No title

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm Basic garm...

Author: Emerald

296 downloads 515 Views 8MB 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

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Basic garment pattern generation using geometric modeling method

Basic garment pattern generation

Sungmin Kim Faculty of Applied Chemical Engineering, Chonnam National University, Kwangju, South Korea, and

Chang Kyu Park Department of Textile Engineering, Konkuk University, Seoul, South Korea

7 Received January 2006 Revised July 2006 Accepted July 2006

Abstract Purpose – The generation of individually fit basic garment pattern is one of the most important steps in the garment-manufacturing process. This paper seeks to present a new methodology to generate basic patterns of various sizes and styles using three-dimensional geometric modeling method. Design/methodology/approach – The geometry of a garment is divided into fit zone and fashion zone. The geometry of fit zone is prepared from 3D body scan data and can be resized parametrically. The fashion zone is modeled using various parameters characterizing the aesthetic appearance of garments. Finally, the 3D garment model is projected into corresponding flat panels considering the physical properties of the base material as well as the producibility of the garment. Findings – The main findings were geometric modeling and flat pattern generation method for various garments. Originality/value – Parametrically deformable garment models enable the design of garments with various size and silhouette so that designers can obtain flat patterns of complex garments before actually making them. Also the number and direction of darts can be determined automatically considering the physical property of fabric. Keywords Geometry, Modelling, Garment industry, Fashion design Paper type Research paper

Introduction Three-dimensional computer-aided design has become one of the most indispensable elements in modern industries. It is very difficult to find any design process that is not aided by CAD systems in traditional manufacturing processes of machinery, aircraft, and watercraft and most engineers take it for granted nowadays (Lee, 1999). Of course, such a trend is steadily spreading over garment industries. However, 2D CAD systems are still prevalent and therefore 3D design of garment is difficult. This is partly because of the difficulties in mathematical modeling of the fabric material that. However, a more significant problem is that designers and pattern makers do not think the CAD method to be better than the conventional manual method (Stjepanovic, 1995, Aldrich, 1992). Although such a phenomenon has always been with the initial development period of other purpose CAD systems, it is more serious as well as natural in garment industry because the quality evaluation of the product depends more on human aesthetic sense than on the optimum physical performance. Certainly, there are garments that need functionality more than the aesthetic appearance such as tight-fitting underwear or protective garments, CAD systems will be powerful tools in designing and manufacturing such mass-customized high-functional garments.

International Journal of Clothing Science and Technology Vol. 19 No. 1, 2007 pp. 7-17 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710717017

IJCST 19,1

8

There are two major basic pattern generation methods in conventional garment design. One is the flat pattern process, where patterns are designed two-dimensionally and the other is the 3D draping method, where a flat fabric is directly formed into a garment on a mannequin. Although it is more difficult to get garment patterns by the latter method, it is more appropriate to make well-fit patterns than the former method and recent studies on automatic pattern generation have focused on the latter method. In that case, some cut lines called as darts are necessary if the garment has undevelopable patches. McCartney et al. tried to develop 3D surfaces using darts and to implement such techniques into pattern generation (McCartney et al., 1999; Collier and Collier, 1990; Heisey et al., 1990a,b). However, full 3D human body data were not easily accessible then that it was difficult to make practical garment patterns. Recently, due to the advancement in non-contact measurement technology, 3D body data became affordable and many studies have been made on this topic (McCartney et al., 2000; Daanen and Water, 1998; Au and Yuen, 1999; Paquette, 1996). Kim et al. tried to generate garment patterns by combining the 3D data of body and garment model. They also developed an interactive pattern design system using a parametrically deformable body model made from the 3D body scan data (Kim and Park, 2004; Kim and Kang, 2002). However, there have been some problems that it is difficult to get the ideal shape of garment by such methods and only the basic patterns can be obtained. In this study, a garment is divided into two zones, one for fit zone and the other for fashion zone. The fit zone is modeled by digitizing the body scan data so that it can provide optimum fit as well as the ideal shape of the garment. The fit zone can be reshaped and resized using basic anthropometric parameters that various sized fit zone can be modeled from a single body scan data. The fashion zone is modeled using parameters that determine the aesthetic appearance of the garment that users can design garments with various silhouette intuitively. Finally, a flat pattern projection algorithm was developed to make flat pattern pieces from three-dimensionally modeled garment considering the physical properties as well as the producibility of fabric. Fit zone modeling Interactive modeling of fit zone from 3D body data A garment can be divided into two zones to facilitate the 3D pattern generation. One is the fit zone, and the other is the fashion zone. The fit zone of a garment is in direct contact with the body surface and is responsible for the comfort of a garment that it must be designed to correctly reflect the shape of underlying body. In our previous research, the fit zone was modeled automatically from the body scan data, however, it was somewhat difficult to get a desirable shape of the fit zone with this method especially for slacks (Kim and Park, 2004; Kim and Kang, 2002). In this study, a fit zone was modeled by capturing and reconstructing the quadrilateral gird structure marked on the surface of a mannequin using a multi-joint coordinate measurement system as shown in Figure 1. The surface model of a fit zone was constructed as a B-Spline surface taking the measured grid points as their control points as shown in Figure 2. The shape of fit zone was then refined to have a smooth shape by adjusting the position of each control point manually. Zebra striping technique was applied to verify the smoothness of the surface visually as shown in Figure 2. Some examples of the fit zones prepared by this method are shown in Figure 3.

Basic garment pattern generation 9

Figure 1. Schematic diagram of fit zone modeling

Figure 2. B-Spline surface based fit-zone modeling

IJCST 19,1

10

Figure 3. Examples of fit zones

However, it is certainly meaningless to model the fit zone through the complicated method described above if it should be generated for each body scan data. Therefore, it is necessary to change the shape and size of the fit zone in a parametric way. The shape parameters of the fit zone defined in this study are listed in Table I. The initial values for parameters in Table I are extracted from the originally obtained fit zone model. Then the original fit zone model is deformed according to the user-supplied parameters. The deformation of the fit zone is made by changing the height, the width, and the circumference of each key cross-section of the fit zone as shown in Figure 4. Fashion zone modeling The fashion zone of a garment is usually not in direct contact with the body but is draped freely to make the aesthetic appearance. The aesthetic appearance of a garment

can be evaluated by many factors such as its overall silhouette or the number of folds at its bottom. In our previous research, only the tight fitting basic garments could be designed because the garment patterns were drawn directly on the surface of the body model. So far, costly trial-and-error methods have been used to obtain the desirable aesthetic appearance of a garment because it is very difficult to draw 2D patterns to accommodate to the shapes of imaginary 3D garments. In this study, the fashion zone of a garment is formed parametrically in 3D and projected into 2D pieces that complex patterns can be generated easily from the intuitive 3D previews. Parameters used to design the aesthetic appearance of a garment in this study are listed in Table II. For example, the bottom shape of a skirt can be changed using two parameters including the number and amplitude of bottom fold. A sine curve of specific period and Bodice Parameter Bust girth Front length girth Neck girth Side length Armhole depth Shoulder angle (8)

Skirt Default (in) 33.15 7.42 7.45 4.53 4.89 25

Parameter Waist girth Abdomen girth Hip girth Skirt girth Side length

Default (in) 27.32 33.83 37.25 37.36 13.30

Basic garment pattern generation 11

Slacks Parameter Default (in) Waist girth Abdomen Hip girth Thigh girth Side length

27.32 33.83 37.25 22.12 13.30

Table I. Shape parameters for fit zones

Figure 4. Parametric deformation of fit zone

Parameter

Bodice

Default value Skirt

Height Bottom width (percent) Bottom depth (percent) Front insert Side insert Bottom rounding Number of bottom fold Amplitude of bottom fold

6.97 in 80.0 80.0 0 0 0 0 0

11.8 in 90.0 90.0 0 0 0 0 0

Slacks 27.5 in 80.0 80.0 Table II. Shape parameters for fashion zones

IJCST 19,1

12

amplitude can be formed using two parameters as shown in Figure 5. Assuming the initial shape of a skirt to be an n-sided cylinder, the shape of its bottom edge can be modified by changing the distance between each edge and the centroid of bottom plane according to the amplitude of sine curve at corresponding angle. As the various 3D shapes of garments can be previewed easily in 3D with this method, it would be possible to obtain appropriate patterns for specific garment more easily. Garments with various shapes can be designed easily by connecting the fit zone and fashion zone together as shown in Figure 6. Flat pattern generation A garment model usually consists of undevelopable surfaces that some cutting lines called the “darts” must be introduced to generate the flat patterns. In this study, darts are defined in either the lateral or the longitudinal direction of a pattern because those directions are approximately parallel to the major UV axes of the B-Spline surface of each pattern. In this method, darts are created along the mesh structure of the surface that it is easy to create as many darts as needed. This method seems acceptable because the darts are usually defined as such in the traditional manual pattern design method. Examples of the defined darts are shown in Figure 7. The strain reduction method was used to flatten the patterns with darts. First, the mesh structure of each pattern was reorganized considering the darts and then projected onto the 2D plane. As each mesh element on a pattern undergoes a certain deformation during the projection, a considerable amount of strain is induced in the flattened pattern. An edge can be defined between every two node on the pattern as shown in Figure 8. By comparing and equalizing the lengths of corresponding edges in

Amplitude of a fold

2p Number of folds

Modulation

Figure 5. Schematic diagram of fashion zone formation

Basic garment pattern generation 13

Figure 6. Examples of designed garments

Figure 7. Examples of dart definition

IJCST 19,1

14

Figure 8. Schematic diagram of strain reduction

2D and 3D pattern iteratively, residual strain in the projected pattern can be reduced. By this strain reduction method, resulting flat patterns can be obtained. By considering the physical properties of the fabric material such as maximum elastic allowance and maximum shear angle allowance during the strain reduction, even the optimum shapes of patterns for different fabric material can assumed to be obtained. Some examples of flat patterns are shown in Figure 9. Strain reduction is terminated after the rate of change in total strain reached a predefined threshold value. The distribution of residual strain is visualized on the flattened patterns to guide the users for the better placement of darts. However, as the number and positions of darts are determined entirely by the user and therefore the physically optimum number and positions of darts cannot always be obtained by this method. Of course, there are certain algorithms to determine the optimum cutting lines for the flattening of undevelopable surfaces (McCartney et al., 1999, 2000, Wang et al., 2004). But the applicable range of those methods are usually limited and the cutting lines suggested by those methods usually have the arbitrarily curved shapes and it is not desirable to introduce such complicated cutting lines to garment patterns because of the limitations not only in the design but also in the producibility of the garment. In this study, an automatic dart generation algorithm was developed to make approximately linear darts along the mesh structure as described in the manual dart definition method. Assuming the rectangular topology of a pattern, darts can be initiated from one of its four edges and can be drawn inward perpendicular to each edge as shown in Figure 10. The principle of this automatic dart

