

Management of the Data and the Results A certain relation among the seven parameters exists under every measurement condition, so regression analysis is needed after management. First, every reading of the parameters is normalized. If the fabric sample number is i (1 to 50), the parameter number is j (1 to 7) and the test condition is k (1 to 3), then the normalized parameter is X1' =

Z k = C1 + C7

Xi − X . σi

X ji – X ji X 1i − X 1i + ... + C j + ... σ 1i σ ji X7i − X7i . σ 7i

Coefficient Cj is presented in Table I. Linear regression of objective evaluation values M1 and M2 are individually done by using these variables, so a basic evaluation parameter value is obtained:

In the formula, Xi is the actual measured value of the number i sample of a certain parameter at a certain condition. The average value of the sample of 50 is Xi; standard deviation of the 50 samples is σ1. Kinds of normalized parameter, at the condition of natural light source, are linearly composed unrelated new parameters Z1, Z2, Z3, Z4, Z5, Z6, Z7, according to main-composition analysis method. The two unrelated variables Z1 and Z2 with the biggest contribution rates are selected among the parameters. After the parameters of plane polarized light source are individually normalized, the ratio of parameters between vertical and parallel polarization directions is calculated. Regarding the ratio as an original variable, according to main-composition

M = b0 + b1Z1 + b2Z2 + b3Z1p + b4Z2p. Coefficient b2 is presented in Table II.

Conclusion (1) According to the above-mentioned, among all kinds of fabrics’ single basic evaluation parameters, M1 is most closely related to Z2 and mainly indicates the total reflection brightness of a fabric’s surface. M2 is closely related to Z1 and mainly indicates the

Cj

Z1

Z2

Z1p

Z2p

C1 C2 C3 C4 C5 C6 C7

0.42860 0.43363 0.42125 0.41175 0.35086 0.38007 0.11714

–0.094753 –0.10620 –0.022726 0.24229 0.51213 0.19947 0.78642

0.35324 0.15823 0.34388 0.15585 0.56875 0.30242 0.54102

–0.49043 –0.09045 –0.39303 0.57164 0.00302 0.49753 0.15062

Total contributing rate

0.6933

0.9052

0.4082

0.7423

Relevant parameter G0 G–25 G–65 Gd Gm Gw Gp

Table I. Coefficient Value Cj

M1 M2

b0

b1

b2

b3

b4

Composed coefficient

Determining coefficient

6.40000 3.96000

0.27549 0.80277

1.10633 0.70811

0.37152 0.04656

0.12531 0.03700

0.93032 0.88555

0.80549 0.78419

Table II. Coefficient Value b2

30

VOLUME 6 NUMBER 2/3 1994

M2 –

10

2.

-

5

–

8

3.

-

6

–

6

-

4.

2 –

4

3

1 -

4

7

–

2

5.

–

8

-

–

6

-

4

–

-

–

-

–

-

2

M1

6.

10

Key: 1= 2= 3= 4= 5= 6= 7= 8=

Pure cotton fabrics Mercerized cotton fabrics Bast fabrics Pure wool fabrics Silk satin fabrics Silk plain fabrics Wool-lick chemical fibre fabrics Silk-lick chemical fibre fabrics

7. 8.

9.

Figure 2. Space Co-ordinates

brightness difference rate of a fabric’s surface from different directions.

10.

(2) There is obvious regionalism of the trait value of fabrics on the two-dimensional space coordinate diagram (Figure 2). Not only the quantity but also quality of the lustrous degree, to a certain degree, are indicated. (3) Fabric lustre is a complicated concept. M1 is still affected by colour. Dyestuffs of deepercolour fabrics have high absorbability to incidental light and the value of M1 is low. On the contrary, the “colour degree” should be researched.

Quality Model”, Journal of Textile Machinery, No. 25 , Japan, 1969, p. 207. Inagaki, K. and Akakawa, N., “The Relationship between the Sight-Feeling Evaluation of Fabric Lustre and Physical Expression Methods”, Journal of Textile Machinery, No. 25, Japan, 1969, p. 215. Minzhe, L. and Mu, Y., “The Fabric Lustre and Lustrous Feeling”, Journal of Northwest Institute of Textile Science and Technology, No. 4, 1991, p. 9. Mizutani, “The Theory about Lustre Measurement and Fabric Lustre”, Buying Industry, No.12, Japan, 1984. p. 697. HaiQuan, Z., Ling, Z. and Mu, Y. “Discussing of Fibre Optical Trait”, Journal of Textile Research, Vol. 9 No. 2, 1988, China, p.14. ShiXin, X., “A Study of Fabric Lustre and Fabric Lustre Instrument”, Journal of Textile Research, Vol. 9 No. 9, 1988, China, p. 196. Hang, X., “Fabric Lustre”, Dyeing, No.1, 1988, p. 44. Ling, Z. and Mu, Y., “A Study of the Nature of Fibre Reflection and Perviousness to Light”, Journal of Northwest Institute of Textile Science and Technology, No. 1, 1986, p. 8. Guangji, W., YiGin, Z., Jinbou, Y. and FanChao, K., “The Lustre Measurement of Worsted Goods”, Journal of Textile Research, Vol.13 No. 12, 1992, China, p. 14. Mu, Y., Minzhe, L., and SuChan, J., “A Study of the Measurement of Textiles’ Lustre”, Journal of Northwest Institute of Textile Science and Technology, Vol. 2 No. 3, 1991.

Further Reading Inagaki, K. and Akakawa, N., “The Mechanism Model of Fabric Lustre”, Journal of Textile Machinery, No. 29, Japan, 1971, p. 217. Nakasato, H. and Meguro, M., “The Sight-Feeling Inspection of Fabric Lustre and Reflection Character”, Journal of Textile Machinery, No. 25, Japan, 1969, p. 241. RaoTing, Z. and Ketai, F., Guild of Poly-Statistic Analysis, Beijing Science Press, 1984, pp. 322-39.

n References 1. Inagaki, K. and Akakawa, N., “The Observation of Single-Fibre of Polarization Light Reflection

31

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Influence of Wool Content on Properties of Blended Fabrics Ji˘rí Militk´y and Vladimír Bajzík Department of Textile Materials, Textile Faculty, Technical University, Liberec, Czech Republic

5-sulphoisophthalic acid. The fibre properties are presented in Table I. The combed intimately blended yarns were constructed. The resulting yarn fineness was 25 × 2 tex and twist levels were 520 turns per metre. Blended yarns with 0, 15, 30, 45 and 100 per cent of wool content were prepared. The plain weaves were fabricated from these yarns. The resulting areal weights, the used setts of warp and weft and the fabric thickness are summarized in Table II. Fabrics were finished by usual woollen type technology. Details about fabric production and finishing are described in [4].

Introduction Wool/polyester blended fabrics are widely applicable in the production of materials for clothing purposes. The main problem is to find such combinations of each component so that the resulting fabrics may have suitable utility properties, surface properties and handle. For constructing of optimally blended fabrics it is necessary to predict the influence of wool and polyester components on the above-mentioned properties of the final textiles. There are a lot of publications which discuss the effect of wool/polyester composition in blended fabrics on selected utility properties[1-3]. The resulting dependencies usually are in the graphic form and are not applicable for designing purposes. Obviously, the examined fabrics are not prepared in comparable conditions, either. In this article, especially prepared plain weave with similar areal weight 200gm–2[4] were used. The standard utility properties, surface properties, total hand value and thermal permeability of fabrics were determined. An attempt to use a simple generalized mixing rule to describe the properties of blended fabrics was made.

Selected Utility Properties A lot of properties connected with fabrics applicability were measured. With respect to the aim of this work, only such properties were selected for which the wool content significantly changed the measured quantities (95 per cent confidence intervals of means for various wool levels were not crossed). It means that

Material

Fibre type

Wool fibres (W) and modified high-shrinkable polyester fibres named Velana S (P), are used for the preparation of woollen type fabrics. Velana S is a modified fibre containing the sodium salt of

Fineness (tex)

Wool (W)

0.63

21.2

49.9

Velana (P)

0.47

36.8

54.1

Table I. Properties of Used Fibres

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 32-36, © MCB University Press, 0955-6222

32

Break Tenacity elongation (mN.dtex–1) (%)

VOLUME 6 NUMBER 2/3 1994

Wool content (%) 0 15 30 45 100

Sett (10 cm–1) Warp 176 177.6 188.8 186 202

Weft

Areal weight (gm–2)

Thickness (mm)

152 152 155 159 180

195.8 201 208 202 201

0.41 0.5 0.6 0.63 0.52

Table II. Geometric Fabrics Characteristics

statistically significant differences which are influenced by wool content exist among mean values of properties.

Objective Handle Evaluation For objective handle evaluation the Kawabata KES system was adopted. From the resulting properties the total hand values THV were computed according to Kawabata’s methodology[5]. A simple program in spreadsheet software Quattro Pro for THV computation was created. Resulting THV values are given in Table III.

Tenacity (T) Textile strength was measured on an Instron tensile tester in accordance with the test standard. Strength is expressed as break force of a 5 cmwide sample. The resulting mean values are given in Table III. Break Elongation (Be ) Break elongation was measured together with strength. Resulting mean values of relative break elongation (per cent) are given in Table III.

Surface Properties of Textiles It is well-known that organoleptic properties of fabrics are highly correlated with surface properties. Therefore the method for assessment of the fabric roughness and friction resistance by using a special adaptor to the Instron tensile tester was selected[6]. The principle is to register the force course S [mN] needed to move the metal disk along the fabric surface which is fixed vertically on the metal desk. This force course presents many local minima SLi and local maxima Sui. The surface roughness can be calculated as a mean difference between maxima and minima

Abrasion (A) Fabrics were abraded according to test standards on the Accelerator device. Abrasion was characterized by relative loss in weight (per cent). Results are summarized in Table III. Heat Permeability (Hp ) Heat permeability was assessed on a special device ALAMBETA. The mean values of heat permeability Hp [Wm–2 K–1] are summarized in Table III.

Wool content (%)

T (N)

Be (%)

A (%)

0 15 30 45 100

810 684 600 523 303

33.3 35.1 34.8 29.4 30.3

2.9 6.0 7.9 8.7 9.6

R = KR (Σi SUi – Σι SLi)

(mN).

(1)

THV

Hp (Wm–2K–1)

R (mN)

U (mN)

2.823 2.307 2.695 2.202 3.168

31.3 29.5 29.5 26.2 31.1

90.58 96.16 99.44 102.36 86.7

158.49 176.53 228.4 233.97 220.2

Table III. Experimentally Determined Average Values of the Selected Fabric Properties

33

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Here KR is proportionality coefficient (recount to the force unit). The second surface characteristic is average friction resistance defined as U = KF (Σi SUi + Σi SLi)

(mN)

If variance of measurement of local extremes (SU, SL) is equal to σ2, then variances of a surface characteristic σ2s lie in the interval 2σ2 < σs2 < 4σ2[6]. It was determined on the basis of previous experiments that σs2/σ2≈2. The average values of R and U for individual fabrics are presented in Table III.

(2)

where KF has the same meaning as constant KR in the equation (1).

10

32

9 Heat permeability (W.m. K)

31

Abrasion (per cent)

8 7 6 5 4

29

28

27

3 2 0.00

30

0.20

0.40

0.60

0.80

26 0.00

1.00

0.20

Wool (per cent)

0.80

1.00

0.80

1.00

0.80

1.00

36 35 Break elongation (per cent)

750

650 Tenacity (N)

0.60

Wool (per cent)

850

550

450

350

250 0.00

0.40

34 33 32 31 30

0.20

0.40

0.60

0.80

29 0.00

1.00

0.20

0.40

0.60

Wool (per cent)

Wool (per cent) 105

240

225 Friction resistance (mN)

Roughness (mN)

100

95

90

210

195

180

165

85 0.00

0.20

0.40

0.60

0.80

150 0.00

1.00

0.20

0.40

Wool (per cent)

0.60

Wool (per cent)

Figure 1. Dependence of Some Properties of Wool/Polyester Blended Fabric on Wool Content

34

VOLUME 6 NUMBER 2/3 1994

Tenacity Tenacity dependence on relative wool content was described as means of the general mixing rule (equation (4)), as by means of the logarithmic mixing rule (equation (5)). It was calculated on the basis of statistical tests that the logarithmic mixing rule is more convenient. Corresponding residual variance is s = 56.4. The model of logarithmic mixing has the form

Property Prediction of Blended Fabrics It is evident that fabric properties will depend on the content of single components for constant parameters of fabric construction, in some way. The simplest model of linear mixing is based on the hypothesis that the influence of a given fibre type on blended fabric property is directly proportional to its content. If the value of blended fabric property is denoted as PB, the same property of pure wool fabric as PW and property of pure polyester fabric Pp, the linear mixing rule has the form PB = WPW + (1–W)PP

PB = 303.04W × 803.97(1–W). As 95 per cent confidence intervals of both parameters PW and PP cover experimentally determined values PWe = 303 (for W = 1) and PPe = 810 (for W = 0), we may use experimentally determined boundary strength-for-strength prediction of blended fabrics based on the logarithmical mixing rule.

(3)

where W is the relative content of wool fibres (0 < W < 1). This equation is not applicable in cases when dependence PB on W is either concave or convex. The simple model of the general mixing rule can be used in these cases: m

m

m

PB = WPW + (1–W)PP

n

(4)

Abrasion dependence on relative wool content is concave

where m is the constant of mixing. This equation describes convex (for m < l) and concave (for m > l) type dependencies. A special case for m = 0 is the so-called logarithmic mixing rule, which has the form PB = pWWPP(l–W).

n It is interesting that this rule was proven for twoand-two twill weave, too. On the other hand, for other types of polyester fibres, the logarithmic mixing rule is only a rough approximation.

(5)

Generally, the problem of estimation of parameters m, PW and PP leads to the problem of non-linear regression. Program ADSTAT for IBM personal computers was used for parameters estimation.

Break Elongation Break elongation dependence on relative wool content was modelled by a relation (4). On the basis of non-linear least squares the following model was determined:

Results and Discussion

PB–2.2 = W29.9–2.2 + (1–W)34.4–2.2.

Testing of the convenience of mixing models for particular properties was realized on the basis of experimental dependence from Figure 1. It is perceptible that results of friction resistance, roughness, handle and heat permeability in pure wool fabrics are markedly different from the tendency of other points. This tendency may be connected with the fact that used fabrics have various cover factors. The cover factor calculated for pure wool fabric is 80.8 per cent and for pure polyester fabric is only 73 per cent. Therefore values for 100 per cent wool roughness, friction resistance and heat permeability were omitted in the calculations. The THV values are not dependent on wool content in a systematic way and therefore are not analysed here.

Corresponding residual variance is equal to 7.47. We may notice that experimental values PWe and PPe lie in the 95 per cent confidence interval of the model parameters. Abrasion It is evident that abrasion dependence on relative wool content is concave; this means that parameter m in equation (4) will be positive. By the means of least squares this model was found: PB4.5 = W9.894.5 + (1–W)2.894.5. Corresponding residual variance is equal to 0.31. Similarly, as in the last cases, experimental

35

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Parameters

PW

PP

r

Roughness

117

91.3

0.98

Friction resistance

342

157.7

0.99

Heat permeability

21.2

31.42

0.93

Conclusion It was verified that a simple general mixing rule is possible for modelling the dependence of selected blended fabric properties on relative wool content. Even though the rule is formal, we may use it easily for prediction purposes and for the designing of optimum blends.

Table IV. Parameters of Linear Mixing

n

values PWe and PP lie in the range of pe confidence intervals of model parameters for W = 1 and W = 0.

References

The Other Performance Characteristics The linear mixing rule for the description of dependence of roughness, friction resistance and heat permeability on relative wool content was used. Model parameters PW and PP estimated by means of least squares are shown in Table IV. Correlation coefficients are presented here too. Tests of the significance of correlation coefficients show that a linear mixing model on level α = 0.05 is significant for roughness and friction resistance only. Critical significance level α = 0.071 resulted for heat permeability. Therefore heat permeability appears to be statistically independent of wool content.

1. Sayere, J.F., Modern Textiles Magazine, No. 4, 1956, p. 38; No. 5, 1956, p. 58. 2. Hargreaves, H.A. and Brooke, B.I., Journal of Textile Institute, No. 2, 1963, p. 112. 3. C˘irli˘c, J., Doctoral Thesis, MTI Moskva, 1965. 4. C˘irli˘c, B., Thesis, TU Liberec, 1983. 5. Kawabata, S., The Standardization and Analysis of Hand Evaluation, 2nd ed., HESC, the Textile Machinery Society of Japan, 1980. 6. Militk´y, J., Bajzík, V. and Sto˘cková, H., Proceedings of the Conference on Textile Testing 90, Hradec Králové, 1990.

36

VOLUME 6 NUMBER 2/3 1994

Features and Characteristic Values of Fabric Compressional Curves Mitsuo Matsudaira and Qing Hong Kanazawa University, Japan

Introduction

Theoretical Model

Compressional property of fabric is one of the most important properties of fabric basic mechanical properties and is closely related to fabric handle. That is the compressional property is concerned with the softness and fullness of fabric and also with the surface smoothness of fabric. Further, the compressional property is related to fabric structure, the surface property of fibres and/or yarns, and the lateral compressional property of fibres and/or yarns. Characteristic values of fabric compressional property which are concerned with fabric handle are proposed by Kawabata[1] such as: LC (linearity of compression); WC (compressional energy); RC (compressional resilience); T0 (fabric thickness at 0.5 gf/cm2); and Tm (fabric thickness at 50 gf/cm2). However, as LC, WC and RC are calculated from integral values of compressional curves (the relationship between pressure and fabric thickness), there is a case in which similar characteristic values are obtained from two different compressional curves. T0 and Tm are parameters of compressional deformation and cannot represent compressional curves. Therefore all the information concerning the compressional property of fabric is not always included by the parameters such as LC, WC, RC, T0 and Tm. In this article, new characteristic values which can express fabric compressional curves more sufficiently are proposed and the relationship between the compressional property and those characteristic values is investigated.

Fabric Cross-section Model of Compression The process of compressional deformation of fabric is assumed to obey the following model, as shown in Figure 1. Fabric is a kind of assembly of yarns and/or fibres and is composed of fibres and space (air) between fibres. In the process of fabric compression, space between fibres decreases considerably and the space is supported by the internal stress of fibres and the frictional force between fibres. If fabric is compressed by a compression plate, the compression plate first comes into contact with hairs and/or protruding fibres on the fabric surface. This region corresponds to the deformation from (a) to (b) in Figure 1 and could be considered to be an elastic region. Compressional force increases linearly with fabric thickness. If the compressional force increases further, the force overcomes inter-yarn and/or inter-fibre static frictional force and slippage of fibres takes place. As the result, space between fibres and fabric thickness decreases. This region corresponds to the deformation from (b) to (c) in Figure 1. If the space decreases sufficiently, then lateral compression of fibre itself begins. This region corresponds to the deformation from (c) to (d) in Figure 1 and could be considered to be the initial elastic region of fibre itself. Features and Regression Curves of Compression Curves According to the assumed model of fabric compression above, the compressional process of fabric is considered to consist of three parts. The first step of the compressional curve is the linear

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 37-43, © MCB University Press, 0955-6222

37

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

(a)

(b)

(c)

(d)

Figure 1. Cross-section Model of Fabric

approximated curve is drawn using the equations (1)-(5) and the result is shown in Figure 2. As the shape of this curve is quite similar to the real curve, it is considered that regression constants ai and bi are characteristic values of fabric compressional property which includes features of fabric compressional curves. Regression constant bi is related to deformation (thickness change), and is considered to be more important. The value of b1 is affected by hairs and protruding fibres on fabric surface, and considered to be the initial compressional modulus of fabric. The value of b2 is concerned with radius of curvature at the second step of the compressional curve. If b2 becomes small, the radius of curvature becomes large, then the amount of inter-yarn and/or inter-fibre slippage become large. As a result, the fabric becomes soft in compression. If b2 is large, vice versa. The value of b3 is related to the initial elastic modulus of fibre in lateral compression. The value of b4 is concerned with recoverability of fibre lateral

region and is assumed to follow the regression line (1). The second step is the non-linear region and is assumed to follow the exponential curve[2,3] (2). The third step is the linear region and is assumed to follow the regression line (3). The first step of the recovery curve corresponds to the region of elastic recovery and is assumed to follow the regression line (4). The second step corresponds to the region of recovery from interfibre and/or inter-yarn friction and is assumed to follow the exponential curve (5). The third step is the region at which instantaneous recovery is impossible. These regression curves are shown as follows: y

= a1 + b1x

(1)

y

= a2 exp(b2x)

(2)

y

= a3 + b3x

(3)

y' = a4 + b4x

(4)

y' = a5 exp (b5x)

(5)

where y compressional force (gf/cm2)

50

y' recovering force (gf/cm2) (3)

40

x deformation (mm) P gf/cm2

ai, bi regression constants. Characteristic Values Which Can Represent Shapes of Compressional Curves It is easy to understand fabric compressional property from the graph of relationship between compressional force and fabric thickness. However, it is difficult to compare several fabrics objectively or to analyse fabric handle objectively from such a graph[1]. Therefore it is very important to choose characteristic values (mechanical parameters) of fabric compressional property which can represent the shapes of the compressional curves perfectly. A typical

(4)

30

20 (5)

(2) 10 (1) 0

0

0.1

0.2

0.3

∆T (mm)

Figure 2. Compression Curves Calculated Using the Equations (1)-(5)

38

VOLUME 6 NUMBER 2/3 1994

compression. The value of b5 is considered to be recoverability which overcomes inter-yarn and/or inter-fibre frictional force in fabric.

Method of Regression Regression lines and/or curves were obtained by the least squares methods. The boundary between each step was decided by the point at which the correlation coefficient between experimental points and corresponding points on the regression curve showed the maximum value. The pressure range of the first step examined here was from 0 to 0.5-5.0 gf/cm2 depending on fabric sample. The third step was from 30-40 gf/cm2 to Pmax.

Experiments Method of Experiments Compressional curves of fabric are obtained from the KES Compression Tester. Data are taken into a personal computer using an A/D module. Data are read when the pressure becomes larger than 0.01 gf/cm2. Fifty data are averaged to obtain a datum. The recovering process starts after the difference between two successive data becomes more than –0.01 gf/cm2. All the experiments are carried out at 20 ± 0.3°C temperature and 65 ± 3 per cent relative humidity.

Results and Discussion Observation by Microscope Fabric surface under compression was observed by microscope using sample No. 7 in Table I. It is clearly seen that surface hairs and/or protruding fibres are bent and compressed easily by a little pressure, corresponding to step 1 (from (a) to (b) in Figure 1). Then yarn thickness was measured and it was shown that yarn thickness changed (increased) mostly from pressure 1.0 to 5.0 gf/cm. This corresponds to step 2 (from (b) to (c) on Figure 1). The change of yarn thickness was plotted with pressure and is shown in Figure 3. Fibres move a lot in the initial region of the second step. It was clear that fibres did not move in the third step. Assumed model above agrees well with these observations.

Samples Samples were collected to cover a wide area of fabrics, such as end-uses (men’s suits, women’s fine dresses, sportswear, industrial uses, etc.), fibre materials (wool, cotton, silk, polyester, nylon, etc.), fabric structure (woven, knitted, bonded, non-woven, etc). A total of 146 samples were collected. Then 12 samples were chosen for investigation from those samples by hand from the point of different compressional softness. Outlines of samples are shown in Table I. Sample no. Structure

Fibre material

Year count (tex)

End-uses

Both: 25 (1/40) Both: 14.8 (40/1)

Ladies’ suits Outdoor sports

Both: Nylon; 5.6 (50d/17f) × polyurethane; 4.4 (40d) Both: 10 (90d) Both: 15 (135d) Both: 23.3 (210d) Both: 20.8 (2/48) Both: 2.3 (21d) Both: Cu; 9.8 (60/1) × Ac; 8.3 (75d) Both: 9.8 (60/2) Warp 13.3 (120d) Weft: 19.7 (30/1) Warp 8.3 (75d) Weft: 13.3 (120d)

Swim suits

1 2

Ponte-de-roma Plain weave

3

Tricot

4 5 6 7 8 9

Plain weave Satin weave Plain weave Twill weave Satin weave Twill weave

10 11

Tubular Plain weave

Wool (100%) Aramide (55%) + polyester (45%) Nylon (80%) + polyurethane (20%) Polyester (100%) Polyester (100%) Polyester (100%) Wool (100%) Silk (100%) Cupra (55%) + acetate (45%) Cotton (100%) Rayon (100%)

12

Plain weave

Cupra (100%)

Table I. Outlines of Fabric Samples

39

Ladies’ dresses Ladies’ suits Ladies’ suits Men’s suits Ladies’ dresses Jackets Sportswear Ladies’ dresses Blouses

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

sample 1 is smaller than that of sample 2. This means that wool fibres are softer than aramide and polyester fibres in lateral compression. In the comparison between samples 3 and 4, the slope of sample 3 is smaller than that of sample 4. This means that nylon and polyurethane fibres are softer than polyester fibres in lateral compression. These results are in accordance with the lateral compressional modulus of single fibres measured recently by Kawabata[4].

200

Pressure (gf/cm2)

50 40 30 20 10

n

0 0

10

∆d (µm)

20

30

The second step agrees well with exponential curves

Figure 3. Changes of Yarn Thickness by Fabric Pressure

n

Features of Compressional Curves Four representative examples of experimental and regressed compressional curves are shown in Figure 4. Sample 1 is the thickest and the softest fabric in compression. Sample 2 is thick and the hardest fabric in compression. Samples 3 and 4 are both fine fabrics, having similar compressional deformation. However, their compressional curves are different from each other. Regression constants and conventional mechanical parameters for these samples are shown in Table II. The first step of compressional curves is regressed well by linear equations. The range of the first step of compression curve is different in each case. In the comparison between samples 1 and 2, deformation of thickness in step 1 of sample 2 is larger than sample 1 and compressional modulus is smaller. In the comparison between samples 3 and 4, the deformation in the first step is similar. However, the slope of sample 3 is larger than sample 4 and initial compressional modulus is larger. The second step of compressional curves agrees well with exponential curves. Radius of curvature of sample 1 is small and the deformation is large. This is caused by the large amount of slippage between yarns and/or fibres because of loose fabric structure. Sample 2 has a smaller radius of curvature and the deformation is smaller. This is caused by the small amount of slippage between yarns and/or fibres because of compact fabric structure. In the comparison between samples 3 and 4, the radius of curvature of sample 3 is larger than that of sample 4 and is elastic and softer in compression. The third step of the compressional curves is regressed well by linear equations. The slope of

The first step of the recovery curves is regressed well by linear equations. The difference of slope between the third step of the compressional curve and the first step of the recovery curve is large for sample 1. Hysteresis of sample 1 is large and that of sample 4 is relatively small. The second step of the recovery curves agrees well with exponential curves. Radius of curvature of sample 1 is larger than that of sample 2 and shows a large amount of deformation at the recovery process. This means that sample 1 is recoverable in compressional deformation. In the comparison between samples 3 and 4, the radius of curvature of sample 3 is larger (b5 is smaller in Table II) than that of sample 4. This means that sample 3 is more elastic in the recovery of compressional deformation. Experimental curves and regressed curves agreed quite well also for the other eight samples and the assumed model proposed above was proved to be valid. Analysis of Characteristic Values by Principal Component Analysis As compressional property of fabric is an overall property, it is difficult to express it solely by one characteristic value. Although some characteristic values are necessary to express whole compressional property, those characteristic values usually have correlation with one another. Therefore principal component analysis[5] is a useful method by which to draw the features of fabric compressional property.

40

VOLUME 6 NUMBER 2/3 1994

50 Sample 1

Sample 3

Sample 2

Sample 4

P gf/cm2

40

30

20

10

0 50

P gf/cm2

40

30

20

10

0 0

0.1

0.2

0.3

∆T (mm)

0

0.1

0.2

∆T (mm)

Figure 4. Calculated and Experimental Compression Curves

the results are shown in Table III with eigenvalues and proportions. It is clearly shown that more than three-quarters of the information is explained by the first principal component and a further 20 per cent is explained by the second principal component. As accumulated proportions of first and second principal components are 96 per cent, almost all

From the assumed model mentioned above, regression constants b2 and b5 are mostly concerned with shapes of compressional curves of fabric. Conventional parameters; LC, WC, RC and T0 – Tm are also very important characteristic values for understanding fabric compressional property[1,6]. So these six characteristic values are used for variables and Parameter b1 b2 b3 b4 b5 LC WC RC T0 T0 – Tm

Unit

Sample 1

Sample 2

Sample 3

Sample 4

– – – – – – gf cm/cm2 per cent mm mm

14.18 24.12 760.3 171.7 31.25 0.666 0.234 60.35 1.282 0.242

10.58 34.80 1,540.0 311.6 55.79 0.488 0.152 51.78 0.439 0.250

38.90 28.06 979.6 189.7 48.06 0.753 0.166 63.37 0.761 0.120

29.24 42.21 1,363.0 246.8 57.07 0.697 0.127 69.67 0.390 0.112

Table II. Regression Constants and Conventional Characteristic Values of Fabric Compressional Property

41

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

component z2 has large coefficients of x3(LC) and x5(RC) as shown in equation (7). If x3 and x5 increase, compressional curve becomes more linear and hysteresis between compressional and recovery curves decreases.

information about fabric compressional property can be explained by these two components. Eigenvectors of each principal component are shown in Table IV. The first and second principal components are expressed as follows:

n

First principal component: z1 = 0.438x1 + 0.437x2 + 0.292x3 – 0.438x4 + 0.357x5 – 0.461x6. ● Second principal component: z2 = – 0.270x1 – 0.305x2 + 0.685x3 – 0.280x4 + 0.534x5 – 0.035x6. ●

Compressional and recovery curves are considered to consist of three steps

(6)

n

(7)

In the first principal component, if x6(T0 – Tm), x4 (WC) become smaller and x1(b2), x2(b5) become larger, then z1 becomes larger. If x6 and x4 are smaller, thickness deformation and integrated value are smaller. If x1 and x2 become larger, the radius of curvature of compressional and recovery curves becomes smaller and fabric stress increases rapidly with increase of compressional deformation. These fabrics are hard in compression. Sample 2 corresponds to this case. Fabrics are soft in the opposite case and sample 1 corresponds to this. The second principal

Eigenvalue

Proportion (%)

Accumulated (%)

1

4.5731

76.2

76.2

2

1.1867

19.8

96.0

3

0.1430

2.4

98.4

4

0.0955

1.6

100.0

5

0.0016

0.0

100.0

No.

Fabrics having these features will show good elasticity and recoverability in compression. Sample 3 corresponds to this case.

Conclusion In order to clarify features and characteristic values of fabric compressional curves, a compressional model of fabric was proposed and its validity was investigated precisely. The following results were obtained: ●

Table III. Eigenvalues and Proportions

Parameter b2 b5 LC WC RC T0 – Tm

(x1) (x2) (x3) (x4) (x5) (x6)

The compressional and recovery curves are considered to consist of three steps, respectively. The first and third steps of the compressional curve and the first step of the recovery curve are approximated by linear equations. The second steps of the compressional and recovery curves are both regressed by exponential curves. The third step of the recovery curve is the region at which instantaneous recovery is impossible. Calculated compressional and recovery curves agreed very well with the experimental curves.

z1

z2

z3

z4

z5

0.4383 0.4369 0.2917 –0.4382 0.3567 –0.4611

–0.2698 –0.3049 0.6850 0.2803 0.5341 0.0348

–0.1703 0.1369 0.5951 –0.0247 –0.7445 –0.2081

–0.5667 0.3745 0.1874 0.5445 –0.0554 0.4518

–0.5275 0.6159 0.0457 –0.2825 0.1428 0.4901

Table IV. Eigenvectors of Each Principal Component

42

VOLUME 6 NUMBER 2/3 1994

Characteristic values proposed here are strongly related to features of the shapes of fabric compressional curves. Especially, characteristic value b2 at the second step of the compressional curve is concerned with fabric structure and inter-yarn and/or interfibre friction. Characteristic value b5 at the second step of the recovery curve is concerned with recoverability from deformation caused by inter-yarn and/or interfibre friction in fabric. ● According to the principal component analysis, the first and second principal components can explain almost all the information about fabric compressional property. The first principal component is related to the amount of deformation and fabric softness. The second principal component is related to fabric elasticity and recoverability in compression.

References

●

1. Kawabata, S., The Standardization and Analysis of Hand Evaluation, 2nd ed., The Textile Machinery Society of Japan, Osaka, 1980. 2. Kawabata, S., Niwa, M. and Kawai, Y., The Journal of the Textile Machinery Society of Japan, Vol. 31, 1978, p. T74. 3. Kawabata, S. and Niwa, M., Journal of the Textile Machinery Society of Japan, Vol. 31, 1978, p. T88. 4. Kawabata, S., Journal of the Textile Institute, Vol. 81, 1989, p. 432. 5. Okuno, T., Kume, H., Haga, T. and Yoshizawa, T., Methods of Multivariate Analysis, Japan Science and Technology Union Publishing Co., Tokyo, 1981, p. 159. 6. De Jong, S., Snaith, J.W. and Michie, N.A., Textile Resource Journal, Vol. 56, 1986, p. 759.

43

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

An Analysis of the Fluttering of a Membrane Suspended in Air Flow Hirokazu Minami, Yasuo Okuda and Sumio Kawamura Osaka City University, Japan

inversely deformed in the wind-tunnel[1]. The directions of the uniform air flow are from left to right. As a fundamental study, the initial membrane shape, which has l, and the sag, Sa, between the leading and trailing edges and the membrane shape fluttering in the uniform horizontal air flow of speed U, modelled as shown in Figure 1. The origin of the co-ordinates is the point of the leading edge, and the directions of the x- and yaxes are the flow direction and vertical direction respectively. The following expression, which describes the position of the membrane element at the time t, h(x,t) shown in Figure 1, can be used for an approximate expression to the fluttering configuration with the sinusoidal wave satisfying the supporting condition:

Introduction At this symposium last year, the authors presented a report[1] of experimental investigation on the characteristics of the fluttering of a membrane, a PTFE-coated glass fibre fabric for the practical use of membrane structures, suspended in windtunnel flow. In the experiment, the membrane which had been statically suspended with a sag between the leading and trailing edges was suddenly transferred to a state of fluttering when the flow speed, U, gradually increased and then reached the value defined as lower critical flutter speed, UFL. Subsequently, during further increase of U the state of the fluttering was sustained and subsequently transferred to a static stable state, that is the state of the membrane being vertically deformed upwards, when U reached the value defined as upper critical flutter speed, UFU. Pursuant to the above experiment the authors have been performing a theoretical analysis on the fluttering of an object. The method of the analysis we are now applying is a fundamental approximate method. The explanation on the method and the results of comparison between the theoretical and experimental results are presented in this article. Then the effectiveness of the method is discussed.

h = a{sin(kx – ωt ) + sin(ωt )}

(1)

where a is the amplitude, ω the circular frequency and k = n(π)/l (where n = 2, 4, 6,…). Therefore

l U

y

The Method of Analysis

h x

The Model of Analysis The configurations of a PTFE-coated glass fibre fabric membrane are suspended, fluttering and

Sa Length, l

Figure 1. Modelling for Analysis

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 44-50, © MCB University Press, 0955-6222

44

VOLUME 6 NUMBER 2/3 1994

the wave period, nw, equals 2(π)/ω and the wave phase speed, Vw, equals ω/k. The magnitude of the amplitudes near the leading and trailing edges observed in the experiment[1] were not equal to that near the centre between supports, but equation (1) expresses equal magnitudes throughout the length. In the experiment the amplitude near the leading edge was less than in the central region of the membrane, and also somewhat less near the trailing edge. But it can be said that equation (1) roughly approximates the experimental a configuration.

upper side, the following relation is obtained using equation (2): Φ 2 – Φ1 = ∫0xγ m (ζ , t )dζ

(0 < x < l ).

( 4)

The pressure, ∆p, working on the membrane element, is obtained by the unsteady Bernoulli’s equation as the value of the difference between the pressure pl at the lower side and p2 at the upper side of the membrane. The expression is ∆p = p1 – p2 = ρU (u2 – u1 ) + ρ

∂ (Φ 2 – Φ1 ), (5) ∂t

where ρ is the density of the air flow. Then, substituting equations (2) and (4),

The Expression for the Pressure on the Membrane

∆p = ρUγ m ( x, t ) + ρ

According to the result of the observation on the fluttering in the experiment, it can be said that, though the amplitude was not low compared with the wave length, the amplitude was at the same time not so high. So, in the present analysis, it is considered that the relation between the amplitude and the pressure created by the flow may be approximately derived as a proportional relation. The air flow is assumed to be a perfect fluid flow. The membrane elements move across the xaxis. But, as an approximation, an assumption that the pressure and the flow speed along the element are considered to be the values given on the x-axis is introduced. The flow velocity components in x and y directions at a position in the x-y plane are denoted as U + u and v respectively. u and v are the variable velocity components in an unsteady flow field. u1 and u2 denote the values of u on the lower side and upper side of the membrane element respectively. vi denotes the mean value of the y-directional velocity components on the lower and upper side of the element. As the thickness of the membrane is considered to be infinitesimal, the strength of the vortex per unit length, γm (where the positive direction is clockwise), can be described in the domain 0 < x < l as:

γ m ( x, t ) = u2 – u1.