Basic garment pattern generation 15

Figure 9. Examples of flattened patterns with user-defined darts

generation method is described in detail in our previous research (Kim and Kang, 2002). The suitability of the automatically generated darts can be verified by the visualized result of residual strain distribution. Results and discussion As shown in the above examples, garments with various sizes and silhouettes could be designed prior to the actual prototyping process. Therefore, the method developed in this study can be used to reduce the waste of both time and cost in conventional trial-and-error based processes. Once designed, garments are managed by a database system as the templates that users can design a new garment from on a similar one stored in the database to save their time. Conclusion In this study, a basic pattern generation system has been developed which generates flat garment patterns by modeling and flattening the 3D garment models. For this, a garment is divided into two zones including the fit zone and the fashion zone. The fit zone was modeled by digitizing the body scan data so that its size and shape could be modified in a parametric way. The fashion zone was modeled by parameterizing the

IJCST 19,1

16

Figure 10. Example of automatic dart formation

aesthetic appearance of garment silhouette so that the users can design various garment intuitively. Also a database management system for garment shape templates has been developed so that the users can design various garments ranging from basic items to fancy ones rapidly. Finally, a flat pattern projection algorithm has been developed to make flat patterns considering the physical properties as well as producibility of garment. Automatic pattern generation system will be a necessary tool to meet the future trend in manufacturing industries such as the mass-customization and the quick-response system. References Aldrich, W. (1992), CAD in Clothing and Textiles, Oxford BSP Professional Books, London. Au, C. and Yuen, M. (1999), “Feature-based reverse engineering of mannequin for garment design”, Computer-Aided Design, Vol. 15 No. 1, pp. 751-9. Collier, B. and Collier, J. (1990), “CAD/CAM in the textile and apparel industry”, Clothing Textile Research Journal, Vol. 8 No. 3, pp. 7-12. Daanen, H. and Water, G. (1998), “Whole body scanners”, Displays, Vol. 19 No. 3, pp. 111-20.

Heisey, F., Brown, P. and Johnson, R. (1990a), “Three-dimensional pattern drafting, part 1: projection”, Textile Research Journal, Vol. 60 No. 11, pp. 690-6. Heisey, F., Brown, P. and Johnson, R. (1990b), “Three-dimensional pattern drafting, part 2: garment modeling”, Textile Research Journal, Vol. 60 No. 12, pp. 731-7. Kim, S. and Kang, T. (2002), “Garment pattern generation from body scan data”, Computer-Aided Design, Vol. 35, pp. 611-8. Kim, S. and Park, C. (2004), “Parametric body model generation for garment drape simulation”, Fibers and Polymers, Vol. 5 No. 1, pp. 12-18. Lee, K. (1999), Principles of CAD/CAM/CAE Systems, Addison-Wesley, Reading, MA. McCartney, J., Hinds, B. and Seow, B. (1999), “The flattening of triangulated surfaces incorporating darts and gussets”, Computer-Aided Design, Vol. 31 No. 4, pp. 249-60. McCartney, J., Hinds, B., Seow, B. and Dong, G. (2000), “An energy based model for the flattening of woven fabrics”, Journal of Materials Processing Technology, Vol. 107 No. 1, pp. 312-8. Paquette, S. (1996), “3D scanning in apparel design and human engineering”, IEEE Computer Graphics and Applications, Vol. 16 No. 5, pp. 11-15. Stjepanovic, Z. (1995), “Computer-aided processes in garment production”, International Journal of Clothing Science & Technology, Vol. 7 No. 2, pp. 81-8. Wang, C.C.L., Wang, Y., Tang, K. and Yuen, M.M.F. (2004), “Reduce the stretch in surface flattening by finding cutting paths to the surface boundary”, Computer-Aided Design, Vol. 30, pp. 665-77. Corresponding author Chang Kyu Park can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Basic garment pattern generation 17

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,1

Prediction of the air permeability of woven fabrics using neural networks

18 Received March 2006 Revised August 2006 Accepted August 2006

Ahmet C¸ay Ege University, Textile Engineering Department, Izmir, Turkey

Savvas Vassiliadis and Maria Rangoussi Technological Education Institute of Piraeus, Department of Electronics, Athens, Greece, and

Is¸ık Tarakc¸ıog˘lu Ege University, Textile Engineering Department, Izmir, Turkey Abstract Purpose – The target of the current work is the creation of a model for the prediction of the air permeability of the woven fabrics and the water content of the fabrics after the vacuum drying. Design/methodology/approach – There have been produced 30 different woven fabrics under certain weft and warp densities. The values of the air permeability and water content after the vacuum drying have been measured using standard laboratory techniques. The structural parameters of the fabrics and the measured values have been correlated using techniques like multiple linear regression and Artificial Neural Networks (ANN). The ANN and especially the generalized regression ANN permit the prediction of the air permeability of the fabrics and consequently of the water content after vacuum drying. The performance of the related models has been evaluated by comparing the predicted values with the respective experimental ones. Findings – The predicted values from the nonlinear models approach satisfactorily the experimental results. Although air permeability of the textile fabrics is a complex phenomenon, the nonlinear modeling becomes a useful tool for its prediction based on the structural data of the woven fabrics. Originality/value – The air permeability and water content modeling support the prediction of the related physical properties of the fabric based on the design parameters only. The vacuum drying performance estimation supports the optimization of the industrial drying procedure. Keywords Air, Permeability, Neural nets, Porosity, Drying, Modelling Paper type Research paper

Introduction Textile fabrics consist of interlaced yarns or fibers. They are complex materials and their structure is porous. They permit the flow of the air through the constituting materials: yarns and fibers. The discussion about the air permeability of the textile International Journal of Clothing Science and Technology Vol. 19 No. 1, 2007 pp. 18-35 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710717026

The authors would like to acknowledge Clothing Textile and Fibre Technological Development SA, in Athens, Greece for the use of the Shirley air permeability tester and Go¨khan Tekstil San. Tic. A.S¸., in Denizli, Turkey, for the kind supply of the woven fabrics. The present work has been partially supported by the “Archimedes 1” EPEAEK Research Project in TEI Piraeus, co-financed by the EU (ESF) and the Greek Ministry of Education.

fabrics deals mainly with the airflow perpendicular to their respective neutral level or the actual level formed by the warp and weft yarns, as it is shown in Figure 1. Air permeability of fabrics is of great significance. Depending on the air flow rate through the fabric under a given pressure drop, textile fabrics get their specific character in terms of thermal properties, or more globally in terms of comfort. Comfort of the textiles is a complex property affecting intensively the acceptance and the performance of the final product. In parallel, the air permeability has a significant influence on the drying behavior of the fabrics and especially in the procedure of vacuum drying. Using vacuum during method, the water content of wet fabrics is removed by suction. Highly porous structures of wet fabrics permit the flow of the air through the mass of the fabrics, resulting into a faster and more effective drying. In the opposite, dense structures block the air flow and subsequently the drying procedure is getting slower, costing in terms of time and energy. The role of the air permeability of the fabrics has an increased importance in the case of the vacuum drying technology. Typically vacuum drying holds the role of a pre-drying stage, positioned before the main drying unit, which operates by heating the wet fabric (stenter). The behavior of fabrics during vacuum drying varies depending on the fabric type and especially on the micro-structural characteristics of the fabric. The comfort of the textile products and the efficiency of specific processing stages of the fabrics, e.g. vacuum drying, are strongly influenced by the air permeability properties. Thus, the prediction of the air permeability fabric becomes of high interest, reflecting on the performance of the end products and the cost of the production procedure. Prediction of the behavior of a fabric during vacuum drying, therefore, becomes equivalent to prediction of its air permeability. Air permeability depends both on the material(s) of the yarn and on the structural parameters of the fabric. In general, the relationship between the air permeability and the structural parameters of the fabric is complex. Theoretically, the calculation of the air permeability is based on the calculation of the porosity of the fabric. As it is explained in the next section, however, there exist known difficulties in determining basic porosity calculation parameters, such as the shape factor, thus restricting its use in practice. Indeed, if the weaving pattern of the fabric is exceptionally loose, the porosity can be estimated from

Air permeability of woven fabrics

19

Air flow Weft

Warp

Figure 1. Airflow through a fabric

IJCST 19,1

20

microscopic images of the fabric, by measuring the area of the void space between the yarns. In the case of dense fabrics, however, which represent the most common case in practice, yarns are compressed and jammed. As a result, there does not exist a straightforward and standardized method for the measurement of the porosity. The task becomes even more complex if the fabric porosity has to be calculated rather than measured, in order to serve as a basis for the prediction of air permeability. In practice, the drawback of predicting the vacuum drying performance (water content remaining after pre-drying) of a fabric based on the air permeability values is exactly the need to know these values. If they are to be obtained by measurement, then this presupposes the construction of the fabric. In this case, however, there is no real need to model the efficiency of the vacuum drying; water content can be exactly measured, instead, in a small-scale experiment using fabric samples. Of practical interest is to predict the efficiency of the vacuum drying before the actual production of the fabric takes place, based on technical characteristics of the fabric during the design phase. This information would support a realistic and optimized planning of the production process. This is why prediction of the air permeability of a fabric before its construction becomes essential. In the present work, our aim is to obtain a reliable practical method for the prediction of the air permeability of the fabric during its design phase. Such a prediction will in turn yield an estimate of the behavior of the fabric during vacuum drying (remaining water content) and consequently of the energy consumption expected during the actual production of the fabric. To this end, we investigate both the porosity path and the structural parameters path to the prediction of air permeability of a given fabric type. Following the reasoning given in the next section, we adopt the later path. Furthermore, we investigate a linear and a nonlinear approach to establish a relation between air permeability and micro-structural parameters of the fabric. In the nonlinear context, we propose the use of an artificial neural network (ANN) of the generalized regression type, which is shown to yield satisfactory air permeability prediction values after a fairly modest training period. The network performance is verified on a set of real field measurement data. Air permeability of textile fabrics The fluid flow through textiles depends mainly on the shape arrangement and size distribution of the voids through which the fluid flows. Other critical factors affecting its permeability are the fabric thickness and the applied pressure. Fluid permeability is a complicated physical phenomenon due to the fibrous and non-even structure of the fabric. Thus, the precise definition of the shape characteristics via geometrical calculations is difficult. For that reason the existing research work on the air permeability is limited mainly on fabrics made of monofilament. In a similar sense, the yarns as basic structural elements of the fabrics are approached as wires (Backer, 1951; Rushton and Griffiths, 1971; Perry and Green, 1984; Gooijer et al., 2003a, b). The fact of the fluid flow between the fibers rise questions such as the resistance of entire fiber assemblies to fluid flow (Gooijer et al., 2003a, b; Leisen and Farber, 2004). Darcy’s law fluid describes the flow through a porous structure by a relationship between the pressure gradient along the flow path of a fluid and the average velocity of the fluid on a macroscopic scale. It mainly describes that the flow rate of fluids through porous structure is inversely proportional to the pressure gradient (Zhong et al., 2002):