∂h ∂h +U . ∂t ∂x

(6)

Now, since it has been considered that the vortex γm is distributed along the membrane, the circulation, Γm, exists around the whole membrane, and this circulation varies in time. Being governed by Lagrange’s theorem on vorticity, the vortex must be shed at the trailing edge corresponding to the variation of the Γm during unit time. The strength of the shed vortex, per unit length in the domain l < x is denoted as γw γm and γw. Distributed along the x-axis induce the velocity components at the positions of the membrane in the y direction. The sum of these velocity components equals the vi in equation (3). This relation is induced as vi =

1 l γm 1 ∞ γw dξ + dξ . ∫0 ∫ 2π ξ – x 2π l ξ – x

( 7)

The γm in the above relation can be obtained as follows by the inversion formula of the Cauchy integral[2]:

γm =

1  l – x  0.5 l  ξ  {2 ∫0   π x   l –ξ

 + ∫l∞ 

ξ    ξ – l

(2)

0.5

γw dξ}. x –ξ

0.5

vi dξ x –ξ (8)

The final expression for ∆p is obtained by substituting equation (8), into which equation (3) has been substituted, into equation (6). Here, by the condition that γw occurring at the trailing edge during unit time equals γm shed during the same time, the following relation exists:

vi equals the sum of the velocity of the membrane itself and the y-directional component of U which is induced by the slope of the membrane. Therefore vi =

∂ x ∫ γ m (ζ , t )dζ . ∂t 0

(3)

Denoting the variable velocity potential as Φ1 at the lower side of the membrane and as Φ2 at the

–

45

∂ l ∫ γ m dx = γ w (l, t )U . ∂t 0

( 9)

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

And so γw flew out from the trailing edge without the variation of the strength therefore the following relation comes out: x – l γ w (l , t ) = γ w  x , t = .  U 

static wave state at USB, has no choice but to transfer to a dynamically stable state, that is the fluttering state. The lift around the membrane per unit width is described as ρUΓm by the KuttaJoukowski theorem. USB is given by the following expression derived from an equilibrium condition between the lift, the membrane weight per unit width and the vertical supporting forces per unit width:

(10)

Using the expressions here,

γ m = γ mx ( x ) exp(iωt ), Γm = Γ0 exp(iωt ), Γ0 = ∫0l γ mx dx,

U SB = {0.5π (mgL + akT (1 – β ) /(1 + a 2k 2 )0.5 ( ρak ∫0l H c dx )}0.5

γw is obtained with relations (9) and (10) as iω γ w = Re  – Γ0 exp(iω ((l – x ) / U )) exp(iωt ),  U  (11) where

where L is the total membrane length. This L is derived by expressing the suspended shape using a hyperbolic function as L = ∫0l [1 + {κSr sin h(κ ( x / l – 0.5))}2 ]0.5 dx, (14)

Γ0 = aG2 / G1,

where Sr=Sa/1, the sag ratio, and κ = 2.6339. In equation (13), m is mass unit area of the membrane surface, g is the acceleration of gravity and Hc is obtained as

l – x  0.5 ω G1 = π – ∫0l  g2 ( x )dx, U  x  G2 =

l 2 ∫0l  

– x  0.5 g1 ( x )dx, x 

l ∫0 

ξ  g1 ( x ) =   l –ξ

0.5

l – x Hc =   x 

1 x –ξ

g2 ( x ) =

ξ    ξ – l

0.5

0.5

l ∫0 

ξ    l –ξ

0.5

1 cos( kξ )dξ . x –ξ

And β is –1 for n = 1, 3, 5, … and is +1 for n = 2, 4, ... Then it can be seen that USB is independent of the membrane stress T (Tensile force per unit membrane width) when the mode of wave is of n = 2, 4, ... When T has been given, the amplitude a can be approximately calculated from

{(Uk – ω ) cos( kξ ) + ω}dξ , ∞ ∫0 

(13 )

1 l –ξ  sin ω  dξ. (12) x –ξ  U 

The above-mentioned method for the induction of the expression for ∆2 is generally prepared for a problem in aero-elasticity and has been already applied by Thwaites[3] and Yamamoto[4] who analysed the fluttering of a sail, Tamada[5], who analysed the flapping of a fabric, and Kunieda[6] who analysed the fluttering of a membrane roof.

a = 2{ Ll (T / E – l / L + 1)}0.5 /(π n ),

(15 )

where E is the tensile elastic modulus of the membrane. The Response to the Flutter The authors apply Hamilton’s principle for proceeding with the approximate analysis which has been adopted. The energy terms, that is the values relating to the membrane at the time t, are the kinetic energy, Tw, the strain energy relating to bending, UM, and to extension, UT, the potential energy worked by the gravity force during time in which the membrane travels from the position in the suspended state to that in the fluttering state, WG, and the work achieved by the pressure, Wp. Setting the domain of integration as one period, nw, the expressions of the principle are

The Flow Speed, USB , at the Static Equilibrium of Forces It may be considered that the theoretical critical flow speed, USB, at which the membrane suspended between supports is transferred to a fluttering state is the speed when the state of static equilibrium between vertical forces comes into existence. The vertical forces are the whole weight of the membrane, the supporting forces and the lift. When U grows larger than USB , the lift begins to overcome the weight of the membrane so that the supporting forces are neglected. Thus, if the membrane continues to sustain such a wavy configuration, the membrane, having been in the

I ( a) = ∫0nw (Tw – U M – UT + WG + Wp )dt,

46

(16)

VOLUME 6 NUMBER 2/3 1994

The Results of the Analysis

dI = 0. (17) da Considering that the oscillation displacement of the membrane is not small and assuming that the stress T is distributed uniformly throughout the membrane and is constant in the period, each energy term is described as Ra =

 ∂h  2  Tw = 0.5m ∫0l 1 +     ∂x   UM =

0.5

The present analysis was performed on PTFEcoated glass fibre fabric, referring to the experimental condition previously reported[1]. The material constants are m = 1.268 N sec 2 / m 3 , D = 0.00331 N m, E = 3.69 × 10 5 N/m.

2

 ∂h    dx,  ∂t 

2 2 2  ∂h   l  ∂ h D∫0  2  1 +     

Though, in the previous experiment, the oscillation mode similar to that of n = 4 was observed on some flow speed and supporting conditions, most of the observed modes were near to or could be considered equal to that of n = 2. So for the present analysis the mode of n = 2 in equation (1) is used. Some integrations are included in the aforementioned equations relating to pressure. For their numerical integrations Gauss’s integral formula is applied. For obtaining the sufficient convergence of the numerical integrations, the domain of the integrations is finely subdivided, especially more finely near the singular points, and in each subdivided domain Gauss’s formula is applied.

–1.5

dx,  ∂x  where D is the bending stiffness per unit width of the membrane,  ∂x 

2

0.5    2  l   ∂h     U T = 0.5 EL ∫0 1 +    dx  / L – 1 ,    ∂x        0.5

2   l  ∂h   WG = – mg ∫0 1 +    ∂x     [ h – S a {cos h (κ ( x / l – 0.5 )) – 2}]dx ,



 ∂h  Wp ∫0l ∫0t ∆p1 +  

The Result of USB Figure 2 shows USB – Sr relations calcuated by equation (13). For reference, USB variations on the other cases of n, on an assumed condition of T = 0, are also shown in this Figure. It is seen that USB depends only on sag ratio Sr and mode n. This means that the wave configuration of a mode

0.5

2

  ∂h     dτdx.  ∂x    ∂t  The following approximations are used for numerical calculations:   ∂h  2  1 +    ∂x  

0.5

2   ∂h   1 +    ∂x  

–1.5

2

 ∂h  = 1 + 0.5  ,  ∂x  2

 ∂h  = 1 – 1.5  .  ∂x 

PTFE-glass fabric 10

USB (m/sec)

In the case where U is a certain value, so the amplitude is obtained from the condition (17) into which a value of frequency is given, that the response curves, frequency-amplitude curves, can be given. Ra is a non-linear function of a. Therefore the relaxation technique is used as follows. Denoting the amplitude known after the jth calculation as aj, the unknown amplitude for the next iterative calculation is written as

n=4

u

n=2 x

5

n=1

a j +1 = a j + ∆a,

y

0

(18)

0

0.1

n=3

0.2

Sr

0.3

where dR   ∆a =–  Ra/  a    da   a 

.

Figure 2. Relation between Sag Ratio (Sr) and Flow Speed (USB) at Static Balance

( 19)

= aj

47

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

PTFE-glass fabric l=1.780m Sa = 0.227m

40 20 0

∆ p (KN/m 2 )

U

Figure 5. For reference in this Figure, the examples of the pressure distribution are also shown. The flow speed region in which the flutter occurred in the previous experiment[1] in this case was between 11(UFL) and 18(UFU)m/sec. In the experiment, the xdirectional bearing forces which continuously worked with somewhat contant magnitude through one period were measured. These forces per unit width of the membrane at the leading edge were roughly 20 to 100 N/m. Though the other type of bearing force, which was larger than the above stated forces and rather impulsive, was also measured, the endurance of this type of force was a relatively small portion in one period of the flutter oscillation. The somewhat contant bearing forces stated above can be considered to be approximately equal to the membrane stress during the oscillation because the slopes of the membrane at the leading edge are small. Referring to the experimental values of the somewhat constant bearing force, the response curves in Figure 4 theoretically show that the fluttering occurs in the wide range of U, 5.2(USB) to about 40 m/sec. However, it seems that the fluttering in the experiment does not occur as expected or as the theory predicts. As to the reason for this, it can be considered that U greater than USB may be required on the process of the gradual increase of U from zero up to the onset of the fluttering. There are also differences of the state of the air flow between experiment and theory. The differences are the existence of the small turbulence in the wind-tunnel flow and the existence of the effect of side edge of the membrane adjoining the vertical side plates with

–20

Figure 3. Statically Balanced Wave (n = 2) and the Pressure Distribution

at static equilibrium can be attained as a theoretical result at the same value of USB, for instance, in either case of l = 1 m with Sa = 0.2 m and l = 10 m with Sa = 2 m. Rapid increase of USB begins when Sr begins to diminish at less than about 0.06. The USB corresponding to n = 2 and the experimental condition, that is 1 = 1.78 m with Sa = 0.227, is given by the curve in Figure 2 as 5.2 m sec. The pressure distribution along the membrane in the stable state corresponding to this flow speed is shown in Figure 3.

28 26 24 22 20 18 16 14

The Result of the Flutter Response The response curves on the same condition about l and Sa as in Figure 3 are shown in Figure 4 using U as a parameter. These curves indicate the relations between the period nw and stress T which are transformed using equation (15) from the relations between circular frequency ω and amplitude a obtained by equation (17). The wave shape variations during one period are shown in

PTFE-glass fabric Q = 1.780 m S a =0.227 m n =2

8

5.22

U = 40m/sec

T (KN/m)

2

0

0

12

30

10

1

0.5

1.0

1.5

Figure 4. Fluttering Response Curves (Period (nw) – Membrane Stress (T) Relations)

48

n w (sec)

2.0

∆p(KN/n2)

VOLUME 6 NUMBER 2/3 1994

U Time = 0

1 0

PTFE-glass fabric l =1.780 m Sa = 0.227 m U = 14 m/sec nw = 0.72 sec T = 190 N/m

x(m)

1.78

–1

= n w /8

1

= n w /4

0 –1

= 3 n w /8

1

= n w /2

0 –1

= 5 n w /8

1

= 3 n w /4

0 –1

= 7 n w /8

1

=nw

0 –1

Figure 5. Fluttering Mode (n = 2) and the Example of Pressure Distribution

small clearance. In fact, as stated above, the experimental UFL did not coincide with the USB and were measured as a higher value. The response curves in Figure 4 show a characteristic that they begin to be more steep in the region U over 20 m/sec. It is considered that when the membrane is fluttering in the flow of U in such a region it is very easy for the membrane to be transferred by some small disturbance from the fluttering state to an unstable state at which the stress, namely the amplitude, increases to infinite. This transition to the unstable state is interpreted as a phenomenon that in the previous experiment the membrane inversely deformed at the critical flow speed UFU. The UFU in the experiment was 19 m/sec. This value is considerably smaller than the values, 30 to 40 m/sec, at which the response curve becomes almost vertically straight. It seems that the reason was a transition caused by some variation in the oscillation and flow state in the tunnel experiment. So equation (17) is non-linear relating to the amplitude, when the amplitude and frequency are not independent of each other. Therefore the period cannot be determined without appointing the value of the membrane stress T on the response curve. However, the values of the period determined on the response curves corresponding to the experimental stress values previously stated do not differ so greatly from the values that are similarly determined assuming arbitrarily T = 0.

Thus the approximate value of the period can be determined assuming T = 0 on the response curves. Figure 6 shows the flutter frequencies, f, measured in the previous experiments. And Figure 6 shows the f theoretically determined by the response curves in Figure 4 citing the previously stated relatively constant experimental stress values and also the assumed value ON/m. The corresponding phase speed of the fluttering wave, Vw(= ω/k), was calculated and is shown in

f(Hz)

PTFE-glass fabric l = 1.780 m S a = 0.227 m n = 2

2

1

0

10 Experiment With experimental T With T = 0

U (m/sec)   

0

Analysis

Figure 6. U-f Relations by Experiment and Analysis

49

20

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

other assumptions. The experimental results with which the theoretical results were compared were those that have been measured on the oscillations under conditions not completely the same as those of the analysis, such as the wind-tunnel flow having small turbulence and the specimen having finite width. Considering these differences, it could be concluded that the present analysis method is effective for approximate theoretical investigation on the flutter characteristics of a membrane suspended in uniform flows.

PTFE-glass fabric l = 1.780 m S a = 0.227 m n = 2

Vw /U

0.2

0.1

0

10 Experiment With experimental T With T = 0

U (m/sec) 20

  

0

n

Analysis

Figure 7. U-Vw/U Relations by Experiment and Analysis

References 1. Minami, H., Okuda, Y. and Kawamura, S., “Observation of the Fluttering of Membrane in Wind-tunnel Flow”, Proceedings of the 21st Textile Research Symposium, August 1992, pp. 97-100. 2. Muskhelishvili, N.I., Singular Integral Equations, P. Noordhoff, Groningen, Holland, 1953. 3. Thwaites, B., “The Aerodynamic Theory of Sails, I”, Proceedings of the Royal Society , London, Vol. 261A, 1961, pp. 402-22. 4. Yamamoto, K. and Ishimaru, K., Journal of the Japan Society of Aeronautics and Astronautics, Vol. 36 No. 412, 1988, pp. 23341 (in Japanese). 5. Tamada, A., Proceedings of the 21st Japan Society of Physics, 1966, p. 90 (in Japanese). 6. Kunieda, H., “Flutter of Hanging Roofs and Curved Membrane Roofs”, International Journal of Solids Structure, Vol. 11, 1975, pp. 477-92.

Figure 7 with the measured values. The experimental values of f and Vw in these Figures show large variations and peak points between the critical values UFL and UFU. In the previous experiment on the conditions of U corresponding approximately to the peak point, flutter waves having higher modes of n > 2 were observed. On the other hand, the theoretical values of f show a relation almost proportional to U, and therefore that the theoretical VW is constant to U. However, it could be seen in viewing the range of all U plotted in the Figures that these theoretical values roughly coincide with the experimental values.

Conclusion The analysis method presented here is based on the assumption that the method uses an expression of fluttering shape differing slightly from the experimental shape and based on the

50

VOLUME 6 NUMBER 2/3 1994

Non-recovery of Futon Padding after Repeated Compression Sachiko Sukigara Faculty of Education, Niigata University Hiroko Yokura Faculty of Education, Shiga University, and Masako Niwa Department of Clothing Science, Nara Women’s University, Japan Introduction Futon Padding using wool fibres has become popular in Japan over the last few years. The advantage of using wool fibre arises from its superior ability to absorb and release moisture vapour. On the other hand, some problems arise in the decrease of bulk or resilience while using Futon. The purpose of this article is to investigate the phenomenon of non-recovery of wool Futon padding after compression deformation and develop a suitable experimental technique to evaluate it. Simulation tests were also carried out on a miniature model Futon and its results are compared with those of the compression experiments.

Fibre

Diameter (µm)

Length (mm)

Crimp (per cent)

A

Wool

31.8

47.2

36.0

B

Wool

28.8

46.4

24.0

C

Wool

25.1

52.2

29.9

D

Wool

34.8

61.5

27.3

E

Wool

27.4

40.4

29.0

F

Wool

30.5

37.4

46.4

G

Wool

29.7

41.4

45.5

R4

PET

45.0

R6 Cotton 18.1a 9.6b diameter along the long axis of ellipse b diameter along the short axis of ellipse

a

Experimental Compression Properties of Futon Padding Sample preparation. Seven types of wool fibre assembly and pure polyester and pure cotton fibre assembly were used in this study, as shown in Table I. These samples were dried in an oven at 100±2°C for seven hours, then transferred to the conditioned room at 20°C, 65 per cent RH more than 24 hours before the experiment. Compression test. Compression properties of loose wool have been studied extensively from

Table I. Samples

both experimental[1] and theoretical[2] points of view. From these studies, there are three important parameters to consider, the modulus of compression, which is a function of the mass structure, the volume at zero pressure, which might be called filling capacity, and fibre slippage. It is very difficult to obtain the accurate volume at zero pressure because it is influenced

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 51-56, © MCB University Press, 0955-6222

51

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

by the handling of the sample. In this study, initial apparent fibre density (ρA, g/cm3) was chosen as 0.017, which is equivalent to the commercial Futon used under the body. Fibre assembly weighing about 8g was placed in the cell with a cross-sectional area of 63.617cm2. This cell was then set in the Handy compression tester (Karo Tech). Samples were compressed until ρA reached 0.017 g/cm3 and then compression and release cycles were carried out at 1mm/sec. The maximum displacement of the compression plate was 10mm. From this curve, characteristic values such as P0 (pressure observed at ρA = 0.0l7g/cm3) and Pm-P0 were obtained as shown in Figure 1. Also compression energy WC (gf cm/cm2) and resilience RC (per cent) were calculated. Compression tests were carried out three times for each sample as shown in Figure 2. The shape of the first curve is very different from the following curves, so a third curve was used for the analysis. Pressure of 2gf/cm2 and 5gf/cm2 were applied to the sample of ρA = 0.0l7g/cm3, then the compression test was carried out in the same manner. Creep test. A preliminary experiment was designed to decide the creep time. Samples(7cm diameter) were set in the creep tester at the apparent fibre density of 0.017g/cm3, then a 20gf/cm2 weight was applied for 16 hours at 20°C 65 per cent RH and then removed. This pressure of 20gf/cm2 was chosen as the average pressure reported when a man weighing 45 to 65kg lies on the Futon[3]. After removing the weight, the sample was then left in the conditioned room at 20°C 65 per cent RH for 24

P

P (gf/cm2) Sample R4 1

δT

2 3

15

9 cm 10

5

0 2

4

6

8

δT

10 mm

Figure 2. Examples of Pressure Thickness Change Curves

hours and the creep test was again carried out in the same manner. During this experiment, changes of thickness were continuously recorded. The relationship between thickness and creep time was found to be very linear during this time, so creep time was set for three hours. The following characteristic values were obtained: T(0), initial thickness at fibre density of 0.017; T(1), thickness after 1 minute creep; ID, thickness change occurred from T(0) to T(1); T(l80), thickness after 180 minutes under the weight; T(f), thickness after the release of the weight; IR, recovery of thickness from T(180) to T(f) as shown in Figure 3. Figure 4 shows the 20 gf/cm2

P (gf/cm2)

T (0)

Pm

T (1)

3 hours

P = 20gf/cm2 Pm – P0

50

T (0) DR

Thickness (mm)

40

P0

0

δT

10 mm

30 ID

IR

20

T (1) 10

Note: P0: Pressure observed at V = 0.017g/cm3

0 0.1

1

10

100

1,000

t (min)

Figure 1. Characteristic Values Obtained from a Compression Curve. P0 is the Pressure Observed at ρA = 0.017g/cm3

Figure 3. Thickness Changes during the Creep Test

52

10,000

VOLUME 6 NUMBER 2/3 1994

1.00

T(t )/T (1)= – R In t + 1.0 28˚C

0.98

P =3.5 gf/cm2

0.96 Futon

T (t)/T (1)

0.94

Paper: 9 mg water/cm2

0.92 0.90 0.88 One cycle 0.86

Seven days, 20˚C, 65 per cent RH

Seven hours' compression

0.84 0.82

17 hours' release the weight 20˚C, 65 per cent RH 2 3 4 One day

0.80 1

10

100

1,000

Figure 4. Normalized Creep Curve, Thickness as a Function of Time

5

6

Seven days

Figure 5. Procedure of Model Futon Experiment

example of thickness change against creep time. The vertical axis represents the ratio of the thickness at times t and 1 minute and characteristic value R which is the slope of this curve was obtained.

observed for sample B and a small one for samples F and G in the range of pressure selected in this study. The modulus of compression is influenced by the structure of fibre assemblies which might be effected by fibre diameter, crimp and fibre length. In Figure 7, values of Pm-P0 are plotted against fibre crimp at the initial fibre density of 0.017 and 0.025. It is seen that crimpy wool shows a higher modulus than uncrimpy wool samples for both fibre densities. Values of Pm-P0 depend on the initial fibre density as shown in this Figure. Figure 8 shows the relationship between fibre crimp and RC. Crimpy fibre assembly tends to show lower resilience when compared with uncrimpy fibre assemblies. Values of RC at the density of 0.025 were larger than those at 0.017 for samples except samples B and F.

Reduction of Thickness of Model Futon after Repeated Compression A 15 × 30cm model Futon was made containing 13.5g of padding covered by gauze fabric. This is equivalent to 0.15 times the size of commercial Futon. Initial thickness was measured under the pressure of 3.5gf/cm2. Wet paper containing 9mg water/cm2 was placed on top of the Futon, then a 3.5gf/cm2 weight was placed on it for seven hours in the conditioned chamber at 28°C. Water content of the Futon was observed at about 15 per cent during this time. After the release of the weight, the model Futon was left in the conditioned room at 20°C, 65 per cent RH for 17 hours. Thickness of the Futon was measured under pressure at 3.5gf/cm2 before and after applying the weight. This experiment was carried on seven days continuously and then samples were left in the conditioned room at 20°C 65 per cent RH for seven days. A second and third series of this experiment were carried out using the same sample. These procedures are shown in Figure 5.

60

20˚C,65 per cent RH GF

R4

A

DC E RG

B

◆

50

P(gf/cm 2 )

40

◆

30

◆ ◆ ◆

20

◆ ◆ ◆

10

Results Compression Test The relationship between pressure and apparent fibre density is shown in Figure 6. There is not much difference in the observed pressure at the density of 0.017 among wool samples and less than 2gf/cm2. A large density change was

◆

0 0

0.02

0.04

0.06

0.08

0.1

0.12

ρ A (g/cm 3 )

Figure 6. Relationship between Pressure and Apparent Fibre Density

53

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

20

20 gf/cm2

ID

F

18 16

T (0)

IR

T (1)

P m - P 0 (gf/cm 2 )

14 G

12

1 min

180 min

T (1)-T (180)

10 A

8

1

E

0.9

6 C

4

0.8

D

2

Thickness (T/T(0))

B

0 10

15

20

25

30

35

40

45

50

Fibre crimp (per cent)

ρA =0.017 g/cm3 ρA =0.025 g/cm3

0.7 0.6 0.5 0.4 0.3 0.2

Figure 7. The Effect of Fibre Crimp on (Pm-P0) at the Initial Fibre Density of 0.017 (■) and 0.025 (l)g/cm3

0.1 0 A

B

ID/T(0)

C

D

IR/T(0)

E

F

G

R2

R4

R6

(T(1)-T(180))/T(0)

Figure 9. The Ratio of Thickness Changes to the Initial Thickness during the First Cycle Creep Test

Creep Test Figure 9 shows the ratio of thickness changes, ID, IR and T(1)-T(180) to T(0) during the first cycle creep test. For samples B and R6, large thickness changes occurred under the weight of 20gf/cm2 but also indicated good recovery after the release of the weight. The thickness change that occurred during the creep test was very small, having a ratio of less than 0.4, and recovered to almost the initial thickness after the 24 hours’ release of the weight. The obvious relationship was not observed between the R and IR, ID. This indicates that the recovery of fibre assembly is mainly influenced by the fibre-to-fibre friction rather than the viscoelastic property of fibre. Figure 10 shows the influence of the fibre crimp on the ratio, IR/T(0) for both the first and second cycles.

Immediate recovery from the strain, IR/T(0), for wool fibre assemblies decreases with increasing fibre crimp. This result is coincident with the tendency observed in the relationship between RC and fibre crimp as shown in Figure 8. Thickness Change of Model Futon Figure 11 shows the thickness of the model Futon under the pressure of 3.5 gf/cm2 on the first day and the last day for each series of measurements. It is obvious that the initial reduction of thickness had a large influence on the following thickness

80 0.8

70

r 2 =0.8767 0.7

D 0.6

E

B 50

IR/T(0)

RC(per cent)

60

A 40

C

0.5

F 0.4

30

G 0.3

20 10

15

20

25

30

35

40

45

50

0.2

Fibre crimp (per cent)

10

ρA = 0.017 g/cm ρA = 0.025 g/cm

15

20

25

30

35

Fibre crimp (per cent)

IR/T(0):1st IR/T(0):2nd

Figure 8. The Effect of Fibre Crimp on Compression Resilience RC

Figure 10. The Effect of Fibre Crimp on IR/T(0)

54

40

45

50

VOLUME 6 NUMBER 2/3 1994

change after repeated compression. Cotton fibre assembly shows a large recovery during the drying process at 20°C 65 per cent RH for 17 hours. Figure 12 shows the relationship between fibre crimp and the ratio of initial thickness T(0) to thickness after seven, 14 and 21 days. It is seen that crimpy fibre assembly shows less thickness change when compared with the uncrimpy fibre assemblies.

30

Thickness (mm)

25

20

15

Discussion It is reported that short, fine, crimpy wool has significantly more bulk than long, coarse, uncrimpy wool[l]. In this study, fibre crimp is the only parameter which influenced the compression property of Futon paddings. The influence of fibre diameter and length on the non-recovery of compression property was not clear. The large hysteresis observed in the pressure-density curves was attributed mainly to fibre friction rather than to the viscoelastic property of the fibre. Thus there was no clear relationship between R and IR/T(0). Recovery of Futon padding after compression consists of two parts, that is the immediate recovery and the recovery occurring with the passage of time. These parts of the recovery are largely influenced by reversible and irreversible fibre-to-fibre slippage. The irreversible fibre slippage produces the decrease in volume after compression. It was shown, as the results of model Futon for samples B and C, the reduction of thickness observed after first compression being larger than the other samples. This might be due to the effect of irreversible slippage.

10

5 1

1'

7

7'

8

8'

14

14'

15

15'

21

21'

Day F G R-4 R-6

A B C D E

Note: 1,8,15 indicate the first day and 1',8',15' indicate the last day for each cycle

Figure 11. Thickness Change for Model Futon under the Pressure of 3.5gf/cm2 0.8

P = 3.5 gf/cm2 0.7

T/T(0)

0.6

n

0.5

The immediate recovery was investigated from parameters ID and RC

0.4

0.3

n

r 2 = 0.88 0.2 10

15

20

25

30

35

40

45

50

Crimpy and relatively coarse wool fibre assemblies such as Samples F and G are closely packed and thus the effect of irreversible slippage on the compression and release cycle is small. Thus the thickness change of model Futon after repeated compression was also found to be small. The reversible fibre slippage is considered to be related to both the immediate and the later recoveries. The immediate recovery was investigated from parameters ID and RC. Fine and uncrimpy samples showed larger values of

Fibre crimp (per cent)

T (7)/ T (0) T (14)/ T (0) T (21)/ T (0) T (0) :Initial thickness T (7) :Thickness after seven days T (14) :Thickness after 14 days T (21) :Thickness after 21 days

Figure 12. The Effect of Fibre Crimp on the Ratio of Initial Thickness to That after Seven, 14 and 21 Days

55

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

ID and RC than did relatively coarse and crimpy samples. It might be considered that the reversible fibre slippage occurred at the fibre assembly which has fewer fibre-to-fibre contact points. The recovery after the release of the weight is largely influenced by the surrounding atmosphere. Moisture in the fibre assembly is considered to stop the fibre slippage. Thus fibre density increases with an increase in the moisture content. Results obtained from the model Futon show that the release of moisture has a large influence on the recovery of thickness of Futon, especially for cotton sample R6. Moisture inside the Futon is affected by the changes in humidity in the room. During this experiment, all samples had the same environmental history and almost the same water content. Further study is needed to investigate the different phenomena between wool and cotton samples in terms of moisture content. To improve the performance of Futon, fibre blend was necessary.

of Futon. Futon padding, which consists of crimpy fibre, has large apparent fibre density and shows less reduction of thickness compared with those made from uncrimpy fibres. An increase in the apparent fibre density of padding increases the compression resilience. The initial bulkiness is not necessarily related to the recovery from the compression. The moisture inside and outside the Futon has a large influence on the recovery process. Fibre blend might be useful to improve the performance of Futon.

n

References 1. Van Wyk, C.M., Journal of the Textile Institute, Vol. 37, 1946, pp. T285-92. 2. Dunlop, J.l., Carnaby, G.A. and Ross, D.A., WRONZ Communication No. 28, 1974. 3. Yasuda, T., Tanaka, M., Kamitani, Y. and Tanijiri, S., Japanese Resource Association of Textile End-Uses, Vol. 3, 1962, pp. 310-15.

Conclusions Fibre crimp was found to be an important parameter to be considered in the non-recovery

56

VOLUME 6 NUMBER 2/3 1994

Measurement of Water Absorption Perpendicular to Fabric Plane in Two- and Multi-layered Fabric Systems Morihiro Yoneda, Yuko Mizuno and Junko Yoneda Nara Women’s University, Kitauoya-nishimachi, Nara, Japan Introduction

to make the meniscus cause water absorption by fabric specimens. In the measurement, a fabric specimen of 64mm diameter attached to stainless steel plate (Figure 2(b)) is placed on the meniscus of the filter paper to cause spontaneous water uptake by capillary action. At the lower side of the cylinder, a pressure sensor is attached to detect the reduction of water column pressure caused by water absorption by the fabric specimen and the small change of pressure signal is magnified by amplifier and traced by x-t recorder for the time elapsed. Figure 3 shows an example of the measurement of transient water absorption perpendicular to the fabric plane. Ordinate is the amount of water absorption, M(t) (g/cm2) and the abscissa is time, t (s). Figure 4 shows M(t) plotted against the square root of time, t. As shown here, M(t) is well regressed linearly by the square root of time.

Water absorption perpendicular to fabric plane by capillary action is an important property in clothing comfort and hygiene. Layering fabric is an effective way of controlling water absorption properties in fabric systems and an analysis of these systems may be useful for designing comfortable clothing. We have already developed a pressure sensor method to measure water absorption perpendicular to fabric plane with good precision and easy operation[l]. In this article, the results of measurement of water absorption in two- and multi-layered fabric systems using the apparatus based on the pressure sensor method are reported. The results are analysed by non-linear diffusion model using finite element method (FEM).

Measurement Apparatus and Method Details of the pressure sensor method apparatus are shown[1]. Figure 1 shows the schematic diagram of the measurement system. This system is composed of the main part and the reservoir and the two parts are connected by a flexible rubber tube. Figure 2(a) shows the main part of the system. The main part is a cylinder made of stainless steel. The cylinder is filled with ionexchanged water and equipped with a PTFE perforated plate and filter paper at the upper side

Attachment plate with specimen Reservoir Flexible tube Main part

Water

Support Sensor

Figure 1. Measurement System

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 57-64, © MCB University Press, 0955-6222

57

Support

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

(a)

(b)

Drain Stainless steel vessel

Hook

Perforated PTFE plate

Filter paper

Plate

Porous rubber

Fabric specimen

Water

To open air Pressure transducer

Side view

Top view

64

φ

Sensor face

Figure 2. (a) Main Part of the Measurement System; (b) Attachment Plate for Fabric System

Sample

M (x10–2 g/cm2)

4

Details of samples used for single-layer measurement are shown in Table I. The fabric specimens are composed of 14 kinds of plain weave fabrics which are made from various fibre materials (including intimate-blend yarn fabrics). Table II shows the fabric construction and the basic properties of the specimens. The samples used here are relatively thin and lightweight fabrics. The measurements are carried out at a constant temperature and humidity in a room at 20°C 65 per cent (Relative Humidity) RH.

2 Sample number 9 CU

0

0

5

10

15

Time (s)

Figure 3. An Example of the Measurement of Water Absorption

M (x10–2 g/cm2)

4

Sample Fibre no. code composition (%)

Yarn

1C 2R 3S 4W 5 RY 6 CU 7 AC 8T 9P 10 A 11 V 12 C/P

Spun Spun Filament Spun Filament Filament Filament Filament Filament Spun Spun Spun

Yarn count (tex) Warp Weft

M max

Cotton Ramie Silk Wool Rayon Cupra Acetate Triacetate Polyester Acrylic Vinylon Cotton/ polyester 13 C/R Cotton/ ramie 14 W/A Wool/ acrylic

2 Sample number 9 CU

kv

r = 0.97

0

0

1

2 √t

10

s1/2

Figure 4. Plot of M against √t to Derive kv and Mmax

Water absorption properties can be evaluated by the water absorption rate constant kv (g/cm2 √s), slope of the straight line and the maximum water absorption Mmax(g/cm2).

Table I. Details of Samples

58

:100 :100 :100 :100 :100 :100 :100 :100 :100 :100 :100 :35/ 65 :50/ 50 :30/ 70

16.0 19.0 8.5 21.0 8.0 16.0 8.0 13.0 5.5 19.0 20.0 12.5

16.0 20.0 8.5 21.0 13.0 16.0 11.0 16.0 8.0 19.0 20.0 12.5

Spun

7.0

11.0

Spun

43.0

43.0

VOLUME 6 NUMBER 2/3 1994

Sample no. code

Weave density Warp Weft (cm–1)

1C 2R 3S 4W 5 RY 6 CU 7 AC 8T 9P 10 A 11 V 12 C/P 13 C/R 14 W/A

56.0 32.0 44.5 26.0 44.5 37.0 42.0 42.0 43.0 24.0 24.0 56.0 40.0 31.0

Thickness (mm)

Weight (mg/cm2)

0.493 0.413 0.372 0.492 0.186 0.545 0.283 0.223 0.094 0.673 0.607 0.475 0.412 1.210

12.7 12.3 6.9 11.3 7.5 11.4 6.7 9.7 5.2 8.5 11.7 11.4 5.8 24.3

28.0 27.5 36.0 24.5 31.0 34.0 31.0 23.0 32.0 20.0 20.0 27.0 32.5 21.0

Air Thermal resistance conductivity (kPa s/m) (W/m k) 1.186 0.075 0.436 0.111 0.156 0.141 0.346 4.400 0.258 0.042 0.266 0.745 0.039 0.287

0.0142 0.0152 0.0103 0.0133 0.0139 0.0118 0.0116 0.0130 0.0114 0.0115 0.0128 0.0129 0.0120 0.0135

Moisture regain (%)

Porosity (%)

9.35 5.95 6.32 12.05 12.54 11.02 8.15 7.30 0.00 1.98 3.86 3.24 8.02 4.72

83.3 80.1 86.1 82.5 66.6 86.1 82.1 66.1 60.0 89.0 84.9 83.4 90.8 83.3

Table II. Fabric Construction and Basic Properties of Samples

Results for Single-layer Fabrics

(2) wool (W) and wool-acrylic blended yarn fabrics (W/A). C/P fabrics deviate on the higher kv and lower Mmax side and this shows that C/P fabrics have a relatively higher water-transmission ability. In contrast, W and W/A fabrics deviate on the lower kv and higher Mmax side and this shows that they have a relatively higher water-retaining ability.

The results of the water absorption measurement for single-layer fabrics are shown in Figure 5. The ordinate of the scattered chart is maximum water absorption Mmax and the abscissa is water absorption rate constant kv. For almost all materials, Mmax increases with increasing kv. It is noted that there are two exceptions: (1) cotton-polyester blended yarn fabrics (C/P); and

Preparation of Two-layered Fabrics and Notation for These As mentioned in the previous section, the water absorption properties of single-layer fabrics were measured using 14 samples of plain weave fabrics made from various fibre materials. After some preparatory experiments on two-layered fabric systems, wool (W), cotton (C), polyester (P) and vinylon (V) were selected as representative. Water absorption properties of two-layered systems were measured using a combination of these four samples. Two fabrics were stitched together at five points within a fabric plane using thin polyester sewing thread. Figure 6 shows a schematic diagram of the measurement of the two-layered system and its notation. For example, P/C means the system in which the upper layer is polyester (P) and the under layer is cotton (C), which contacts the meniscus directly. It is expected that the under layer may play an important role in water absorption in a twolayered system.

6 20˚C

Mmax (x10–2 g/cm2)

W/A

A

4

CU V C W

2

R

S AC T

C/P

C/R RY

P 0 0

5

kv

10 (x10–3

15

g/cm2 s1/2 )

Figure 5. Result of kv and Mmax for Single-layer Fabric Made from Various Materials

59

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Results for kvI, water absorption constant at the first stage, kvII at the second stage and Mmax, maximum water absorption are shown in Table III. From this Table, it is understandable that a combination of two different fibre materials shows various effects on water absorption behaviour. As an example, the combination effect of cotton (C) and polyester (P) is put into the form of graph for kv and Mmax in Figures 8(a) and (b), respectively. This is because they are very different in fibre properties when reacting with water; these include surface properties and water absorption properties; cotton is hydrophilic and polyester is hydrophobic.

P/C Polyester Cotton

Water

Figure 6. Schematic Diagram of Two-layered Fabrics and the Notation for These

n

For Mmax the additive law may hold

Results for Two-layered Fabrics Measurement was carried out for eight combinations selected from W, C, P and V samples, considering which material is the under layer. As a result, transient water absorption behaviour is divided into three types as follows. Typical results are shown in Figure 7(a) and (b): (1) Single straight-line type (P/C, P/W) (Figure 4, Type 1). (2) Two-straight-line type. The slope of the first stage is greater than the second stage (W/C, W/V) (Figure 7(a) Solid line, Type 2). (3) Two-straight-line type. The slope of the first stage is smaller than the second stage (C/P, W/P, C/W, V/W) (Figure 7(b) Solid line, Type 3).

n From Figure 8(a), it is clear that kv of P/C is governed by cotton at the under layer and kv of the first stage of C/P is governed by polyester at the under layer and there is a transition into the second stage which is governed by the cotton. It is interesting that cotton and polyester intimate blend yarn fabric has the maximum kv value in this group. As for Mmax (Figure 8(b)), it is expected that the additive law may hold and this may be proven by the coincidence in Mmax between C/P and C + P. This fact holds for other systems. This coincidence does not hold in some systems, it may be explained by the imperfection

(a)

(b) 6

Sample: W/C

Sample: C/W

M (x10–2 g/cm2)

M (x10–2 g/cm2)

6

4

2

0

4

2

0 0

2

4 √t

exp

6

8

10

0

4

2

√

s1/2

FEM

exp

Figure 7.

60

FEM

6 s1/2

8

10

VOLUME 6 NUMBER 2/3 1994

Sample

KvIa Mean

CV (%)

Type 1

P/C P/W

10.07 4.62

10.1 2.8

Type 2

W/C W/V

9.81 12.50

11.0 1.4

7.60 2.84

Type 3

C/P W/P C/W V/W

1.86 0.75 5.03 5.39

14.5 22.6 14.5 6.8

8.16 2.35 9.35 12.02

a b

KvII Mean

Mmaxb Mean

CV (%)

CV (%)

3.70 3.14

3.4 4.7

3.0 11.4

5.47 5.42

1.5 2.7

22.4 25.3 7.5 11.6

2.96 1.89 5.68 6.01

3.3 8.0 2.7 8.4

× 10–3 g/cm2 s1/2 g/cm2

Table III. Results of Water Absorption in Two-layered Fabrics

(a)

(b)

Polyester/cotton

Polyester/cotton

10.1

Cotton/polyester I II

2.38

Polyester

7.2

0.5

Cotton and polyester

2.54

Cotton-polyester

13.65 0

2.96

Cotton

8.16

Cotton Polyester

3.7

Cotton/polyester

1.86

5

10

2.88

Cotton-polyester

15

2.05 0

1

k v (g/cm 2 s 1/2 )

2

3

4

M max (g/cm 2 )

Figure 8. Combination Effect of Two-layered Fabrics Cotton (C) and Polyester (P), (a) kv (b) Mmax

where u = concentration, z = position, τ = time, D = diffusion coefficient, D0 = initial diffusion coefficient and σ = non-linearity. Because this type of diffusion equation cannot be solved analytically in the case of two-layered body problem, we solve this numerically by using the

of contact between the two layers. Thus Mmax can be discussed using the simple additive law but, in the case of kv, the situation is expected to be more complicated because it is concerned with the rate process. This will be discussed using a non-linear diffusion model.

Simulation of Water Absorption in Twolayered Fabrics by a Non-linear Diffusion Model

∂u =0 ∂z

We have already shown that water absorption perpendicular to fabric plane is well explained by non-linear diffusion model[l]. Therefore it is expected that a non-linear model may also be applicable to the two-layered system. A model for the analysis is shown in Figure 9. The diffusion equation is as follows:

Layer 2

D2

σ2

l2

Layer 1

D1

σ1

l1

u=1 Flux

Figure 9. A Model for the Analysis of Two-layered Body

61

z

0

Water

∂u ∂  ∂u  =  D(u)  ∂τ ∂z  ∂z  D(u) = D0 (1 + σu)

1

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

finite element method (FEM)[2]. The diffusion equation is formulated using FEM as follows: [C ]

elements is 20 and the number of nodal points is 21. To realize a two-layered body model, material assignment number (MAN) is used. Each element has a MAN to identify to which material it belongs. Using MAN, each set of material constants is introduced to [K] matrix when it is constructed. Figures 11(a) and (b) show an example of the results of calculation for concentration distribution in a two-layered body. Figure 11(a) is the case where D1 (lower side) > D2 (upper side). It is clear that the second layer is the bottleneck to diffuse. Figure 11(b) is the case where D1 < D2. It is clear that the first layer is the rate-determining process of diffusion. Figure 12 shows the validity of Finite Element approximation in the single layer body. Ordinate is the normalized value of absorption and the abscissa is the Fourier number or nondimensional time (F0 = Dt/l 2). Solid line denotes the strict solution obtained from the method of moment[l] and an open circle denotes the FEM solution. Calculation is carried out for various non-linearity, σ as parameter. The agreement between strict and FEM solutions is fairly good and it is concluded that this method is applicable to the two-layered body problem if the degree of non-linearity is not so strong. In the application of the method, the conversion factor from the physical space to computer space has to be decided so that physical behaviour is the same for

∂ {u} + [ K ]{u} = {F} ∂t

where [C] = capacitance matrix; [K] = conductance (diffusion coefficient) matrix; (F) = flux vector; and (u) = nodal point concentration (solution) vector. Formulation for the time domain is carried out using Crank-Nicholson method[3] as follows:  1 [ K ]( n +1) + 1 [C ] {u}( n +1) 2  ∆t 1 1 =  – [ K ]( n ) + [C ] {u}( n ) + {F}.  2  ∆t Because of the concentration dependence of the [K] matrix, this matrix equation cannot be solved, that is the [K] matrix at the next step depends on the concentration at the next step, which is unknown and yet to be solved. To overcome this problem, we use the quasi-linearizing method[4] which consists of two steps to obtain a new solution vector as follows:  1 [ K ]( n ) + 1 [C ] {u}(∗) 2  ∆t 1 1 =  – [ K ]( n ) + [C ] {u}( n ) + {F}  2  ∆t

(1)

 1 [ K ](∗) + 1 [C ] {u}( n +1) 2  ∆t 1 1 =  – [ K ]( n ) + [C ] {u}( n ) + {F}.  2  ∆t

(a)

u (z)

1

(2)

4

3

0.5 1

Figure 10 shows the one-dimensional finite element used in this analysis. The number of

0

0

0.5

z Layer 2

Layer 1

20

2 1

3 2

4 3

5 4

19 20 5

1

18 19 21 Flux

6

u (z)

1

z=0

z =0

5

0.5

4 2

Material assignment MAN

1

(b)

Element Node

5

2

3

1

1

1

1 Layer 1

1

2

2

2

2

0

Layer 2

0

0.5 Layer 1

z Layer 2

Figure 11. Concentration Distribution in Two-layered Body (a) D1>D2 (1:10); (b) D1
Figure 10. One-dimensional Element Used for the Analysis and the Way of Material Assignment

62

1

VOLUME 6 NUMBER 2/3 1994

3.5

1.5

0.5

1

1

0

M(t)

M (t) / M max

3

σ=0

0.5

4

2

0.5

0

0 0

0

1

0.5

√ F0

Strict solution Finite element method (FEM)

Figure 13. Effect of D and σ on Transient Water Absorption

Figure 12. Comparison between Strict Solution (Method and Moment) and Finite Element Method for Single Layer Fabric 1 2 3 4 0

the same Fourier number. The values of D and σ used in the calculation are summarized in Table IV[1]. The broken line in Figures 7(a) and (b) shows the results of calculation of absorption in a twolayered body using FEM and the solid line and the open circle show the observed value. Figure 7(a) is the case that D1(lower side) > D2 (upper side) and Figure 7(b) is the case that D1 < D2. It is shown that in each case the agreement is fairly good, considering the simplicity of the approximation. It is noted that the deviation between the two becomes larger after the flextion point. It is possible that interface between the two layers may affect the rate process but further elucidation will be needed to confirm this hypothesis. Thus transient water-absorption Sample D no. ×10–4 cm2/s 0.13 2.75 0.21 3.82

D0 ×10–4 cm2/s 0.08 1.33 0.10 2.08

1

0.5

F 0 (= Dt / l 2 )

P C W V

1

σ nd

Thickness mm

0.9 2.3 2.3 1.7

0.094 0.517 0.475 0.607

D1

D2

σ1

σ1

1 0.2 1 0.2 1

0.04 3 0.04 3 –

3 0 0 3 0

0 3 3 0 –

Table V. Data for Figure 13

behaviour is explained by applying non-linear diffusion model using FEM. Figure 13 and Table V show an example of calculation of the combination effect of diffusion coefficient, D, and non-linearity, σ. It is clear that water absorption behaviour changes in a complicated manner for various combination of D and σ. This result indicates the possibility of controlling the transient water-absorption behaviour in a two- or multi-layered fabric system.

Results for Multi-layered Fabrics In order to investigate the effect of thickness on kv and Mmax, water absorption in a multilayered system of similar fabrics (two to five layers) were measured for six various fibre materials. Figure 14 shows the relationship between Mmax and number of layers. Mmax is almost proportional to number of layers for all materials. This fact indicates that, in the case of Mmax, the

Table IV. Parameters Used in the Calculation in the Non-linear Diffusion Model

63

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

measurement is carried out correctly and the interface effect is not recognized. Figure 15 shows the relationship between kv and the number of layers. The value of kv increases with an increasing number of layers. This fact may be related to the mechanism of water flow in fibre assembly. Therefore further investigation, especially on the physical meaning of nonlinearity, σ, will be needed to explain this fact.

20

M max (x10–2g/cm2)

15

10

Conclusion Water absorption perpendicular to the fabric plane in two- and multi-layered fabrics was measured using the pressure sensor method. The results were obtained as follows: (1) Water absorption perpendicular to the fabric plane can be measured and evaluated using pressure sensor method with good precision and easy handling. (2) In single-layer fabrics, water-absorption properties can be evaluated by kv, the water absorption rate constant, and Mmax, the maximum water absorption. (3) In the measurement of two-layered fabrics, the following results were obtained: ● Transient water-absorption behaviour is divided into three types for various combinations of fibre materials. ● The absorption behaviour in two-layered fabrics is explained by a non-linear diffusion model using Finite Element Method. (4) In the measurement of multi-layered fabrics made of the same fibre materials, the following results were obtained: ● Mmax is almost proportional to the number of layers. ● kv increases with an increasing number of layers.

5

0 1

2

3

4

5

Number of layers Cotton Silk Wool

Rayon Cupra Acryl

Figure 14. Mmax vs. Number of Layers 4

k v (x10–2g/cm2 s1/2 )

3

2

n

1

References 1. Yoneda, M. and Niwa, M., Sen-i Gakkaishi, Vol. 49, 1993, pp. 243-53. 2. Zienkiewicz, O.C., The Finite Element Method in Engineering Science, McGraw-Hill, Maidenhead, 1971. 3. Yagawa, M., Numerical Analysis of Heat and Flow (Japanese), The Japan Machinery Society edition, Corona, Tokyo, Japan, 1986, p. 45. 4. Murata, K., Basic Mathematics (Japanese), No. 1, 1993, p. 19.

0 1

2

3

4

5

Number of layers Cotton Silk Wool

Rayon Cupra Acryl

Figure 15. kv vs. Number of Layers

64

VOLUME 6 NUMBER 2/3 1994

The Bending Properties of Multi-ply Worsted Yarns Xiaoming Tao CSIRO, Division of Wool Technology, Belmont, Australia only one. Let the single yarn have a radius of as and twist Ts1. When single yarns are folded and twisted, fibres in the single yarn will follow the paths of coaxial doubly wound helices. Treloar[10] has developed a mathematical description of the coaxial double wound helices for multi-ply yarns as shown in Figures 1 and 2. The present analysis is based on the same geometry. Assume the doubly wound helix starts at the origin of the co-ordinate system, then the co-ordinates of the fibre axis can be written as:

Introduction In the study of yarn bending properties, the discrete fibre approach has been adopted by a number of workers[1-5]. The classical model of a single yarn consists of many coaxial helical fibres, with the maximum helical angle on the yarn surface. The minimum bending rigidity of the yarn is regarded as the summation of the rigidity of all coaxial helices without considering the interfibre interaction. Small deflections of a helix have been investigated by means of analytical techniques by Timoshenko[6] and Livesey and Owen[3]. Leaf[1,2] has obtained solutions for large deflections by using a computer method based on the analysis of Love[4] and Konopasek and Hearle[7]. Ly has shown that Livesey and Owen’s equation is accurate to 2 per cent, even for large deflections[8]. Up to date, the treatment of single worsted yarns using a discrete fibre approach has yielded a good prediction of yarn bending rigidity as a function of fibre properties (diameter, ellipticity and length) and yarn twist. However, the predicted flexural rigidity of twofold worsted yarns has been found to be smaller than the observed values[9]. This discrepancy is due to the combination of single and double twists of fibres or an increased fibre-interacton in the multi-ply yarn structure. The analysis reported here is an extension of the earlier work to multi-ply yarn structure in order to investigate the effect of multi-ply structure on the yarn bending properties. It provides a basis for the mathematical description of the bending behaviour of multi-ply yarns in the absence of interfibre interactions.

x(φ) = apcos(φ/λ) + rscos(φ/λ)cosφ x(φ) = – rs sinφcosθpsin(φ/λ) – ap – rs.

(1)

y(φ) = apsin(φ/λ) + rssin(φ/λ)cosφ x(φ) = + rs sinφcosθpcos(φ/λ).

(2)

z(φ) = ±ap(φ/λ)cotθp + rssinφsinθp.

(3)

Equations 1-3 take into consideration the two possible twist combinations, i.e. the Z-Z and Z-S twists. The coefficient λ relates the angular position of a point on the single yarn axis to the angular position of the fibre with respect to the single yarn axis. It is expressed by Treloar[10] as: λ = 2πTs1(L1/L)ascosecΘs.

(4)

L1/L, the ratio of the single yarn length before and after plying can be found from the equation by Carnaby[1]: L1/L=

Geometry of Multi-ply Yarn

tan –1 2πTs1as 2 –1 tan Tas 1 – tan 2 2

1 – tan 2

(5)

where

Doubly Wound Helix In a worsted yarn, normally, two or more identical single yarns are plied but here we will consider

This work has been funded by the Australian Wool Research and Development Corporation. The author wishes to thank Joanne Spence for technical assistance.

International Journal of Clothing Science and Technology, Vol. 6 No. 2/3, 1994, pp. 65-72, © MCB University Press, 0955-6222

65

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

yarn was used for evaluation of this work. The fibre does not seem to follow any periodical pattern over the 25mm of length plotted.

Y P´ D´ φ C A´

B φp

Directional Cosine Matrix As shown in Figure 4, let (X, Y, Z) be a rectangular cartesian co-ordinate system fixed in space and (x,y,z) another co-ordinate system moving with the fibre axis, such that the two axes of the moving system coincides with the principal axes of the fibre cross section (u and v) and the third axis coincides with the tangential direction w. The moving and fixed co-ordinates are related by an orthogonal transformation that can be written using a directional cosine matrix of 3 × 3 elements:

X

O

[–(ap +rs),o]

Figure 1. The Co-ordinate System for a Doubly Wound Helix

Multi-ply yarn axis

Single yarn axis

 x   ux  y =  v    x  z  w x

uy vy wy

uz   vz  wz 

X  Y .    Z 

(6)

The directional cosine matrix can be written in terms of the three Euler’s angles:  ux   vx w x 

A fibre

uy vy wy

uz   cosψ  vz  =  – sin ψ  wz   0

cos Θ 0 – sin Θ   0 1 0     sin Θ 0 cos Θ 

0 0  1 

 cos Φ sin Φ 0   – sin Φ cos Φ 0  ( 7)    0 0 1 

where Ψ, θ and Φ represent the angular twist in the fibre, inclination of the fibre axis, and the angular rotation of the fibre around the Z axis, respectively. If the fibre has no twist initially before spinning then Ψ = 0. {wx, wy, wz} represents the unit tangential vector of the fibre axis. Thus the following relationships can be applied to equation 7. dx dx dφ wx = = = sin Θ cos Φ ds dφ ds

Path of a fibre in a multi-ply yarn

Figure 2. Path of a Fibre in a Multi-ply Yarn

T = 2πTs1 ( L1 / L ) ±

sin ψ cosψ 0

sinθ p cos θ p ap

where + is used for the Z-Z, and – for Z-S twist combination. Generally speaking, unlike the helix characterizing a single yarn, there is no repeat unit for the doubly wound helix in a plied yarn within the length of interest. Figure 3 shows a set of projection plots of a surface fibre in a worsted two-ply yarn calculated from equations 1-3 using the yarn parameters given in the later section. The

wy =

dy dy dφ = = sin Θ sin Φ ds dφ ds

wz =

dz dz dφ = = cos Θ. ds dΦ ds

(8)

Then the triangular values in equation 7 can be obtained as:

66

VOLUME 6 NUMBER 2/3 1994

Y

Z Z 0.3 17.5

0.2

–0.6

–0.5

–0.4

–0.3

–0.2

0.1

15 12.5

X

10

–0.1

20 15 10

7.5

–0.1

5 –0.2

5

2.5 X

–0.3 –0.5

–0.4

(a)

–0.3

–0.2

–0.1

´Y –0.3 –0.2

–0.1

0.1

0.2

0.3

(c)

(b)

Figure 3. Calculated Projection Plots of a Surface Fibre in a Two-ply Worsted Yarn

Z

dz dφ 2 sin Θ = 1 – cos Θ = 1 – ( ) dφ ds 2

(0 ≤ Θ ≤ π ) dy dφ dx dφ dφ ds dφ ds sin Φ = and cos Φ = sin Θ sin Θ

θ

( 9)

z

(10) x

where ds dx dy dz = ( )2 + ( )2 + ( )2 ≥ 0. dφ dφ dφ dφ

ψ

y

Y φ

Curvature and Twist Components Following Love’s derivation[4], the curvature components in the principal directions of the cross section of fibre, p,q are expressed using equation 7: dvy dv dv dφ p= (wx x + wy + wz z (11) ds dφ dφ dφ dw y dφ dw x dwz q= (u x + uy + uz ds dφ dφ dφ

(12)

and the twist component, r, is: du y dφ du du r= ( vx x + v y + vz z . ds dφ dφ dφ

(13 )

X

Figure 4. A Deformed Fibre in a Fixed Co-ordinate System (X,Y,Z) and a Moving System (x,y,z)

(b) follows a single helical path and the surface fibre (a) is a doubly wound helix. The curvature and twist components of the surface fibre vary along the fibre length and have a period which coincides with a revolution in the single yarn (φ = 2π). The central fibre has a much higher constant curvature in the major principal direction (xdirection) than the fibre on the surface. The curvature in the minor direction (y-direction) and twist (z-direction) components are complicated. The curvature component in the y-direction oscillates along fibre for the fibre at the surface and remains zero for the fibre in the centre. At the origin the twist of surface fibre (4.43mm–1) is much higher than that of the central fibre (1.06mm–1). In some part of the fibre, the twist of the surface fibre is lower than that of the central fibre.

Unlike a single helix whose curvature and twist are constant along the fibre, or the angular position of the fibre axis, the curvature and twist components of the doubly wound helix vary with the angular position of the fibre in the single yarn. Figure 5 shows the calculated curvature and twist components of two fibres plotted against angular position. In each diagram, (a) represents a fibre on the single yarn surface and (b) is a fibre at the centre of the single yarn. The central fibre

67

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Mu + A( p – po ); Mv = B( q – qo );

(b)

po

Mw = C(r – ro ).

(14)

4

Since the only external couple, M, acts in a direction parallel to the X-axis: and equations

3

Mu = – Mux ; Mv = – Mvx ; Mw = – Mw x (15)

2 1

Equation 14 can be rewritten using (15) as

(a)

p = Mux / A + po ; q = – Mvx / B = qo ;

∅ 2

4

6

8

10

12

q = – Mw x / C + ro .

(a)

qo

Using Konopasek and Hearle’s development[7], the following equations between the co-ordinates, the directional cosines and curvature/twist components are adjusted with respect to the angular position of the fibre in the single yarn:

0.15 0.10 0.05

(b) ∅ 2

–0.05

4

6

8

10

(16)

dx

12

dφ

–0.10

= wx

ds dφ

;

dy dφ

= wy

ds dφ

;

dz dφ

= wz

ds dφ

. (17 )

–0.15 (a)

dux

(b)

dφ du z

ro

(a)

5

dφ

ds

= (rvx – qwx ) = (rv z – qw z )

dφ ds dφ

du y

;

dφ

= (rv y – qw y )

ds dφ

.

;

(18 )

4

dvx

3

dφ dv z

2 (b)

dφ

1

ds

= ( pwx – rux ) = ( pw z – ru z )

dφ ds dφ

;

2

4

6

8

10

dwx

12

dφ dw z

(c)

Figure 5. Calculated Values of Initial Curvature (po and qo) and Twist (ro) Components of a Surface Fibre (a) and a Central Fibre (b) in a Single Strand

dφ

= (qux – pvx ) = (qu z – pv z )

ds dφ ds dφ

dφ

= ( pw y – ru y )

ds dφ

.

;

(19 )

∅ 0

dv y

;

dw y dφ

= (qu y – pv y )

.

ds dφ

;

(20 )

The 12 differential equations with 12 boundary conditions yield the solution of the co-ordinates and directional cosines of the deformed fibre axis. All the terms used in equations 17-20 are dimensionless. The equivalent dimensioned terms are defined by the following equations:

Bending of a Fibre in a Doubly Wound Helix Governing Equations and Boundary Conditions Consider a fibre bent by an external couple M parallel to the X axis. If po, qo and ro are the initial curvature and twist components of the doubly wound helix before bending, then the three components of the moment in the three directions follows the Euler-Bernoulian relation at any cross section of the fibre:

x ′ = xl; y ′ = yl; z ′ = zl; p ′ = p / l; q ′ = q / l; r ′ = r / l; as′ = as l; a ′p = a pl; rs′ = rsl; A′ = EIo A; B′ = EIo B; C ′ = EIo C; M ′ = ( EIo / l ) M s ′ = sl

68

VOLUME 6 NUMBER 2/3 1994

where E is the Young’s modulus of the fibre, Io, the moment of inertia of a circular cross-section, l is the length of fibre of interest.

A dr r

The Bending Rigidity of a Doubly Wound Helix When a doubly wound helix is bent by a couple M in a plane parallel to the axis of the doubly wound helix, the dimensionless bending rigidity, Bh, is defined as

O

OA=as OA=as sinθp

Bh = MR where R(=R'1) is the dimensionless radius of curvature of the bent neutral axis of the helix in the plane of bending. Unlike the simple case of a single helix studied by Leaf[1], the deformed fibre is not symmetrical about its midpoint of a repeated unit, therefore the neutral axis cannot be easily identified. Hence this geometrical approach can not be used to determine the bending rigidity of the doubly wound helix. However, the radius of curvature of the deformed helix can be determined from the dimensionless work, W, done by the dimensionless M on the axial length of helix (z) of the undeformed helix on the equation: 1 z W= M . 2 R

Figure 6. Integration Over the Normal Cross-section of a Strand in a Multi-ply Yarn

The bending rigidity of a doubly wound helix is EIoBh. Integration over the cross-section of a single strand, perpendicular to the plied yarn axis after the multipfication of the bending rigidity of a doubly wound helix by the number of fibres in an annulus section, one can derive the bending rigidity of a k-ply yarn: as

By′ = k ∫ no cos Θ sin φ p EIo Bh ds.

s

0

(22)

0

Equation 22 can be written in the same form as Livesey and Owen’s[3] equation using a transfer equation arising from the doubly wound helix, Ct:

The total strain energy, u, of the deformed doubly wound helix is written in dimensionless form as u = ∫(

B

By′

Mu2 Mv2 Mw2 ds + + ) dφ 2 A 2 B 2 C dφ

Tex y

where the dimensionless axial length, z, can be calculated from equation 3. By equating the work done to the total strain energy, one derives the following expression for the dimensionless bending rigidity of the doubly wound helix: 1 z Bh = M 2 . (21) 2 u

Ct =

=

NEIo Ct where Tex y

2 as ∫ Bh cos Θrdr as2 0

(23)

where cosθ = wx, defined by equation 8, N is the number of fibres in yarn cross-section and Texy is the yarn tex.

Evaluation of the Theory The Bending Rigidity of Multi-ply Yarn

Calculation of Yarn Bending Rigidity The first approximation of the yarn bending rigidity is the summation of the bending rigidity of all constituent fibres assuming they are all straight, parallel to the yarn axis and have the same diameter. Equation 23 defines a transfer function translating the straight and parallel fibres into the doubly wound helices. Furthermore, Ly and Denby[14] have shown that for a wool worsted single yarn, other three corrections are required, that is, the corrections for fibre diameter distribution, fibre length distribution and fibre cross-sectional shape. If you mean fibre diameter is in microns, the fibre density in g/cm3, the

It is assumed that the bending rigidity of multi-ply yarn is the summation of the bending rigidity of all constituent coaxial doubly wound helices, i.e. the interaction between fibres will be ignored. Let the number of fibres per unit area in a normal cross-section of a parallel fibre bundle be no. Consider an elliptical annulus formed by a crosssection normal to the plied yarn axis as shown in Figure 6, the number of fibres in the annulus is nocosθsinθpds where the area is given by: ds =

2πrdr sin θ p

69

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

numerically by using a subroutine called NSOLVE of the software MATHEMATICA. The bending rigidity of single doubly wound helices is determined by equations 3, 15 and 21. The dimensionless bending rigidity of a fibre at the centre of the single yarn strand is 0.96 while that of a fibre at the surface of the single yarn strand is 0.999. Equation 23 determines the transfer function for the doubly wound helical structure (M=1).

Young’s modulus of fibre in millinewton/mm2, equation 23 can be extended in terms of the first approximation, the transfer function and the three correction factors: By Tex y

= 62.5 × 10 –12 Ed 2 / ρ C f Cl Ce Ct

where Cf is the factor for the fibre diameter distribution, Cl for the ineffective fibre ends, Ce for the ellipticity of the fibre cross-section and Ct the transfer function for the plied yarn structure.

Other Three Correction Factors A correction is needed for the coefficient of variation, skewness and kurtosis of fibre diameter distribution[5]. Based on the measurement of fibre diameter distribution of 41 commercial tops, Lunney and Irvine[13] derived relations between the fibre diameter, coefficient of variation, skewness and kurtosis. It was found that for tops with average diameter ranging from 19 to 32 microns, the correction factor Cf can be treated as a function of mean fibre diameter alone with less than 2 per cent error[5]. In our calculation following Ly[5], Cf is taken as 1.29, based on the measured mean fibre diameter and the measured coefficient of variation of fibre diameter distribution in the experimental yarn. The cross-sectional shape of wool fibres is generally elliptical with the ratio of major and minor radius ranging from 1 to 2. Australian merino wool has an average ellipticity of 1.3[8]. If bent about its major axis, this elliptical fibre is approximately 20 per cent more flexible than a circular fibre with the same cross-section area. However, the bending of a fibre in a yarn is much complicated because of its helical configuration. Ly and Denby have shown[14] that the correction factor for fibre ellipticity increases with yarn helical angle and decreases with the bending moment. It ranges from 0.94 for θs = 10˚ and M = 4 to 1.02 for θs = 50˚ and M = 1. To match the yarn parameter and M the value used was 0.97. When yarn samples were clamped at a gauge length of 10mm, the fraction of fibre ends in the segment was 20/1, where 1 is the mean fibre length in millimetres. Half of the total ends are leading ones which may take up any position in the yarn

Yarn Properties An evaluation of the theoretical analysis has been carried out using a commercial two-ply wool worsted knitting yarn. The yarn has been stored for more than four years and the time is sufficient for the fibres to be set in the doubly wound helices. Table I provides specifications of the yarn sample. The diameter of the double yarn, dy, was measured using a microscope. No yarn tension was applied during measurement. Following Riding’s correction for the radius of the ply axis[12], ap = as= dy/4. The twist angles for the plied and single yarns were derived from the following: tanΘs = 2πasTs ; tanθp = 2πapTp. Other data used for the calculation were: E = 4 × 106 mN/mm2, ρ = 1.31 g/cm3. The yarn bending rigidity was measured by a KESF bending tester using a gauge length of 10mm. Yarn samples were wound on a board to keep them parallel to each other. Approximately 20 yarns were taken for each sample. Three samples were tested. Calculation Procedure for the Transfer Function Ct The helix starts from the origin thus x = y = z = 0 when φ = 0. The directional cosines at the origin are calculated from equation 7 using equations 810. The initial curvature and twist components of the doubly wound helix are determined from equations 11, 12 and 13. The solutions to the 12 first order differential equations 17-20 with 12 known boundary conditions have been obtained

Texy

Ts (tpm)

Tp (tpm)

θs

θp

as (mm)

ap (mm)

Lf (mm)

df(µ)

CVf (%)

R74/2

439 (Z)

192 (S)

25.2˚

11.6˚

0.171

0.171

60.2

20.0

22.4

Table I. Specifications of the Yarn Sample

70

VOLUME 6 NUMBER 2/3 1994

First approximation (10–2mNmm2/tex) 7.33

Transfer function Cf 0.98

Cl

Ce

Cd

0.89

0.97

1.29

Predicted value (10–2mNmm2/tex) 8.00

Measured value (10–2mNmm2/tex 7.98 (SD=0.24)

Measured/ predicted 1.00

Table II. Results

depending on spinning, the other half being trailing ends which tend to be displaced outside of the yarn. Using Ly and Denby’s approximation[14], one-third of the total fibre ends are inside the yarn, thus contributing to the bending rigidity. The coefficient of the effective fibre ends is given by: 1 20 Ct = 1 – . 3 l

2.

3.

4.

Results Table II lists the terms used in the evaluation and the results. The ratio of the measured and predicted values of yarn bending rigidity is 1.00.

5.

Conclusion

6.

A mathematical model dealing with the largescale bending of multi-ply yarns has beeen established using the geometry of doubly wound helices. The analysis is based on the discrete approach of yarn structure without taking into consideration interfibre friction. Thus it may be used to predict the minimum bending rigidity of yarn, bending and torsion strain as well as position of the deformed fibres in a multi-plied yarn structure. Wool worsted yarns have, generally, a lower level of the fibre friction compared with yarns made of other fibres. The present analysis, therefore, may lead to a quantitative prediction of yarn-bending behaviour. Preliminary evaluation of the model is encouraging because of the reasonable agreement between the predicted and measured values of yarn bending rigidity. Further work is under way to assess the accuracy in predicting yarn bending rigidity over a range of yarn parameters.

7. 8.

9.

10.

11.

12.

n 13.

References 1. Leaf, G.A.V., “The Bending Behaviour of a Helical Filament Part I: The Rigidity of the

71

Helix”, Journal of the Textile Institute, Vol. 70, 1979, p. 323. Leaf, G.A.V., “The Bending Behaviour of a Helical Filament Part II: Curvature, Twist, and Strain Energy”, Journal of the Textile Institute, Vol. 70, 1979, p. 330 Livesey, R.G. and Owen, J.D., “Cloth Stiffness and Hysteresis in Bending”, Journal of the Textile Institute, Vol. 55, 1964, T516 Love, A.E.H., A Treatise on the Mathematical Theory of Elasticity, 4th ed., Dover Publications, New York, NY, 1944. Ly, N.G., “The Contribution of the Fibrediameter Distribution to the Bending Rigidity of Yarns”, Journal of the Textile Institute, Vol. 74, 1983, pp. 228-30. Timoshenko, S., “Strength of Materials”, Part II, van Nostrand, New York, NY, 3rd ed., 1956, p. 296. Konopasek, M. and Hearle, J., Fibre and Scence Technology, Vol. 5 No. 1, 1972. Ly, N.G., “The Bending of a Helical Wool Fibre”, Journal of the Textile Institute, Vol. 74, 1983, pp. 228-416. Ly, N.G., “Predicting the Bending Properties of Wool Worsted Yarn”, in Postle, R., Kawabata, S. and Niwa, M. (Eds), Objective Evaluation of Apparel Fabrics, Textile Machinery Society of Japan, Osaka, 1983, pp. 529-38. Treloar, L.R.G., “The Geometry of Multi-ply Yarns”, Journal of the Textile Institute, Vol. 47, 1956, T348. Carnaby, G.A., The Tensile Behaviour of TwoPly Yarns, WRONZ Communications, No. C90, May 1984. Riding, G., “A Study of the Geometrical Structure of Multi-ply Yarns”, Journal of the Textile Institute, Vol. 52, T366, 1961. Lunney, H.W.M. and Irvine, P.A., “Measurements with the CSIRO Fibre Fineness Distribution Analyser”, Textile Results Journal, Vol. 52, 1982, p. 217.

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

as θs Ts1, Ts Tp φp φ

14. Ly, N.G. and Denby, E.F.,. “Bending Rigidity and Hysteresis of Wool Worsted Yarns”, Textile Results Journal, Vol. 54, 1984, p. 180.

Further Reading

λ s L1, L

Platt, M.M., Klein, W.G. and Hamburger, W.J., “Mechanics of Elastic Performance of Textile Materials Part XIV: Some Aspects of Bending Rigidity of Singles Yarns”, Textile Results Journal, Vol. 29, pp. 611-27, 1959. Postle, R., Carnaby, G.A. and de Jong, S., The Mechanics of Wool Structures, Ellis Horwood Ltd, New York, NY, 1988.

A,B C M p,r,q ux,uy,uz vx,vy,vz wx,wy,wz x,y,z φ, Φ, Ψ E Io

Appendix. Principal Symbols Employed θp ap θs rs

Ply helix angle Ply helix radius Helical angle of fibre in single yarn Radius of fibre helix

72

Single yarn radius Helical angle of single yarn Single yarn twist before and after plying Ply yarn twist Angular position of point on single yarn axis Angular position of point on fibre with respect single yarn axis Parameter defining relation between φ and φp Length of fibre Length of single yarn axis before and after plying Principal bending rigidity of fibre Torsional rigidity of fibre Bending moment Curvature and twist components of fibre

Directional cosines of the fibre axis Co-ordinates of the fibre axis Euler angles of the fibre axis Young’s modulus of the fibre Moment of inertia of a circular cross section

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

A Model for Protective Clothing Effects on Performance Paul S. Adams Center for Ergonomics, University of Michigan, USA and Ann C. Slocum Department of Human Environment and Design, Michigan State University, USA W. Monroe Keyserling Center for Ergonomics, University of Michigan, USA Received 2 February 1993 Accepted 3 March 1994

of garment preference and suitability, the results are often not transferable to workers wearing other clothing systems or performing dissimilar tasks. Just as systematic research on fabric characteristics has led to predictive modelling of heat stress and thermal comfort[16-19], a systematic approach is needed to predict other performance changes due to PPC. However, before quantitative models can be developed, a better understanding is needed of how worker performance is affected by various garment characteristics.

Introduction Protective clothing is routinely worn by millions of workers to protect against everyday conditions such as dirt and grime, or adverse weather, as well as against environmental hazards; e.g. chemical, biological, thermal and physical agents. Although protective garments are designed to enhance worker comfort or safety, they can negatively affect worker performance in several ways, including: ● heat stress[1-4]; ● reduced task efficiency[5-8]; ● reduced range-of-motion[9-12]. Discomfort and reduced efficiency may also lead to rejection of personal protective clothing (PPC)[13], thus increasing the risk of worker injury or disease[14,15].

Uses of Performance-based Models The ability to predict performance changes due to PPC attributes would assist with the design and selection of protective clothing. Performanceeffect models would enable better evaluation of garments and ensembles of various sizes, materials and designs in the context of a variety of tasks and applications. Ultimately, the worker would benefit through reduced garment impediment, lowered physiological costs and improved productivity.

Need for a Systematic Approach Most of the research dealing with effects of PPC on the wearer has focused on fabric properties and their effects on heat stress, comfort, tolerance time and worker acceptance. Nearly all of this research has been task and/or garment specific. Although such studies answer specific questions

Purpose and Organization The purpose of this article is to present a framework for understanding the effects of PPC on worker performance and to propose a conceptual model that can lead to the prediction of these effects based on garment properties. This article is organized into four sections with the following objectives:

International Journal of Clothing Science and Technology, Vol. 6 No. 4, 1994, pp. 6-16, © MCB University Press, 0955-6222

6

VOLUME 6 NUMBER 4 1994

(1) Introduce a relational model for understanding the negative effects of protective clothing on worker performance. (2) Present a framework for assessing gaps in the research literature and for suggesting future research aimed at quantifying the impediment a garment imposes on workers. (3) Propose a conceptual model called the Garment Impediment Index (GII) model based on measurable garment parameters. (4) Present three approaches for developing a GII model.

expansion of the triad model, taking into account differences among workers and considering the general problem of performance decrements associated with PPC. In this expanded model, thermal balance is affected by four causal factors: clothing, task requirements, environmental conditions, and worker traits. Three of these factors: clothing, task requirements and worker characteristics, also determine changes in garment form and position that accompany movement. The processes of maintaining thermal balance and changing garment form cause immediate effects on movement capability, physiological balance, and sensory feedback. These immediate effects may in turn produce the net effects of reduced productivity, increased physiological strain, and reduced comfort. A detailed version of the relational model is shown in Figures 3-5. This model illustrates how specific aspects of the four causal factors may combine to affect the mechanisms associated with thermal balance and changes in garment form, and ultimately net effects.