B¼

mt n DP

ð1Þ

where B is the permeability of the material, m the viscosity of the fluid, t the thickness of the fabric, n is the volumetric velocity of the fluid and DP is the pressure drop across the fabric. The value of B depends on the nature of the porous media and the pore size. Since, the voids in a porous media affect the media permeability, the Kozeny-Carman equation was developed to provide a description of fluid flow based on the filtration properties. One form of this equation is (Zhu and Li, 2003): B¼

13 K 0 S 20 ð1 2 1Þ2 1

ð2Þ

where K0is the Kozeny constant, S0 is shape factor and 1 is the porosity. B is the permeability of the porous media. The shape factor is defined as following: S0 ¼

Surface area of solid phase Volume of solid phase

ð3Þ

Porosity is defined as the ratio of the projected areas of the pores to the total area of the material. It is also defined as the ratio of the void to the total volume (Hsieh, 1995): 1¼12

ra rb

ð4Þ

where ra is the fabric mass density (g/cm3) and rb is the fiber mass density (g/cm3). For the application of the above laws, it is necessary the definition of the shape factor. The non-uniform surface and microstructure as well as the deformability limit essentially the successful approach of the shape factor through geometrical calculations. In addition, the shape factor calculation does not take in account the voids between fibers and it is quite difficult to determine it for these fibrous structures. In the literature has never appeared any detailed study of the Kozeny constant for textile fabrics. In light of the above discussion, we hereafter abandon the porosity path to air permeability prediction and proceed to investigate the relation between the air permeability of the fabrics and its major structural parameters. Given the woven fabric construction process, three parameters are intuitively more attractive for our purposes, namely, the warp yarn density, the weft yarn density and the mass per unit area. Besides, their influence on the woven fabric properties, these parameters are advantageous in that their values either are predefined in the weaving process (warp and weft densities) or can be easily and accurately measured (mass per unit area). Our aim is therefore to establish a general relation between air permeability and these three parameters and then use it to predict air permeability values for given fabric types. As already mentioned in the Introduction section, this would be a result of great practical interest in the industrial fabric production context, because prediction of the air permeability and related fabric properties before mass production of the fabric would facilitate the design of fabrics of optimized performance.

Air permeability of woven fabrics

21

IJCST 19,1

22

Experimental set of data In order to form an experimental basis for the investigation of the permeability/structural parameters relation, a set of different samples of woven fabric have been produced. All fabric types used are made of 100 percent cotton yarns and they are of plain weave structure. The warp and weft yarn linear densities are Ne 40 and Ne 50, respectively. The warp and weft yarn densities of the samples vary in incremental steps. We have combined six different weft densities and five different warp densities to produce a set of 6 £ 5 ¼ 30 different fabric types in total. Subsequently, the air permeability of the above fabrics was measured under constant pressure drop. Let us note here that, rather than being directly measured, the air permeability is in fact estimated via the measurement of air velocity. Air permeability in (cm3/s/cm2) was measured via the standard BS5636 method, using the Shirley FX 3300-5 air permeability tester. For each one of the 30 fabric types constructed, the air permeability measurement was repeated across five different samples of the given fabric type, thus producing a total of 30 £ 5 ¼ 150 measurements. The five measurements of each fabric type were then averaged to produce an average air permeability value, in order to reduce non-systematic measurement errors in it. The structural fabric parameters, the respective air permeability values (averages across five samples) and the standard deviation of the measurements are given in the Table I. As it can be seen in Table I, air permeability values decreases when warp and weft yarn densities increase. This is a reasonable behavior because the dimensions of the pores through which air flows are getting smaller when moving from looser towards tighter fabric types, where resistance to the airflow is higher. Linear approach: multiple linear regression model Although the relation under investigation is clearly nonlinear, due to the complex dynamics involved in the underlying physical phenomena, a multiple linear regression is still worth as a first attempt, in order to: . arrive to a crude linear approximation of the relation sought; and . obtain a clue as to the relative importance of the three intuitively chosen parameters within this relation. The results of a multiple linear regression analysis of the data in Table I are given in Table II. Data in columns C1, C2, C3 contain weft densities, warp densities and mass per unit area data, respectively, while column C4 contains average air permeability values. In Table II, the p-value in the analysis of variance section is less than 102 3, showing that the regression per se is indeed significant, i.e. there does exist an influential relation between the dependent variable (air permeability) and the three independent variables selected here. The R 2 value is 84 percent (adjusted R 2 is 82.2 percent), showing that 84 percent of the variability in the air permeability data is “explained” by the linear combination of the specific three independent variables. This percentage is high enough to indicate that a linear relation could certainly be used as a crude approximation of the true relation and, at the same time, low enough to justify the investigation of nonlinear alternatives that might explain the remaining variability in the data.

Sample no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Warp yarn density (ends/cm)

Weft yarn density (picks/cm)

Mass per unit area (gr/m2)

Average of air permeability (cm3/s/cm2)

54 54 54 54 54 54 57 57 57 57 57 57 60 60 60 60 60 60 63 63 63 63 63 63 66 66 66 66 66 66

20 25 30 35 40 45 20 25 30 35 40 45 20 25 30 35 40 45 20 25 30 35 40 42 20 25 30 35 40 42

101.0 110.3 119.5 127.3 136.7 144.1 101.2 111.1 120.5 130.7 142.0 146.3 107.9 118.7 129.5 141.4 151.2 153.7 109.0 119.1 129.1 148.5 150.3 154.4 111.2 123.0 134.0 144.4 155.0 160.2

45.74 27.02 15.68 8.76 4.37 2.90 47.94 27.52 14.84 8.99 3.98 2.68 33.98 17.01 9.75 4.56 2.12 1.70 27.68 15.24 7.23 3.22 1.89 1.61 27.28 14.42 6.90 3.19 1.65 1.38

STD of air permeability 1.80 2.44 0.68 0.39 0.15 0.40 1.83 0.78 0.97 0.48 0.24 0.12 1.16 0.79 0.41 0.41 0.07 0.05 0.55 0.64 0.13 0.42 0.08 0.05 1.68 1.39 0.21 0.17 0.06 0.05

Linear approach: modified multiple linear regression model Under a closer examination, sequential sum-of-squares analysis of variance shows that variable C1 (weft density) alone explains a 3825.8/4285.9 or a 89.9 percent of the variability in the air permeability values, while variable C2 (warp density), if incorporated in the linear combination along with C1, contributes a further 430.6/4285.9 or a 9.9 percent. Further, incorporation of the variable C3 (mass per unit area) does not contribute significantly to the linear combination that already contains variables C1 and C2 (99.5/4285.9 or 0.02 percent only). This result prompts the modification of the multiple linear regression in order to exclude warp density from the set of the independent variables. The results of the modified multiple linear regression on the data of Table I are given in Table III. Data in columns C1, C2 contain weft densities and warp densities, respectively, while column C4 contains average air permeability values. In Table III, the p-value in the analysis of variance section remains less than 102 3, showing that this modified regression is still significant, even without warp density in the

Air permeability of woven fabrics

23

Table I. Structural parameters and corresponding air permeability measurements for 30 fabric types

IJCST 19,1

24

Table II. Results of multiple linear regression analysis on the data of Table I

The regression equation is C4 ¼ 105 2 0.679 C1 2 0.392 C2 2 0.355 C3 Predictor Coef SE coef T P Constant 105.04 16.36 6.42 0.000 C1 20.6793 0.7280 20.93 0.359 C2 20.3920 0.5694 20.69 0.497 C3 20.3551 0.3655 20.97 0.340 S ¼ 5.59416 R 2 ¼ 84.0 percent R 2 (adj) ¼ 82.2 percent Analysis of variance Source DF SS MS F Regression 3 4285.9 1428.6 45.65 Residual error 26 813.7 31.3 Total 29 5099.6 Source DF Seq SS C1 1 3825.8 C2 1 430.6 C3 1 29.5 Unusual observations Obs C1 C4 Fit SE fit 1 20.0 45.74 34.42 2.54 7 20.0 47.94 33.17 2.05

P 0.000

Residual 11.32 14.77

St resid 2.27R 2.84R

Note: R denotes an observation with a large standardized residual

model. The R 2 value is in the same level as before (83.5 or 82.2 percent adjusted), which means that the exclusion of warp densities did not impair the model. Note, however, that the p-values of the three regression coefficients in the Predictor section are now fairly lower than the corresponding p-values in Table II. This means that the modified regression with two independent variables yields a more reliable and robust model.

Table III. Results of modified multiple linear regression analysis on the data of Table I

The regression equation is C4 ¼ 111 2 1.38 C1-0.893 C2 Predictor Coef SE coef Constant 111.09 15.11 C1 21.3764 0.1233 C2 20.8933 0.2406 C3 20.3551 0.3655 S ¼ 5.58835 R 2 ¼ 83.5 percent R 2 (adj) ¼ 82.2 percent Analysis of variance Source DF SS Regression 2 4256.4 Residual error 27 843.2 Total 29 5099.6 Source DF Seq C1 1 3825.8 C2 1 430.6 C3 1 29.5 Unusual observations Obs C1 C4 1 20.0 45.74 7 20.0 47.94

T 7.35 211.16 23.71 20.97

P 0.000 0.000 0.001 0.340

MS 2128.2 31.2

F 68.15

P 0.000

SE fit 2.35 1.98

Residual 10.41 15.29

SS

Fit 35.33 32.65

Note: R denotes an observation with a large standardized residual

St resid 2.05R 2.93R

In either case, there still remains a 16 percent of the variability in the air permeability data that are not linearly explicable. Seeking an enhanced model, explicit or implicit, that will capture the nonlinear aspects of the relation under question, we propose here to investigate ANN as one of possible alternatives.

Air permeability of woven fabrics

Nonlinear approach: artificial neural networks ANNs are algorithmic structures derived from a simplified concept of the human brain structure. Their various types have already been successfully employed in a wide variety of application fields (Haykin, 1998). Their major functionalities are: . Function approximators. This functionality is exploited in system input-output modeling; and . Classifiers. This functionality is exploited in pattern recognition/classification problems (Lippman, 1987).

25

Under their function approximator form, ANNs have served as a powerful modeling tool, able to capture and represent almost any type of input-output relation, either linear or nonlinear (Majumdar, 2004). The shortcoming of such an ANN-based modeling solution is that the model is implicit. Indeed, rather than formulating an explicit input-output analytic expression, either linear or nonlinear, an ANN processes examples of inputs and outputs, to capture and store knowledge on the system. It can subsequently simulate the system, or predict output values; yet, it cannot offer a closed-form description of the system. Internally an ANN contains a number of nodes, called neurons, organized in layers and interconnected into a net-like structure. Neurons can perform parallel computation for data processing and knowledge representation (Brasquet and Le Cloirec, 2000). Weighted averaging, followed by linear or nonlinear (thresholding) operations, with the possibility of feedback between layers, constitutes the main processing operation in an ANN. The acquired knowledge is stored as the weight values of the nodes. The architecture of a single layer of nodes for a sample ANN is shown in Figure 2 (Matlab, 2005). There exist S nodes in this layer. In the i-th node is computed a linear combination of a vector of R inputs [p1, . . .pR] weighted by weights [wi1, . . . wiR] and a bias term bi is added. The scalar value produced undergoes a transformation by a generally nonlinear function f( · ) to yield the corresponding i-th output, for all i ¼ 1, 2, . . . ,S. Sigmoid, log-sigmoid, hard-limiter or even linear types of functions are employed, resulting in accordingly varying properties of the network. Generally, the ANN architecture is characterized by: . A large number of simple, neuron-like processing elements. . A large number of weighted connections between elements. The weights of these connections store the “knowledge” of the network. The adaptive adjustment of these weight values, while sliding down an error surface constitutes the “learning process” of the network. . Highly parallel and distributed control. . An emphasis on learning internal system representations automatically (Rich and Knight, 1991).

IJCST 19,1

Input

Layer of neurons

W1,1 P1

26 1

P2

n 1

Σ

ƒ

a 1

b 1 n 2

Σ

ƒ

a 2

P3 1 PR WS,R

Figure 2. A sample ANN architecture (a single layer of S nodes)

Σ

b 2 nS

ƒ

aS

bS 1 a = f (Wp+b)

An ANN model is specified by its topology, node characteristics and training rules. These rules specify an initial set of weights and indicate how weights should be adapted to improve the performance of the network by minimizing the error between actual and ideal outputs, when the network is presented with a set of known inputs. A variety of different network models have already been proposed and used in practical applications, such as the Perceptron, the Multilayer Perceptron, the Hopfield Network, the Carpenter/Grossberg classifier, the Back-propagation network, the Self-organizing Map, the Radial Basis Function Network, etc. (Ramesh et al., 1995; Lippman, 1987). In terms of functionality, an ANN undergoes a training phase, where a rich set of examples (pairs of input-output values), called the training set, is presented to the network. The weights in the nodes of the network are iteratively optimized by a gradient-type algorithm, so as to produce correct outputs for all inputs in the training set. Subsequently, the network enters the so-called test phase, where it is asked to produce outputs for unknown inputs presented to it. Decisions are made (i.e. outputs are produced) on the basis of the experience gained by the network over the training set. During the test phase, the network is expected to “generalize” successfully, i.e. to exploit the experience stored in its elements in order to make correct decisions on incoming data that are similar but not identical to the training set data. In the area of the textile and clothing technology often the phenomena are of complex character and nonlinear behavior. Some problems have been solved using ANN already since the decade of the 1990s (Stylios and Sotomi, 1996; Stylios and Parsons-Moore, 1993). Thus, the efficiency of the ANN for the treatment of complex problems in the textile and clothing sector has been already been controlled and found satisfactory.

Generalized regression neural networks Selection of the appropriate type of ANN depends on the peculiarities of the specific problem. Several types of ANNs are promising candidates for the problem under consideration. In fact, any neural network that can take on the form of a universal function approximator will do. These are networks whose output layer nodes operate linearly and can therefore produce practically any real value in the output – as opposed to the limited output values possible when nonlinear/thresholding nodes are employed in the output layer. The major network families in this class are: . Multilayer Perceptrons. . Radial Basis Function Networks (more specifically: GRNN) (Chen et al., 1991). . Elman Networks (Elman, 1990).

Air permeability of woven fabrics

27

Among the three competing classes we select GRNNs, because they are known to approximate arbitrarily closely any desired function with finite discontinuities, and although they require an increased number of nodes in their architecture, they can be designed in far less time than necessary for the design and training of, e.g. a multiplayer perceptron. A GRNN contains two layers of neurons, each consisting of N neurons, where N is the cardinality of the training set (number of the input – output pairs available). The first (hidden) layer consists of radial basis function (RBF) neurons while the second (output) layer consists of linear neurons of special structure. Owing to the later, the network can produce any real value in its output. Figure 3(a) shows a RBF neuron sample transfer function. An RBF neuron “fires” or produces an output of 1 when its input lies on or fairly close to its central coordinates. Owing to its symmetry and finite radial support (controlled by a spread parameter), it responds only to local inputs, i.e. those befalling within a neighborhood of controlled radius. Figure 3(b) shows a linear neuron sample transfer function. The linear layer of a GRNN is special in that its weights are set equal to the output values contained in the training set. When a value equal or close to 1 is produced by the first layer and propagated to the second layer, it produces – by means of an inner product operation – a significantly non-zero value at a certain output and practically zero values at the rest of the outputs. During the design phase of the GRNN, the centers of the RBF neurons in the first layer are set equal to the input vectors in the training set. During the test phase, when an unknown input close (similar) to one of the training set inputs appears, the corresponding RBF neuron (or neighboring set of neurons) fires a value equal or close to 1. The second layer performs an inner product between the vector of the first layer outputs and the vector of training set outputs. Training set outputs matched with a

a

+1

1.0

n

0.5 0.0

0 −0.833

+0.833 (a)

n

−1 (b)

Figure 3. A radial basis (a) and a linear; (b) neuron transfer function

IJCST 19,1

28

values close to 1 in this inner product contribute significantly to the output, while the contribution of the rest is negligible. This is how the GRNN generalizes from the training set to unknown inputs. The major shortcoming of a GRNN is that both its layers consist of a number of nodes equal to the vectors in the training set. Given the fact that the training set should be rich enough to be representative of the input space, this number may grow large for a given application. As a counter-balance, however, it has been proved that the GRNN can be designed to produce zero error for any given training set in a very short time. The main fabric construction parameters, warp and weft yarn densities and mass per unit area, are used as the input vector. Therefore, each input – output pair in the training set contains a vector of three inputs (warp density, weft density, mass per unit area) and a single output (air permeability). As mentioned in Experimental set of data section, we have averaged air permeability measurements across sets of five samples for each fabric type. Therefore, our data set consists of 30 independent combinations of values of the three independent variables (warp density, weft density and mass per unit area) along with the respective averaged air permeability per case. We intend to use 24 out of the 30 cases for training and the remaining six cases for testing. This means that the GRNN architecture will have 24 neurons in each of the two layers. Furthermore, in order to reduce the dependency of the results on a specific partition of the data into training and testing sets, we perform a five-ways cross validation test, i.e. we partition the data in five different ways and finally average the results across the five partitions. Figures 4 – 8 show the results from the five different experiments of the five-ways cross-validation. In each figure, the upper plot shows results where the training set is used as test set as well, while the lower plot shows results where the test set was distinct from the training set. Within each plot, stars (*) and circles (o) indicate measured and estimated air permeability values, respectively, while dots ( · ) indicate the ANN performance error (difference between measured and estimated air permeability values). As it can be seen in Figures 4–8, the proposed neural network approximates the air permeability values with a very low output error, regardless of the specific data partition into training and test sets. Error (sum-of-squares) values, tabulated in Table IV, are indeed very low. Moreover, the (appropriately scaled) average error produced by the ANN approach (33.68 *30/6 ¼ 168.4) compares favorably to the respective error produced by the multiple linear regression approach (813.7 or 843.2 for the modified regression). The later is a very encouraging result, as it justifies the extra effort required by the nonlinear approach. Indeed, in terms of the multiple linear regression, it can be claimed that the ANN structure proposed here leaves unexplained only the (168.4/5099.6 ¼ ) 3.3 percent of the total variability in the air permeability data, thus offering a five times better result than the respective 16 percent of the linear regression approach. An error pattern apparent in all lower plots of Figures 4 –8, associates looser fabric types (towards the left side of the horizontal axis) with error values higher than the respective error values for dense fabrics (towards the right side of the horizontal axis). For loose fabrics the pores (between the yarns) are bigger so the yarn mobility is higher, thus the pore dimensions become bigger because of the deformation during the airflow. On the contrary, dense fabrics have very small pores and high resistance to airflow and can preserve their compactness during airflow.