A Relational Model of Factors Affecting Performance To develop a quantitative index for predicting garment effects on performance, it is necessary to understand how worker performance is affected by protective clothing. A model is proposed that identifies relationships among clothing parameters, task requirements, environmental conditions and worker characteristics, and their effects on performance. The model is based on work by Nunneley[2] who analysed heat stress associated with protective clothing. Nunneley introduced the heat stress triad shown in Figure 1, arguing that heat stress may result from one or more of three factors: ● work rate; ● clothing; ● environment. The importance of these factors was recently confirmed[20]. Nunneley’s triad can also be applied to effects other than heat stress, such as reduced productivity and comfort, and increased physiological strain. Figure 2 presents an

Four Causal Factors Garment properties. To understand how protective clothing affects performance, it is necessary to identify those garment properties that potentially affect worker performance and to quantify their contributions. Figure 3 illustrates how garment attributes are integrated to yield garments with the properties that are of interest in this study. The list of garment properties in

Task requirements

Environmental conditions Worker characteristics

Clothing properties

Changes in garment form and position

Thermal balance

Immediate effects: Decreased movement capability Disturbance of physiological balance Decreased sensory feedback

Work

Environment

Net effects: Reduced productivity Physiological strain Reduced comfort

Clothing

Figure 2. Overview Model for Causes of Negative Performance Effects on Workers Wearing Protective Clothing. Effects on the Worker and His/Her Productivity Result from Clothing Properties, Task Requirements, Worker Characteristics and the Environment

Source: [2]

Figure 1. Heat Stress Triad[2]

7

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Garment subcomponents

Garment components

Fibre Yarn Construction Finish Seams Openings Fasteners Construction Ease

Immediate effects

Garment properties

Fabric

Stiffness Hand* Coefficient of friction Vapour permeability Insulation

Design (style)

Bulk/compression Weight Ventilation Stretch

Functional events

Changes in garment form

Expansion Compression Displacement

Ease

Size

Net effects Reduced productivity: Completion time

Changes in ease

Bending

Length Girths

Visual impairment Auditory impairment Movement compromised: Accuracy decreased

Quality

Speed decreased

Work rate

Force/exertion increased Range-of-motion/ distance decreased Physiological response: (increases) Energy expenditure/ oxygen uptake

* Hand is a tactile property and probably does not directly affect performance Thermal imbalance: Heat build-up and sweating

Figure 3. Relationships Among Garment Subcomponents and Garment Properties

Tolerance time Work rate

Blood pressure Fatigue

Heat transfer

Heart rate

Fluid transfer

Surface and core temperature Sensory/Perceptual feedback: Thermal discomfort

Task requirements

Physiological strain:

Reduced comfort

Wetness Rubbing/chafing

Movement

Garment properties

Localized pressures Environmental conditions

Speed

Harmful agents

Force

Space constraints

Ventilation Vapour permeability

Distance

Radiant energy

Direction

Conductive thermal energy

Coefficient of friction Stretch

Thermal balance: Heat build-up and sweating Heat transfer

Weight Insulation

Bulk Stiffness Design ease Size ease

Restriction

Accuracy

Figure 5. Intuitive and Known Relationships Among Functional Events and Net Effects

Convective thermal energy

Fluid transfer Changes in garment form

Humidity

heat is considered a by-product to be expelled or transferred away from the body. Worker movement also causes clothing to move and change form. Clothing must slide across the skin (displace), stretch (expand), fold(bend), and bunch up (compress) as the body moves[21]. These mechanisms all resist changes in garment form, with the level of resistance determined by the garment characteristics. Resistance to change in form imposes additional force requirements on the wearer and may compromise movement capability. Worker characteristics. Differences among workers in three characteristics help determine the effects of PPC on performance: anthropometry, physiology and motivation. Anthropometry affects how well a garment fits and, along with physical strength, the relative difficulty of a task for a given individual. Physiologically, the rate of metabolic heat generation and level of sweating often vary substantially among individuals performing the same task[23]. Motivation affects the rate and duration of work and the choice of movements involved. Environmental conditions. Environmental conditions often mandate the use of PPC, but they may also affect the wearer’s performance as shown in Figure 4. Work space may be limited and mobility in close quarters may be constrained by increased effective anthropometry, i.e. worker dimensions as measured over the exterior of a

Worker characteristics Physiology

Displacement

Anthropometry

Expansion

Motivation

Compression Bending Changes in ease due to filling

Figure 4. Relationships Among Four Causal Factors and Functional Events

Figures 3 and 4 is based on selected literature[11,21,22] and an analysis of forces working within garments during movement. Task requirements. The second of the four contributing factors is the task or work requirements. Movement is required in performing work, assuming a task is not purely cognitive. Movements may be described in terms of speed, direction, force, accuracy, and magnitude, as shown in Figure 4. Magnitude of movement or distance may be measured by the change in joint angle or the distance moved. Work requirements determine what movements must be made, as well as the characteristics of those movements. Movement involves the contraction of muscles and the subsequent generation of metabolic heat. In most instances, this metabolic

8

VOLUME 6 NUMBER 4 1994

garments without compromising worker protection. Specifically, this requires working with the garment properties listed in Figure 3. If the immediate effects in Figure 5 are known, then it may be possible to predict net effects. This concept has been demonstrated by: ● the use of motion study to estimate task completion[32,35]; ● the direct dependence of tolerance time on heart rate and other physiological measures[4,28]; ● the correlation of sensory feedback with comfort[31]. Therefore the key to reducing negative net effects is to reduce negative immediate effects. Reducing negative immediate effects necessitates a better understanding of the relationships between garment properties and immediate effects. Consequently, both the garment properties and the immediate effects need to be quantified and their relationships investigated.

protective ensemble. Ambient radiant, conductive and convective thermal energy may impose additional heat load on the worker and inhibit the transfer loss of metabolic heat. High humidity may also prevent sweat evaporation. Immediate Effects Clothing properties, task requirements, environmental conditions and worker characteristics combine to cause three types of immediate effects on the PPC wearer, as illustrated in Figure 5. First, movement may be compromised. Compared to a semi-nude working condition, movement speed, accuracy, and range of motion (ROM) may be reduced while muscular exertion requirements may be increased[9,24-26]. In addition, the ability to receive visual and auditory feedback may be compromised[13,27]. Second, physiological responses may occur, such as increases in heart rate, blood pressure, body surface and core temperatures, oxygen uptake or energy expenditure, and fatigue[1,4,28]. Third, the wearer may experience unpleasant sensations such as thermal discomfort, skin wetness, rubbing or chafing, localized pressures, and restriction[14,29-31].

A Framework for PPC Research on Worker Performance

Net Effects Ultimately, immediate effects may lead to reduced productivity, increased physiological strain, and reduced comfort. Productivity effects may include longer task completion times, slower work rates, and poorer product or service quality, including operator errors[8,26,32,33]. Physiological strain can severely limit work times and work rates. For the sake of health and safety, wearers of PPC often have work-rest cycles imposed and may have work sessions terminated prior to task completion due to physiological constraints on body temperature, heart rate, blood pressure, and fatigue[4,28,34]. Finally, unpleasant sensations cause discomfort. Comfort is a complex issue with multiple facets. Although it is uncertain whether comfort directly affects productivity, it is clearly an important factor in achieving worker acceptance of PPC[15,31].

One approach for developing a model of PPC effects on worker performance is to study systematically the relationships among the identified garment properties and the immediate effects. Figure 6 introduces a matrix of independent garment properties and dependent performance measures taken from Figures 4 and 5. The framework in Figure 6 may be used to view previous work and to identify areas needing further study. A survey was conducted of 118 protective clothing studies contained in a database of approximately 300 papers pertaining to the ergonomic effects of protective clothing and equipment[36]. (See Notes (2) and (3) to Figure 6 regarding papers not included in the survey). Studies included in the survey were reported in refereed journals, conference proceedings, and government technical reports for the period 1957 to 1992. The database emphasized work in areas other than heat stress and comfort, although a sampling of papers dealing with these aspects was also included. The dependent and independent variables were noted for the studies surveyed. Each cell of the matrix was then coded to indicate the number of studies that utilized a particular measure to assess effects of the corresponding garment parameter. Studies that did not isolate or quantify specific garment parameters were included in the “Properties confounded” column. Confounding

Understanding Effects of Garment Properties Changing task requirements or environmental conditions can have dramatic effects on productivity, physiological strain and comfort. Whenever practical, environmental conditions and work requirements should be engineered to preclude the need for PPC. Assuming the task and environment cannot be changed, the objective becomes one of minimizing any deleterious effects associated with wearing protective

9

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Garment property Performance measure

Weight

Stiffness

Bulk

Coefficient of friction

4-6

7-9

10-12

Style

Stretch

Size

Vapour Ventilation permeability

Insulation

Properties confounded

Completion time

Objective measures

Work rate/ movement time Performance quality Range of motion

Heart rate Energy expenditure O2 Uptake/ Ventilation rate Skin moisture

Subjective measures

Body temperature

Key:

Psychophysical quantification Comfort

None

1-3

13+

Notes (1) Survey included 118 studies in refereed journals or proceedings, or in government technical reports available through NTIS. Publication dates ranged from 1955 to 1992. (2) Survey did not include papers dealing primarily with respirators, gloves, shoes, masks, SCBA, or load carriage (n = 64). (3) Studies designed to measure parameters or properties as the dependent variables were not included, i.e. ventilation, compressibility, vapour resistance and insulation, if temperature or another variable was not explicitly stated (n = 27). (4) Confounded column included studies in which: (a) Garment specifications were not given. (b) Garment parameters were not controlled. (c) Comparisons were made between competing garments or configurations without any property differentiation. (d) Only one garment type was used (treatment versus a nude control). (5) Studies in which one parameter was specified, but all others were confounded, are included in both the appropriate property column and confounded column. (6) Heart rate dependent variable also included studies where blood pressure or tolerance time was used. (7) Energy expenditure dependent variable also included studies where data are given as oxygen uptake or where ventilatory volume changes are reported. (8) Temperature included rectal and/or skin temperature. (9) Psychophysical variable studies asked subjects to quantify a specific response variable; i.e. estimate the level of a variable such as rating of perceived exertion. (10) Comfort studies asked subjects to report level of “comfort” in a general sense without quantifying the level of a specific aspect of comfort.

Figure 6. Number of Studies out of 118 Reviewed that Isolated or Defined a Given Garment Property and the Corresponding Dependent Measure

among garment parameters and insufficient garment data preclude the use of such studies in development of a predictive model based on clothing characteristics. Two points are apparent from Figure 6. First, the “Properties confounded” column indicates that the preponderance of previous work has been directed towards the important but short-term goal of comparing competing garments or simply

measuring effects, rather than relating those effects to specific garment attributes. The exceptions have been those properties which are known to affect heat stress: permeability, insulation and ventilation, as well as characteristics that affect comfort. As noted previously, models have been developed for predicting heat stress that were based on studies of garment attributes.

10

VOLUME 6 NUMBER 4 1994

Second, Figure 6 shows that performance effects are largely unexplored for many of the garment parameters. Relatively few studies have attempted to quantify the effects of various parameters on those aspects of work performance that are not directly related to heat stress (i.e. heart rate, ventilation rate, skin wettedness and body temperature.)

Garment Impediment Indices Defined Garment impediment. Garment impediment occurs when clothing impairs a worker’s performance. With respect to productivity, garment impediment may be defined as the difference between work output when semi-nude (or wearing very comfortable, non-restrictive clothing) and output when wearing a protective garment. Productivity impediment may be expressed as reduction in rate of task performance, increase in time for task completion, or reduction in quality of output. Impediment can also be described as a percentage reduction in performance capability, e.g. a 10 per cent decrease in walking speed may be defined as a 10 per cent impediment for this measure of performance. Garment Impediment Indices. A Garment Impediment Index (GII) is simply an index or measure of a performance change that results from wearing a protective garment. GIIs may be defined as response functions of the garment properties listed in Figure 7. For example, the GII for task completion time may be defined as follows: GIICT = f(weight, stiffness, bulk…, and insulation).

Development of a Garment Impediment Index Using the framework in Figure 6, a conceptual model is proposed called the Garment Impediment Index model (GII model). The GII model relates garment properties to a set of performance measures, as shown in Figure 7. The underlying concept behind the GII model is as follows: Actual = Ideal – Impediment, where Actual = Actual performance while wearing PPC Ideal = Performance capability without PPC Impediment = Effect of PPC.

Garment properties Coefficient Bulk

of friction

Vapour Style

Stretch

Size

Ventilation

permeability

Insulation

Completion time

G11CT

Work rate/ movement time

G11MT

Performance quality

G11Q

Range of motion

G11ROM

Heart rate

G11HR

Energy expend./ O2 Uptake/ ventilation rate

G11EE

Skin wettedness

G11SW

Subjective measures

Body temperature

G11TEMP

Psychophysical quantification

Garment impediment indices

Objective performance measures

Weight Stiffness

G11RPE

Comfort

G11COMF DWT

DSTIFF

DBULK

DFRIC

DSTY

DSTR

DSIZE

DVENT

DVPERM

DINSUL

G11COMPOSITE

Performance decrements

Figure 7. Garment Properties and Performance Measures Combine to Form Garment Impediment Indices (GIIs) and Performance Decrements (DYs)

11

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

environmental conditions present, the task being performed, and the characteristics of the worker. Some garments may mechanically restrict movement but not affect tolerance time, whereas others may restrict evaporative cooling so effectively that heat stress becomes a major concern, even though no mechanical effect is evident. However fatigue, either local or general, is likely to accompany both mechanical and physiological factors.

Garment effects are described by a set of GIIs, with each GII indicating the cumulative effect of the various garment parameters on a specific performance measure. For example, one GII reflects the decrement in task performance as indicated by output rate, another GII indicates the physiological response of heart rate increase, and yet another reflects psychophysical response as indicated by a change in the Rating of Perceived Exertion[37]. Ignoring interaction effects, this may be stated mathematically as:

n

Total garment effect = wCTGIICT + wMTGIIMT + wQGIIQ + … + wCOMFGIICOMF

Response functions should incorporate multiple performance measures

where GIIx = garment impediment index for specific performance measure x, and wx = weighting factor for measure x, such that ∑wx = 1.0. Weighting factors are functions of the task being performed and will be discussed later. With interactions, total impediment = ∑(wGIIs) + interactions. A proposed set of GIIs is shown in the right-hand column of Figure 7.

n It is likely that PPC affects some tasks more than others; e.g. a simple hand cranking task may be much more sensitive to mechanical impediment than a walking task. Whether or not PPC impediment is sufficient to cause a performance problem is clearly a function of the task requirements. Therefore response functions need to be defined in terms of the requirements for the tasks being performed, including descriptions of the movements involved and their speeds. Accounting for task differences (and the importance of individual GIIs) is accomplished by weighting the various GII terms in the model to reflect their relative contribution to total garment effect. One of the task requirements that should receive special consideration is task duration. Physiological and psychophysical performance measures are typically time dependent, i.e. physiological and comfort factors may become more limiting as the duration of work in PPC increases. Owing to the complexity of potential PPC effects, response functions should incorporate multiple performance measures to reflect the mechanical, physiological and psychophysical factor effects. For example, the response function for a fire-fighter turnout coat should consider the garment’s effect on range-of-motion, body temperature, heart rate and comfort, as well as work rate. Failure to use multiple measures in this case would probably result in inaccurate predictions of performance capability. Use of a single indicator assumes correct identification of the constraining performance variable, and multiple performance variables may actually combine to limit work capability. Measuring only

Model Parameters Figure 7 also summarizes the garment properties and response parameters that potentially compose a GII COMPOSITE model for a specified task. The GIIx and Dy terms are defined as follows: GIIx = GII for the change in performance measure x. Dy = Performance decrement due to garment property y. The individual GIIx s are scalar quantities, whereas the Dy s are vectors. GIIs may be expressed as coefficients corresponding to percentage decrement. For example, if GIIEE = 0.20, then a worker wearing a specified garment would expend 20 per cent more energy while performing a task than he/she would if he/she were performing the task without the garment. For a psychophysical measure of overall comfort, a GIICOMF = 0.20 could indicate that ratings of overall body comfort with the garment were two points lower on a ten-point scale than ratings given without the garment, assuming that the psychophysical measure used is a true ratio scale. Need for Multiple Performance Measures Consistent with the relational model presented earlier (Figure 2), the effect of PPC on a wearer’s task performance or physiological response varies with the type of garment being worn, the

12

VOLUME 6 NUMBER 4 1994

one aspect of performance could therefore result in underestimates of garment impediment and overestimates of performance capability.

Approaches for Developing GIIs The GII model provides a theoretical construct for understanding garment effects on task performance. Relationships among dependent measures and individual garment properties must be quantified before the model can be practically useful. At present, it is not clear which properties are significant, what the relative importance is of each property, and what effect differences in tasks and wearers have on the response variables. Interaction effects are also not well understood. Three experimental approaches are proposed for developing the GII model.

Overall Garment Impediment Index The overall garment impediment index, GIICOMPOSITE, is described by combining the complete set of individual-parameter GIIs. GIICOMPOSITE provides an indication of the combined effect that a specified garment has on a worker when performing a specific task. GIICOMPOSITE may be defined in two ways. First, it can be defined with respect to dependent response variables, as presented in the total impediment equation above. Restating, GIICOMPOSITE can be given as a function of the individual GIIs: GIICOMPOSITE = f (GIICT, GIIMT, GIIQ, GIIROM, GIIHR, GIIEE, GIISW, GIITEMP, GIIRPE, GIICOMF) + error. (GIIRPE = GII for Borg’s Rating of Perceived Exertion[37,38], a commonly used psychophysical scale.) Assuming the performance measures are not equally important, it becomes necessary to weight the component GII terms to reflect their individual contributions. If a dependent variable is not relevant for a particular task, e.g. performance quality, then the GII weighting factor is zero and the corresponding GII is essentially excluded from the model. If equal contributions exist, then all of the weightings would be equal. As stated earlier, weightings of the component GIIs in the GIICOMPOSITE function depend upon the work situation. Second, GIICOMPOSITE may be defined with respect to independent garment parameters. An example for this case is GIICOMPOSITE = f (DWT, DSTIFF, DBULK, DFRICTION, DSTY, DSTR, DSIZE, DVENT, DVPERM, DINSUL, DINTERACTIONS) + error,

Single Performance Measure/Multiple Garment Properties First, it may be possible to stratify the GII model by garment property. Under this approach, experiments are conducted within individual cells across a single row in Figure 7. Since some garment factors are typically confounded, e.g. bulk and weight, the challenge is to isolate properties and then identify those that have the greatest effect on a specific performance measure. This may be done by selecting garments that differ in only one or two parameters. For example, a study designed to isolate weight effects might use a single pair of coveralls with variably weighted sleeve, torso and leg segments. Segment weights could be designed to add mass without affecting the other mechanical properties, such as bending and compression at the joints. This approach would effectively control the coefficient of friction, since the garment surface contacting the skin would remain unchanged. The single performance measure/multiple garment property approach might be used when comparing garments for a task where interest lies primarily with one specific response measure. For example, range-of-motion might be measured in a study of garments that will be worn by maintenance workers who frequently utilize extreme postures.

where DINTERACTIONS = performance decrement due to interaction effects. As with the previous GIICOMPOSITE model, the weightings of the component terms in the function depend upon the work requirements. When GIICOMPOSITE is defined using independent garment variables, each component DY term incorporates all of the performance measures of interest. For example, DWT represents the decrement in overall performance due to garment weight, as indicated by the mechanical, physiological and psychophysical measures combined.

Multiple Performance Measures/One Garment Property A second approach is to work within a single column in Figure 7 using multiple response measures while studying one specific garment property. The objective of this approach is to determine a garment variable’s effect on multiple performance measures (whereas the intent of the previous approach was to determine the effects of multiple garment variables on a single performance measure). The multiple performance/single property approach is appropriate when comparing similar garments

13

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

which substantially differ in only one property and which will be worn when performing complex tasks; e.g. comparing different designs of fire-fighter coats that are made from the same materials. Variables of interest in this example might include range-of-motion, movement time, energy expenditure, heart rate, body temperature, and comfort.

References 1. Rodahl, K. and Guthe, T., “Physiological Limitations of Human Performance in Hot Environments, with Particular Reference to Work in Heat-Exposed Industry”, in Mekjavic, I.B., Banister, E.W. and Morrison, J. B. (Eds), Environmental Ergonomics: Sustaining Human Performance in Harsh Environments, Taylor & Francis, Basingstoke, 1986, pp. 2269. 2. Nunneley, S.A., “Design and Evaluation of Clothing for Protection from Heat Stress: An Overview”, in Mekjavic, I.B., Banister, E.W. and Morrison, J. B. (Eds), Environmental Ergonomics: Sustaining Human Performance in Harsh Environments, Taylor & Francis, Basingstoke, 1986, pp. 87-98 (also available from the National Technical Information Service, NTIS: AD-A196 438). 3. Fine, B.J., The Effect of Heat and Chemical Protective Clothing on the Ability of a Group of Female Soldiers to Sustain Performance of Military Cognitive Tasks, T7-88, US Army Research Institute of Environmental Medicine, 1987 (available NTIS: AD-A192 596). 4. Van de Linde, E.J.G. and Lotens, W.A., “Restraint by Clothing Upon Fire-fighters’ Performance”, Proceedings 1983: International Conference on Protective Clothing Systems, 1981, pp. 195-204. 5. Alexander, M. and Laubach, L., “The Effects of Personal Protective Equipment Upon the Arm-Reach Capability of USAF Pilots”, Proceedings Reprint of the Interagency Conference on Management and Technology in the Crew System Process, 1973, pp 225-33. 6. Bachrach, A.J. and Egstrom, G.H., Human Engineering Considerations in the Evaluation of Diving Equipment, Naval Medical Research Institute, 1974 (available NTIS: AD-A011 680). 7. Dunlap & Associates, Inc. and Activity US Army General Equipment Test, Development of a Methodology for Measuring Infantry Performance in Digging Hasty Fighting Positions, ASATECOM Project No. 8-3-770001, US Army General Equipment, 1965 (available NTIS: AD-467 159). 8. King, J.M. and Frelin, A.J., “Impact of the Chemical Protective Ensemble on the Performance of Basic Medical Tasks”, Military Medicine, Vol. 149 No. 9, 1984, pp. 496-501. 9. Huck, J., “Protective Clothing Systems: A Technique for Evaluating Restriction of Wearer Mobility”, Applied Ergonomics, Vol. 19 No. 3, 1988, pp. 185-90.

Multiple Performance Measures/Multiple Garment Properties A third approach is to vary several garment properties and collect multiple sets of response data simultaneously. This approach requires selection of specific cells from multiple rows and columns within Figure 7. Exploratory factor analysis can then be used to consolidate variables into subgroups that are relatively independent and to generate hypotheses about relationships in the reduced data set. Confirmatory factor analysis can be performed to test hypotheses on specific variables and to estimate factor scores[39]. This multivariate approach is perhaps best suited for garment development studies, where multiple configurations and applications need to be considered. It is also a means of dealing with confounding among garment properties.

Conclusion Predicting garment effects on worker performance is difficult because relationships among garment properties and human responses are not well understood. This article has introduced a framework for viewing previous studies. A conceptual model was proposed that provides a systematic approach for studying the effects of PPC properties on various aspects of worker performance. Once the effects of various garment parameters on performance are understood and quantified, models can be developed that will enable engineers to predict the PPC effects on worker performance, even prior to garment purchase or task assignment. Developing such performance effect models will require a multidisciplinary perspective, incorporating the fields of textile and clothing science, ergonomics, work physiology, industrial engineering, and statistics. The understanding gained from these efforts should ultimately facilitate development of more comfortable and less impeding garments. It may also be possible to adapt the systematic approaches outlined in this article when studying the effects of gloves, respirators and other types of personal protective equipment.

14

VOLUME 6 NUMBER 4 1994

10. Saul, E.V. and Jaffe, J., The Effects of Clothing on Gross Motor Performance, EP-12, US Army Quartermaster Research and Development Center, 1955 (available NTIS: AD-066 180). 11. Nicoloff, C., Effects of Clothing on Range of Motion in the Arm and Shoulder Girdle, EP-49, Environmental Protection Research Division, Quartermaster, 2957 (available NTIS: AD-142 863). 12. Alexander, M., McConville, J.T. and Tebbetts, I., Revised Height/Weight Sizing Programs for Men’s Protective Flight Garments, AMRL-TR79-28, Aerospace Medical Research Laboratory, 1979 (Available NTIS: AD-A070 732). 13. Bensel, C.K., Teixeira, R.A. and Kaplan, D.B., The Effects of US Army Chemical Protective Clothing on Speech Intelligibility, Visual Field, Body Mobility, and Psychomotor Coordination of Men, NATICK/TR-87/037, US Army Natick Research, Development and Engineering Center, 1987 (available NTIS: AD-A188 478). 14. Rosenblad-Wallin, E., “Development of Protective Clothing Based on the Demands of the User and His Environment”, Proceedings 1983: International Conference on Protective Clothing Systems, 1981, pp. 327-33. 15. Abeysekera, J.D.A., “The Need for National and International Ergonomics Standards for Personal Protective Devices”, in Mital, A. (Ed.), Advances in Industrial Ergonomics and Safety I, Taylor & Francis, Basingstoke, 1989, pp. 809-16. 16. McCullough, E.A., Jones, B.W. and Huck, J., “A Comprehensive Data Base for Estimating Clothing Insulation”, ASHRAE Transactions, Vol. 91, 1985, pp. 29-47. 17. Lotens, W. A. and Havenith, G., “Calculation of Clothing Insulation and Vapour Resistance”, Ergonomics, Vol. 34 No. 2, 1991, pp. 233-54. 18. Parsons, K.C., “Protective Clothing: Heat Exchange and Physiological Objectives”, Ergonomics, Vol. 31 No. 7, 1988, pp. 991-1007. 19. Umbach, K.H., “Physiological Tests and Evaluation Models for the Optimization of the Performance of Protective Clothing”, in Mekjavic, I.B., Banister, E.W. and Morrison, J. B. (Eds). Environmental Ergonomics, Taylor & Francis, Basingstoke, 1988, pp. 139-61. 20. Fan, J. and Keighley, J.H., “An Investigation on the Effects of: Body Motion, Clothing Design and Environmental Conditions on the Clothing Thermal Insulation by Using a Fabric

21.

22.

23.

24.

25.

26.

27.

28.

Manikin”, International Journal of Clothing Science and Technology. Vol. 3 No. 5, 1991, pp. 6-13. Kirk, W.J. and Ibrahim, S.M., “Fundamental Relationship of Fabric Extensibility to Anthropometric Requirements and Garment Performance”, Textile Research Journal, Vol. 36 No. 1, 1966, pp. 37-47. Watkins, S.M., Clothing: The Profitable Environment, Iowa State University Press, Des Moines, IA, 1984. Astrand, P.O. and Rodahl, K., Textbook of Work Physiology: Physiological Bases for Exercise, McGraw-Hill Book Company, New York, NY, 1986. Adams, P.S. and Keyserling, W.M., “Effect of Garment Weight on Arm Movement”, Presentation at the Fourth International Symposium on the Performance of Protective Clothing, 19 June, 1991, Montreal. Adams, P.S., “The Effect of Size and Fabric Weight of Protective Coveralls on Range of Gross Body Motions”, Chapter 4 of “The Effects of Protective Clothing on Worker Performance: A Study of Size and Fabric Weight Effects on Range-of-motion”, Unpublished doctoral dissertation, The University of Michigan, 1993. Gregoire, H., Call, D.W., Omlie, L.C. and Spicuzza, R.J., “An Automated Methodology for Conducting Human Factors Evaluation of Protective Garments”, Proceedings of the Human Factors Society 29th Annual Meeting, 1985, pp. 916-19. Welsh, K.W. and Vaughan, J.A., “Visual and Optical Assessment of Gas Protective Face Masks”, AGARD Conference Proceedings No. 255. Operational Helicopter Aviation Medicine, 1978 (available NTIS: AGARD-CP255). Pandolf, K.B., Allan, A.E., Gonzalez, R.R., Sawka, M.N., Stroschein, L.A. and Young, A.J., Chemical Warfare Protective Clothing: Identification of Performance Limitations and Their Possible Solution, US Army Research Institute of Environmental Medicine, 1987 (available NTIS: AD-A177 871).

29. Denton, M.J., “Fit, Stretch, and Comfort”, Textiles for Comfort, Third Shirley International Seminar, 1971. 30. Clulow, E.E., “Protective Clothing and Comfort”, Proceedings of Shirley Institute Conference on 21 October, 1982, 1983.

15

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

31. Cowan, S.L., Tilley, R.C. and Wiczynski, M.E., “Comfort Factors of Protective Clothing: Mechanical and Transport Properties, Subjective Evaluation of Comfort”, in Mansdorf, S.Z., Sager, R. and Nielsen, A.P. (Eds), Performance of Protective Clothing: Second Symposium, ASTM STP 989, American Society for Testing and Materials, 1988, pp.32-42.

35. 36.

32. Harris, D.W., A Degradation Analysis Methodology for Maintenance Tasks, Masters Thesis, Georgia Institute of Technology, 1985 (available NTIS: AD-A155 073).

37.

33. Kelly, T.L., Englund, C.E., Ryman, D.H., Yeager, J.E. and Sucec, A.A., The Effects of 12 Hours of MOPP IV Gear on Cognitive Performance Under Non-Exercise Conditions, NHRC Report No. 88-6, Naval Health Research Center, 1987 (available NTIS: AD-A192 527). 34. Konz, S.A., Rohles, F.H. and McCullough, E.A., “The Effectiveness of Water Cooling under Protective Clothing at Temperatures between 22˚C and 55˚C”, Proceedings 1983:

38.

39.

16

International Conference on Protective Clothing Systems, 1981, pp. 215-21. Niebel, B.W., Motion and Time Study, Richard D. Irwin, New York, NY, 1988. Adams, P.S., “Framework for PPC Research on Worker Performance”, abstract with data table in Proceedings of The Second International Symposium on Consumer Environmental Issues: Safety, Health, Chemicals and Textiles in the Near Environment, 1993 (in press). Borg, G., “Perceived Exertion as an Indicator of Somatic Stress”, Scandinavian Journal of Rehabilitative Medicine, Vol. 2 No. 3, 1970, pp. 92-8. Borg, G.A.V., “Psychophysical Bases of Perceived Exertion”, Medicine and Science in Sports and Exercise, Vol. 14 No. 5, 1982, pp. 377-81. Tabachnick, B.G. and Fidell, L.S., Using Multivariate Statistics, Harper & Row, New York, NY, 1983.

VOLUME 6 NUMBER 4 1994

Medical Clothing The Stress Relaxation and Shrinkage of Pressure Garments Frency Sau-fun Ng Yip Institute of Textiles and Clothing, Hong Kong Polytechnic, Hong Kong Received 20 November 1993 Accepted 14 March 1994

knitted Lycra fabrics, mainly of power net or sleeknit structure. The last fabric sample, Tubigrip, was a weft-knit cotton elastic material having a rubber yarn laid in horizontally. All the samples selected were at the time used by hospitals or burns units in the UK and/or in Hong Kong for making pressure garments.

Introduction The existing systems for making pressure garments have been reported in an earlier article[1]. Currently, elastomeric fabrics used by most pressure garment manufacturers and the burns units of hospitals are purchased from specialist fabric producers. Many types of elastic fabrics have been used for making pressure garments, but none of them was specially designed for the end use. A study by Cheng et al.[2] demonstrated that there was a gradual decline in garment-scar interface pressure when patients wore pressure garments over a period of time. Views of doctors, therapists and patients support their findings: they comment that the tension of the fabric is timedependent, so that slackening is bound to occur in pressure garments when patients wear them over a period of time. Elastic deterioration in the stretch fabric used affected the clinical effectiveness of the pressure garments, additional problems being garment fitting subsequent size alteration. In this study, comparisons of the stress relaxation and shrinkage properties were made on six selected fabrics collected from the United Kingdom and/or Hong Kong, in order to have a better understanding of the stress relaxation behaviour of the fabrics used for making pressure garments.

Experimental Investigation of the stress decay, after a long period of extension and repeated washings, was made. The design of the tests was determined by performance-in-use criteria. As the pressure garment had to be worn continuously for about 23 hours a day, the fabric was correspondingly under stress at a given extension level over a prolonged period of time before such tension was released. Measuring Stress Relaxation In order to examine the behaviour of the load required to hold the fabrics in a fixed state of elongation over a prolonged period of time, i.e. 23 hours, cut strip tests were used. The specimens used were 5 × 15cm in dimension, with a gauge length of 10cm. The specimens were mounted in an Instron Tensile Strength machine (model 1026), manually under zero load. Each of the tested fabric specimens were extended by 5, 10, 15 and 25 per cent respectively, in both lengthwise and widthwise directions. This range of extension was determined on the basis of use: existing pressure garments are only stretched to relatively low levels of extension. Even though the stretch

Selecting Fabric Samples to Be Tested Six different fabrics were examined in this study and these are described in Table I. Three of the selected fabrics (#25324, #23883, and #24832) were supplied by a major manufacturer in the United Kingdom. Another two fabric samples (#24206 and 22804) were obtained from Hong Kong. These five fabric samples were warp-

This article is extracted from an unpublished MPhil dissertation. The author acknowledges the supervision, assistance and guidance of staff of the De Montfort University, Hong Kong Polytechnic, Prince of Wales Hospital, Hong Kong, Leicester Royal Infirmary, Jobst Institute Inc., Pan Med Ltd, Penn International, and others for having made this work possible.

International Journal of Clothing Science and Technology, Vol. 6 No. 4, 1994, pp. 17-27, © MCB University Press, 0955-6222

17

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Fabric particulars Supplier Gauge Fabric weight Composition

Breaking load (lengthway) (widthway) Breaking extension: (lengthway) % (widthway) %

#25324 #23883 UK UK (Raschel warp (Raschel warp knitted on 56 knitted on 56 gauge) gauge) 217 gm per 300 gm per sq. metre sq. metre 67 d. tex 78 d. tex nylon and 470 nylon 420 d. tex elastane d. tex elastane

#28432 UK (Raschel warp knitted on 56 gauge) 270 gm per sq. metre 56 d. tex nylon 480 d. tex elastane and includes 16% of 100 Nm cotton

#24206 Hong Kong (Raschel warp knitted on 48 gauge) 220 gm per sq. metre 70 Denier Nylon SD. 420 Denier Elastomer

#22804 Hong Kong (Raschel warp knitted on 56 gauge) 190 gm per sq. metre 70 Denier Nylon SD. 280 Denier Elastomer

Tubigrip Hong Kong 1 × 1 weft circular knit 245 gm per sq metre Single 22s cotton yarn with a natural rubber thread laying in at every fourth feeder

61 kg 43 kg

82 kg 95 kg

58 kg 75 kg

48 kg 48 kg

54 kg 55 kg

– 10 kg

360 310

400 280

360 280

472 312

440 320

– 320

Table I. Examination of Six Different Fabrics

direction of the existing Lycra pressure garments was mainly in the lengthwise direction (which is the run-in direction of the Lycra yarn), this study involved the testing of both lengthwise and widthwise directions for better comparison of the fabric characteristics. In the case of the Tubigrip fabric, as its elasticity is chiefly in the width direction and a much higher extension is applied in practical use, the range of extensions for the test of Tubigrip was increased to 25-100 per cent and it was tested in the widthwise direction only. In all cases the specimens were extended at a constant rate of 200mm/min. The clamp width of the machines was 5cm (flat faces), and the tension loan cell used was 5kg. Based on a fixed extension rate, the tension against time was recorded at fixed time intervals for a period of 23 hours to study the stress relaxation behaviour of the fabrics. The time intervals were as follows: ten minutes, 30 minutes, one hour, three hours, six hours, and 23 hours. From the raw data recorded from the testing the percentage tension loss, the behaviour of the load changes over various periods of time was calculated and load/extension curves were drawn. As the load under stress was very unsteady (stress decay rapid) during the first ten minutes, the starting point of the graph was ten minutes. The test results are presented in Table II, in Figure 1 and Figure 5 (for the Tubigrip).

Test direction and stretch %

Load (g force) 10 mins 23 hrs

25324

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

305 135 560 285 790 500

280 95 510 240 725 425

25 40 50 45 65 75

8 30 9 16 8 15

23883

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

204 98 382 248 618 610

190 84 358 228 560 517

14 14 24 20 58 93

7 14 6 8 9 15

28432

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

200 105 425 315 640 680

158 85 385 275 572 551

42 20 40 40 68 129

21 19 9 13 11 19

24206

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

195 220 315 400 535 640

162 140 257 290 410 375

33 80 58 110 125 265

17 36 18 28 23 41

22804

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

290 435 480 750 595 1,020

260 320 420 560 500 760

30 115 60 190 95 260

10 26 13 25 16 25

Tubigrip widthway 25 widthway 50 widthway 75 widthway100

117 172 230 250

112 170 210 225

5 2 20 25

4 1 9 10

Fabric

Elongation After Stretching After the specimens had been stretched for 23 hours the tension was released, then the length of

Loss

% Loss

Table II. Twenty-three Hours Stress Relaxation Test (Before Wash)

18

VOLUME 6 NUMBER 4 1994

the specimen was measured under zero stress after it had relaxed 30 minutes, as well as after 23 hours, in order to find the elastic recovery of the materials. Test results are shown in Tables X and XI.