Air permeability of woven fabrics

True(*), RB approximation(o), error(.) (Test set = training set) Air permeability

60 40 20

29

0 −20

0

5

10 15 20 Fabric sample index (test set of 24) True(*), RB approximation(o), error(.) (Test set <> training set)

25

Air permeability

60 40 20 0 0

4 5 2 3 Fabric sample index (test set of 6)

1

6

7

Figure 4. Air permeability estimated (approximated) by the GRNN (partition No. 1)

True(*), RB approximation(o), error(.) (Test set = training set) Air permeability

60 40 20 0 −20

0

5

10 15 20 Fabric sample index (test set of 24) True(*), RB approximation(o), error(.) (Test set <> training set)

25

Air permeability

60 40 20 0 0

1

2 3 4 5 Fabric sample index (test set of 6)

6

7

Figure 5. Air permeability estimated (approximated) by the GRNN (partition No. 2)

IJCST 19,1 Air permeability

30

True(*), RB approximation(o), error(.) (Test set = training set) 60 40 20 0 −20

0

5

10 15 Fabric sample index (test set of 24)

20

25

True(*), RB approximation(o), error(.) (Test set <> training set)

Figure 6. Air permeability estimated (approximated) by the GRNN (partition No. 3)

Air permeability

60 40 20 0 0

1

2 3 4 5 Fabric sample index (test set of 6)

6

7

True(*), RB approximation(o), error(.) (Test set = training set) Air permeability

60 40 20 0 −20

0

5

10 15 20 Fabric sample index (test set of 24) True(*), RB approximation(o), error(.) (Test set <> training set)

25

Figure 7. Air permeability estimated (approximated) by the GRNN (partition No. 4)

Air permeability

60 40 20 0 0

1

2 3 4 5 Fabric sample index (test set of 6)

6

7

Air permeability of woven fabrics

True(*), RB approximation(o), error(.) (Test set = training set) Air permeability

60 40 20

31

0 −20

0

5

10 15 20 Fabric sample index (test set of 24) True(*), RB approximation(o), error(.) (Test set <> training set)

25

Air permeability

60 40 20 0 0

1

2 3 4 5 Fabric sample index (test set of 6)

6

7

Figure 8. Air permeability estimated (approximated) by the GRNN (partition No. 5)

Figure 9 shows the strong correlation between the measured and the predicted by the ANN air permeability values. Air permeability and vacuum-drying performance The next step in the process is to exploit air permeability values in order to predict vacuum drying performance (remaining water content) for a given fabric type. Although there certainly exists a relation between these two quantities, this is again of a complex form. Owing to this relation, however, the air permeability prediction error propagates to the water content prediction value. Since, the relation is not available in an analytic form, experimental vacuum-drying tests, where the remaining water content is measured after drying, are necessary in order to study the error propagation. The tests are performed in a laboratory-scaled vacuum drying unit that was designed and constructed for the purposes of this study. Pre-drying tests using the vacuum drying principle are performed at a constant pump power of 1.5 kW for all fabric samples. Each one of the five different samples of every one of the 30 fabric types presented in Section 3 undergoes the following processing: The fabric is first Data partition no. 1 2 3 4 5 Average error (sum-of-squares)

Error (sum-of-squares) 16.62 22.64 17.10 17.52 94.49 33.27

Table IV. Sum-of-squares error across five partitions and average

IJCST 19,1

Figure 9. Correlation between measured and ANN-predicted air permeability values

Measured air permeability

32

60 50 40 30 20 10 0 0.00

10.00

20.00

30.00

40.00

50.00

60.00

Modelled air permeability

immersed in a water bath and then dried using the vacuum drying procedure. The water content of the fabric type is extracted by averaging across the five sample water content measurements of this type, in order to reduce non-systematic measurement errors in the water content values. These averages for each fabric type are tabulated in Table V (second column). Corresponding averages of air permeability measurements from Table I are repeated in Table V for convenience (fourth column). Figure 10 shows graphically the correlation between the two sets of averaged measurements of the respective quantities, along with a fitted curve based on the least-squares principle. This experimental correlation indicates a structured relation of the two quantities. The fitted curve in Figure 10 shows that the relation is nonlinear, as a result of the complex physical phenomena involved in the process. Could this relation be put into an analytic form, it would offer a viable alternative for the estimation of the water content, thus bypassing the industrious and non-standard porosity calculation. For the purposes of the present study, rather than seeking to formulate an analytic relation, we adopt the experimental relation defined by the graphical fitted curve in Figure 10, as a mapping from air permeability to water content values. For each one of the 30 fabric types, we thus obtain a set of water content prediction values from the air permeability prediction values produced by the GRNN. These are given in Table V (third column). Figure 11 shows the experimental correlation between predicted (via the procedure described above) and measured water content values. The fitted line lies on the diagonal of the plane, showing that the relation between them is identity. Moreover, scatter around the diagonal is limited, showing that the air permeability prediction error is not magnified when propagating to the water content prediction values. These satisfactory experimental results allow us to argue that it is indeed both meaningful and viable to predict fabric performance during vacuum drying (water content) through air permeability prediction, on the basis of an ANN modeling tool trained on a few basic structural parameters of the fabrics, such as the warp and weft density and mass per unit area.

Sample no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Water content – measured (percentage remaining)

Water content – predicted (percentage remaining)

Air permeability (cm3/s/cm2)

53.65 45.65 40.34 38.21 35.66 32.35 50.53 44.91 40.93 37.49 33.62 32.33 50.19 45.68 39.31 36.00 33.80 32.04 47.89 44.52 40.20 35.72 33.40 32.31 47.77 43.59 37.66 37.09 35.69 29.49

51.83 47.96 42.44 38.36 35.26 34.20 51.93 47.99 42.15 38.53 34.95 34.02 49.97 43.09 38.93 35.39 33.55 33.24 47.89 42.21 37.29 34.45 33.39 33.17 47.83 41.74 37.05 34.41 33.21 32.99

45.74 27.02 15.68 8.76 4.37 2.90 47.94 27.52 14.84 8.99 3.98 2.68 33.98 17.01 9.75 4.56 2.12 1.70 27.68 15.24 7.23 3.22 1.89 1.61 27.28 14.42 6.90 3.19 1.65 1.37

Conclusions For woven textile fabrics of a certain composition, it is known that the efficiency of the vacuum-drying water extraction process is primarily affected by the air permeability of the fabric. In order to estimate the vacuum-extraction efficiency, we propose to predict air permeability of the fabric based on three basic structural parameters of the fabric, namely the warp density, weft density and mass per unit area. An ANN of the generalized regression type (GRNN) is proposed and employed to approximate (predict) air permeability values, when fed in the input with the structural parameter values of a given fabric type. The performance of the proposed neural network is tested on real field data, with very satisfactory results. Specifically, the neural network prediction error is shown to be five times lower than the corresponding error produced by a multiple linear regression applied on the same data. Furthermore, the proposed method has been employed for the prediction of the vacuum-drying process efficiency (water content remaining in the fabric after vacuum drying) with equally satisfactory results. The practical implication of both the above results is that the behavior of a specific fabric type in the vacuum dryer – and therefore the expected energy

Air permeability of woven fabrics

33

Table V. Water content measurements (averaged across five samples) and corresponding air permeability of the fabric types in Table I

IJCST 19,1

50 45

34 Figure 10. Experimental correlation between air permeability and water content after vacuum drying (averaged measurement values)

Air permeability

40 35 30 25 20 15 10 5 0 25

30

35

40 45 Water content (%)

50

55

60

Figure 11. Relation between measured and predicted values of water content (percent) after vacuum extraction

modelled data of water content after vacuum extraction

55 50 45 40 35 30 25 20 20

30 40 50 measured data of water content after vacuum extraction

60

consumption during this stage of the fabric production process – can be accurately predicted in the fabric design phase, before actual production takes place, thus allowing for optimized production planning. References Backer, S. (1951), “The relationship between the structural geometry of a textile fabric and its physical properties, part IV: intercise geometry and air permeability”, Textile Research Journal, Vol. 2, pp. 703-14. Brasquet, C. and Le Cloirec, P. (2000), “Pressure drop through textile fabrics-experimental data modeling using classical models and neural networks”, Chemical Engineering Science, Vol. 55, pp. 2767-78. Chen, S., Cowan, C.F.N. and Grant, P.M. (1991), “Orthogonal least-squares learning algorithm for radial basis function networks”, IEEE Transactions on Neural Networks, Vol. 2 No. 2, pp. 302-9. Elman, J.L. (1990), “Finding structure in time”, Cognitive Science, Vol. 14, pp. 179-211.

Gooijer, H., Warmoeskerken, M.M.C.G. and Wassink, J.G. (2003a), “Flow resistance of textile materials, part I: monofilament fabrics”, Textile Research Journal, Vol. 73 No. 5, pp. 437-43. Gooijer, H., Warmoeskerken, M.M.C.G. and Wassink, J.G. (2003b), “Flow resistance of textile materials, part II: multifilament fabrics”, Textile Research Journal, Vol. 73 No. 6, pp. 480-4. Haykin, S. (1998), Neural Networks. A Comprehensive Foundation, 2nd ed., Prentice-Hall, Englewood Cliffs, NJ. Hsieh, Y.L. (1995), “Liquid transport in fabric structures”, Textile Research Journal, Vol. 65 No. 5, pp. 299-307. Leisen, J. and Farber, P. (2004), “Micro-flow in textiles”, NTC Project F04-GT05, National Textile Center Annual Report, available at: www.ptfe.gatech.edu/faculty/leisen/F04-GT05.html (accessed 14 December 2004). Lippman, R.P. (1987), “An introduction to computing with neural nets”, IEEE ASSP Magazine, pp. 4-22. Majumdar, P.K. (2004), “Predicting the breaking elongation of ring spun cotton yarns using mathematical, statistical and artificial neural network models”, Textile Research Journal, Vol. 74 No. 7, pp. 652-5. Matlab (2005), MATLAB 7 R14, Neural Network Toolbox User’s Guide, The MathWorks Inc., Natick, MA. Perry, R.H. and Green, D.H. (1984), Perry’s Chemical Engineering Handbook, 6th ed., McGraw-Hill, New York, NY, pp. 5-16. Ramesh, M.C., Rjamanickam, R. and Jayaraman, S. (1995), “The prediction of yarn tensile properties by using artificial neural networks”, Journal of Textile Institute, Vol. 86 No. 3, pp. 459-69. Rich, E. and Knight, K. (1991), Artificial Intelligence, McGraw-Hill, New York, NY, pp. 487-524. Rushton, A. and Griffiths, P. (1971), “Fluid flow in monofilament filter media”, Trans. Inst. Chem. Eng., Vol. 49, pp. 49-59. Stylios, G. and Parsons-Moore, R. (1993), “Seam pucker predictions using neural computing”, International Journal of Clothing Science and Technology, Vol. 5 No. 5, pp. 24-7. Stylios, G. and Sotomi, J.O. (1996), “Thinking sewing machines for intelligent garment manufacture”, International Journal of Clothing Science and Technology, Vol. 8 Nos 1/2, pp. 44-55. Zhong, W., Ding, X. and Tang, Z.L. (2002), “Analysis of fluid flow through fibrous structures”, Textile Research Journal, Vol. 72 No. 9, pp. 751-5. Zhu, Q. and Li, Y. (2003), “Effects of pore size distribution and fiber diameter on the coupled heat and moisture transfer in porous textiles”, International Journal of Heat and Mass Transfer, Vol. 46, pp. 5099-111. Further reading Keeler, J. (1992), “Vision of neural networks and fuzzy logic for prediction and optimization of manufacturing processes”, Applications of Artificial Neural Networks III, Vol. 1709, pp. 447-56. Corresponding author Savvas Vassiliadis can be contacted at: [email protected] To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Air permeability of woven fabrics