Test Results and Discussion Test results are shown in Figures 1 to 5, and Tables II to XI. Examination of the individual load/extension curves indicates that for both lengthwise and widthwise directions, an increase in the amount of extension will give rise to a greater slope (tension loss/logtime) to the curve, which means the rate of loss in tension is increased. This is the general physical property of elastic fabric. However, test results in Tables II to V show that there is no correlation between the percentage of load loss and the amount of stretch. This indicates that it would be more appropriate to use the slope (as per Table VIII and Figures 1 to 5) to compare the behaviour of stress relaxation, because the slope can indicate the average loss of stress relaxation, while the figures in Tables II to V can indicate only the loss of load between the start and finish of the test. All the fabrics examined, except the Tubigrip, were found to have similar elastic properties in the lengthwise direction in terms of the loss in

Tests After Washing The same tests were repeated on the same specimens after repeated washing. The specimens were immersed in five litres of water at 30˚C with 0.1 per cent commercial washing powder (Persil). The six fabric samples were hand washed together for a period of two minutes, and they were air-dried flat and tested again (repeat of tests (1) and (2)); the operation was repeated after first wash, fifth wash and the tenth wash. See Figures 2 to 5; and Tables III and VIII for the test results. Shrinkage after Washing After the specimens were washed and dried, the original marks of gauge length on the specimens were measured for shrinkage. Test results are shown in Table IX.

Fabric

Test direction and stretch %

Load (g force) 10 mins 23 hrs

Loss % Loss *% Loss

Fabric

Test direction Load (g force) and stretch % 10 mins 23 hrs

Loss % Loss *% Loss

25324

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

305 152 525 330 745 470

280 116 480 290 680 400

25 36 45 40 65 70

8 24 9 12 9 15

8 14 14 +2 14 20

25324

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

328 225 544 436 770 580

295 160 505 360 712 496

33 65 39 76 58 84

10 29 7 17 8 14

3 +19 10 +26 10 1

23883

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

218 128 440 330 645 680

198 100 400 285 576 550

20 28 40 45 69 130

9 22 9 13 11 19

3 +2 +5 +15 7 10

23883

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

218 165 402 420 635 730

198 135 368 305 600 572

20 30 34 115 35 158

9 18 8 27 6 22

3 +38 4 +23 3 6

28432

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

195 136 425 390 640 650

172 112 380 295 565 534

23 24 45 95 75 116

12 18 11 24 12 18

14 +7 11 6 12 21

28432

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

242 150 418 340 670 630

214 120 358 255 575 485

28 30 60 85 95 145

12 20 14 25 14 23

+7 +14 16 19 10 29

24206

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

155 175 288 440 480 555

120 125 230 325 380 400

35 50 58 115 100 155

23 29 20 26 21 28

38 43 27 19 29 38

24206

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

155 110 307 300 450 570

120 80 275 217 340 410

35 30 32 83 110 160

23 27 10 28 24 28

38 64 13 46 36 36

22804

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

190 320 400 555 600 960

175 195 330 380 505 700

15 125 70 175 95 260

8 39 18 32 16 27

40 55 31 49 15 31

22804

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

230 344 376 670 550 750

218 230 325 540 440 560

12 114 51 130 110 190

5 33 14 19 20 25

25 47 32 28 26 45

Tubigrip

widthway 25 widthway 50 widthway 75 widthway 100

136 180 215 265

125 170 200 230

11 10 15 35

8 6 7 13

+7 1 13 8

Tubigrip

widthway 25 widthway 50 widthway 75 widthway 100

130 170 225 260

120 160 205 220

10 10 20 40

8 6 9 15

+3 7 11 12

*% Loss of stress relaxation after 23 hours when compared with unwashed sample.

*% Loss of stress relaxation after 23 hours when compared with unwashed sample.

Table III. Twenty-three Hours Tensile Stress Relaxation Test (After 1st Wash)

Table IV. Twenty-three Hours Tensile Stress Relaxation Test (After 5th Wash)

19

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Tension (G) 350 –

Tension (G) 500 –

300 –

●

✕

✕ ●

●

●

●

✕

✕

250 – 200 –

+

+

✕

400 –

●

✕ ✕

300 – +

+

+

+

150 –

200 –

100 –

●

❋

50 – 0– 0

0.5

1

1.5

2

2.5

3

0– 0

3.5

●

❋

100 –

0.5

1

●

❋

1.5

●

●

❋

❋

2

2.5

❋

3

3.5

Log time (minutes)

Log time (minutes)

Graph 1: 5 per cent lengthway before wash

Graph 2: 5 per cent widthway before wash

Tension (G) 600 –

Tension (G) 800 –

✕

● ●

500 –

✕ ❋

400 –

+

●

✕

●

●

✕

✕

❋

❋

❋

+

❋

+

+

+

✕

✕

●

✕

600 –

✕

✕

✕

✕

❋

+

400 –

300 –

❋

200 –

●

+

200 –

❋ ●

+

❋ ●

❋

●

+

❋

●

+

+

●

+

100 – 0– 0

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

1

1.5

2

2.5

3

3.5

Log time (minutes)

Log time (minutes)

Graph 3: 15 per cent lengthway before wash

Graph 4: 15 per cent widthway before wash

Tension (G) 1000 –

Tension (G) 1200 – ✕

800 –

1000 –

● ●

●

●

✕

● ●

❋

+

600 –

✕

❋

+ ✕

✕

✕ ✕

800 –

❋

❋

+

+

✕

✕

❋ +

✕

❋

❋

✕

+

+

✕

600 –

●

400 –

+❋ ●

❋

+ ●

400 – 200 –

0– 0

❋ +

❋ +

●

●

❋ + ●

200 –

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

1

1.5

2

2.5

3

3.5

Log time (minutes)

Log time (minutes)

Graph 5: 25 per cent lengthway before wash

Graph 6: 25 per cent widthway before wash

Key: ● 25324

+ 23883

❋

28432

24206

✕

22804

Figure 1. Before Wash

20

Tubigrip

VOLUME 6 NUMBER 4 1994

Tension (G) 350 –

Tension (G) 350 – ●

300 –

✕ ●

●

●

300 –

●

✕

●

250 –

250 – +

200 –

❋ ✕

✕ ✕

+

+

+

+

✕ ❋

✕ ❋

+

✕ ❋

✕ ❋

✕ ❋

✕

200 –

150 –

✕

150 –

●

❋

+

100 –

100 –

50 –

50 –

0– 0

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

1

Log time (minutes)

●

❋ +

●

●

❋ +

1.5

●

❋

❋

+

2

●

❋

+

2.5

+

3

3.5

Log time (minutes)

Graph 7: 5 per cent lengthway after first wash

Graph 8: 5 per cent widthway after first wash

Tension (G) 600 –

Tension (G) 600 – ✕ ●

500 – + ❋ ✕

400 –

●

+ ❋ ✕

● ●

●

+

+

500 –

●

✕ ✕ ✕

+ ❋

❋

✕

✕

✕

400 –

+

❋

❋

❋

✕

+ ●

✕

300 –

300 –

200 –

200 –

100 –

100 –

0– 0

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

1

Log time (minutes)

1.5

❋ + ●

2

+❋

❋ +

●

2.5

●

3

3.5

Graph 10: 15 per cent widthway after first wash

Tension (G) 1000 –

Tension (G) 800 –

✕ ● ●

+ ❋

●

✕

●

● ●

+ ❋

✕ ✕

+ ❋

+ ❋

✕

✕

+ ❋

+ ❋

✕

✕

800 – + ❋

600 –

✕

400 –

●

+ ❋

●

400 – 200 –

0– 0

+ ●

Log time (minutes)

Graph 9: 15 per cent lengthway after first wash

600 –

✕

❋

+ ●

✕

✕

+ ❋

+ ❋

✕

+ ❋

●

●

+ ❋

● ●

200 –

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

Log time (minutes)

+ 23883

❋

28432

1.5

2

2.5

3

3.5

Log time (minutes)

Graph 11: 25 per cent lengthway after first wash

Key: ● 25324

1

24206

Graph 12: 25 per cent widthway after first wash

✕

22804

Figure 2. After First Wash

21

Tubigrip

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Tension (G) 350 –

Tension (G) 400 – ●

●

●

●

300 –

350 –

● ●

✕ ✕

300 –

250 –

❋ ✕

+

200 –

❋ ✕

+

❋ ✕

❋ ✕

+

+

✕ ✕

❋ ✕

✕

250 –

❋ ✕

+

+

✕

● ●

200 – 150 –

+

150 – 100 –

❋

+ ❋

●

● ●

+

●

+ ❋

❋

+ ❋

+ ❋

100 –

50 –

50 –

0– 0

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

1

Log time (minutes)

1.5

2

2.5

3

3.5

Log time (minutes)

Graph 13: 5 per cent lengthway after fifth wash

Graph 14: 5 per cent widthway after fifth wash

Tension (G) 600 –

Tension (G) 700 – ●

✕ ●

●

●

●

500 – ❋

400 –

+

❋

✕

✕

+

❋ +

+

✕

✕

✕

600 –

●

✕

✕

✕ ✕

500 –

❋

❋

+

❋ +

✕

● +

400 –

✕

300 –

❋

300 –

●

+ ❋

●

●

+

●

+

❋

●

+

❋

+

❋

❋

200 –

200 –

100 –

100 –

0– 0

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

1

Log time (minutes)

2.5

3

3.5

Tension (G) 800 –

Tension (G) 800 – ●

●

●

● ●

❋

+ ✕

❋

+ ✕

❋

+

+❋

✕ +

●

+

❋

600 –

+

✕

+ ✕

+

❋

✕ ❋

●

❋

●

✕

+ ✕

+ ✕

+

❋

●

●

❋ ● ● ❋

✕

✕

400 –

400 –

200 –

0– 0

2

Graph 16: 15 per cent widthway after fifth wash

Graph 15: 15 per cent lengthway after fifth wash

600 –

1.5

Log time (minutes)

200 –

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

Log time (minutes)

+ 23883

❋

28432

1.5

2

2.5

3

3.5

Log time (minutes)

Graph 17: 25 per cent lengthway after fifth wash

Key: ● 25324

1

24206

Graph 18: 25 per cent widthway after fifth wash

✕

22804

Figure 3. After Fifth Wash

22

Tubigrip

VOLUME 6 NUMBER 4 1994

Tension (G) 350 –

●

Tension (G) 250 – ●

●

●

✕

●

●

●

✕

300 – ●

●

✕

●

✕ ●

✕

200 –

250 –

●

❋ ✕

200 – +

✕

❋ ✕

❋ ✕

❋ ✕

❋ ✕

+

+

+

+

150 –

150 –

❋ ✕

+ ❋

+

+

+

❋

❋

+

+

❋

100 –

+

❋ ❋

100 – 50 – 50 – 0– 0

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

1

Log time (minutes) Graph 19: 5 per cent lengthway after tenth wash

●

❋

+

400 –

✕

●

❋

+ ✕

●

●

❋

❋

+

+

✕

❋

❋

❋

+

✕

+

+

✕

✕

100 –

2

2.5

3

❋

+

●

✕ ✕

✕

❋

+

❋

●

+ ●

✕

❋

+●

❋

+

●

0– 0

3.5

0.5

1

Log time (minutes)

1.5

2

2.5

3

3.5

Log time (minutes) Graph 22: 15 per cent widthway after tenth wash

Graph 21: 15 per cent lengthway after tenth wash

Tension (G) 1000 –

Tension (G) 1000 – ●

❋

●

❋

●

●

+

+

✕

✕

✕

❋ ✕

❋

❋

+ ✕

600 –

❋

+

+ ❋ ✕

●

❋

+

+

800 –

●

+

+

❋ ✕

+

❋

+

❋

✕

●

+

●

❋ ✕

✕

●

●

●

●

✕

●

✕

400 –

400 –

200 –

200 –

0– 0

3.5

300 –

100 –

1.5

●

400 –

200 –

1

✕

●

200 –

600 –

3

✕

500 –

●

300 –

800 –

2.5

Tension (G) 600 –

500 –

0.5

2

Graph 20: 5 per cent widthway after tenth wash

Tension (G) 600 –

0– 0

1.5

Log time (minutes)

0.5

1

1.5

2

2.5

3

0– 0

3.5

0.5

Log time (minutes)

+ 23883

❋

28432

1.5

2

2.5

3

3.5

Log time (minutes)

Graph 23: 25 per cent lengthway after tenth wash

Key: ● 25324

1

24206

Graph 24: 25 per cent widthway after tenth wash

✕

22804

Figure 4. After Tenth Wash

23

Tubigrip

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

tension against log time. But in the widthwise direction, the UK fabrics, and especially #25324, were much better in elastic properties than the two samples from Hong Kong (see Table VIII).

Fabric

n From the test results shown in Table VI which compare the load after the specimen is stretched before and after washing, it may be observed that the tested fabrics (except the Tubigrip) respond differently if stretched lengthwise and widthwise at the same extension. The three UK fabrics lengthways gave higher tensions than widthwise, but the two Hong Kong fabrics had higher tension widthwise. The Tubigrip performed best in elastic properties particularly before washing: the slopes of the curves range from –0.6 to 7.1 when the extension was up to 50 per cent. However, the slopes change to a relatively higher value when the extension was up to 100 per cent, especially after repeated washing. This indicates that a much greater loss in tension will be shown when the fabric is stretched to 100 per cent extension.

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

344 235 525 410 825 570

285 175 472 340 680 448

59 60 53 70 145 122

17 26 10 17 18 21

7 +30 16 +19 14 10

23883

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

178 150 450 435 610 820

160 135 415 325 560 720

18 15 35 110 50 100

10 10 8 25 8 12

22 +38 +9 +31 9 +18

28432

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

226 125 500 450 675 735

200 95 450 355 590 590

26 30 50 95 85 145

12 24 10 21 13 20

0 10 +6 +13 8 13

24206

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

200 100 290 350 435 640

145 80 252 285 360 425

55 20 38 65 75 215

28 20 13 19 17 34

26 64 20 29 33 34

22804

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

215 230 375 530 500 740

192 170 325 400 430 560

23 60 50 130 70 180

11 26 13 25 14 24

34 61 32 47 28 45

Tubigrip

widthway 25 widthway 50 widthway 75 widthway 100

110 145 215 245

95 130 180 200

15 15 35 45

14 10 16 18

19 24 22 20

Table V. Twenty-three Hours Tensile Stress Relaxation Test (After 5th Wash)

The stress relaxation test was repeated on the same specimen after repeated washing. Even though the specimen shrank by a certain percentage after washing, it was stretched to the original specified extension. If the loads used for stretching the same specimen are compared before and after washing (see Table VI), then it may be found that the loads after washing are similar to, or higher than, the load before washing. This is due to the specimens having shrunk to a smaller size but then having stretched back to the original extension. For the two fabrics from Hong Kong the shrinkage was comparatively low (see Table IX), therefore the load required to stretch the fabric after washing was not much higher than that before washing. The Tubigrip is an exceptional case; even though the fabric had shrunk 4-5 per cent the load after washing was similar to or even smaller than the load before washing, which indicates that the elastic of the rubber may deteriorate after repeated washing.

250 – ✕

✕ ❖

✕ ✕

❖

200 – ❋

❋

❖

❖

❋

❋

❋

●

●

+

●

+

+

●

❖ ❋

150 – ●

+ 100 –

●

+

+

50 –

0– 0

0.5

1

1.5

2

2.5

3

3.5

Log time (minutes)

Key: ●

25 per cent 0 Washes

❋ 50 per cent 0 Washes ✕

75 per cent 0 Washes 100 per cent 0 Washes

Loss % Loss *% Loss

*% Loss of stress relaxation after 23 hours when compared with unwashed sample.

Tension (G) 300 –

❖

Load (g force) 10 mins 23 hrs

25324

n

The tested fabrics respond differently if stretched lengthwise and widthwise at the same extension

Test direction and stretch %

+ 25 per cent 10 Washes 50 per cent 10 Washes

❖ 75 per cent 10 Washes 100 per cent 10 Washes

Figure 5. Tubigrip (Before and After Wash)

24

VOLUME 6 NUMBER 4 1994

Fabric 25324

Test direction and stretch % lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

305 135 560 285 790 500

305 152 525 330 745 470

328 225 544 436 770 580

344 235 525 410 825 570

23883

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

204 98 382 248 618 610

218 128 440 330 645 680

218 165 402 420 635 730

178 150 450 435 610 820

28432

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

200 105 425 315 640 680

195 136 425 390 640 650

242 150 418 340 670 630

226 125 500 450 675 735

24206

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

195 220 315 400 535 640

155 175 288 440 480 555

155 110 307 300 450 570

200 100 290 350 435 640

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

290 435 480 750 595 1020

190 320 400 555 600 960

230 344 376 670 550 750

215 230 375 530 500 740

Tubigrip widthway 25 widthway 50 widthway 75 widthway 100

117 172 230 250

136 180 215 265

130 170 225 260

110 145 215 245

22804

cent extension, but for 25 per cent extension the elongation increased to 3 per cent.

Load (g force) Before After After After wash 1st wash 5th wash 10th wash

Shrinkage From the test results in Table IX, the shrinkage of the elastic net fabrics was between 0.5 and 2 per cent in the lengthwise direction, and 2.5 and 11 per cent in the widthwise direction. It is obvious that the fabrics are more stable lengthways, probably due to the elastomeric yarn running chiefly in this direction. Comparing the shrinkage in the widthwise direction, the Tubigrip and the three United Kingdom fabrics exhibited relatively higher levels than the Hong Kong fabrics. In the case of the five Lycra fabric samples which had been stretched up to 25 per cent, and Tubigrip which was stretched up to 100 per cent, the shrinkage percentage was more or less the

Fabric 25324

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

8 30 9 16 8 15

8 24 9 12 9 15

10 29 7 17 8 14

17 26 10 17 18 21

23883

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

7 14 6 8 9 15

9 22 9 13 11 19

9 18 8 27 6 22

10 10 8 25 8 12

28432

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

21 19 9 13 11 19

12 18 11 24 12 18

12 20 14 25 14 23

12 24 10 21 13 20

24206

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

17 36 18 28 23 41

23 29 20 26 21 28

23 27 10 28 24 28

28 20 13 19 17 34

22804

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

10 26 13 25 16 25

8 39 18 32 16 27

5 33 14 19 20 25

11 26 13 25 14 24

Tubigrip widthway 25 widthway 50 widthway 75 widthway 100

4 1 9 10

8 6 7 13

8 6 9 15

14 10 16 18

Table VI. Comparing Tables II, III, IV, V for the Load (g force) After Stretched Ten Minutes

Elongation After Stretching The test results as shown in Tables X and XI indicate that the amount of widthwise elongation after being stretched and relaxed increases with the number of washings. This indicates that the stretch and recovery properties of the fabrics in the widthwise direction is affected by repeated washing. However, in the lengthwise direction, based on the experimental washings, the stretch and recovery properties showed no significant change in up to ten washes. The residual elongation of the specimen after being stretched and relaxed was higher when the amount of stretch is increased. This occurred both in the widthwise and lengthwise directions, but the results were greater in the widthwise direction. For example, fabric #25324 (before washing) when stretched widthways and relaxed had an elongation of only 0.5 per cent for 5 per

% Loss of load Test direction Before After After After and stretch % wash 1st wash 5th wash 10th wash

Table VII. Comparing Tables II, III, IV, V for the Loss of Load Before and After Washing

25

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Fabric

Slope Test direction Before After After After and stretch % wash 1st wash 5th wash 10th wash

Fabric

After After Test 1st wash 5th wash direction and stretch % Length % loss Length % loss

After 10th wash Length % loss

25324

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 15 25

–10.3 –18.3 –21.4 –21.4 –28.3 –34.2

–12.6 –16.7 –23.7 –17.1 –26.9 –32

-16 –30.6 –18.6 –34.3 –32.5 –38

–20 –28.4 –22.9 –31.4 –51.5 –47

25324

lengthway lengthway lengthway widthway widthway widthway

5 15 25 5 15 25

9.9 9.9 9.9 9.4 9.25 9.2

1 1 1 6 7.5 8

9.85 9.85 9.85 9.2 9.15 9.15

1.5 1.5 1.5 8 8.5 8.5

9.85 9.85 9.8 9 8.9 8.9

1.5 1.5 2 10 11 11

23883

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

–5.7 –7.7 –11.4 –10.6 –28.6 –40

–8 –13.7 –18.6 –20.6 –29.1 –56.6

–11.4 –14.3 –15.7 –54.3 –11.4 –61.5

–8.3 –6.3 –15.1 –47.1 –22.3 –48.6

23883

lengthway lengthway lengthway widthway widthway widthway

5 15 25 5 15 25

9.95 9.95 9.95 9.45 9.45 9.45

0.5 0.5 0.5 5.5 5.5 5.5

9.95 9.95 9.95 9.35 9.4 9.4

0.5 0.5 0.5 6.5 6 6

9.95 9.95 9.95 9.25 9.3 9.3

0.5 0.5 0.5 7.5 7 7

28432

28432

lengthway widthway lengthway widthway lengthway widthway

5 5 15 15 25 25

–17.7 –7.7 –17.1 –16.6 –30.6 –54.3

–10.9 –12 –20 –41.7 –34.3 –48.6

–11.7 –12 –25.1 –38.6 –41.7 –69.4

–10.3 –13.4 –18 –45.7 –34.3 –61.7

lengthway lengthway lengthway widthway widthway widthway

5 15 25 5 15 25

9.95 9.95 9.95 9.45 9.45 9.5

0.5 0.5 0.5 5.5 5.5 5

9.85 9.9 9.9 9.4 9.35 9.4

1.5 1 1 6 6.5 6

9.8 9.85 9.85 9.25 9.25 9.3

2 1.5 1.5 7.5 7.5 7

24206

lengthway widthway lengthway widthway lengthway widthway

5 –17.8 5 –40 15 –22.9 15 –52.9 25 –66.6 25 –122.9

–15.1 –25.1 –25.7 –55.7 –47.7 –80

–14.3 –14.3 –14.3 –31.4 –47.7 –76.3

–23.7 –10.9 –17.1 –35.7 –29.4 –85.7

lengthway lengthway lengthway widthway widthway widthway

5 15 25 5 15 25

9.95 9.95 9.95 9.75 9.75 9.75

0.5 0.5 0.5 2.5 2.5 2.5

9.9 9.9 9.9 9.7 9.7 9.7

1 1 1 3 3 3

9.85 9.85 9.85 9.6 9.6 9.6

1.5 1.5 1.5 4 4 4

22804

lengthway widthway lengthway widthway lengthway widthway

5 –15.1 5 –57.7 15 –28.6 15 –91 25 –40.6 25 –111.4

–7.4 –58.8 –30 –79.7 –37.1 –130.8

–5.1 –50.6 –22.9 –55.2 –57.1 –91.8

–9.4 –23.2 –22.9 –60 –33.4 –89.1

lengthway lengthway lengthway widthway widthway widthway

5 15 25 5 15 25

9.95 9.95 9.95 9.75 9.75 9.8

0.5 0.5 0.5 2.5 2.5 2

9.95 9.95 9.95 9.55 9.55 9.55

0.5 0.5 0.5 4.5 4.5 4.5

9.95 9.95 9.95 9.5 9.5 9.5

0.5 0.5 0.5 5 5 5

–0.6 –1.4 –0.6 –5.2 –7.7 –8.5 –8.6 –15.2

–2.9 –5.2 –9.7 –16.5

–4.3 –7.1 –19.1 –21.4

Tubigrip widthway 25 widthway 50 widthway 75 widthway100

9.5 9.55 9.6 9.6

5 4.5 4 4

9.55 9.6 9.65 9.6

4.5 4 3.5 4

9.6 9.55 9.55 9.55

4 4.5 4.5 4.5

24206

22804

Tubigrip widthway 25 widthway 50 widthway 75 widthway 100

Table IX. The Shrinkage in Length After Repeated Washing

Table VIII. Comparing the Slope Loss Tension/Log Time Before and After Repeated Washing

those of the Hong Kong fabrics. Since different Lycra fabrics may have different elastic properties in lengthwise and widthwise directions, it is important to identify in which direction the fabric will be cut for the making of pressure garments. Obviously, for effective compression, the run-in direction of the Lycra yarn in the fabric will be used as the stretch direction of the pressure garment. The shrinkage of the fabric range, from 0.25 to 2 per cent lengthways, was within the normal acceptable limits and would therefore not appreciably affect the size of the pressure garments after washing. However, the shrinkage in the widthwise direction was quite large; for example, the shrinkage of the sample #25324 was up to 11 per cent after being washed ten times. It was also found that samples from the UK had comparatively higher shrinkage in the widthwise direction. Shrinkage of this order appears to be beneficial as it compensates for losses in elasticity during washing, so the load under stress can maintain more or less the value after washing due

same regardless of the degree to which they had been stretched. These test results show that the percentage of shrinkage was not affected by the amount of stretch. It was also observed that the percentage of shrinkage increased as the number of washes increased. Exceptions were the Tubigrip, and fabrics #23883 and #22804, which showed no difference lengthways after repeated washings.

Conclusion All the tested fabrics with the exception of the Tubigrip exhibited similar overall characteristics. From the point of stress relaxation, the lengthwise was better than the widthwise direction. The rate of stress relaxation was similar in the lengthwise direction, but widthwise the three United Kingdom fabric samples were much lower than

26

VOLUME 6 NUMBER 4 1994

Elongation % After After After 1st wash 5th wash 10th wash

Fabric and stretch %

Relaxation time

Before wash

25324 5

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

0.5 0 1.7 0 2 1

1 0.5 1 1 0.5 0.5

1 1.5 1.5 1 1 0.5

1.5 0.5 1.5 0.5 2 1.5

25324

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

0 0 0.5 0.5 1 0.5

0 0 1 0.5 2.5 1

0.5 0 1 0.5 2.5 1

0.5 0 1 0.5 1.5 1

23883

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

0.5 0.5 1 0.5 1 0.5

0 0 1 0.5 2 1

0.5 0 0.5 0 2 1

0.5 0 1.5 0.5 2 1.5

28432

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

1 0 2 1 2 1

0.5 0.5 1.5 0.5 2 0.5

0 0 1.5 1 2 1

0.5 0 1 0.5 3 2

24206

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

0 0 0.5 0.5 1 0.5

0.5 0.5 1 0.5 1.5 0.5

0 0 0.5 0 0.5 0

0.5 0 0.5 0.5 1.5 0.5

22804

15 25

23883 5 15 25

28432 5 15 25

24206 5 15 25

22804 5 15 25

Fabric and stretch % 5 15 25

5 15 25

5 15 25

5 15 25

5 15 25

Tubigrip 25

Table X. The Elongation Percentage After Stretched (At Lengthway) for 23 Hours and After Relaxation at Zero Stress

50 75 100

to the shrinkage in the fabric. It is believed that if the shrinkage is higher than the elongation after being stretched, a more constant pressure could be provided throughout the life of the garment. Tubigrip behaves completely differently from the other tested fabrics. It is very good in elastic properties but in the widthwise direction only. This behaviour is mainly related to its special fabric construction resulting from the natural rubber yarns which are inserted in widthwise direction. The Tubigrip loses the least percentage of load if the stress is within the elastic limit, but on the other hand its initial stress is very low. The test results agree with the principle of Meredith[3] that the natural rubber exhibits less power than Lycra but is less prone to stress relaxation. Moreover, the Tubigrip is obviously more affected by repeated washings, as the rate of tension loss increases by a factor of two or three times after a few washes. The amount of tension will probably decrease after further washing, which would indicate that the fabric is not durable in washing.

Elongation % After After After 1st wash 5th wash 10th wash

Relaxation time

Before wash

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

2 0.5 2 1 6 3

2.7 1.1 5.4 3.3 8.2 5.5

4.3 3.3 6 4.4 9.3 6

6.6 6.1 9 7.3 10.1 7.9

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

0.5 0 1.5 1 3.5 2

2.6 1.6 6.3 3.7 10.1 7.4

3.2 2.1 6.5 5.5 10.6 6.4

2.7 1.1 3.2 2.7 9.1 4.3

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

1 0 1.5 0.5 1.5 1.5

1.6 1.1 5.8 2.6 5.8 3.2

3.2 1.2 3.7 2.4 5.9 4.3

3.2 1.6 4.9 0.5 8.1 5.4

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

2.5 0.5 2 1 7 4

2.6 1.5 5.6 3.1 8.2 3.6

4.1 3.1 6.2 4.6 10.3 7.7

4.2 2.1 6.6 4.2 4.2 2.6

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

0.5 0 3 1.5 5.5 3.5

1.5 1 3.6 2.1 6.6 3.6

2.6 1.6 7.8 1.6 8.9 6.3

3.2 2.1 6.3 2.1 8.4 6.3

30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs 30 mins 23 hrs

0 0 1.5 0.5 4 2 4 3

0 0 2.6 1.6 5.8 3.6 7.8 4.2

2.6 1.6 3.1 1.6 4.7 1.6 8.3 5.2

2.5 2.1 3.1 1.6 4.7 3.7 8.9 6.3

Table XI. The Elongation Percentage After Stretched (At Widthway) for 23 Hours and After Relaxation at Zero Stress

References 1. Ng Yip, S.F., “Medical Clothing – A Tutorial Paper on Pressure Garments”, International Journal of Clothing Science and Technology, Vol. 5 No. 1, 1993, pp 17-24. 2. Cheng, J.C.Y., Evans, J.H., Leung, K.S., Clarke, J.A., Choy, I.T.C. and Leung P.C., “Pressure Therapy in the Treatment of Post-burn Hypertrophic Scar – A Critical Look into its Usefulness and Fallacies by Pressure Monitoring”, Burns, Vol. 10, 1983, pp. 154-63. 3. Meredith R., Elastomeric Fibres, Merrow, 1971.

27

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Development of a Commercial Made-to-Measure Garment Pattern System J.P. Turner Hollings Faculty, The Manchester Metropolitan University, UK Received 24 July 1993 Accepted 12 April 1994

Objectives The main objectives of the research project were: (1) Development of a microcomputer-based system for automatic construction and plotting of “made-to-measure” ladies’ garment patterns. (2) Implementation of the developed system, including comprehensive trials and utilization in the commercial environment of the host company, in particular for bridalwear, the product of this company.

Introduction This paper describes the stages of development and implementation of a PC-based CAD system for producing Made-to-Measure (MTM) garment patterns in the commercial environment of a bridalwear manufacturer. The system, which was aptly named MICROFIT, was based on standard PC technology and existing pattern grading software whose source code was available with the following advantages: (1) The hardware with its MS-DOS operating system was readily available and inexpensive, with the prospect of becoming increasingly cheap and powerful into the foreseeable future. (2) Some existing PC-based software routines used for pattern grading could be utilized. (3) None of the programming restrictions imposed by a standard CAD package would impede development. (4) The majority of companies and colleges worldwide already have PCs installed and would be able to utilize the software if they so wished. The project of one year’s duration was supported by a Royal Society/SERC Industrial Fellowship and the host company was Creation Bridalwear of Wigan.

Preliminary Investigations Within the environment of the host bridalwear company, namely Creation of Wigan, all aspects of made-to-measure pattern design and construction, size charts used, and the methods used to take measurements for order form entry and to cut patterns to fit a customer were observed and noted. This was necessary in order to understand how to simulate the hand methods used and implement them in a computerized system. The methods used by the company, although tried and proven, showed shortcomings in the pattern cutting area. In particular, patterns were cut on the large side on the principle that cut patterns can be reduced (taken in) after the first fitting, but that it is not possible to add on material already cut off. Erring on the small side entails wasteful and costly recutting of expensive cloth. Additionally, the occasional human error occurred and this was more likely with new staff who had not completed their training. The time taken to draft and cut a set of individual made-to measure patterns could take up to two hours. The time and cost to train a new pattern cutter was

International Journal of Clothing Science and Technology, Vol. 6 No. 4, 1994, pp. 28-33, © MCB University Press, 0955-6222

28

VOLUME 6 NUMBER 4 1994

considerable. A six-month period, with the constant attention of the proprietor or another experienced cutter, was required for a new employee. Bearing in mind all the above points, the objectives of implementation of a computerized system were therefore expanded as follows:

Software Development Bodice patterns chosen for initial experimentation were of a style which accounted for more than 50 per cent of the company’s customer orders. The computer program logic developed was one that closely mirrored the hand drafting and fitting methods of the company’s pattern cutters. It was, therefore, amenable to modification according to advice from the staff as and when required. Briefly the method adopted is as follows. A base size set of patterns is input through the digitizing board. This pattern set is graded according to the company’s size chart which has also been entered into the computer system. These together form the base information for the pattern style. Production of a set of patterns takes place following input through the keyboard of the individual customer’s measurements directly from an order form. Rules are built into the system so that either a set of drafting rules are followed or drafting adjustments to the fitting are applied automatically to the base graded patterns. The MTM patterns are displayed on the computer screen and subsequently plotted out onto paper using a suitable layout. The COMPAQ computer was able to draft a set of made-to-measure bodice patterns and form a layout in about 40 seconds. (A modern PC with 486 processor would be able to reduce this to about ten seconds.) An order queuing logic allows many orders to be processed in succession without operative interference. The whole process of measurement entry, pattern processing and pattern plotting averages about six minutes per customer order giving the system the capability of producing about 300 customer made-to-measure pattern sets in a 30-hour week. Off-line customer order entry and plotting could, of course, increase this capacity many times.

(1) to produce patterns which would need minimal or no alterations at first fitting of the customer; (2) to produce patterns extremely rapidly, thus increasing the capacity of pattern cutting massively; (3) to produce a user interface which was friendly and foolproof, with a very short training period (of the order of days rather than weeks or months); (4) to deskill the pattern cutting area completely so that problems of staff turnover became minimal; (5) to eliminate human error in pattern production. The benefits to the company of a computer system which fulfilled the above objectives would immediately be obvious in terms of pattern accuracy, productive capacity, human resources and training and direct costs of labour and material.

System Hardware The MICROFIT hardware used in the project comprised: ●

IBM compatible PC COMPAQ 386 16MHz processor with 40 Mbyte hard disc, dual 5.25” floppy drives and EGA colour monitor. (This computer was available for use at the outset of the project and was adequate for the development work.)

●

Hewlett Packard Thinkjet Inkjet printer for file listings.

●

Hewlett Packard HP7475A A3 size plotter for miniature reference markers.

●

Calcomp 1044 AO size plotter for full-scale patterns and markers.

●

GTCO 60” x 40” digitizer for base pattern input.

The Principal of Made-to-Measure Drafting Utilized Computerized made-to-measure pattern drafting has been described by Turner[l], who has outlined the methods used by Burtons (The Centaur Clothes Group) at Goole and Hepworths (whose system was subsequently moved to Complan Computer Bureau) for the production of men’s suits. He also developed computer programs for the drafting of ladies’ tailored jackets and skirts and men’s suits using traditional tailor’s code methods. Similar programs which follow the instructions defined by Aldrich[2] for the generation of pattern blocks are described by Jo and Harlock[3,4] in their work on the

A current MICROFIT System would utilize a 486 PC with 100 Mbyte hard disk, 3.5” microdisk drive and SVGA monitor.

29

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

development of an educational garment design system. Made-to-measure pattern production may be carried out by either of two methods, namely: (1) Making alterations to standard graded sizes of the garment as in Gerber (used by Centaur Clothes) or Camsco (used by Complan). (2) Drafting a block or styled pattern to measure by programmed instructions. The MICROFIT method chosen for the production of bridalwear patterns combines the strengths of both of these methods by using drafting rules (instead of grade rules) which in combination with a size chart can produce all the standard sizes automatically and can equally produce any one-off made-to-measure garment to a set of personal measurements. The versatility of the alteration method is retained whereby the base shapes for a style may be entered by digitization, while the drafting instructions laid down in drafting rule tables modify the base shapes in girth and length according to the input measurements. The rules applied at each point on the pattern pieces are functions of the measurements rather than the numerical values as used in grading. Table I shows a bridalwear size chart which spans the sizes 8 to 20. Notice that the measurements are all in imperial inches, indicated by an I entered under the column headed “U” (units). An M in this column would indicate metric centimetres. Inch units are still commonly used in the UK made-to measure business. Notice also that under column “a”, the abbreviated single letter names B, W, H, C etc. represent the incremental values of BUST, WAIST, HIPS, CROSS CHEST etc. These abbreviated single letter names are used in the drafting rule tables.