35

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,1

A model for the twist distribution in the slub-yarn

36

Key Laboratory of Science and Technology of Eco-textiles Ministry of Education, Southern Yangtze University, Wuxi, China

Yuzheng Lu, Weidong Gao and Hongbo Wang

Received May 2006 Accepted June 2006

Abstract Purpose – This paper seeks to build a mathematical model to deduct the twists distribution in the slub-yarn. Design/methodology/approach – The model, considering the shearing modulus and polar moment of inertia of the yarn, influenced by the yarn density, is established based on the bar torsion model. The twists distribution in the slub-yarn is concluded based on the analysis of the results. The rules were verified via an indirect method, by testing the breaking strength of the slub-yarn. Findings – The results of the analysis showed that twists in every section of the slub-yarn are in inverse proportion to the square of the line density of the corresponding section. Slub length is a key factor to the twist in the base yarn, and the increase of the slub length will increase the twist; while the multiple is the key factor to the slub twist, and the enhancement of the multiple will decrease the twist significantly. Research limitations/implications – The analysis cannot be verified via a direct method. A new, more conceivable direct method should be utilized to test the result. Originality/value – The paper builds a base for the research of the mechanical properties of slub-yarn and gives directions to the slub-yarn production. Keywords Yarn, Dissipation factor, Shearing, Modelling Paper type Research paper

1. Introduction Slub-yarn is a kind of fancy yarn, whose slub appearance is gained via the variation of the yarn linear density during the spinning process, and because of its special appearance, has been widely used in a variety of garments. The mechanical properties of slub-yarns were analyzed based on mathematical statistics (Grabowska, 2001), but the distribution of the twists in the slub-yarn was not mentioned. In this paper, the twist distribution is deduced based on a mechanical model, which builds a base for the analysis of the mechanical properties of slub-yarns.

International Journal of Clothing Science and Technology Vol. 19 No. 1, 2007 pp. 36-42 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710717035

2. The mechanism of the uneven distribution of twists Taking slub-yarns produced in ring-spun machine for instance, just like the common yarns, twists are acquired through the rotation of the spindle, equaling to the ratio of the rotation speed of the spindle to the linear velocity of the front roller. Thus, without taking the transfer of twists into account, for the slub-yarn produced by the overfeeding of the back-mid roller method, twists in any part of the slub-yarn would be equal. (Both the rotation speed of the spindle and the linear velocity of the front roller are invariable.) Figure1 (1) shows this kind of uniform distribution of twists.

However, in actual twisting process, the diameter of the base yarn is smaller than that of the slub, thus the polar moment of inertia and torsional rigidity of the two parts have great differences. Suppose the twists of the two parts were equal, the torque in the slub would be stronger than that in the base yarn. While the yarn in a steady state requires that the torque in all cross sections of the slub-yarn must be equal, which yields the decrease of twists in the slub and the increase of twists in the base yarn, and the twists transfer result is well shown in Figure 1(a).

Twist distribution in the slub-yarn 37

3. The mathematical and mechanical model for the twist distribution The torsion of a subject is consistent with the following equation: Ml GI

ð1Þ

a M ¼ GI l

ð2Þ

a¼ For yarn, it can be expressed as: T¼

where a is the torque angle, l is the bar length, G is the shearing modulus, T is the twist of a yarn, M is the leverage of the yarn cross section and I is the polar moment of inertia. It has been proven that the yarn density dy increases linearly with the twist multiple (Barella, 1950), and dy affects the shearing modulus Gy and the polar moment of inertia Iy of the yarn, so before making a model for the twist distribution, it is necessary to analyze the impact of dy on Gy and Iy. Suppose the cross section of the yarn with uniformly distributed fibers is circular, the yarn density on certain cross section is dy, the fiber density is df, and the filling rate of the yarn w is defined as follows:

w¼

dy df

ð3Þ

3.1 The relationship between Gy and dy When there is a shearing load placed on the yarn, both the fibers and pores in the yarn will be deformed. Since, the pores cannot bear any load, all shearing load is acted on the fibers. If the shearing load is relatively small, the shape of the yarn would not change, thus the volume ratio of fibers to pores in the yarn will remain unchanged, and the pores as well as the fibers would have the same strain r. For yarn, the following expression can be acquired:

Figure 1. Twist distribution of slub-yarn

IJCST 19,1

38

Gy ¼

Q Ay r

ð4Þ

where Ay is the area of cross section of yarn, Q is the shearing load. The sum of the shearing load acted on the fibers is equal to Q. Suppose that every fiber has the same shearing modulus Gf, the following expression would be gained: Gf ¼

Q Q ¼ Af g wAy g

ð5Þ

where Af is the sum of the area of all the fibers in certain yarn cross section. From expressions (4) and (5), the following equation can be deduced: Gy ¼ wGf

ð6Þ

3.2 The relationship between Iy and dy The diameter of the yarn dy and f are the key factors to Iy. The area of the pores in yarn cannot be the effective integral area. For the uniform distribution of fibers in the yarn, before calculating the integral, the area differential coefficient dAy should be multiplied by w (Figure 2). Z Z p I y ¼ r 2 dAy w ¼ w r 2 dAy ¼ d4y w ð7Þ 32 Ay

Ay

The relationship between dy and dy can be expressed as: sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ Nt 4N t ¼ dy ¼ 2 1; 000pdy 1; 000pwdf

Figure 2. The polar moment of inertia of the yarn cross section

ð8Þ

where Nt is line density of the yarn (tex), and from expressions (7) and (8), the relationship between Iy and dy can be expressed as: Iy ¼

N 2t 2 £ 106 pwd2f

ð9Þ

39

3.3 Rules of the twists distribution From expressions (2), (6) and (9), the following expression can be deduced: N 2t Gf T ¼ M 2 £ 106 pd2f

Twist distribution in the slub-yarn

ð10Þ

For a certain kind of slub-yarn, Gf and df is a constant, and the torque M in any cross section along the yarn axis is equal. So the twist T in any cross section varies inversely as the square of line density of the yarn. TN 2t ¼ CðconstantÞ

ð11Þ

4. Impact of parameters of slub-yarn on twist distribution Slub-yarn is composed of two parts: base yarn and slubs. Its appearance is influenced by the length and linear density of each constitute part (Testore and Minero, 1988; Wang and Huang, 2002). Since, the twist is inversely proportional to the square of linear density, twists should be equal as long as having the same linear density; moreover, the increased linear density leads to the decreased twists. Figure 3 showed the appearance of a regular slub-yarn in a cycle. On the assumption that the slub-yarn is periodic, and according to conservation rule of twists, in a cycle, the twist number would be the product of the cycle length and the design twist, in which the design twist is determined by the spinning machine. For ring-spun machine, the design twist is related to the configuration of the twist gear. Suppose the transition section is not present, the relationship mentioned above can be expressed as follows: T 0l0 þ T 1l1 þ T 2l2 þ · · · þ T nln ¼ T

n X

li

ð12Þ

i¼0

T 0 N 2t0 ¼ T 1 N 2t1 ¼ T 2 N 2t2 ¼ · · · ¼ T n N 2tn

ð13Þ

where T is the design twist, T0 is the twist of the base yarn, l0 is the sum of all base yarn length in a cycle, Nt0 is the linear density of the base yarn, n is the number of slubs with

Figure 3. The appearance of a slub-yarn in a cycle

IJCST 19,1

different linear density, Nti (i ¼ 1 to n) is the linear density of the slub, and Ti (i ¼ 1 to n) is the twist of the slub with the linear density Nti, li (i ¼ 1 to n) is the sum of the length of slub with the linear density Nti in a cycle. All parameters mentioned above, except Ti (i ¼ 0 to n), can be controlled in the manufacturing processing, so from expressions (12) and (13), the following expression can be obtained:

40 T

n X

lj

j¼0

Ti ¼ N 2ti

n X

ðl j Þ

N 2tj

ð14Þ

j¼0

Because Nt0 is the smallest in Nti (i ¼ 0 to n), twist of the base yarn T0 is always larger than T, for the thickest slub in yarn with the largest linear density Nti, Ti must be smaller than T. When the slub-yarn is not periodic, a long section of the yarn can be chosen, turning the nonperiodic slub-yarn into periodic one, and then expression (14) can be used to analyze the twist of the chosen section of the slub-yarn. In order to simplify the analyses, the regular slub-yarn only with one slub in its cycle is supposed for analysis. Then expression (14) can be rewritten as following: T0 ¼

Tðl 0 þ l 1 Þ 1 þ l 1 =l 0 ð1 þ RÞ ¼T 2 ¼ T N 1 þ XR2 1 þ ll 10 N t20 N 2t0 Nl 02 þ Nl 12 t0

T1 ¼

t1

t1

Tðl 0 þ l 1 Þ 1 þ l 1 =l 0 1þR ¼T 2 ¼ T 2 1þ R N t X l 2 l0 l1 0 2 1 X 1 þ N t1 N 2 þ N 2 l1 N 2 t0

t1

ð15Þ

ð16Þ

t0

where R ¼ l 1 =l 0 ; X ¼ ðN t1 =N t0 Þ . 1; and X is the multiple. Expression (15) indicates that R is a key factor to T0 while X affects T0 slightly, and T0 increases significantly with the increase of R. Expression (16) reveals that X is a key factor to T1 while R rarely has impact on T1, and the increased X leads to the significant decrease of T1. In practical production, R reflects the length of the slub, thus the control of R can avoid T0 exceeding the critical twist. While the control of X can adjust T1, in case that twists of the slub is too low to make the yarn failure. 5. Experiment verification The model we built up is for analyzing the twist distribution in slub-yarn. To verify its veracity, the most direct method has been used to measure the twists of base yarns and the slub, respectively, however, because the boundary between base yarn and slubs is not easy to identify, the untwist-retwist method cannot obtain the twists of each part correctly. As the twist affects the strength of the yarn directly, an indirect method was employed to verify the twist distribution by, respectively, testing the breaking strength

of the slub-yarn and the corresponding common yarn. The testing results are detailed in Tables I and II. All the tested yarns are spun from the same roving. Results in Tables I and II clearly show that breaking strength of the slub-yarn is stronger than that of the common yarn with the same twist design. It is believed that basic yarns in slub-yarn acquire more twists than the common yarn, and before reaching the critical twists, the increased twists lead to the high breaking strength. The breaking strength of the slub yarn is nearly the same as that of the common yarn with the same calculated twist. It is conceivable that the breaking strength of the slub would be stronger than that of the base yarn in such specific processing conditions, leading to the base yarn becoming the weak point of the yarn. Hence, it is proved that twists of base yarn are approximately in line with the calculated twists. Since, the slub-yarns spun in experiments are regular, just with one slub in a cycle, therefore, according to conservation rules of twist, twists of the slubs should be equal to the calculated twists.