* SSS 7 sizes nnndddddddddddda U 001BUST BI 002WAIST WI 003HIPS HI 004CROSS CHEST CI 005CROSS BACK XI 006F B LENGTH FI 007B B LENGTH LI 008BUST DEPTH DI 009SLV LENGTH SI 010WRIST RI 011SKIRT LENGTH KI

sssss 8 mmmmm 30 24 32 11.5 12 16 15 9.5 26 5.75 40.5

sssss 10 mmmmm 32 26 34 12 12.5 16.5 15.5 10 26.5 6 41

Tables II and III show two examples of an individual person’s measurements as taken in the bridalwear shop and entered on to an order form. The base size 14 measurements are also shown so that the incremental difference can be calculated and shown in the last column of each table. Table IV shows a drafting rule table. Notice that all the rules, both for length x and girth y, are functions of length or girth measurements, or alternatively zero (0) indicating no movement. The pattern in Figure 1 is an example from a bridal dress bodice and the rule number applied at each cardinal point is one of the drafting rules in Table IV. For example, Rule 2 is ∆x = 0, ∆y = C/2. This means that the point moves by half the incremental value of the cross chest measurement in the girth direction (vertically), but does not move in the length direction. For the personal measurement set of Table II, C = –0.5" and C/2 = –0.25", and the points redrafted by rule 2 move one quarter of an inch vertically downwards due to the narrower cross chest. Similarly, Rule 3 is ∆x = D, ∆y = B/8. From Table II, D = –0.5" and B = 1", B/8 = 0.125". Thus the bust depth is half an inch less than on the standard size and the point is moved 1/2" to the left, while moving also 1/8" vertically upwards. Rules 4, 6 and 14 are interpreted using the same method. The new draft through these points for this individual customer’s measurements is seen superimposed over the size 14 in Figure 2 where it can be seen that there is not a great difference between the new pattern and the original 14. These changes, however, make a significant difference to the fit of the eventual garment on the customer. In the case of the pattern produced from the personal measurements in Table III, Figure 3 shows very great differences between the new

sssss 12 mmmmm 34 28 36 12.5 13 17 16 10.5 27 6.25 41.5

Table I. Bridalwear Size Chart in Imperial Inches

30

sssss 14 mmmmm 36 30 38 13 13.5 17.5 16.5 11 27.5 6.5 42

sssss 16 mmmmm 38 32 40 13.5 14 18 17 11.5 28 6.75 42.5

sssss 18 mmmmm 40 34 42 14 14.5 18.5 17.5 12 28.5 7 43

sssss 20 mmmmm 42 36 44 14.5 15 19 18 12.5 29 7.25 43.5

VOLUME 6 NUMBER 4 1994

Measurements in inches BUST WAIST HIPS CROSS CHEST CROSS BACK FRONT BODICE LENGTH BACK BODICE LENGTH BUST DEPTH SLEEVE LENGTH WRIST SKIRT LENGTH

Personal measurements

Size 14

37 29.5 38 12.5 14 17 16.5 10.5 26.5 6.3 43

36 30 38 13 13.5 17.5 16.5 11 27.5 6.5 42

Increment B W H C X F L D S R K

= = = = = = = = = = =

37 29.5 38 12.5 14 17 16.5 10 26.5 6.3 43

– – – – – – – – – – –

36 30 38 13 13.5 17.5 16.5 11 27.5 6.5 42

= = = = = = = = = = =

1 0.5 0 –0.5 0.5 –0.5 0 –0.5 –1 0.2 1

Table II. First Example of Individual Measurements

pattern and the base size. In particular, the girth of bust and hips is large (between sizes 18 and 20 on the size chart), whereas the waist girth is relatively very small (between sizes 10 and 12 on the size chart). The waist suppression on the side panel is particularly severe. A good fit would be impossible from a standard size garment.

operative would easily learn to perform the task of order processing within a two-hour training session.

System Enhancements The system was enhanced to cover the full range of bodice patterns and all the sleeves and skirts produced by the company. The software allows bodices, sleeves and skirts to be mixed and matched so that the pattern pieces of the chosen styles in each may be brought together and formed into a lay-plan. Automatically formed layplans ready for cutting are produced on the system by programming the order of pattern insertion and manipulation. This saves the time of an operative moving the pattern pieces on screen interactively until an efficient layout is formed. It also allows uninterrupted automatic processing of a queue of orders.

Results of Initial Development Phase During this phase nearly 100 orders of the most common style were processed by the system. The results showed that all the objectives of computerization were achievable: patterns fit the customer more closely; the sewing machinists can allow smaller seam allowances; the pattern cutter’s role of pattern drafting has been almost eliminated and the skilled designer/pattern cutter employed by the company can now concentrate on creative designing. The computer programs are menu driven and extremely user friendly – a new

Measurements in inches BUST WAIST HIPS CROSS CHEST CROSS BACK FRONT BODICE LENGTH BACK BODICE LENGTH BUST DEPTH SLEEVE LENGTH WRIST SKIRT LENGTH

Personal measurements

Size 14

41 27 43 14 14.5 17.5 16.5 11 28.5 6.7 42.5

36 30 38 13 13.5 17.5 16.5 11 27.5 6.5 42

Table III. Second Example of Individual Measurements

31

Increment B W H C X F L D S R K

= = = = = = = = = = =

41 27 43 14 14.5 17.5 16.5 11 28.5 6.7 42.5

– – – – – – – – – – –

36 30 38 13 13.5 17.5 16.5 11 27.5 6.5 42

= = = = = = = = = = =

5 –3 5 1 1 0 0 0 1 0.2 0.5

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

nnnxxxxxxxxxxxxyyyyyyyyyyyy 001 0 0 002 0 C/2 003 D B/8 004 F B/8 005 F H/8 006 F 0 007 0 X/2 008 L 0 009 L B/8 010 L H/8 011 S R/2 012 S –R/2 013 0 K 014 C/5 0 015 C/5 B/8 016 K 0 017 C 0 018 D 0 019 D B/8 020 F –W/8 021 D –B/8 022 0 –C/2 023 L W/4-B/8

FRLIN 39705 14+5432A

Figure 2. Personal Measurements from Table II Superimposed on Size 14 Front Panel

The size chart itself must have correct measurements throughout the whole size range. This is important for the production of standard sizes, if these are required, because the software produces these directly from the size chart measurements, rather than by grading. One of the

Table IV. Drafting Rules for Bridalwear

The Need for a Disciplined Approach The MICROFIT computer system, while having the capability of producing very accurate patterns which fit well at first fitting, imposes responsibilities and disciplines on the company in order to achieve this. The base size pattern set of each must reflect exactly the base size measurements in the size chart. The measurements taken of individual customers must be carried out in a consistent way. This necessarily means that if more than one employee has this responsibility, they should all be trained to use exactly the same repeatable measuring method.

FRLIN 39705 14+5433A

2 2 3

+ 14 zero reference point

4

FRLIN 39705 12

4

SDFRT 39703 14+5433A

6

Figure 3. Personal Measurements from Table III Superimposed on Size 14 Front Panel and Side Panel

Figure 1. Bodice Front Panel

32

VOLUME 6 NUMBER 4 1994

MICROFIT system developed during this Industrial Fellowship. The MICROFIT system combines customer order entry and measurement validation, with pattern processing, lay-planning and plotting all on the same PC, thus providing an inexpensive and compact solution for the small and medium-sized company. The versatility of MICROFIT makes it a suitable basic system for other made-to-measure garments such as ladies’ jackets, skirts and trousers, men’s suits and specialized wear such as wet and dry suits for divers. For college use, MICROFIT can be added to standard PC based grading systems in order to familiarize staff and students with made-tomeasure techniques.

strengths of such a system is that it can produce standard sizes from any size chart, whether this be a company or national chart, and in either inches or centimetres. The setting-up of a completely new style on the computer does require a skilled pattern technician. This, of course, only happens each season when new styles are introduced to the company’s range. In the case of bridalwear the pattern range is very traditional and stable.

Summary The PC-based system MICROFIT has been developed and implemented in a bridalwear company in order to automate the production of made-to-measure brides’ and bridesmaids’ dress patterns. The system addresses well the needs of the company in that the patterns produced are accurate, the day-to-day skills of pattern production are eliminated, the system itself is relatively easy to operate and the potential production from even a slow PC is more than sufficient for the company’s pattern output requirements. The company has now purchased its own hardware and is operating the system with all the benefits envisaged in the original objectives.

n References 1. Turner, J.P., A Computerised Technique for the Production of Clothing Patterns, PhD Thesis, UMIST, 1986. 2. Aldrich, W., Metric Pattern Cutting, Bell and Hyman Ltd, London, 1985. 3. Jo, J.S. and Harlock, S.C., “Developing an Educational CAD System for Garment Design”, International Journal of Clothing Science and Technology, Vol. 2 No. 1, 1990, pp 16-20. 4. Jo, J.S. and Harlock, S.C., “Developing an Educational CAD System for Garment Design”, International Journal of Clothing Science and Technology, Vol. 2 No. 2, 1990, pp 23-30.

Conclusion While made-to-measure software is available on Gerber, Lectra and other commercial CAD/CAM garment systems, these are more costly and more complex to operate than the PC-based

33

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

COMMUNICATIONS

Influence of a Spinning Process on Spun Yarn Quality and Economy of Yarn Production Momir Nikolic′ and Janez Cerkvenik University of Ljubljana, Slovenia v

Zoran Stjepanovic University of Maribor, Slovenia

Introduction

Mechanical Models of Staple Yarn Formation

Flat textiles (woven and knitted fabrics) for textile purposes have to fulfil certain mechanical, physical, textural, aesthetical, and economic demands as well as some special requirements for technical end-uses. Their design and construction need an interdisciplinary co-operation of spinners, weavers, knitters, garment manufacturers and fashion designers due to the wide range of their properties. The choice of yarns for different purposes was very simple until the year 1963 because only ring spinning was used for the spinning of cotton type staple yarns. Today it is much more difficult to choose the suitable yarn for the optimal quality and economic effect of a certain final product since rotor and air-jet spinning are also commercially widespread. The purpose of the article is to present the essential differences between the ring, rotor and air-jet spun yarns according to their structure, mechanical and physical properties, which are to be taken into account while planning and designing flat textiles.

A thorough knowledge of the formation models of the yarns in question is necessary for the determination of differences in their micro- and macrostructure, mechanical and physical properties. Mechanical Model of Ring-spun Yarn Formation Figure 1 shows the model of yarn formation by the ring spinning frame. The roving is thinned by draw frames, fixed (twisted) and at the same time wound to the tube by the ring-traveller-spindle system. The torsion energy (i.e. twist) of the ring-spun yarns is transmitted from outside of the continuous fibre bundle to its interior by helical twisting of fibres at the spinning triangle. Fibres in the yarn are maximally straightened, paralleled, longitudinally orientated and axially tensioned. It is characteristic for the ring-spinning process that fibres in the yarn evenly take over helical twisting around the yarn axis. The torsion energy is therefore equally distributed across the ring-spun yarn cross-section (Plate 1). Fibres in the yarn skin are more axially stressed due to the higher length of helical twisting. These fibres take over the radial stress while the yarn core fibres bear axial stress. The radial pressure is

International Journal of Clothing Science and Technology, Vol. 6 No. 4, 1994, pp. 34-40, © MCB University Press, 0955-6222

34

VOLUME 6 NUMBER 4 1994

1 2

Tm

S

Z

Tm

3

Bw/PES 50/50 Tt = 20 tex

Plate 1. The Twist Distribution across the Ring-spun Yarn Cross-section

4

Mechanical Model of Rotor-spun Yarn Formation The rotor spinning is characterized by the formation of the conical bundle of fibres in the rotor groove. The conical bundle is consequently attached to the open rotating yarn end (Figure 2). Fibres in the yarn structure are less straightened, less axially oriented and less parallelized because of the fibre transport to rotor by a turbulent air jet. Smaller axial stress of fibres, which are gathering in the rotor groove, enables the conical yarn ring to impart twist from inside of the yarn to

Key: 1-Roving 2-Spinning triangle 3-Yarn balloon 4-Traveller–ring

Figure 1. The Model of Ring-spun Yarn Formation

transferred from fibres in the skin towards the core fibres with the radial stress increment. Such distribution of radial stress in ring-spun yarn enables all fibres to bear the axial forces. This is the reason for the high strength of ring-spun yarns, i.e. the high efficiency of substantial strength of a fibre bundle. Free movement of the fibre bundle composing the yarn in the area between the take-away rollers and the traveller and also the inertial field of forces active in the yarn balloon cause stronger hairiness with a higher number of fibres protruding from the yarn. The structure of ring-spun yarn has the following good properties due to the continuous connection between the fibres, high radial stress of skin fibres and the twist transfer from outside to the interior of the yarn cross-section: ● high number of straightened fibres, parallelization and orientation of fibres with helical twisting from outside to the interior; ● relatively closed yarn structure and therefore smaller friction coefficient; ● increased yarn hairiness; ● low yarn rigidity; ● worse insulation properties; ● medium abrasion resistance; ● less pronounced pilling effect.

5

A 4 B 3

1 2 2

Iε

5 1

Mt

ωG

ωG 3 Key: 1 – Rotor 1 – Conical bundle of fibres 3 – Open yarn end 4 – Doffing tube 5 – Rotor-spun yarn A-B – The area of false twist transfer

Figure 2. OE Rotor-spun Yarn Formation Model

35

δ

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

its exterior. This is just the opposite to the principle of ring spinning. The yarn open end rolling around its own axis in the rotor groove makes the ad-spinning and the twist transfer to the yarn core fibres possible. The further yarn rolling in the rotor groove at the distance 1E and the yarn rotation around the rotor axis causes further ad-spinning of fibres with smaller radial stress. Yarn core fibres cannot avoid the helical twisting, which causes the increased adhesion and compression forces among them. This results in higher twist accumulation in the rotor yarn core and tougher handle of products made of rotor yarns. Plate 2 shows the structure and distribution of twist in rotor-spun yarn.

Tm

S

Z

The arrangement of fibres and outwards twist transmission results in bilateral structure of a rotor-spun yarn, therefore composed of a helically twisted fibre bundle (i.e. yarn core) and an outer layer of fibres, which is less twisted or even twisted in the opposite direction in comparison with the yarn core. The surface layer of fibres in the rotor-spun yarn skin is built of a thin layer of fibres nearly without twist or with twist of the opposite direction to the yarn core fibre twist and the wrap fibres, which form typical transverse strips along the rotor-spun yarn length. Counter-wrapping of a part of fibres in the yarn skin causes false twist in the area between the rotor wall and doffing tube (area A-B in Figure 3). While shaping a balloon the rotating yarn reaches the area below the transfer tube, used for the feeding of individual fibres. Certain protruding fibres, which should reach the rotor groove, are instead attached to the rotating part of the yarn and wound around it in a shape of a transverse strip. Important properties of rotor-spun yarns resulting from the outwards twist transfer are the following: ● less parallelized, straightened and longitudinally oriented fibres with helical twisting in the yarn core; ● coincidental arrangement of skin fibres with lower twist, which can have a different direction from the core fibre twist;

Tm

Bw/PES 50/50

Tt = 20 tex

Plate 2. Structure and Distribution of Twist in the Rotor-spun Yarn

1 a b

N1 N2

3

2 W

F

C

a

b

B

H

a

b

N1

Z

4 T

S Key: 1 – back roller 2 – condenser 3 – front roller Z 4 – twisting device N1 – first nozzle N2 – second nozzle a – individual skin fibres b – continuous fibre bundle composing the yarn core T – the area of air-jet spun yarn formation

Figure 3. The Scheme of Air-jet Spun Yarn Formation

36

N2

VOLUME 6 NUMBER 4 1994

● ● ● ● ● ● ●

● ● ●

rotor-spun yarn cannot be completely unwound; lower tensile strength of a rotor-spun yarn; more rigid handle of a yarn and flat textiles; low abrasion resistance of a yarn skin and the whole yarn; higher diameter of a yarn of a same fineness; better insulation properties; better Uster evenness; the rotor-spun yarn is less resistant towards the axial abrasion with a sharp edge because of the bad wrapping in the surface layer of fibres; the rotor-spun yarn is less hairy; its surface is rougher and has a higher friction coefficient than a ring-spun yarn; pilling effect is very pronounced.

Fre

Fmax

100 – 80 – Per cent

●

60 – 40 – 20 – 0– 0

10

20

30

40

50

60

70

80

90

100

Pvl per cent

Figure 4. The Dependence of the Tensile Force of Air-jet Spun Yarn on the Percentage of Wrap Fibres

percentage of fibres wrapped at a higher axial stress causes the formation of air-jet spun yarn with higher tensile strength, lower hairiness and higher rigidity (the yarn handle is rougher). Ten to 25 per cent of wrapped fibres are recommended for the formation of air-jet spun yarn of adequate tensile strength. Plate 3 shows the microscopic appearance and distribution of torsion energy in the cross-section of an air-jet spun yarn. Air-jet spun yarn is less elastic and has rougher handle than ring- and rotor-spun yarns due to strongly wrapped fibres around the untwisted core. Also the air-jet spun yarn behaves differently during axial abrasion than ring-spun yarn because of a different principle of structure fixation by means of transverse wrap fibres, which enables suitable adhesion between the fibres in the untwisted core. The movement of a sharp edge along the core axis of a ring-spun yarn causes the movement of

Mechanical Model of Air-jet Spun Yarn Formation Air-jet spinning is a hybrid between the ring and OE-rotor spinning. The air-jet spinning requires that the majority of fibres stay in a spinning triangle as a continuous bundle of parallelized fibres, from which a yarn core is formed. Few fibres remain beside the spinning triangle (Figure 3). The false twist is applied to the fibres in the spinning triangle by one or two aerodynamic nozzles. When a spinning device is equipped with two nozzles, the second one (N2) imparts the Stwist to the core fibres. The first nozzle, whose twisting efficiency is lower, only partially affects the yarn core and twists the yarn skin fibres in the Z-direction in the shape of transverse strips around the yarn core. S-twist is applied to the yarn core and to the twisted skin fibres in the lower part of the first nozzle. The core fibre twist is increased after passing the second nozzle, while the surface transverse strips unwind somewhat. The yarn core is unwound in the Z-direction in the lower part of the second nozzle and the wrapping of skin fibres around the untwisted yarn core intensified. Air-jet spun yarn is composed of a twistless bundle of fibres as a core and individual skin fibres, which helically wrap the untwisted yarn core. The mechanical and physical properties, texture and rigidity of air-jet spun yarn mostly depend on the share of fibres wrapped around the untwisted yarn core (Figure 4). A smaller percentage of wrap fibres and simultaneous loose wrapping around the untwisted yarn core enables the formation of airjet spun yarn with lower tensile strength, increased hairiness and smaller rigidity. A higher

Tm

S

Z

Tm

Bw/PES 50/50 Tt = 20 tex

Plate 3. Structure and Distribution of Twist in Air-jet Spun Yarn

37

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

●

1 — Tm

● Tm max

Tm max

0

● Tm min

● 2 —

●

Key: 1-Ring-spun yarn

●

2-Air-jet spun yarn

●

Figure 5. The Influence of Axial Friction on Twist Distribution

●

twists and the increase of twist in the area, which is not in touch with a sharp edge (marked by an arrow in Figure 5). The twist decreases in the area in touch with the sharp end, when the yarn is weakened and loosened. This is evidence that ring-spun yarn is less resistant towards abrasion and has increased pilling. The transverse strips of wrap fibres in the structure of air-jet spun yarn resist axial disposition with friction and higher stress. This imparts higher resistance to abrasion and less pronounced pilling to air-jet spun yarn compared with ring-spun yarn. The ring-spun yarn is more easily elongated at axial loading due to the easier change of twist angle in the yarn (Figure 6). The elongation of air-jet spun yarn at axial loading is smaller because of the transversely wrapped fibres hindering the change of the twist angle. This behaviour of air-jet spun yarn is the reason for lower contraction of fabrics and better dimensional stability of end products. The characteristics of an air-jet spun yarn as a consequence of its specific structure are as follows:

● ●

high degree of straightening, parallelization and axial orientation of fibres in a twistless yarn core; tensile strength and rigidity of a yarn depend on the number of transversely wrapped strips; air-jet spun yarn cannot be completely unwound; its tensile strength is lower than the tensile strength of a ring-spun yarn and higher than the tensile strength of a rotor-spun yarn; lower yarn hairiness; high resistance towards abrasion and axial friction with a sharp edge; better Uster characteristics compared with ring-spun yarn; increased yarn rigidity; better resistance to pilling; rougher texture and higher coefficient of friction compared with ring-spun yarn.

Comparison of Yarn and Fabric Properties The article is intended for weavers, knitters, garment manufacturers and textile designers. It displays differences in micro- and macrostructure and in mechanical and physical properties of yarns formed by ring, rotor and air-jet spinning. We have compared eight quality parameters for ring, rotor and air-jet spun yarns of the same linear density and fibre content (20 tex × 1, 50 per cent polyester/50 per cent cotton) (Figure 7). The differences in yarn properties are presented by graphs. The highest values of the quality parameters are located at the circular line of Figure 7. They diminish as they approach the circle centre. The

Yarn strength

high

Ring of yarn count

1

8

2

good

finer

1

Evenness

good

δ

Weavability

F

F

3 Minor 7

slubs

fewer

δ good

fewer

2

Quality control 6 during spinning

α

4

Hairiness

F

F

2 - Air-jet spun yarn

fewer

Key:

α

Key: 1 - Ring-spun yarn

5

Air-jet spun yarn Ring-spun yarn OE rotor spun yarn (11)

α,δ - Yarn twist angle

Figure 6. The Influence of Axial Loading on Twist Angle

Figure 7. Quality of Staple Yarns

38

Major slubs

VOLUME 6 NUMBER 4 1994

comparison of eight quality parameters shows that the air-jet spun yarn line approaches the circle line the most, therefore indicating that it has the best overall performance of the three yarn types. The values in Figure 7 are not absolute, but relative to the sequence of decreasing values. The quality of fabrics woven from the three mentioned yarn types of the same linear density and fibre content (20 tex × 1, 50 per cent polyester/50 per cent cotton) were compared using 12 different fabric quality parameters (Figure 8). The estimation of yarn suitability for the manufacture of flat textiles was based upon their handle, surface smoothness and appearance evenness. It can be deduced from the comparison of 12 quality parameters that fabrics made from air-jet spun yarns possess the best properties. Fabrics made from ring- and rotor-spun yarns follow (Figure 8). The peculiar properties of fabrics made of airjet spun yarns are the following: ● reduced pilling; ● good abrasion resistance; ● good air permeability; ● good washing properties; ● less wrinkle during wash and wear; ● reduced wash contraction of fabrics. Good properties of air-jet spun yarns compared with ring- and rotor-spun yarns enable the designers to have the optimal choice of yarns according to the product end-use.

1 12

good

Spinning speed m/min

No.802H MJS

2

Dyeability

soft

Smoothness 5 smooth

9

even appearance

good

Air permeability

8

6 7

Wash and wear

No.802MJS

OES

For choice of an optimal spinning technology, the yarn quality that must be achieved in the endproduct, has to be the decisive criterion. After the required quality is achieved, we can compare the production and economic parameters of different spinning processes. Figure 9 compares a spinning speed and a spinning count for three treated spinning techniques. The ring-spinning machine has a production speed ten times lower than the air-jet spinning machine (MJC 802H) and a production speed five times lower than an OE-rotor spinning frame (OES). The ring-spinning frame allows the production of coarse, medium-fine and fine yarns. On an OE-rotor spinning machine we can economically produce coarse and medium-fine yarns. An air-jet spinning machine enables the production of quality medium-fine and fine yarns. The comparison of running costs for different spinning machines (maintenance, power and labour), calculated in ¢/LB of yarn, is shown in Figure 10. According to Figure 10 we can assume that production costs are lowest with the air-jet spinning machine (MJS 802H). The OE-rotor spinning machine (OES) and ring-spinning machine follow. The finer the yarn count, the higher is the difference in production costs between the compared spinning techniques. The estimation of the economy of yarn production between the compared spinning techniques is shown in Figure 11 on the basis of a comparison of eight production-technological and economic criteria.

4 Stiffness

good

150

Comparison of Economic Parameters

stronger

10

t up

Figure 9. Spinning Speed versus Spinning Count for Different Spinning Techniques

good

Wash

cen

200

0 Ne 20 30 35 40 45 50 60 70 80 90 100 Yarn count

Wear out

3

per

Ring frame

stronger

Shrinkage 11

50

50

Tear strength

little

250

100

Tensile strength

stronger

Pilling

300

Defect

wrinkle

Key: Air-jet spun yarn Ring-spun yarn OEs-rotor spun yarn (11)

Figure 8. Quality of the Fabrics Woven from (a) Air-jet Spun Yarn (b) Ring-spun Yarn (c) OE Rotor-spun Yarn

39

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

80

the relationship of the needed floor space, calculated in ¢/LB, between the compared spinning techniques (MJS:RING:OES) is 1:1.32:1.14. On the basis of a comparison of eight productiontechnological and economic criteria in circular graphs we can assume that air-jet spinning is the most economical of the compared spinning processes.

●

⊃

2

( /LB)

1 -DES 2 -Ring frame 3 -MJS 802 4-MJS 802H Maintenance

70

Power

Labour 2

60

1 2

50

1 40 2

1

30

3

20 10

3

1

2

3 1

3

3

4

4

4 4

4

0 Ne20

Ne30

Ne40

Ne50

Conclusion

Ne60

The analysis of the mechanical models of ring, rotor and air-jet spun yarns shows the essential differences in the yarn formation. Different principles of fibre alignment and fixation decisively influence the micro- and macrostructure as well as the mechanical and physical properties of staple yarns. The estimation of productivity, spinability and economy of production between the ring spinning, OE-rotor and air-jet spinning processes was given. The information given in the article can be of use to weavers, knitters, garment manufacturers, finishers and designers of flat textiles, for the optimal choice of yarns according to the demands of a final product.

Figure 10. Comparison of Running Costs for Different Spinning Machines Productivity

higher

1

Energy consumption 8

2

Manipulation cheaper

lower

higher

Maintenance 3 costs

Flexibility 7

lower

unnecessary

Floor reinforcement

Lower 6

4 5

Required place

Material and preparatory costs

n

smaller

Key:

Bibliography

Air-jet spun yarn (MJS) Ring-spun yarn OES rotor spun yarn

Brockmanns, K.I., Textilbetrieb, Vol. 1 No. 2, 1982, p. 41. Deussen, H., Chemiefasern Textilind, Vol. 9, 1984, p. 622. Deussen, H., Schlafhorst – Dokumentation, No. 28, 1988. Klein, W., The Technology of Short-staple Spinning, The Textile Institute, Manchester, 1987. Luenenschloss, I., Kompen, W. and Rossback, D., Chemiefasern Textilind, Vol. 24 No. 76, 1974, p. 917. Luenenschloss, I., Kompen, W. and Rossback, D., Chemiefasern Textilind, Vol. 31 No. 83, Matsushima, A. and Manning, W.A., Textile World, Vol. 139, April 1989, p. 49. Neckar, B., Textil, Vol. 44, 1989, p. 450. Neckar, B., Prize, Praha, 1990. Oxtoby, E., Spun Yarn Technology, Butterworths, London, 1987. Technical documentation of the Murata, Schlafhorst and Rieter companies. Uster Statistics, Uster News Bulletin, No. 36, 1989.

Figure 11. Economy Diagram for Compared Spinning Techniques

On the basis of a comparison of productiontechnological and economic criteria between the compared spinning techniques the following findings could be indicated: ● the productivity of the air-jet spinning machine is twice as high as the productivity of the OE-rotor spinning machine and up to ten times higher than the productivity of a ring-spinning machine; ●

labour costs by air-jet spinning are up to 60 per cent lower than by ring spinning;

the air-jet spinning process enables up to 56 per cent lower maintenance costs in comparison with the OE-rotor spinning process; ● OE-rotor and air-jet spinning processes enable up to 10 per cent lower costs for raw materials; ●

40

VOLUME 6 NUMBER 5 1994

Comparison between Standards for Seam-woven Fabric Properties Determination A.M. Manich Spanish Council for Scientific Research, Research and Development Center of Barcelona, Spain, J.P. Domingues Universidade da Beira Interior, Covilhá, Portugal, and R.M. Saurí Spanish Council for Scientific Research, Research and Development Center of Barcelona, Spain Received 28 March 1994 Accepted 21 July 1994

Materials and Methods Standards Used The British, American and French standards were used as well as the International Wool Secretariat (IWS) and Renault methods. The British Standard used was the 1970s version instead of the more recent 1988 version because the latter was very similar to the American standard and the 1970s standard is still used by industry. The first four standards examine the behaviour of sewn fabrics. The Renault method substitutes the effect of the seam for the action of a line of perpendicular thrusting needles into the fabric.

Introduction The making-up garment process has taken on a greater importance nowadays. Therefore woven fabrics’ seam slippage and seam strength have received more attention. At the April 1992 Meeting of the International Wool Textile Organisation (IWTO) Yarn and Fabric Group, a paper was presented for discussion about the determination of seaming properties of fabrics, prepared by H. Camiou[1]. In 1990, the Spanish Council for Scientific Research (CSIC) and the Universidade da Beira Interior undertook a study on the properties of the seam-woven fabric assemblies of wool and blended fabrics. The first work carried out was a comparison between existing standards and methods for seaming properties of fabrics determination[2-6]. Six fabrics, very different in composition and structure, were taken to obtain results that could facilitate the comparison.

This article was presented as a paper at the IWTO Istanbul Conference held in May 1993, and the authors want to thank the IWTO Publication Committee which has accepted and recommended its publication. The authors also want to thank Ms C. Martínez and R. Mateu for their technical work. They also want to express thanks for the financial support received from the Spanish and Portuguese governments through the PB 90-0097 CICYT Project and Ciencia Program of the Junta Nacional de Investigacao Científica e Tecnológica.

International Journal of Clothing Science and Technology, Vol. 6 No. 5, 1994, pp. 7-14, © MCB University Press, 0955-6222

7

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

The British standard and the American standard calculate the strength necessary to cause a 6mm slippage in the seam. For fabrics with warp and weft of different colours, the British standard sets a 3mm seam slippage. The American standard also gives the fabric and seam-breaking strength. The French standard and the IWS method determine the seam slippage produced applying a load that, according to each method, depends on the mass per unit area and final application of the fabric. The Renault method gives the strength that causes yarn slippage beginning in the fabric, and the highest strength reached by the line of thrusted needles that submits the fabric sample to a traction test. Standards and methods that produce a seam perform the 1.01.01 seam type[7] and the 301 stitch type[8].

(5) Jacquard fabric with polyester textured continuous filament warp yarns and wool weft yarns for upholstery; (6) reprocessed fibres, plain-weave fabric for low quality women’s outerwear garments.

Fabrics Used The six selected fabrics were the following:

Results and Discussion

Fabric Characteristics Mass per unit area, warp and weft yarn density, yarn linear density and yarn twist have been calculated. Besides fabric breaking, strength and strain using the strip method have been calculated. Portuguese standards have been used. Results obtained are shown in Table I. Sewing Threads In Table II sewing thread characteristics are included, which has been selected according to the corresponding standards.

Sewing Slippage Tests In Table III, results obtained by the different used standards are shown for every direction of test, U being warp direction and T-weft direction. Results obtained are identified according to the following codes: (1) British standard: BS load applied in warp and weft direction (daN); (2) French standard: AFNOR slip produced in warp and weft direction (mm) (instead of slipping or not slipping statement);

(1) wool/polyester 55/45, purple colour plainweave fabric, for women’s outerwear garments; (2) cotton/polyester 70/30, blue colour plain weave fabric, for shirts; (3) wool/polyester 77/33, twill weave fabric for men’s outerwear garments; (4) cotton 100 per cent twill weave fabric, for worker’s garments;

Fabric Characteristic Mass (g/m2) Warp yarns/cm Weft yarns/cm Warp (tex) Weft (tex) Warp turns/m Weft turns/m Warp breaking strength Warp breaking elongation Weft breaking strength Weft breaking elongation

A 276.25 16.6 14.6 88.2 86.1 367.4 367.7 63.2 40.8 51.7 46.6

B

C

122.81 40.8 25.4 21.1 18.9 703.4 702.9 83.8 19.5 55.2 21.0

175.1 21.9 20.8 38.39a 37.46a 716.3a 702.7a 27.0 24.1 25.3 35.8

D 165.98 54.0 25.0 26.2 22.0 720.6 742.7 90.2 14.6 37.6 14.6

Notes: Fabric breaking strength in warp and weft directions in (daN) Fabric breaking elongation in warp and weft directions in percentage a Mean value of different colour yarns Table I. Fabric Characteristics

8

E 156.62 120.0 28.2 9.3 23.9 632.8 245.6 150.4 32.7 39.1 17.3

F 350.0 6.0 4.6 349.0 361.0 91.3 99.2 32.0 16.0 16.0 24.6

VOLUME 6 NUMBER 5 1994

1

2

Sewing yarn used 3

4

5

C/e Cotton 46 3 375.6z 1.49 5.21

G/m Polyester 32 3 443.5z 1.36 13.92

A/s Polyester 43 3 369.9z 2.05 14.07

C/d Cotton 73 3 372.3z 1.93 7.40

C/c Cotton 30 3 495.3z 0.95 6.54

Characteristic Supplier Composition Lin.D. (tex) Number plies Ply twist (tpm) Breaking strength (daN) Breaking elongation (%) Table II. Sewing Yarn Characteristics

(3) IWS method: IWS slip produced in warp and weft direction (mm);

●

(4) ASTM Standard: ●

ASTMt Fabric breaking strength in warp and weft direction (daN);

●

ASTMc seam breaking strength in warp and weft direction (daN);

Correlations and Comments Besides comments made about some aspects of the methodology of each standard or method of test, the relationships existing between them were examined. The results obtained were presented graphically. Linear correlation coefficients and their significance levels were also calculated.

ASTMd necessary load to open the seam 6mm in warp and weft direction (daN). (5) Renault method: ●

●

British Standard The determination of the necessary load to cause the standard fixed seam slippage, depends on the fabric structure and on the operator skill. When breaking occurs before reaching 6mm seam

Renaulti load necessary to begin the yarn slippage into the fabric in warp and weft direction (daN);

Standard

Direction

BS (daN)

U warp T weft U T U T Ut Uc Ud Tt Tc Td Ui Ur Ti Tr

AFNOR (mm) IWS (mm) ASTM (daN)

Renault (daN)

Renaulti highest strength reached in warp and weft direction (daN).

A

B

Fabric C

22.5 18.9 4.2 5.3 6.67 6.33 64.4 51.3 28.2 53.5 45.2 26.5 22.4 49.4 22.6 41.3

25.5 17.0 1.63 1.83 2.67 3.67 59.6 33.9 29.0 43.9 38.9 20.5 26.7 44.55 26.2 39.4

14.6 12.1 4.22 4.5 6.33 7.67 32.4 27.9 27.0 28.5 22.6 21.4 16.4 22.45 18.75 22.15

D

E

F

23.2 11.7 3.0 4.6 5.0 6.67 59.7 33.1 33.1 25.1 30.5 12.8 20.0 34.0 15.8 25.2

22.3 7.3 3.9 10.67 4.67 7.0 168.6 80.0 43.0 33.2 33.2 6.1 16.65 24.85 6.4 14.9

– – – – – – 29.17 26.83 7.5 17.33 20.5 1.5 4.42 17.52 5.9 14.5

Notes: U: warp; T: weft; t: fabric breaking strength; c: seamed fabric breaking strength; d: load necessary to produce a seam slippage of 6 mm; i: load at which begins yarn slippage; r: maximum reached strength Table III. Seaming Properties Obtained with Different Standards

9

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

opening, interpretation of results could be confusing. The correlation coefficients between this standard and AFNOR and IWS standards were –0.71 and –0.67 respectively. These values do not reach the 1 per cent signification level. These results seem logical because standards are based on different measuring principles.

French Standard Using this standard, a classification between slipping and not slipping seams is done, testing them under a load of 4 daN or 8 daN, and at a speed that depends on the type of fabric. In relation to the other standards tested, the level of obtained information is scarce. Neither seamslipping extent nor information about the kind of breaking are given. The selection of the samples is a little complicated, carrying out the tests on ten specimens in each direction. Time available for seam-opening reading could be excessive and relaxation phenomena have been observed. In our case the seam-opening extent was recorded immediately after a fixed load was reached. The relationship with the British standard has already been commented on in a former paragraph. The obtained correlation coefficient with the IWS method has been 0.66 (Figure 2). In spite of working under the same principle, this result shows the existent discrepancies between both methods considering the load to be applied according to the type of fabric. Nevertheless, if the result obtained with the fabric E in weft direction is rejected, the existence of a strong relationship between these methods could be observed. The correlation coefficients obtained with the three parameters of the standard ASTM were –0.20 with ASTMt, –0.19 with ASTMc, and –0.60 with ASTMd. None of them reached the 5 per cent signification level. Also here the effect of the existing differences between both principles of measuring can be seen. The correlation coefficients obtained with the results supplied by the Renault method were –0.86 with Renaulti and –0.60 with Renaultr (see Figure 2). The first coefficient is significant at the 1 per cent level, therefore it could be stated that there exists a certain relationship between both considered parameters.

n

The selection of the samples is a little complicated n Comparing results obtained with BS and ASTM standards the correlation coefficient with fabric breaking strength (r = 0.56) and with seambreaking strength (r = 0.54) were not significant. On the other hand, the correlation coefficient obtained with the parameter ASTMd (see Figure 1) was 0.83. This value is significant at 1 per cent level. This could explain the fact that the two standards operate under the same principle. Comparing the results obtained with the two parameters given by the Renault method, a 0.73 correlation coefficient with Renaulti is obtained (Figure 1), not significant at the 1 per cent level, and of 0.76 with Renaultr (Figure 1) which reaches the 1 per cent significance level. Therefore, the existence of a certain relationship between both parameters could be stated.

ASTMd, Renaulti , Renaultr (daN) 60 50 40 30

IWS Method With this method the seam opening produced upon samples submitted to 12 daN or to 8 daN according to the weight of the fabric or to 18 daN if fabric is destined for upholstery. The seam-opening determination is not easy, because it depends on the type of fabric and on the operator skill. Here also, relaxation phenomena, depending on the type of fabric, were observed. Because of seam slippage, reading was carried out immediately after reaching the determined load.