Twist distribution in the slub-yarn 41

6. Conclusion . Twists distribution of the slub-yarn can be deduced through the bar torsion model, and the experimental results showed that the calculated twists is so close to the test result. . Slub length is a key factor to the basic yarn twist, while the multiple mainly affects the twists of the slub. In a slub yarn, the basic yarn twist is always larger than the designing twist, and the twist of the slub with the largest linear density must be smaller than the designing twist. . It is necessary to adjust the slub length to avoid basic yarn twist exceeding critical twist. Increased length of the slub leads to the increased twists of the basic yarn. Multiple of the slub affects twists of the slub significantly, and the increased multiple results in the significant decrease in twists of the slub.

Length of Length of Designing twist the slub basic yarn (/10 cm) Multiple (cm) (cm) No. 1 2 3

No. 1 2 3 4

5 5 10

20 20 15

49.9 49.9 49.9

2 3 2

Calculated basic yarn twist (/10 cm)

Calculated slub twist (/10 cm)

58.7 60.7 71.2

14.7 6.7 17.8

Breaking strength CV (cN) percent 467.7 487.6 537.7

7.5 7.0 6.5

Designing twist (/10 cm)

Breaking strength (cN)

CV percent

49.9 58.5 60.1 71.8

389.1 461.5 486.2 541.2

6.5 3.5 7.3 7.1

Table I. Breaking strength of slub yarns

Table II. Breaking strength of common yarns

IJCST 19,1

42

References Barella, A. (1950), “Law of critical yarn diameter and twist influence on yarn characteristics”, Textile Research Journal, Vol. 20 No. 2, pp. 249-58. Grabowska, K.E. (2001), “Characteristics of slub fancy yarn”, Fibres and Textile in Eastern Europe, No. 1, pp. 28-30. Testore, F. and Minero, G. (1988), “A study of the fundamental parameters of some fancy yarns”, Journal of Textile Institute, Vol. 79 No. 4, pp. 606-19. Wang, J. and Huang, X.B. (2002), “Parameters of rotor spun slub yarn”, Textile Research Journal, Vol. 72 No. 1, pp. 12-16. Corresponding author Weidong Gao can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Dimensional characteristics of core spun cotton-spandex 1 3 1 rib knitted fabrics in laundering C.N. Herath and Bok Choon Kang Department of Textile Engineering, Inha University, Incheon, South Korea

Characteristics of core spun cotton-spandex 43 Received May 2006 Accepted August 2006

Abstract Purpose – This paper aims to study the dimensional characteristics such as fabric density variations, dimensional constant parameters, linear and area dimensional changes and spirality angle variations of 1 £ 1 rib knitted structures made from cotton-spandex core spun yarns, under laundering regimes till 10th washing cycle. Design/methodology/approach – Samples of the above fabrics underwent dry, wet and full relaxation treatments and were subjected to standard atmospheric conditions prior to take the measurements. Washing was done in a front loading machine under normal agitation with machine 56 RPM. Each washing regime includes wash, rinse, spin, tumble dry steps. Washing temperature was set at 408C and water intake for washing was 30 l and rinsed with cold water. 0.1 g/l standard wetting agent was used. The mass of the load was maintained constant to 3 kg to keep the material ratio as 1:10. Washing regimes were continued till 10th cycle. Findings – Cotton-spandex rib structures came to a more stable state (minimum energy state) after 10th laundering cycle under the experimental conditions. Cotton did not come to such a state, even after 10th cycle proceeded. ANOVA analysis done under 95 percent confidential level has shown that fabric tightness and relaxation procedures give significant effect on dimensional characteristics of cotton-spandex and cotton rib structures. However, area shrinkage variations of cotton rib fabrics have shown an exception to this. Research limitations/implications – According to the dimensional constant values, evenafter 10th washing cycle, cotton rib structures did not come to a stable position. This should be further investigated to achieve a better stable rib knitted structure. Practical implications – The number of washing cycles can be increased or tumble dry duration can be increased to 120 min. to get a more stable state of cotton rib structures. Originality/value – The results are important for the knitting industry to predict the dimensional behavior of designed knitted fabric under relaxation. These data can be used to set the circular machine parameters to achieve a more stable fabric after laundering. Keywords Cotton, Dimensional measurement, Knitwear, Fabric testing Paper type Research paper

Introduction Geometrical dimensions of knit wear is quite easy to change during laundering and tumble drying (due to its, non balance condition), which results that knit wear will be unfit for use. Dimensional deformation of knit wear is influenced by fabric structure, raw material, structure tightness, etc. (Mikucioniene, 2004, 2001; Quaynor et al., 1999; Wool Science Review, 1971). This work was supported by Inha University Research Grant.

International Journal of Clothing Science and Technology Vol. 19 No. 1, 2007 pp. 43-58 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710717044

IJCST 19,1

44

During relaxation, internal stresses of knitted structures decrease and gradually moves to an stable balance condition with minimum internal energy (relaxation) and at the same time fabric density values are increased (structure consolidation) (Heap et al., 1983, 1985). In this event, two main forces affect to change their non balance state, as resiliency force of a knitted yarn and frictional forces acting on yarn interlocking points of a stitch (Mikucioniene, 2004; Quaynor et al., 1999). Therefore, to achieve lesser shrinkages during relaxation, researchers and manufacturers have been experimented to add certain amount of raw materials with good resiliency power to the basic raw materials. Thus, during laundering, extra mechanical energy is supplied to yarns in the structure through thorough agitation, in addition to the lubrication effect given by water to the yarn interlocking points of a stitch. Those factors result further shrinkages during laundering. Thus, the occurrence of loop jamming may effect the dimensions of relaxed knitted fabrics. It has been theoretically shown that 1 £ 1 rib structures may have more length jamming and however, in tighter structures both length and width jamming will occur or at least be closely approached than their medium and loose structures (Shanahan and Postle, 1973). During relaxation through laundering, 1 £ 1 rib fabrics show lesser shrinkages than plain fabrics and exhibit higher length shrinkages than their width shrinkages (Mikucioniene, 2004, 2001; Heap et al., 1983, 1985). However, 1 £ 1 rib needs at least 5 washing cycles and certain cases it was reported up to 10 cycles to get more stable shrinkages (Mikucioniene, 2004; Quaynor et al., 1999; Wool Science Review, 1971; Heap et al., 1985). In this experiment, 1 £ 1 rib knitted fabrics of core spun cotton-spandex yarns were subjected to full relaxation and washing regimes up to 10 cycles. Thus, in order to observe any effect with increasing cycle intervals, we increase the cycle interval to three between 5th and 8th washing cycles. Otherwise, cycle interval from 1st to 10th washing cycles were kept as 2. Investigated factors were course and rib density (per cm), stitch length, dimensional changes in length and width directions, area shrinkages and spirality angle variations. The results are compared with those for similar fabrics knitted from 100 percent cotton. Experimental Material About 100 percent Cotton (abbreviation: CO) and core spun cotton/spandex (abbreviation: CO/SP) with 93 percent cotton and 7 percent spandex were used to knit 1 £ 1 rib structures in a circular knitting machine in tight, medium and loose tightness factors. Spandex filaments with 40dtex were used in producing of core spun cotton/spandex yarns. Knitting details for CO and CO/SP rib fabrics are as follows: machine diameter – 30 inch; gauge – 24; RPM – 20; number of positive feeders used – 60; number of needles used-1680. Both material structure fabrics were untreated. Method Full relaxation Cut samples were dry relaxed (laid tension free on a flat surface under 218C ^ 1 at a RH of 65 percent ^ 2 for 48 h) and then, wetted out in a stainless steel water bath containing 0.1 g/l of standard wetting agent, maintaining a constant temperature of 388C for 24 h with minimum agitation. Then, washed thoroughly, briefly hydro