20 10

0 6

8

10

12

14

16

18

20

22

24

26

28

30

BS results (daN)

Key : ASTMd Renaulti Renaultr

Figure 1. Results Obtained with BS, ASTM and Renault Methods

10

VOLUME 6 NUMBER 5 1994

IWS (mm)

American Standard This standard gives more objective results to identify this phenomenon. It also gives a graphical registration that allows to compare seam and fabric evolution until both break. With this standard the necessary load to open the seam 6mm is calculated from the graphical registration. Nevertheless the break could be produced before reaching this opening. In this case, results could be ambiguous. The existent correlation between the three parameters obtained is good. So the correlation between ASTMt and ASTMc is of 0.94 (Figure 4) possibly because of the high value obtained by fabric E in warp direction, the correlation between ASTMt and ASTMd was of 0.77 (Figure 4), and both coefficients were significant at 1 per cent level, while the existent correlation between ASTMc and ASTMd was of 0.70, significant at 5 per cent level. For the type of seam, stitch and sewing thread considered in this standard, it seems that there could exist a relationship between fabric strength on one hand and, on the other hand, seam strength and necessary load to open the seam 6mm. The relationship between this standard and British and French standards, and IWS method has been examined. Correlation coefficients between the fabric strength and the two parameters supplied by the Renault method were almost zero. However, if the result obtained with fabric E in warp direction was eliminated, the correlation between fabric strength and results obtained by the Renault method would be very much enhanced (Figure 5). Between the other parameters very low correlations were obtained.

Renault (daN)

12

60

10

50

8

40

6

30 Fabric E weft direction

4 2

20 10

0 0

2

1

3

4

5

6

8 7 AFNOR results (mm)

9

10

0 12

11

Key : IWS Renaulti Renaultr

Figure 2. Results Obtained with AFNOR, IWS and Renault Methods

The relationships existing between this method and the British and French standards have been commented on already. The correlation coefficients obtained with the three parameters yielded by the American standard were very low –0.37 with ASTMt, –0.29 with ASTMc, and –0.44 with ASTMd. These results are according to the existing disagreement between standards measuring principles. In relation to the results obtained when comparing this method with the two parameters yielded by the Renault method, a correlation coefficient of –0.61 (Figure 3) with Renaulti and of –0.48 with Renaultr were obtained. The first of these coefficients approaches the 5 per cent signification level. These are the same results that obtained the highest correlation coefficient as the standard AFNOR results which operate under the same IWS measuring principle.

ASTMc, ASTMd (daN) 100

80 Renaulti (daN)

60

30 25

Fabric E weft direction

40

20 20 15 0

10

0

0

20

40

60

80

100

120

140

160

180

ASTMt (daN)

5

0

1

2

3

4

5

6

7

8

9

Key : ASTMc ASTMd

10

IWS results (mm)

Figure 4. Relation between the Three Parameters Obtained by the ASTM Standard

Figure 3. Results Obtained with IWS and Renault Methods

11

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

standards and IWS and Renault methods was examined. Leaving aside this last method, which is based on a slipping test attaching the fabric sample in an end using a jamb containing a line of thrusting needles, the existing methods could be classified into two groups:

Renaulti, Renaultr (daN) 60 50 40 Fabric E

30

warp direction

(1) The first group includes the French standard and the IWS method which calculate the seam opening produced by a determined load. An acceptable correlation between them is observed, the variations introduced by the different testing conditions applied to the fabrics are considered.

20 10 0 0

20

40

60

80

100

120

140

160

180

ASTMt (daN)

Key : Renaulti Renaultr

(2) The second group includes the British and American standards, based on the determination of the load necessary to produce a fixed slip in the seam. The existing correlation between both methods is very high. Nevertheless, when breaking occurs before reaching the fixed seam slippage, interpretation of results is not easy.

Figure 5. Relation between ASTM Standard and Renault Method

Renaulti (daN) 30 25

In considering the methodology, results obtained by the American standard are more objective. It is the only standard that gives information about the whole process of slipping until the fabric breaks. Bearing in mind the similarity between American standard and 1988 British standard version, the same comments could be made about the latter. Correlation coefficients obtained between results belonging to the first and the second group of standards were very low.

20 15 10 5 0 10

15

20

25

30

35

40

45

50

Renaultr (daN)

Figure 6. Relation between Parameters Obtained by Renault Method

n

Renault Method This is the only method in use that does not use a seam to determine the seaming properties of fabrics. One end of the sample is thrust by a line of subjected needles into a special jaw which substitutes for the seam. The evaluation of the load at which the beginning of the yarn slipping into the fabric is produced relies on the ability of those operator. The correlation obtained between those two parameters supplied by this method is 0.88, significant at 1 per cent level (Figure 6). The correlations between these parameters and those supplied by the other standards have already been examined.

When breaking occurs interpretation of results is not easy n The parameter Renaulti (beginning of slipping) from the Renault method gives a good correlation with results obtained by the British standard. The parameter Renaultr (maximum strength) of the same method gives a good correlation with results obtained by the French standard. It appears that it would be convenient to include some information about the type of failure that occurs during the seam slippage second group standards test.

Conclusions The determination of seaming properties of woven fabrics using British, French, and American

12

VOLUME 6 NUMBER 5 1994

From the results obtained, a method for the comparative assessment of seamed and unseamed fabric behaviour under tensile strain to break is suggested[9]. A shortened version of it is included in the Appendix.

Appendix. Method for the Assessment of Seamed Fabrics Behaviour in Relation to That of the Unseamed Fabric Foreword Bearing in mind results obtained by the most used standards, a method for the determination of both seaming strength and slippage of sewn fabrics is proposed. The knowledge of this behaviour is fundamental in fabrics that must be sewn for its final application. Tests are conducted using a dynamometer that works under CRE (Constant Rate of Elongation). Tensile strength tester must be provided with a computer for autographic strength/strain curve registration. Using computer analysis the maximum reached opening before breaking and breaking strength are given. Load necessary to produce seam slippage levels from 1mm to 6mm in steps of 1mm are given too. For sewing compensation a preload of 5N is used. Although different seam types, stitch types and sewing yarns, could be used, normally 301 stitch types and 1.01.01 seam will be done according to BS 3870 (NP 3801 and NP 3800) classification. Seams normally will be stitched at a distance of 10mm from the edge.

n

References 1. Camiou, H., The International Wool Textile Organisation, Technical Committee, “The Determination of Seaming Properties of Fabrics”, IWTO Yarn and Fabric Group, Test Method under Examination, April 1992. 2. British Standard 3320:1970, “Method for the Determination of Seam Slippage of Woven Fabrics”, 1970. 3. Standard AFNOR G07-117, “Essais Tissus – Pour Vêtements – Méthode d’Appréciation du Glissement des Fils d’un Tissu et Mesure de la Résistance des Coutures” (“Fabric Testing – For Clothes – Yarn Slippage in Fabrics and Seamed Fabric Strengths”).

Purpose Given a test method, the study of resistance and slippage of seamed fabrics could be done under tensile forces using the Grab method until that breaks. A comparison with the unseamed fabric behaviour is obtained.

4. International Wool Secretariat, Test Method 117, “Method of Test for Determining Seam Slippage of Woven Fabrics”. 5. ASTM Standard D434-75, “Resistance to Slippage of Yarns in Woven Fabrics Using a Standard Seam”.

Definitions Seam value: The distance from the edge at which seam is conducted. Seam strength: the peak force to failure the seam. Type of seam failure: Group into seam breaking process could be classified:

6. Méthode d’Essai Véhicules – Renault No. 1127, “Résistance à la Couture; Tissus” (Car Testing Methods – Renault No. 1127, “Seamed Fabric Strengths”).

(1) fabric yarns breaking close to the seam; (2) sewing yarn breaking;

7. Portuguese Standard – NP 3800, “Texteis. Tipos de Costuras. Classificaçao e Terminologia”, (“Textiles, Seam Types, Classification and Definitions”), 1991.

(3) fabric yarn slippage induced by the seam; (4) seam yarn slippage; and (5) a combination of some previous conditions.

8. Portuguese Standard – NP 3801, “Têxteis. Tipos de Pontos de Costura. Classificaçao e Terminologia”, (“Textiles, Stitch Types, Classification and Definitions”), 1991.

Conditioning and Testing Atmospheres For testing and sample conditioning the required conditions are as ISO 139 Standard describes.

9. Domingues, J.P., “Mecanica da Uniao de Estruturas Texteis”, (“Mechanics of the Assembling of Textile Structures”), PhD Thesis, Universidade da Beira Interior, Covilhá, Portugal, 1993, pp. 65-70.

Apparatus Set the clamps of the tensile tester a distance of 75mm apart. Central point of the clamps has to be on the tensile straight line. The face of the jaw clamps shall

13

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

measure 25mm × 25mm and they have to be situated on the same plane. The jaws can hold the specimen without allowing it to slip and shall be designed so that they do not cut or otherwise weaken the specimen. The rate of clamp separation shall be of 300 ± 10mm/min.

sample is rejected and another additional sample has to be tested.

Calculations For each group of warp or weft specimens, mean values of fabric-breaking strength and seam breaking strength are calculated. The breaking strain for fabric and seam also has to be calculated. The seam efficiency parameter SE in percentage is calculated by relating seam-breaking strength SBS and fabricbreaking strength FBS in the following way: SE = (SBS/FBS) * 100

Sampling and Testing Method Samples have to be selected so that they are representatives of the fabric. Three samples in warp way and three samples in weft way have to be cut. Other directions on the fabric could be selected. Three rectangular strips of 300mm × 100mm are cut. They are then folded at a distance of 75mm of their edge and a seam across its width with 10mm of seam value. The strip is cut at the fold and then extended at right angles to the seam. The unseamed test specimen is mounted centrally and a strength/strain curve until fabric breaking is recorded. Using the same sample, the seamed test specimen is clamped in the jaws in the same way as above, ensuring the seam is midway between and parallel to the jaws. Another strength/strain curve until breaking is produced. Strength/strain curves for both fabric and seam must be recorded. The possibility of finding out load levels that causes from 1mm to 6mm seam openings has to be considered. Seam strength and fabric strength are obtained. If the sample slips or breaks nearby the jaw

Report The report should state the following information: (1) Tested material identification. (2) Number of trials conducted. (3) Type of seam failure for every sample. (4) Number of rejected samples. (5) Mean seam-breaking strength and strain. (6) Seam efficiency. (7) Level of strength that produces 1mm to 6mm seam slippages. (8) Sewing yarn characteristics (composition, linear density, twist, and breaking strength and strain).

14

VOLUME 6 NUMBER 5 1994

Automated Fabric Inspection S.Convery, T. Lunney, A. Hashim and M. McGinnity Magee College, University of Ulster, Londonderry, Northern Ireland, UK

Received 27 May 1994 Accepted 28 July 1994

Defect Classification The structure of a typical knitted fabric is shown in Figure 1. A knitting machine will knit one course (horizontal row) of stitches at a time. A vertical column of stitches is known as a wale. It is important to note that the classification of flaws is not standardized within the industry[3]. However three major approaches to describing flaws are: (1) feature tree; (2) numerical description; (3) technical structure. With the feature tree approach, the most widely used, the flaws are “self describing” from the point of view of manual inspection (e.g. dropstitches, rip needles…), as can be seen from the examples in Plate 1. The numerical description approach classifies flaws according to their length and breadth. The technical structure approach classifies flaws by detecting the number of missing stitches in the flaw itself.

Introduction Technological advances have improved the efficiency of most industrial processes. However the textile industry in general, including both knitted and woven fabric manufacture, remains labour intensive in the production, and more especially in the inspection processes. Human inspection imposes limitations, both in terms of rate and consistency on identifying flaws, which is naturally affected by fatigue and boredom. Current human inspection practices require “trained and experienced personnel” but at the same time, results based on uncertainty and bias are often produced[1]. A computerized inspection system eliminates the need for human inspection, increases the quality of the product, and thus leads to better customer satisfaction. Inspection system technologies are available, but the number of flaws that they can reliably detect, and the speeds at which the flaws can be detected, currently limits their deployment. Detection of defects can take place at several stages in the production process: yarn inspection preceding the knitting process, fabric inspection on or immediately after the knitting machine (called “grey goods level”), on finished material after dyeing, or finally on completed garments after assembly. As in all quality assessment procedures, the cost impact of a defect can be greatly reduced by detecting it as early as possible in the production life cycle[2]; thus we will concentrate on detection of defects at “grey goods level”. In the following sections typical fabric defects are described and classified, and possible methods to detect and report them are identified.

Available Systems and Technologies Existing systems in the textile industry in general include both semi-automated and fully automated systems. The purpose of semiautomated systems is to help the human inspector mark and classify the flaw, e.g. a button is pushed every time a fault is detected[4]. These systems speed up the human inspection process, but have very little impact on the reliability and consistency of flaw detection. The authors would like to acknowledge The Interactive Systems Centre who are supporting this work, and also the help of Fruit of the Loom (a large clothing manufacturer) who supplied defect samples and information. Sean Convery is currently being funded by Research and Development funds from the Department of Electronic Engineering, University of Ulster.

International Journal of Clothing Science and Technology, Vol. 6 No. 5, 1994, pp. 15-19, © MCB University Press, 0955-6222

15

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Walewise direction

(a) Dropstitch run

(f) Slub

(b) Needle line

(g) Slub hole

(c) Rip needle

(h) Press-off

(d) Knots

(i) Press-off one bed

(e) Barre

(j) Dropstitches

Coursewise direction

Figure 1. Knitted Fabric Structure

Fully automated systems are showing more promise, particularly for woven fabric. This fabric is more regular in structure and hence flaws which are deviations from this structure are more easily identified. However the development of new techniques and algorithms (see section on Image Processing Activity) is likely to overcome some of the difficulties associated with knitted fabric. Fully automated defect recognition systems use a variety of sensing technologies to acquire data relating to the fabric condition. Sensing methods can be broadly categorized according to whether they are “active” or “passive”. Active sensing requires an illumination source to highlight the fabric, whereas passive sensing does not (see Figures 2 and 3). The principle automated methods used to highlight the fabric being analysed are all active sensing arrangements based on light beam reflection[5], laser beam reflection[1] or video image processing. These automated systems usually employ the simple scanning of a photocell or capacitance measuring device across the surface of the fabric. Such systems will pick up quite coarse defects, such as press-off, but reliable results for finer defects can be affected by the “yarn hairiness” of the wool. Video image processing would appear to be the most promising technique if high reliability is required, since the evaluation of video images of textiles offers a wealth of information which cannot be obtained so readily by any other means[1]. Increasingly popular in “woven” textiles is the use of charge coupled device (CCD) linescan cameras with between 2,048 to 4,096 pixels resolution to build up a detailed image of the fabric for processing. In terms of hardware, such systems are typically made up of a CCD camera, an analogue to digital converter, image storage unit, a digital to analogue converter for input to a monitor, a central processing unit (CPU) to analyse the video data from the camera,

Plate 1.

16

VOLUME 6 NUMBER 5 1994

Source

place over the first section of cloth in which a profile for the cloth is built up. This inspection system for woven fabric has proved to be economically viable with inspection speeds of up to 120m/min. The faults size and its location are recorded and classified into three types (warp, weft and surface faults) and the severity of the fault is indicated.

Sensor

Material

Automated Inspection System Requirements Automated inspection is complicated by the fact that “good” knitted fabric has inherent natural irregularities in its structure, which are regarded as acceptable[2]. Ideally, a system should scan all the fabric and pick up the flaws which have occurred (ignoring naturally occurring discrepancies). The flaws must then be identified and classified by the system in a reliable and deterministic way, so that appropriate action can be taken. Also the inspection system should be as close as possible to the source of the flaw and so should be integrated into, or placed close to, the knitting machine, to facilitate early detection and cut down on fabric wastage. Carrying out detection and classification of flaws in real time is computationally a very demanding process. Analysis of data for inspection purposes requires a large amount of image processing prior to classification. Classification lists for faults using the feature tree approach may contain references for up to 100 different types of flaws adding to the already weighty amount of computations. Hinze et al.[7] have estimated that processing rates of at least 109 operations per second is required, indicating the need for a powerful system. (This calculation is based on: breadth of fabric = 2m, pass speed = 60 m/min, resolution = ∆x = ∆y = 0.2mm, 20 operations per image point.) To be economically viable the cost of the inspection system must be balanced against the long-term savings due to the manpower replaced, and the increase in quality of the resulting fabric[8].

Source

Material

Sensor

Figure 2. Active Sensing Configurations

Sensor

Material

Figure 3. Passive Sensing Configuration

and a mass storage system to store results after processing (see Figure 4). Image processing algorithms can be applied to detect any flaws in real time (see next section). The Uster Visotex system[6] for detecting defects in smooth surfaced, single coloured, woven fabrics, detects flaws by passing the cloth over a constant light source with a slot type aperture. Two or more special types of video CCD cameras are used (depending on the width of fabric) to scan the fabric. A learning process takes

Image Processing Activity In image processing, the sensed image (e.g. by a video camera) is translated into a digital image, i.e. a two-dimensional array of numbers or grey levels, by an analogue to digital converter. This two-dimensional array can then be analysed using image processing techniques to pick out the flaws. Each of the defects (described in Plate 1) can be searched for by utilizing a combination of

17

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Video camera

Video monitor

Image display Fabric under inspection Light box

Frame grabber and store

Output results

Image acquisition

IBM-compatible PC

Image processing

Figure 4. Image Acquisition and Processing System

optimum setup for the detection and classification of flaws. Techniques for detection of flaws are numerous and varied in their approaches. They range from standard image processing filters (e.g. the low-pass filter above) and characterizing patterns using Fourier transforms[9], to texture analysis algorithms, such as the co-occurrence matrix [10,11] and the correlation method (as in Sick’s Video Shrink Inspector[12]), and the learning of the periodic nature of stitches[13]. It must be emphasized that most of these algorithms are very demanding on processing time, hence the need for special purpose powerful computing hardware for complete (i.e. all the fabric is inspected) real time inspection. This would suggest dedicated parallel processing hardware (e.g. transputers), but the question of economic viability of such a system then arises. A lower specification system using a PC with framegrabber board and digital signal processing capability would be adequate for a samplingbased system. Such a system would take a sample of the fabric and process it (detect and classify the flaws) before acquiring the next sample for processing. This latter system should prove sufficient for many manufacturing environments as the flaws tend to repeat themselves and thus they would be picked up by such a sampling system.

standard image processing algorithms and filters. For example, a dropstitch run (see Plate 1(a)) can be detected by applying a low-pass filter to the image. A low-pass filter effectively averages out areas in the image to highlight regions of different light intensities. (e.g. a hole, or normal fabric). In other words the small areas of light between stitches would be discounted, but the larger areas of light caused by a hole would remain, and thus the hole is easily identified (see Plate 2(a)). The dimensions of the flaw can then be used to determine the type of flaw using the numerical description approach to flaw classification. In this case it will report the existence of a dropstitch run. This low-pass filter template is only one example of many possible filter selectors that can be used to detect flaws (see Plate 2(b)) and experimentation may be required to determine an

(a) A low-pass filtered image of a dropstitch run flaw

(b) Pixelated and threshold image of a dropstitch run flaw

Plate 2.

18

VOLUME 6 NUMBER 5 1994

Conclusion

5. Lenka Microsystems BV, Melliand TextiIberichte, Vol. 71 No. 1, 1990, p. 42. 6. Sellweger Uster Ltd, “Automatic Cloth Inspection with the Uster Visotex”, Chemiefasern/Textilindustrie, Vol. 37/89 No. 10, 1987, pp. 1026-28. 7. Hinze, D., Mehlhorn, H. and Burkhardt, S., “Digital Image Processing by Means of a Parallel Computer System for Testing of Fabrics”, Melliand Textilberichte, Vol. 72 No. 12, 1991, pp. 993-4. 8. Knoll, A.L., “Automatic Fabric Inspection”, Textile Institute and Industry, January 1975, pp. 5-9. 9. Wood, E.J., “Applying Fourier and Associated Transforms to Pattern Characterization in Textiles”, Textile Research Journal, Vol. 60 No. 4, 1990, pp. 212-20. 10. Kaasjager, A., “Textile Fabric Monitoring with Image Analysis”, Melliand Textilberichte, Vol. 71 No. 1, 1990, pp. 64-6. 11. Roester, U., “Application of Digital Picture Processing Methods to Automated Cloth Inspection”, Bekleidung und Maschenware, Vol. 28 No. 2, 1989, pp. 58-61. 12. Dusek, Z., “Contactless Thread Density Measurement of Woven and Knitted Fabrics”, Melliand Textilberichte Vol. 72 No. 11, 1991, pp. 917-20. 13. Clark, A., Parui, S., You, Y. and Hashim A., “Detection of Defects on Fabrics”, Institute of Electrical Engineers, Digest No. 1986/48, 8 April 1986.

Automated fabric inspection has proved problematic due to the inherent irregular nature of the material under inspection. However inspection and processing technology has now matured to a level which enables viable automatic inspection systems to be built, although the processing power required for real time inspection is still substantial. What is required is the “inhouse” manufacturing awareness of the potential of these systems to enable them to be developed and deployed. We hope that this article succeeds to some degree in addressing the former, and that an investigation of their economic viability may lead to the latter.

n References 1. Hinze, D. and Viertel, E., “Image Sensor Technology for the Light Industry”, Melliand Textilberichte, Vol. 72 No. 11, 1991, pp. 958-62. 2. Munden, L. and Norton-Wayne, L., “Machine Vision in Textile Manufacture”, Textile Asia, Vol. 19 No. 10, 1988, pp. 85-8. 3. Ward, P.T., “Visual Inspection, Its Automation and Application in the Textile Industry”, PhD thesis, Leicester University, 1986. 4. Shelton, M., “Profit through Inspection”, International Dyer and Textile Printer, Part 8, August 1988, p. 13.

19

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

High Quality Fabrics for Garments Sueo Kawabata Kyoto University, Kyoto, Japan and Masako Niwa Nara Women’s University, Nara, Japan

handle evaluation system, applying these parameters to their tailoring process control[1,2]. During the transfer of the fabric measurement technology to the suit manufacturing engineering, tailoring engineers were interested in predicting making-up of suit appearance on the basis of the fabric mechanical parameters[3]. In addition, the range of the mechanical parameters for high quality suiting has been investigated in connection with fabric mechanical parameters from the engineers’ experience from the point of view of mechanical wear comfort[2]. We now have the following three criteria for predicting fabric quality from different perspectives: (1) Fabric handle. Traditional subjectiveassessment of the quality of fabric. The objective evaluation system of this handle has been fully developed. This is originally based on the hand-touch feeling. Fabric smoothness handle is of primary importance. (2) Suit appearance. This concerns the making-up property of suit which may be predicted on the basis of fabric mechanical properties. Traditionally, this has also been predicted from fabric handle, however, its criterion has not been clear because this criterion is closely related with fabric mechanical properties in the low-load deformation region, where the detection of these mechanical properties by hand is difficult and uncertain. The objective system of this prediction was recently developed.

Received 25 May 1993 Accepted 21 July 1994

Introduction In the past 20 years, progress has been made in fabric objective measurement towards explaining what properties of fabric are important when used as garments. The traditional method of evaluating the fabric quality was the handle judgement of fabric by subjective method. At the first stage of the investigation, the “handle” was investigated in around 1970. The Hand Evaluation and Standardization Committee (HESC) was organized by the authors. The leading experts in handle judgement in worsted weaving and finishing mills were invited to join the committee as committee members. The investigation and the standardization of the subjective handle of men’s suiting was carried out, then the objective method of the handle judgement was developed[1]. This objective method is based on the measurement of the mechanical properties of fabric. The mechanical properties are transformed into the hand values which are the numerical expressions of the intensity of the primary hands. Then these hand values are again transformed into the total hand value (THV) which expresses the quality grading of fabric for suiting. This objective evaluation system is now quite popular and has been widely applied to the fibre and textile industry. When this objective system became widely accepted, some suit manufacturing apparel engineers directed their attention to the fabric mechanical parameters (see Appendix) which were used for deriving primary hand in the objective

The authors express their many thanks to the HESC members for their co-operation in discussions about refinishing plans. Mr Kurihara of Mitsuboshi Worsted Co., who is a member of the HESC, handled the refinishing of the trial fabrics.

International Journal of Clothing Science and Technology, Vol. 6 No. 5, 1994, pp. 20-25, © MCB University Press, 0955-6222

20

VOLUME 6 NUMBER 5 1994

It is now possible to derive the HV and THV from fabric mechanical parameters Xiss as follows, where Xiss is i the mechanical parameter of the fabric (i = l, 2, .... ,16), and Yj is the HV of the jth Primary hand where j = l, 2, 3, ( j = 1, 2, 3, 4 for mid-summer).

(3) Wearing comfort: This property is related to mainly fabric tensile and shearing deformations, and also related to comfortable wear associated with less restraint of human body by fabric. The shape-retention of suit is also covered by this property. The quality of suiting must be designed from the three perspectives by means of scientific methods based on the fabric mechanical parameters.

[xi ]

1 Stiffness Koshi

1 Stiffness

2 Smoothness Numeri

2 Crispness Shari

3 Fullness Fukurami

3 Fullness

Total hand value

In order to derive the fabric property in relation to the making-up property of suit, the fabric mechanical parameters were correlated with the appearance of the tailored suit by tailoring factory experts. In this investigation, the total appearance value (TAV) was defined and evaluated by the experts[1,3]. The grading of the TAV is the same as that of THV. The three components relating to the appearance were defined. These three components are newly created values in this research. The new components correspond to primary hand in the case of handle and are used in the same way as the THV-HV relation. They are: (1) Formability components, S1: The fabric property relating to formability of smooth three-dimensional curve of suit.

Very strong

Strong

10

8

Excellent 5

[C" j ]

Good Appearance

Table I. The Primary Hands

9

[THV] (1)

The primary hand and its numerical expression have enabled us directly to derive the property of high quality fabrics from expert experience. The high quality zone of primary hand is shown in Figure 1, which has been derived based on a selection of good fabrics from commercial fabrics by experts. The practical application of this chart is that the primary hand values Yj and successively derived THV from equation (1) may be plotted on the chart to compare with the good zone. When all primary hand values are inside the shaded zone, then its fabric is a high quality fabric. If some are out of the zone, the fabric has problems.

4 Anti-drape Hari

Primary hand value (for each primary hand)

[C'j ]

where C ij, C' j, C" j : constant coefficients

As mentioned above, fabric handle evaluation has been used for a long time by professional experts in textile mills. From the analyses of the handle, the authors have assumed that there are two types of fabric handle. One handle expresses fabric characteristics such as stiffness, smoothness etc., called the primary hand. The other handle expresses fabric quality such as high quality or poor quality using the expressions “good handle and poor handle”. This handle was named total hand by the authors. The total hand is judged after primary hand judgement. The primary hand consists of three components for winter/autumn and all season suiting, and four for mid-summer suiting as shown in Table I. After the definition of these Primary hands, the intensity of these handles was graded by a number called hand value (HV), and the fabric total hand was graded by a number called total hand value (THV) as shown in Table II.

Primary hand for mid-summer suiting

[Yj ] [Y 2j ]

High Quality Fabric from Fabric Handle

Primary hand

[Cij ]

7 Good 4

Table II. The Grading of Handle

21

Average 6

5 Average 3

4

Weaker

None

3

1

2

Below average 2

0 Poor 1

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Hand value Hand value Stiffness (Koshi)

2

Smoothness (Numeri)

3

0

Fullness (Fukurami)

1

0

1

Total hand value

4

2

3

2

1

–3σ

5

3

6

4

–2σ

5

4

2

5

3

–1σ

7

6

8

7

6

7

8

9

8

9

4

0

1σ

9

10

10

10

5

2σ

3σ

Figure 1. High Quality Zone of Fabric Handle

parameters[1,2]. There are no equations and particular values expressing the excellence of this property, however, they expressed the good zone by directly using the mechanical parameters of tensile and shearing properties as shown in Figure 3. The central part of the chart in Figure 3, which is indicated as “non-control zone”, is the zone of fabrics for which tailoring is easy and no control is necessary in the tailoring process. It should be noted that the good zone and the non-control zone do not coincide. This means that the process control is necessary for high quality fabrics. In this quality range, fabric extensibility and shearing deformation ability are important for fabric to deform with human body motion. Smooth handle is not regarded. This quality from wearing comfort was of course evaluated by the traditional handle judgement, however, it would not allow clear evaluation because of strong influence of smoothness to THV quality.

(2) Elastic component, S2: It is related to the elastic property of fabric, it makes a beautiful and smooth curve of suit with high shape retention ability. (3) Drape component, S3: It is related to beautiful suit silhouette. The derivation of the TAV is also carried out in the same manner as THV, as follows. The mechanical parameters Xi (only tensile, shearing, bending properties and fabric weight are applied in this case) are converted to xi which are the modified mechanical parameters that correlate more closely to TAV than the original Xi 1). [xi ]

[Cij ]

[Sj ]

[C"j ]

[TAV] (2)

[x 2i ]

[C'ij] where [xi]: The modified mechanical parameter, i =1~8 C" j , C' ij , C" j : constant coefficients, j = 1~3.

Design of Ideal Fabrics

The good appearance zone of the components is shown in Figure 2. High THV fabric does not necessarily have high TAV. For this reason we need to use TAV in addition to THV.

We now have clear targets for realizing overall high-quality suiting with properties which satisfy all three of the quality criteria introduced here. We call these fabrics “ideal fabrics”. Since around 1930, much research on fabric mechanics has been carried out in the textile technology field. In the next stage of textile technology, we have to combine this research with the criteria for high quality which have been developed during the past decade. It is very difficult to achieve target properties in fabric design, because

High Performance from Wearing Comfort Tailoring engineers have found good fabrics through experience and have expressed the properties of these good fabrics with a combination of fabric mechanical

22

VOLUME 6 NUMBER 5 1994

Normalized value –3 σ Mechanical parameter X11; EL2

X12; BS2

X21; BP X22; SP X31;3

2 0.03

0.4

0.5

0.6

0.05

0.07

X32;3 √SS/W

4

5

0.04 0.05

1σ

0

6

0.07

7

4

5

1.0

6

0.3

7

0

Elastic potential; S2 Drape ; S3

0

1 0

TAV

0.7

4

2

3

4

2

5.0

30

40

3

4

3

4

6.0

0.5

0.55

0.6

6 5

6

5

6

5

6

Control zone 0.65

0.7

RT 50

EM1

55

3

60

3.5

EM2

2

EM2/EM1

65

70

4

4.5

5

4

6

8

2

1

75

5.5

6

10

3

15 4

G 0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.2

2 HG 5 0.8

1

1.5

1.8

2

High THV and comfort zone

Source: [2, corrected in 1993]

Figure 3. “Tailoring Control Chart” and High Quality Zone from Wearing Comfort

23

50

7.0

5

Non-control zone

LT

7.0

1.0

Figure 2. High TAV Zone for Suit Expressed by the Three Components

Control zone

6.0

2.5

3

1

40 0.7

4.0

5.0

2

0.5

20

2

1

0

0.5

8 9 10

4.0

1

0.4

2.0

3.0

Three basic components of tailorability Formability; S1

30

3.0

0.4

1.5

2.0

0.3

2.0 0.2

3σ

20 0.2

0.7 0.8 0.9 1.0

3

2σ

8 9 10

0.1

0.1

2

√BS/W

–1 σ

3

0.02

X13; SS

–2 σ

2.5

3

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Finished fabric

Target value

After first re-finishing

After second re-finishing

9.98 0.59

8.01 0.59

0.93 2.25

0.73 1.66

0.0148

0.0142

6.02 5.57 5.41

5.71 5.77 5.35

3.59

3.66

4.21

4.28

Tensile EM-2 7.59 13.00 LT 0.62 0.55 Shear G 0.86 0.60 2HG5 1.92 1.30 Surface MMD 0.0166 0.012 Primary handc Stiffness (Koshi) 5.65 4.98a Smoothness (Numeri) 4.96 6.60a Fullness (Fukurami) 4.79 5.73a Total hand THVb 3.24 4.03a Total appearance TAVb 3.44 4.28a a The value predicted from the target value of mechanical parameter b 5 is “excellent” and 1 is “poor” c 10 is “very strong” and 1 is “very weak”

Table III. Result of a Trial (Only Refinishing Was Applied for the Correction of Fabric Property. A Commercial Fabric, Sample No. J-08 Worsted 254g/m2 Was Chosen for This Trial of Trimming)

there are numerous fabric parameters and processing parameters to be controlled. The results of much research has given us an idea of how to come close to the properties in the target zone but not how to achieve them. We have to create guidelines for realizing ideal fabrics in the future. In this stage, however, we are proposing the trimming method as follows:

(1) selection of those fabrics which have properties near the ideal fabric; (2) inspection of the mechanical parameters which deviate from the ideal zone; (3) selection of the important parameters which effectively help to eliminate the deviation under the guidance of equations (1) and (2); Hand value

Hand value Stiffness (Koshi)

2

Smoothness (Numeri)

3

0

Fullness (Fukurami)

0

4

1

2

1

Total hand value

3

2

1

–3σ

5

4

3

5

3

–1σ

7

5

4

2

–2σ

6

6

8

7

6

7

1σ

Key : : Original fabric

: After first trimming finishing

: Target

: After second trimming finishing

Figure 4. Hand Chart

24

8

9

8

9

4

0

9

10

10

5

2σ

10

3σ

VOLUME 6 NUMBER 5 1994

(4) discussion with finishing experts about how to correct the properties by shifting the parameter values to ideal zone; (5) refinishing trials; (6) discussion of the result; (7) if the refinishing method does not sufficiently move the properties to the ideal direction, call weaving experts to redesign the fabric structure; (8) continue this procedure up to the fibre stage if necessary. We are now continuing to use this trimming method to improve fabric quality to the ideal direction. Table III shows the selected mechanical parameters for improving a test fabric and the corrected result of the trimming by refinishing. The corrected primary hand values and total hand values are plotted on the hand chart as shown in Figure 4. The original fabric was commercially produced worsted fabric and its properties nearly all fall into the good zone after refinishing twice.

Industry, Wool Research Organization of New Zealand and the Textile Institute (New Zealand Section).

Appendix: Fabric Mechanical and Surface Parameters Measured under Standard Conditions Parameters Tensile LT WT RT EMa

Bending B 2HB

Concluding Remarks Shearing G

Fibres, not only natural fibres but also synthetic fibres, are valuable materials as a product of natural resources. On the other hand, our standard of living is becoming higher. We must produce only good fabrics and cannot produce poor fabrics even though market requires lower prices. When low price is required we have to produce ideal fabrics or near ideal fabrics at a lower cost. The next stage of textile technology is the age of real application of technologies developed in this century to the production of higher grade fabrics, that is, the ideal fabrics.

2HG 2HG5

Description

Linearity of load/ extension curve Tensile energy Tensile resilience Extensibility, strain at 500 N/m (gf/cm of tensile load) Bending rigidity Hysteresis of bending moment Shear stiffness Hysteresis of shear force at 0.5 deg. of shear angle Hysteresis of shear force at 5 deg. of shear angle

Unit

None N/m (gf cm/ cm2) % None ×10-4Nm (gf cm2/cm) ×10-2N (gf cm/cm) N/m deg. (gf/cm deg.) N/m (gf/cm) N/m (gf/cm)

Compression LC Linearity of thickness curve WC Compressional energy RC

n

Surface MIU MMD

References 1. Kawabata, S. and Niwa, M., “Fabric Performance in Clothing and Clothing Manufacture”, Journal of Textile Institute, Vol. 80 No. 1, 1989, pp. 19-50. 2. Kawabata, S., Ito, K. and Niwa, M., “Tailoring Process Control”, Journal of Textile Institute, Vol. 83 No. 3, 1992, pp. 361-74. 3. Niwa, M. and Kawabata, S., “The Three Mechanical Components of Fabric Relating to Suit Appearance”, in Carnaby, G.A., Wood, E.J. and Story, L.F. (Eds), The Application of Mathematics and Physics in the Wool

SMD

Compression/ none N/m (gf cm/ cm2) Compressional resilience %

Coefficient of friction Mean deviation of coefficient of friction (frictional roughness) Geometrical roughness

Construction T Fabric thickness W Fabric weight/unit area

None None µm mm ×10g/m2 (mg/cm2)

The warp and weft directional values are identified by 1 and 2 respectively such as MMD-1, B-2, etc. a EM is not used for the calculation of hand value Source: [1]

Table AI.

25

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

A Study of Thread Tensions on a Lockstitch Sewing Machine (Part II) F.B.N. Ferreira Department of Textile Engineering, Minho University, Guimares, Portugal S.C. Harlock Department of Textile Industries, The University of Leeds, UK, and P. Grosberg Shenkar College, Ramat Gan, Israel Received 2 April 1993 Accepted 22 November 1993

the seam. For that reason, references to seam balance in the text should be understood as a function of the ratio between both thread lengths. In an ideal balanced seam, the ratio is 1. Using a multiple regression technique, regression equations were derived for the parameters under study. The dependent variable was always the ratio between the thread lengths and the independent variables were the tension peaks that could influence the setting of the stitch, viz. peak 1 on the needle thread and peaks 5 and 6 on the bobbin thread. These tension peaks were chosen after a careful study of their timing and the relationship with the events that occurred during the same period. It was found that tension peaks 2, 3 and 4 had no influence on the balance of the seam. To study the relationship between these variables, regression equations were defined for all the different experiments presented in Part I, establishing the relationship between the thread length ratio; and (1) peak 1; (2) peak 5; (3) peak 6; (4) peak 1 and peak 5 simultaneously; and (5) peak 1 and peak 6 simultaneously. Also, a remark should be made about the qualitative analysis of the seam balance. After a careful analysis of the visual aspect of the seams obtained during the experiments, it was found

Introduction It is a traditionally-held belief that the balance of the seam is a function of both the needle thread tension and the bobbin thread tension. Using the results obtained during the experiments presented in Part I on the different sewing conditions, the relationship between seam balance and the peak tensions generated during a stitch cycle was studied in order to derive an equation to predict the balance of the seam according to the peak tensions developed. The objective in this part of the work was to try to define the thread tension settings, based on the measurement of peak tensions, according to changes in sewing parameters, to ensure the production of balanced seams.

Factors Affecting Seam Balance Experimental Methodology In each test, a defined seam portion was unpicked and the thread lengths were measured in order to establish the thread length ratio (TLR) of needle thread length to bobbin thread length, which was used as a quantitative measure of the balance of International Journal of Clothing Science and Technology, Vol. 6 No. 5, 1994, pp. 26-29, © MCB University Press, 0955-6222

26

VOLUME 6 NUMBER 5 1994

that, if the length ratio between the needle thread and the bobbin thread is higher than 0.85 and lower than 1.15, the visual appearance of the seam is that of an acceptable balanced seam. Therefore, these will be considered as the tolerance limits for seam balance. In the graphs presented in this article, it will be evident that it was not always possible to establish a simple relationship between the balance of the seam and the variables presented on both axes. However, attempts were made to try to identify an envelope of conditions which could contain balanced seams.

by the tensioner. As that tension increases and consequently reaches peak 1, the balance of the seam tended to occur near the upper side of the fabric. The most significant parameter to the balance of the seam was peak 1, in the conditions of the first experiment. Also, in this experiment, it was found the influence on seam balance of the bobbin thread peak tensions was not significant.

Peak 1 450 Peak 1 = (TLR* 100 – 125.75 – 0.84* P5)/(–0.39)

R = 0.86

400

Sewing Conditions As referred to before, the analysis of the relationship between seam balance and the peak tensions will be made using the data already presented in Part I of this article. For that reason, the sewing conditions are the same as presented there.

350 300 250 TLR = 1.0

200 150 0

50

100

150

200 Peak 5

Experimental Results The different experiments made were described in Part I. Processing the data obtained using a multiple regression technique, equations were obtained relating the seam balance and the tension peaks referred to previously. Graphs illustrating the results and the theoretical curve obtained for an ideal balanced seam are presented in Figures 1 to 6, in which peak 1 and peak 5 are considered simultaneously, for the different experiments.

Figure 2.

Peak 1 450 400 350

TLR = 1.0

300

Discussion of Results

250

In the analysis of the relationship between the peak tensions and the balance of the seam, it was found that: ● As expected and stated in previous studies [1-8], the balance of the seam is influenced by the tension applied on the needle thread

200 Peak 1 = (TLR* 100 – 153.36 + 0.07* P5)/0.16

R = 0.87

150 0

50

100

150

200 Peak 5

Figure 3.

Peak 1

Peak 1 450

450

400

400

350

350 300

TLR = 1.0

300

250

250

200

200 Peak 1 = (TLR* 100 – 160.31 – 0.07* P5)/(–0.22)

Peak 1 = (TLR* 100 – 130.52 – 0.53* P5)/(–0.27)

R = 0.96

R = 0.83

150

150 0

50

100

0

150

50

100

150

200 Peak 5

Peak 5

Figure 4.

Figure 1.

27

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

according to the properties of the fabric different values of tension are required in order to obtain balanced seams. ● The same conclusion was reached regarding the effect of sewing thread quality on the balance of the seam. Again, different trends were noticeable according to the sewing thread quality used. This confirms that distinctive regression equations need to be used in order to predict the balance of the seam. These conclusions showed that most of the factors studied significantly influence the balance of the seam. For this reason, even considering that the reasons for some of the results are not very clear and therefore require further study, an effort was made to try to find common conditions in the distinctive experiments, in order to obtain a balanced seam. It was found that if: peak 1 was greater than 250g and lower than 320g; and the difference between peak 1 and peak 5 is greater than 160g and lower than 275g; and the difference between peak 1 and peak 6 is greater than 160g and lower than 275g; and the ratio between peak 1 and peak 5 is greater than 3 then thread length ratios between 0.85 and 1.15 were obtained. This shows that it is possible to define working conditions as a function of the tensions detected on the needle thread and on the bobbin thread which could lead to the production of balanced seams.

Peak 1 A – TLR = 1.0

450 B – TLR = 1.0 350

C – TLR = 1.0

250 A – Peak 1 = (TLR* 100 – 294.21 – 0.19* P5)/(–0.57) R = 0.93 B – Peak 1 = (TLR* 100 – 200.80 – 0.37* P5)/(–0.48) R = 0.90 C – Peak 1 = (TLR* 100 – 117.59 – 0.35* P5)/(–0.21) R = 0.83 150 0

50

100

150

200 Peak 5

Figure 5.

Peak 1

A – Peak 1 = (TLR* 100 – 134.18 – 0.14* P5)/(–0.19) R = 0.95 B – Peak 1 = (TLR* 100 – 170.41 + 0.20* P5)/(–0.18) R = 0.82 C – Peak 1 = (TLR* 100 – 86.17 + 0.25* P5)/0.21 R = 0.34

550

C – TLR = 1.0

450

350 A – TLR = 1.0 250 B – TLR = 1.0 150

0

50

100

150

200 Peak 5

Figure 6.

In the conditions of the second experiment, the dominant factor in terms of seam balance was the bobbin peak tensions. As the tension on the bobbin thread increased, the balance of the seam tended to occur near the bottom side of the fabric. The tensions generated on the needle thread did not significantly influence the balance of the seam. ● The sewing speed does not influence the balance of the seam. No significant correlation was found between this parameter and the balance of the seam. ● The analysis of the effect of the number of plies on the balance of the seam was not conclusive. However, with the results obtained it is possible that distinctive trends with the number of plies do exist and consequently different regression equations should be used to predict the balance of the seam. ● Also, with different fabrics, distinctive trends were noticed on the relationship between seam balance and the tensions generated on both the needle and the bobbin thread. The results obtained confirm that ●

Conclusions According to the conclusions presented before, a control system can be idealized whereby the peak tensions generated on both the needle and the bobbin thread during the stitch cycle can be measured by the transducers and this information can be compared with the working conditions previously defined. If significant differences are detected, an output can be sent to a device that will adjust the tension applied on the needle thread to the correct value in order to obtain a balanced seam. The tension applied on the needle thread used as a control variable does not mean that the tensions generated on the bobbin thread during a stitch cycle can be treated in a passive way. A technique is required to adjust the tension applied on the bobbin thread in a continuous mode. Both the needle thread tension variations and the bobbin thread tension variations should be measured dynamically because, as shown

28

VOLUME 6 NUMBER 5 1994

before, both play an important role on the balance of the seam. Strictly speaking, the limits on peak tension variations should be set for individual sewing threads and fabrics, as indicated by these results. However, it is possible to define a general envelope of conditions within which balanced seams can be obtained – at least for the fabrics tested.

5.

■

6.

References 1. Deery, W.A. and Chamberlain, N.H., “A Study of Thread Tension Variation during the Work Cycle in a Lockstitch Sewing Machine”, Technical Report No. 15, The Clothing Institute, 1964. 2. Ferreira, F.B.N., “A Study of Thread Tensions on a Lockstitch Sewing Machine”, PhD thesis, University of Leeds, 1991. 3. Greenberg, N.G., “An Instrument for Measurement of Thread Dynamic Tension Characteristics during the Sewing Operation”, Clothing Research Journal, Vol. 3 No. 2, 1975, pp. 77-84. 4. Horino, T., Miura, Y., Ando, Y. and Sakamoto, K., “Simultaneous Measurements of Needle Thread Tension and Check Spring Motion of

7.

8.

29

Lockstitch Sewing Machine for Industrial Use”, Journal of the Textile Machinery Society of Japan, Vol. 28 No. 2, 1982, pp. T30-37. Kamata, Y., Kinoshita, R., Ishikawa, S. and Fujisaki, K., “Disengagement of Needle Thread from Rotating Hook, Effects of Its Timing on Tightening Tension on an Industrial Single Needle Lockstitch Machine”, Journal of the Textile Machinery Society of Japan, Vol. 30 No. 2, 1984, pp. T40-9. Kamata, Y., Sakai, T., Onoue, M. and Chatani, Y., “Analysis of Tightening Tension Waves in Single Needle Lockstitch Machine”, Journal of the Textile Machinery Society of Japan, Vol. 39 No. 1, 1986, pp. T7-15 and Vol. 39 No. 6, 1986, pp. T86-96. Matsubara, T. and Jinbo, Y., “Analysis Approach for Stitch Construction and Stitch Tightening of Lockstitch Sewing Machine”, Journal of the Society of Fibre Science and Technology, Vol. 10, 1984, pp. T387-394. Onoue, M., “Influences of the Sewing Conditions of the Lockstitch Sewing Machine for Industrial Use on the Needle Thread Tension”, Journal of the Society of Fibre Science and Technology, Vol. 10, 1984, pp. T395-401.

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

Automated Apparel Processing Computer Simulation of Fabric Deformation for the Design of Equipment S.D. McWaters and T.G. Clapp College of Textiles, North Carolina State University, USA, and J.W. Eischen Department of Mechanical and Aerospace Engineering, North Carolina State University, USA

means of simulating fabric behaviour. The use of simulations allows us to optimize the design of a machine without physically having to change parameters on the machine, thus reducing time and costs. One very important property where simulation would be useful is the bending behaviour of fabric during a process. A computer model has been developed by Dr J.W. Eischen and T.W. McDevitt at North Carolina State University which simulates the behaviour of fabric parts during apparel manufacturing[2]. The model is based on large deflection beam theory and the finite element method to characterize a fabric strip under given loads or displacements. The model will be evaluated in this article and will be compared with experimental research. The design of a fabric feeding device will be used to illustrate how the computer simulation method can be applied.

Received 24 September 1994 Accepted 1 November 1994

Introduction Apparel manufacturing processes are a part of a traditionally highly skilled, labour-intensive industry. The international market is extremely competitive, owing to much lower labour wages. In order to maintain a competitive advantage in the global market, selected apparel assembly operations can be automated. A major focus of apparel automation has been placed on materials handling, because it accounts for 80 per cent of the time needed to manufacture apparel[1]. One of the major limitations of automated equipment for apparel manufacturing is designing a machine which has the flexibility to handle a variety of fabric types and sizes. In order to design equipment to be used for apparel automation, it is necessary to have the ability to predict the behaviour of fabric parts as they are in contact with the machines. It is important to understand fabric bending behaviour, so that these predictions can be made. In order to design manufacturing systems for apparel automation, it is necessary to have a

Computer Model The computer program which was used incorporated numerical methods to arrive at solutions owing to the difficulty of determining “closed form” solutions. This difficulty arises owing to the limpness of fabrics which are The authors wish to thank the Defense Logistics Agency for its support of this research.

International Journal of Clothing Science and Technology, Vol. 6 No. 5, 1994, pp. 30-38, © MCB University Press, 0955-6222

30

VOLUME 6 NUMBER 5 1994

subjected to large deflections and rotations. The model uses a practical method of characterizing the fabric strip, whereby the strip is assumed as an elastic continuum. The model is designed for large displacements and rotations of the fabric strip in two dimensions, allowing stretching and bending of the fabric. The behaviour of the fabric is determined by equations derived from differential equilibrium equations. Using a finite element method to discretize the fabric strip, a final set of equations was obtained to comprise a non-linear program. The program can be used to simulate the behaviour of a fabric strip under certain conditions.

by the user, as in the case of the fabric part. One limitation of this model is that the contact surface is assumed to be frictionless. An additional constraint of the model is that it considers only the behaviour of one fabric ply on a rigid surface, and cannot consider multiple plies of fabric. This model was used to simulate the bending behaviour of four woven fabrics, of which the necessary properties of the model are shown in Table I. Owing to the fact that most woven fabrics possess different bending properties in the warp and weft directions, the fabrics were tested in both directions. Therefore, the results were analysed as if there were eight fabrics tested. In addition to simulating the behaviour of eight fabrics, three contact surfaces were tested with the model, representing the shape of the contact surface of a fabric feeding device, called the bottom feeder. This system picks plies of fabric, one at a time, from the bottom of a stack of fabric parts. Each surface represented in the simulations consisted of a long, flat, horizontal section which was concluded with a curved semicircular section, as shown in Figure 2. The curved portion of the surfaces represents a picking roller in the bottom feeder, which is located beneath the stack of fabric parts. The surfaces were varied by using end surface diameters of 6.033cm (2.375in), 10.16cm (4in), and 15.24cm (6in). Each sample simulated was 30cm long and 1cm wide, and was divided into 60 linear segments, defined by 61 nodal points, for the finite element analysis of the fabric piece. Initially, the fabric strip was positioned along the fiat portion of the contact surface. The rigid contact surface, which was assumed to be frictionless, was broken down into 37 linear segments, defined by 38 nodal points. The first segment characterizes the flat portion of the surface, and the remaining segments define the geometry of the curved component.

n

Certain fabric properties are required for the model n The necessary inputs of the program are: the initial position of the strip; relevant material properties; boundary conditions; and loading conditions. The fabric strip is divided into a number of segments defined by nodal points, as shown in Figure 1. The co-ordinates of these nodal points are specified by the user and entered into the input file. Certain fabric properties are required for the model: fabric thickness, bending rigidity, and the weight per unit area. A set of boundary conditions, based on the support configuration of the fabric element, must be included in the input data file. External loads must also be defined in the input data, and can exist in the following forms: the weight of the fabric itself, assigned forces at any point along the fabric strip, or assigned displacements along the fabric strip. The loads are applied in increments and at time intervals specified by the user. One of the constraints of this model is that it has not been made to accommodate dynamic forces, such as inertial and aerodynamic forces. The model is designed to account for a contact surface for the fabric ply. This surface is also divided into a series of connected nodes defined by co-ordinates provided

Fabric

Bending rigidity (gm – cm2/cm)

1 2 3 4 5 6 7 8

Nodes

Table I. Fabric Properties

Figure 1. Nodal Segments

31

0.864 0.928 1.111 2.294 0.336 0.514 0.143 0.178

Thickness Weight (cm) (gmf/cm) 0.0512 0.0512 0.0471 0.0471 0.0588 0.0588 0.0322 0.0332

0.0244 0.0244 0.0236 0.0236 0.0267 0.0267 0.0145 0.0145

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

δh a) 6.033cm diameter (surface 1)

δv

+ b) 10.16cm diameter (surface 2)

Figure 4. Horizontal and Vertical Displacements of Fabric Edge c) 15.24cm diameter (surface 3) Tip displacement (cm) 20 -

Figure 2. Fabric Contact Surfaces

δy

A similar representation is shown in Figure 3, where only 12 segments are used to define the curved component. The simulations were carried out by applying two loads to the fabric strip: the weight of the fabric itself and a force applied to push the fabric along the surface. For some simulations to converge, it was necessary to increase the number of time steps of the load functions and, initially to apply the loads gently. The results of the simulations are graphically presented as the stepwise displacement of the fabric (s) versus the horizontal (δh) or vertical displacement (δv). Figure 4 shows the reference used to obtain the horizontal and vertical displacements of the tip of the fabric sample as it is moved along the surface. The curves of the graph are defined by co-ordinates produced by the computer model, and are given in increments of 2cm from 0 to 20cm of fabric displacement. The results of the simulation of fabric 1 being moved along surface 1 is shown in Figure 5, and along surface 3 in Figure 6. The horizontal displacement levels out after an inflection point,

10 -

δx

00

10

20

s (fabric displacement) (cm)

Figure 5. Simulated Horizontal and Vertical Tip Displacements (Fabric 1, Surface 1)

Tip displacement (cm) 20 -

δy

15 -

10 -

δx

5-

00

10

20

s (fabric displacement) (cm)

Figure 6. Simulated Horizontal and Vertical Tip Displacements (Fabric 1, Surface 3)

Figure 3. Contact Surface Model

32

VOLUME 6 NUMBER 5 1994

and the vertical displacement increases linearly after an inflection point, as expected. The results of the simulations using the other fabrics and surfaces were found to exhibit the same general trends. When comparing the graphs based upon bending rigidity, the curves of the fabrics with lower bending rigidities have inflections occurring sooner than the fabrics of higher bending rigidities. This result is expected, because stiff fabrics have a larger bending curvature which must be overcome before the fabric will begin to drape vertically over the curved surface (Figure 7). It is also observed that when simulations were run using a surface with a larger-diameter curved end, the displacement curves exhibited later inflections than those using a smaller-diameter curved end surface. In other words, the greater the curvature

diameter of the end surface, the greater the bending curvature of the fabric. The analysis of the results of the finite element computer model suggests that it could be valid for simulating the bending behaviour of a range of woven fabrics. In order to make use of the simulated results, the computer model must be validated by experimentation. Experiments were carried out under similar boundary conditions that were defined for the finite element model simulations. The data obtained from these experiments were then compared to the data produced by the model simulations, and conclusions were made based on these comparisons.

Experimental Validation In order to perform testing, the experiments were set up to resemble the conditions of the computer model simulations. A typical setup of the components used for testing and for recording results is shown in Figure 8. Three surfaces were made from 0.3175cm (0.125in) aluminium sheets which were each rolled on one end to form a curved-end surface with a diameter of 6.033cm (2.375in), 10.16cm (4in), or 15.24cm (6in), as shown in Figure 2. Each surface was affixed by an adhesive to a plywood board for stability and then placed on a box during testing. Other equipment used for the experiments was a Javelin television video camera, a Tokina zoom lens, a Matrox

+ Stiff fabric Limp fabric

Figure 7. Bending Curvature of Limp vs Stiff Fabric

Surface

VGA monitor Vision system

Fabric

Camera

Figure 8. Experimental Setup

33

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

its original position. The recorded images were stored for analysis. Several steps had to be taken to transform the data for comparison with the model simulations. The data consisted of pixel co-ordinates corresponding to the edge of the fabric sample. The only data needed for this research were the co-ordinates of the tip of the fabric sample. Therefore, the appropriate co-ordinates had to be determined from the entire set. Each data set was analysed, and the pertinent co-ordinates were obtained. The data values were further manipulated by subtracting out co-ordinate values of the reference point of the surface (Figure 4). The final transformation of the data was done by scaling the pixel values to obtain horizontal and vertical tip displacements in units of centimetres. As seen in Figure 9, the set of data points of each replicate coincided or fell very close together, therefore showing the low variation among the experimental results. The standard deviations among each set of three points were calculated for this representative case, and were found to be about 1mm (Table II). The plotted results of experimental testing exhibited take characteristics similar to those of the computer simulations, as discussed previously. When the experimental data had been transformed into a format compatible with the theoretical results of the finite element model, observational and statistical analyses could be made. The horizontal and vertical displacements of the tip of the fabric sample, as determined by model simulations and experimental testing, were plotted on one graph for each fabric-surface

vision board, and a standard PC. In addition to the apparatus used, a computer program, written in C language, was used to determine the data. An image was captured and corresponding pixel values were stored on the vision board. The program analysed these values in order to determine pixel co-ordinates of the fabric edge, which were stored in output files. Before the experiments could be performed, the vision system setup had to be calibrated. An image in the vision board was stored as intensities of pixels contained in a rectangular area measuring 512 horizontal pixels and 480 vertical pixels. Intensities were measured as values ranging from 0 to 256, where 0 is black and 256 is white. A threshold pixel value was determined in order for the experimental values to be binarized as black or white. The program analysed pixel values row by row, starting from the top right-hand corner of the image, in order to determine the location of the edge of the fabric, defined by black pixel values. The co-ordinates of the first black pixel value in each row were saved to an output file. In addition to determining a threshold value, horizontal and vertical scales had to be calculated in order to convert from pixels to centimetres. Images of objects of known dimensions were taken, and 6 pixels were found to equal 1cm both in the horizontal and the vertical directions.

n

The horizontal and vertical displacements were plotted on one graph n Three samples of each fabric described in the previous section were cut and tested. Each fabric sample was originally positioned on the surface with its leading edge resting on the tangent point of the curved surface, in the same way as simulated in the model. The camera and attached zoom lens were placed about 10 metres from the fabric sample, which was positioned on the surface, and was focused in order to capture the image. The vision board then retrieved and stored the captured image from the camera in the form of pixel values, from which the appropriate data were extracted and sent to output files. The sample was then pushed 2cm towards the curved end of the surface in order to capture the next image. The fabric continued to be pushed in increments of 2cm over the curved surface, recording images of each step, until it had been displaced 20cm from

Tip displacement (cm) 20 -

δy

15 10 -

δx

50–5 0

10

s (fabric displacement) (cm)

Figure 9. Experimental Horizontal and Vertical Tip Displacements (Fabric 7, Surface 3)

34

20

VOLUME 6 NUMBER 5 1994

Fabric displacement step (cm) 2 4 6 8 10 12 14 16 18 20

Standard deviation (cm) x-displacement y-displacement 0.0981 0.0981 0.0000 0.0924 0.0981 0.0981 0.0981 0.0924 0.0924 0.1905

Tip displacement (cm) 20 -

0.0924 0.0058 0.0981 0.0058 0.0058 0.0000 0.0058 0.0058 0.0954 0.0058

δy

10 -

δx

00

20

10

s (fabric displacement) (cm)

Table II. Standard Deviation among Experimental Data Points

Figure 10. Combined Horizontal and Vertical Displacements (Fabric 1, Surface 1)

combination. Figure 10 shows one of these graphs, where the curves represent the simulated results and the points refer to experimental data. In general, the graphs suggested strong relationships between theoretical and experimental data, as seen in Figure 10.

was capable of explaining the variation of tip deflection of fabric as it is pushed over a curved surface is the case of fabric 6 and surface 1. This exception occurs because this case involves the smallest diameter surface and a very limp fabric. There were several sources of error for the collection of experimental data, as well as the computer model simulations. One possible source of error was the operator, because the fabric sample was manually moved for each step. Other experimental errors were owing to camera resolution and the condition of the fabric samples. Some fabrics were deformed owing to previous folding or position on the roll of fabric. Owing to these factors, some fabric samples showed a curvature, when the computer model predicted none. Therefore, there was a decrease in horizontal deflection of the experimental data while that of

Results Statistical analysis of the simulated and experimental data was necessary to evaluate their relationship further. A Pearson’s product-moment coefficient of determination was determined for the paired data of the simulated results and the average experimental results for each case, and is shown in Table III. Most of the values in the table suggest that there exists a high correlation between the simulated results of the computer model and experimental results. One exception which is not high enough to suggest that the computer model

Fabric 1 2 3 4 5 6 7 8

Horizontal 0.991 0.973 0.860 0.893 0.627 0.000 0.964 0.662

Surface 1 Vertical 0.999 0.998 0.998 0.998 0.997 0.898 0.998 0.985

Surface 2 Horizontal Vertical 0.997 0.988 0.994 0.928 0.801 0.915 0.976 0.961

Table III. Pearson’s Product-moment Coefficients of Determination (R2)

35

0.999 0.999 0.998 0.995 0.994 0.996 0.998 0.998

Surface 3 Horizontal Vertical 0.998 0.994 0.997 0.958 0.957 0.998 0.993 0.999

0.999 0.999 0.994 0.978 0.997 0.998 0.998 0.999

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

the simulation remained constant. Even though these are common situations in real uses of fabric, the computer model did not account for these circumstances. An additional source of error was the values of bending rigidity of the fabrics used, as determined by FAST testing. The FAST system determines one value for bending rigidity and assumes that it is a linear property, when it has been shown to be a non-linear property.

x-displacement (cm) 10 -

SD3 8-

SD2 SD1

6-

Conclusions

4-

After evaluation of the finite element computer model through experimental testing, it was found 20.0

0.5

1.0

1.5

2.0

2.5

Figure 13. Horizontal Displacement – s = 10cm

5

x-displacement (cm)

10

y-displacement (cm) 9-

8SD1

0 2.5

20

7-

2.0 10 Displacement step (cm)

SD2

1.5 1.0 0 0.0

Bending rigidity (gmf-cm2/cm)

0.5

6SD3 5-

Figure 11. Horizontal Displacement – 6.033cm Diameter

40.0

0.5

1.0

1.5

2.0

2.5

Figure 14. Vertical Displacement – s = 10cm

5

x-displacement (cm)

10

6.1cm

0 2.5

20

6.8cm

+

2.0 10 Displacement step (cm)

1.5 1.0 0 0.0

0.5

Bending rigidity (gmf-cm2/cm)

Figure 15. Example of Gap Setting

Figure 12. Horizontal Displacement – 15.24cm Diameter

36

VOLUME 6 NUMBER 5 1994

Horizontal setting

Horizontal setting

Vertical setting

+

Vertical setting

+

Stiff fabric

Limp fabric

Figure 16. Gap Adjustment Bar Position for Stiff and Limp Fabrics

that the model was a valid approximation of the bending behaviour of fabric over a curved surface. In addition to the evaluation of the computer model performed here, further testing needs to be done, where different boundary conditions would be used. Further improvements for the computer model would be to develop the model to accommodate dynamic forces, as are present in fabric handling. The results obtained from the simulations can be used to choose an optimum diameter of the picking roller in the bottom feeder. If the range of location of the gap adjustment bar is known, it can be determined whether or not a fabric ply will successfully pass above the bar. Therefore, the diameter of the picking roller depends on the range of movement of the gap adjustment bar. The model should also be tested for the ability of its results to be implemented in apparel manufacturing processes.

analysed using SAS, which showed surface diameter and bending rigidity to have significant effects on the horizontal displacement of the fabric tip. The displacement step also had a significant effect both on horizontal and on vertical displacement, but this result was obviously anticipated. Figures 11-12 show the effects of fabric step displacement and bending rigidity on horizontal tip displacement of the fabric strips for each surface diameter. A comparison of the graphs shows that an increase in diameter causes an increase in horizontal displacement. In other words, the fabric has less curvature on a larger diameter surface. This effect was also shown for the vertical displacement, where it decreases with an increase in surface diameter. The graphs also demonstrated that an increase in bending rigidity was found to increase the horizontal displacement; however, there was no significant effect on vertical displacement. Therefore, a stiffer fabric has a larger horizontal displacement as it is pushed over a contact surface. It is also important to notice the peaks of the surface plots, which indicate the displacement step where a fabric tip drops and continues to fall straight vertically. As surface diameter increased the peak occurred at a higher displacement step. A larger surface diameter would be more desirable because the fabric tip positions would be very similar for a wider range of bending rigidities and the fabric plies would be more likely to pass above the gap adjustment bar. Using this information, an appropriate diameter and gap setting can be chosen, based on how far a fabric strip will be pushed at the point it is picked from the bottom of the stack. In order for a fabric

n

Increase in diameter causes an increase in horizontal displacement n Statistical analysis was performed to determine the factors which produced significant effects on the horizontal and vertical displacements of the fabric tip. The factors tested were surface diameter (SD), fabric thickness (TH), fabric bending rigidity (BR), and the fabric displacement step (S). The dam points of the computer model output were

37

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

cylinder can be changed. Choosing a surface diameter depends on the allowable adjustments of the gap setting, the bending rigidities, and fabric displacement. As can be seen in Figures 13 and 14, the largest surface diameter causes the horizontal and vertical displacements to have the least variation across a range of bending rigidities. Therefore it would be advantageous to use as large a surface diameter as possible, in order to accommodate a wider range of bending rigidities. The size of the cylinder diameter would also be governed by the ability of the remaining plies of the stack to stay above the gap adjustment bar.

to be successfully picked from a stack, the gap must be properly set to allow passage of the bottom ply only. Therefore, the second ply must remain above the gap adjustment bar during operation of the machine. For example, suppose that the second ply of a fabric, which has a bending rigidity of 0.750 gmf-cm2/cm, has been displaced 10cm from a machine cylinder with a diameter of 10.16cm. Using these values and the graphs in Figures 13 and 14, the horizontal and vertical displacements of the fabric tip would be 6.1cm and 6.8cm, respectively. Therefore, the horizontal and vertical settings of the gap adjustment bar would be set as shown in Figure 15, where the fabric tip lies above the bar. These settings are approximate, but the vertical adjustment should be set no higher and the horizontal adjustment no larger than the estimated settings so that the ply will pass over the bar. Limp fabrics require a smaller horizontal and a lower vertical gap setting than stiff fabrics (Figure 16). If the gap setting adjustment is not capable of accommodating a specified range of bending rigidities, the surface diameter of the picking

n References 1. McDevitt, T.W., “Flexible Fabric Mechanics Analysis Using Large Deflection Beam Theory”, Master’s Thesis, North Carolina State University, 1993. 2. Tait, N., “Materials Handling”, Apparel International, Vol. 22 No. 3. October 1992, pp. 5-9, 11.

38

VOLUME 6 NUMBER 5 1994

COMMUNICATIONS

A Study of Thread Tensions on a Lockstitch Sewing Machine (Part III) Further Stitch Formation Analysis F.B.N. Ferreira Department of Textile Engineering, Minho University, Guimares, Portugal S.C. Harlock Department of Textile Industries, The University of Leeds, UK, and P. Grosberg Shenkar College, Ramat Gan, Israel Introduction

From a frame-by-frame analysis of the video, a mechanism of stitch formation based on a “robbing back” phenomenon was identified. To illustrate better the events observed, some simplified diagrams are presented in Figure 1. When the thread take-up lever completes its upward movement, drawing the slack needle thread through the fabric, the bobbin thread pulls out of the bobbin and is drawn right through the fabric to its upper side (from 0° to 35°), as shown in the diagram presented in Figure 1a. As the fabric continues its movement forward and the needle moves down (Figure 1b), it was evident that the “bobbin thread loop” remained on the upper side of the fabric. Then, as the needle in its movement upwards withdraws from the fabric, the needle thread loop is moved around the bobbin case and the feed dog starts moving the fabric, feeding it for the next stitch. Also during this period, it is evident that the “bobbin thread loop” remains on the upper side of the fabric. When the needle thread loop slips out of the transfer hook (at 355°), it is noticeable that some tension is applied to the “bobbin thread loop” by the contact of the needle thread loop on the bobbin thread. Then, as the take up lever moves upward drawing the slack needle thread and the bobbin thread through the fabric to form a new stitch, the “bobbin thread loop” of the previous stitch starts

Previous investigations have been inconclusive with regard to the exact tension conditions required to produce seam balance. It was certainly not the case that seam balance occurred when the needle thread and bobbin thread peak tensions occurring simultaneously were equal, as might be expected. Therefore, a further investigation was performed to provide an explanation and a theoretical model to predict the conditions for seam balance. Investigative Procedure In order to understand better the events generated during the stitch formation, a video recording was made, capturing the action in the area where the needle penetrates the fabric at slow speed. A camera, JVC CCD 5310E, equipped with a Computar lens with extension tubes (F - 2.8/16) was used to perform the video recording. Theory of Stitch Formation From a careful observation both “in situ” and of the video recording, some facts were noticed, at these very slow speeds. International Journal of Clothing Science and Technology, Vol. 6 No. 5, 1994, pp. 39-42, © MCB University Press, 0955-6222

39

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

moving down, as shown in Figure 1c. The bobbin thread is again drawn right through to the upper side of the fabric. It seems that, as the bobbin thread is pulled up by the needle thread loop, simultaneously the “bobbin thread loop” of the previous stitch is pulled down; the final setting of that stitch being made at this stage (see Figure 1d). This suggests that a “robbing-back” action happens during the lockstitch formation. For this reason, the setting of the stitch is determined not in one stitch formation cycle, but in its cycle and the cycle of the subsequent stitch. Assuming that this “robbing-back” effect only affects the previous stitch, it seems that when the needle thread loop is pulled upwards by the take up lever, the bobbin thread can be pulled both from the previous stitch and from the bobbin. Consideration is now given to the conditions under which “robbing back” will occur or (to be exact) cease to occur. A diagram illustrating the action somewhere between 355° and 35° of the stitch formation cycle is shown in Figure 2, where: T0, …, T5 = tension on the corresponding sections of the needle thread; B0, …, B5 = tension on the corresponding sections of the bobbin thread; A = stitch in formation; B = stitch previously formed. Assuming that at point C, both needle and bobbin thread are fixed and that the needle thread loop at stitch A is moving upwards drawing the bobbin thread through the fabric, the events may be described as follows. When the needle thread loop starts to contact the bobbin thread, after slipping out of the transfer hook and moving up through the fabric, because the bobbin thread loop (of the previous stitch) is slack,

Needle Needle thread

Fabric

Throat plate Bobbin thread

(a)

Needle Needle thread

Fabric

Throat plate Bobbin thread

(b) Needle Needle thread

Fabric

Throat plate Bobbin thread

(c)

Needle Needle thread

Needle C

T0

A

B

T1

Needle thread Fabric

T5

T4

B5

B4

Fabric

Throat plate

T3

T2

B3

B2

Bobbin thread (d)

Bobbin thread

Throat plate

B1

B0

Figure 2. Stitch Formation Diagram

Figure 1. Lockstitch Formation

40

VOLUME 6 NUMBER 5 1994

as observed on the video record, B3 is lower than B2 (B2 is greater than B0, due to friction. In other words, B2 is greater than the tension applied by the tensioner on the bobbin thread) and for that reason the needle thread loop starts to draw bobbin thread from the previous stitch. During this process, the tension on the needle thread is increasing and consequently T0, T1, T2, T3, T4 and T5 are increasing. With the increment of T4, B4 also increases as the force required to move down stitch B increases with T4. As B4 increases so does B3. When B3 reaches a tension equal to B2, then the bobbin thread needed to be fed to follow the movement up through the fabric of the needle thread loop starts to be drawn not from the previous stitch but from the bobbin, until no more bobbin thread is required. This suggests that the balance of the seam will be mainly a function of the amount of bobbin thread drawn from the previous stitch, until the bobbin thread starts to be released from the bobbin.

is reached. This would prevent robbing-back and gives the lowest possible value for T3. Assuming that friction in the needle eye is τ, so that T1 = T0 – τ (6) and because the needle thread is moving upwards T0 – τ = T1 = r2 T2 (7) where r2 ≈ eµπ/2, (8) consequently T2 = (1/r2)(T0 – τ) (9) T3 = (1/r1r2)(T0 – τ) (10) T4 = (1/r2)(T0 – τ) (11) T5 = (r1r2)(T0 – τ). (12) Considering now the bottom tensions, the tensioner will apply a friction force δ and B1 = B0 + δ. (13) However, from direct measurement, the tension detected is B1 and for this reason this equation is not useful. Because the bobbin thread is always moving from the bobbin to stitch A, (14) B2 = r1B1. On the left side of stitch A, the bobbin thread is always pulled to the right and consequently B3 = (1/r1) B2 = B1 (15) B4 = (1/r1) B3 = (1/r1) B1 (16) 2 B5 = (1/r1) B4 = (1/r1 ) B1. (17) The bobbin thread will come off the bobbin (and presumably the loop at stitch B will stop descending) when T2 + T3 = B2 + B3. (18) Equations (9), (10), (14) and (15) yield: r1 + 1 (T0 – τ) = (r1 + 1)B1 (19) r1r2 or T0 = τ + r1r2B1 (20) τ is probably small and consequently: T0 ≈ r1r2B1 (21) Assuming that µ = 0.3, r1 = 2.57 and r2 = 1.60. Replacing in (21), T0 ≈ 4.1 B1. (22) According to this explanation, it means that if T0 is less than 4.1 B1 the loop at stitch B moves to the lower side of the fabric, without removing bobbin thread from the bobbin. If T0 is larger than 4.1 B1 the loop at stitch B stays on the upper side

Estimation of the Needle Thread Tension at which Bobbin Thread Is Released from the Bobbin From the analysis presented, it is necessary to define the relationship between both the needle and the bobbin thread tensions at which the bobbin thread stops being pulled from the previous stitch and starts to be pulled from the bobbin. From Figure 2, when the take up lever removes the slack needle thread and pulls with it the bobbin thread, the loop at stitch A moves up and stitch B moves down. Friction must act to prevent these movements and therefore: T5 > T4 > T3. (1) If the angle of wrap between the yarns with tension T3 and T4 and T4 and T5 is approximately 180° as shown in Figure 2, (2) T5 = r1 T4 and T4 = r1 T3 (3) where r1 = eµπ. (4) At stitch A, the needle thread is probably not moving from right to left and it could move in the opposite way if the limit T2 = r1 T3 (5)

41

INTERNATIONAL JOURNAL OF CLOTHING SCIENCE AND TECHNOLOGY

of the fabric, the thread required for the formation of stitch A being supplied by the bobbin. It seems that, in order to obtain a balanced seam, the timing at which T0 reaches 4.1 B1 is the most significant factor. If this value is reached early, the stitch will be set closer to the upper side of the fabric. If this happens later, the stitch will be set closer to the bottom side of the fabric. This suggests that further research should be made in order to investigate the timing at which this equality occurs and its relationship to the balance of the seam. Using the data available, namely those obtained in the experiments presented in Parts I and II, it was noticeable that, generally, the ratio between needle thread tension and bobbin thread tension more quickly increased, as the tension applied on the needle thread by the tensioner increased. This shows that, if the tension applied on the needle thread increases, the stage at which the previously derived equality occurs will be earlier

in the stitch cycle and consequently the stitch should be set closer to the upper side of the fabric. This conclusion agrees with the experimental results obtained in the previous articles.

Conclusions As previously referred to, it is a traditionally held belief that the seam balance, or the stitch setting, is a function of the tensions on the needle thread and on the bobbin thread. However, from the analysis presented above, instead of the stitch setting occurring during the stitch cycle of the stitch itself, it seems that the setting of the stitch occurs finally only on the formation cycle of the next stitch. From this analysis it is evident that the balance of the seam is a function not only of both the needle and bobbin thread tension values, as was traditionally believed, but also of the stage at which the appropriate relationship between both the needle and the bobbin thread tensions occurs.

42

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

No title

Recommend Documents