extracted for 1 min and tumble dried for 90 min at 708C. The samples were laid on a flat surface in a conditioning cabinet of 218C ^ 1 at a RH of 65 percent ^ 2 for 24 h with free of tensions. Washing and tumble drying Washing was done in a front loading machine under normal agitation with machine 56 RPM. Each washing regime includes: wash, rinse, spin, and tumble dry steps. Washing temperature set at 408C and water intake for washing was 30 liters and rinsed with cold water. 0.1 g/l standard wetting agent was used. The mass of the load was maintained constant to 3 kg to keep the material ratio as 1:10. Washing regimes was proceeded till 10th cycle. After each selected washing cycle, samples were tumble dried for 90 min. around 708C. Samples were then laid on a flat surface in a conditioning cabinet of 218C ^ 1 at a relative humidity of 65 percent ^ 2 for 24 h, before taking the measurements at the selected washing cycles. At every relaxation stage, measurements were taken as follows: SCSL(structural cell stitch length): length of yarn containing 25 ribs unraveled from the fabric sample was measured with “Zweigle” tester by putting 0.5cN/tex weight on the underside. This weight was good enough to remove crimp in the unraveled yarns. Five yarn lengths were measured from each sample and finally the average loop length for 30 yarn samples was calculated. Based on the SKC(structural knit cell) concept, SCSL values p were calculated; Tightness factor (TF) was calculated as machine tightness factor ( tex/[SCSL/nt]), where nt – number of needles in knitting operation in SKC. Spirality angle was measured according to the IWS 276 standard test method. For each average value of CPC, RPC, stitch length, linear shrinkages and spirality, 30 measurements were taken in each tightness factor. All other measurements were taken according to the ASTM standards. Results and discussions Course and rib density variations Course (CPC) and ribs density (RPC) of 1 £ 1 rib CO and core spun CO/SP fabrics from full relaxation to washing up to 10th cycle behave differently. Figures 1 and 2 show these variations with tight (H), medium (M) and low (L) tightness factors (TF). In 100 percent CO rib fabrics, RPC is gradually, but slowly, increases with full relaxation through wash treatments. However, after 5th cycle, RPC does not change significantly. Low and tight structures show some deviation, after 5th cycle, but, it is hardly to give special reason. In CO/SP rib fabrics, tight and medium structures show steep increase of RPC till 3rd wash cycle. After 8th cycle, RPC is not increase significantly. In low tight structures, steep RPC increasing can be seen till 2nd cycle and after that, gradually, increase its’ RPC. Therefore, it is clear that to become to a more complete relaxation state, at least 5th washing cycle is needed. However, 8th or 10th cycle is better for both 100 percent CO and CO/SP 1 £ 1 rib fabrics. Graphs for both CO and CO/SP between cycle intervals two and three do not show any special fluctuations. Therefore, It is hardly seen any special effect by increasing wash cycle intervals to three with 5th and 8th cycles. CPC of 100 percent CO ribs increase till 5th washing cycle and thereafter it will become almost flat. But, CO/SP rib fabrics show the CPC increasing only till 3rd cycle and after that, curve takes almost flat shape. Thus, tighter structures of CO/SP are showed the flat

Characteristics of core spun cotton-spandex 45

IJCST 19,1

H-TF-CO/SP M-TF-CO/SP L-TF-CO/SP H-TF-CO M-TF-CO L-TF-CO/SP

16

15

46 RPC

14

13

12

Figure 1. RPC changes of CO and CO/SP fabrics

11 Full

W1

W3 W5 Relaxation treatment

W8

W10

34 32 30 H-TF-CO/SP M-TF-CO/SP L-TF-CO/SP H-TF-CO M-TF-CO L-TF-CO

CPC

28 26 24 22 20 18

Figure 2. CPC variations of CO and CO/SP fabrics

16 Full

W1

W3 W5 Relaxation treatment

W8

W10

level of CPC after 1st washing cycle. 100 percent CO rib fabrics take at least 5th cycles to come to a stable state and that for CO/SP rib fabrics is 3rd washing cycles. There is no special fluctuations between 5th and 8th wash cycles compared to other wash cycles with cycle interval two. Therefore, it is possible to say that no effect is given by increasing wash cycle interval from two to three on CPC changes.

To find the behavior of CPC and RPC with their stitch lengths, we plotted the following graphs to establish mathematical models. Figures 3 and 4 show these behavior. Reference state were taken as full relaxation and after 10th washing cycle. After full relaxation and 10th washing cycle, both RPC and CPC show a linear geometrical correlation with an intercepts to loop length 2 1. Therefore, we established following mathematical model to describe the behavior of them in full relaxation and after 10th washing cycle. R ¼ a0 þ a1/L and C ¼ a2 þ a3/L; where R-RPC; C-CPC; L-loop length; a0, a1, a2, a3-constants Following Tables I and II gives the values for a0, a1, a2 and a3 constants. Based on these loop geometrical data, we calculated the dimensional constants, which can be used to design a dimensionally stable fabric.

Characteristics of core spun cotton-spandex 47

Dimensional constants (K-values) To calculate dimensional constant values for 1 £ 1 rib structures, accepted following mathematical models associated with structural knitted cell (SKC) concept given in scheme 1 was used. Scheme 1. Dimensional constant calculation formulae CU ¼

KC ; lU

WU ¼

KW ; lU

KP ¼

KC KW

K S ¼ C U xW U xl 2U ¼ K C xK W where CU - course units per fabric length (considered 1 cm),WU - Ribs per fabric width (considered 1 cm), lU - average total length of yarn in the SKC (SCSL). Kc, Kw, Ks and Kp (poisson’s function or loop shape factor) are non dimensional parameters. 16 Full-CO/SP Wash10-CO/SP Full-CO Wash10-CO

14 12

RPC

10 8 6 4 2 0 0

1

3 2 1/Loop length

4

5

Figure 3. RPC changes with loop length2 1 (cm2 1)

IJCST 19,1

40 Full-CO/SP Wash10-CO/SP Full-CO Wash10-CO

35 30

48 CPC

25 20 15 10 5

Figure 4. CPC changes with loop length2 1 (cm2 1)

Table I. Constant values for CPC and RPC models for CO/SP and CO rib fabrics

0 0

Material

Relaxation stage

CO/SP

Full relaxation After 10th wash cycle Full relaxation After 10th wash cycle

CO

2 3 1/Loop length

1

4

5

a0

a1

R 2*

a2

a3

R 2*

4.22 1.87 6.27 2.14

2.31 2.83 1.55 2.56

0.934 0.988 0.999 0.972

1.61 1.55 24.89 21.95

6.89 7.22 5.96 5.36

0.985 0.969 0.979 0.977

Note: R 2 * – regression correlation coefficient

Tables II and III give the calculated K-values for CO/SP and CO 1 £ 1 rib fabrics. Thus, according to the calculation of CV percent of Kp, in CO/SP rib structures, at full relaxation, Kp values has CV percent of 2.73 and but, after 3rd cycle, at 5th and 8th cycles, CV percent are around 0.6-0.75. Then, at 10th cycle, CV percent of Kp reduces up to 0.50. At full relaxation of 100 percent CO shows a CV percent of 5.43 and it varies within the range of 5-5.42 values till 5th cycle. Then, it increases and at 10th cycle, it reaches up to 6.29 percent. Thus, there is no significant effect by increasing cycle interval by 3 between 5th and 8th cycles. Therefore, we assume that CO/SP rib structures became more stable (relaxed) state than CO rib fabrics. Then, relaxation (changing loop shape factor) and consolidation (by increasing CPC and RPC during laundering) of the rib structure occurred at the same time as given in some other literature (Heap et al., 1983, 1985). Thus, CO and CO/SP rib fabrics give different values during full relaxation as well as in further relaxation during washing. It is clear that after full relaxation, both material structures are undergone to further relaxation states – complete relaxation (Quaynor et al., 1999). Almost all the K-values of CO and CO/SP rib structures are increased at 10th washing cycle. Thus, CO/SP rib fabrics show

TF

Ks

Kc

Kw

Kp

L M H L M H L M H L M H L M H L M H

100.53 94.17 88.72 106.15 98.67 93.65 107.63 100.93 93.36 106.68 99.76 93.05 107.03 99.64 90.40 106.86 99.43 90.45

14.68 14.41 14.19 15.50 14.89 14.79 15.7 15.17 14.69 15.61 15.10 14.66 15.67 15.08 14.46 15.64 15.06 14.45

6.84 6.53 6.25 6.84 6.62 6.33 6.87 6.65 6.35 6.83 6.61 6.34 6.82 6.61 6.25 6.83 6.60 6.26

2.14 2.21 2.26 2.26 2.24 2.33 2.28 2.28 2.31 2.28 2.28 2.31 2.29 2.28 2.31 2.28 2.28 2.30

Full relax Wash cycle 1

49 Wash cycle 3 Wash cycle 5 Wash cycle 8 Wash cycle 10

Notes: TF- tightness factor (tex1/2 cm2 1); L - low TF; M - medium TF and H - high TF

TF

Ks

Kc

Kw

Kp

L M H L M H L M H L M H L M H L M H

59.76 60.21 57.86 60.64 59.60 58.38 60.52 61.00 58.79 60.70 61.81 58.80 60.96 61.93 58.79 61.02 61.97 58.83

9.23 9.59 9.57 9.45 9.56 9.75 9.47 9.79 9.84 9.50 9.90 9.78 9.53 9.90 9.96 9.53 9.92 9.98

6.47 6.27 6.04 6.41 6.24 5.99 6.38 6.23 5.97 6.38 6.24 5.93 6.39 6.25 5.90 6.40 6.24 5.89

1.42 1.53 1.58 1.47 1.53 1.63 1.48 1.57 1.65 1.49 1.58 1.65 1.49 1.58 1.68 1.49 1.59 1.69

Characteristics of core spun cotton-spandex

Table II. K-values for CO/SP rib fabrics

Full relax Wash cycle 1 Wash cycle 3 Wash cycle 5 Wash cycle 8 Wash cycle 10

Notes: TF- tightness factor (tex1/2 cm2 1); L - low TF; M - medium TF and H - high TF

higher K-values than that of 100 percent CO rib fabrics, due to its higher CPC, WPC values and higher resiliency power of spandex component in the yarn core. Therefore, material with good resiliency power can give a positive effect on relaxation and consolidation of knitted structures (Mikucioniene, 2004; Quaynor et al., 1999).

Table III. K-values for 100 percent CO rib fabrics

IJCST 19,1

50

Liner dimensional changes Linear dimensional changes in length and width wise in cotton/spandex and 100 percent cotton 1 £ 1 rib fabrics were measured under full relaxation and laundering treatments. These linear shrinkage percentages for each sample were calculated by following formula. Shrinkage percentage ¼ [(L0 2 L1)/L0] £ 100; where L0 – the distance between data lines, before relaxation treatment and L1- same distance, after relaxation. Figures 5-8 show results of CO and CO/SP fabrics during full relaxation laundering. Both rib material structures show higher length shrinkages than their width shrinkages (Quaynor et al., 1999; Heap et al., 1983, 1985; Anand et al., 2002). After 10th wash cycle, CO/SP rib fabrics give higher length shrinkages (around 30-38 percent) compared to their width shrinkages (around 6-13 percent), whereas CO rib fabrics also show higher length shrinkages (around 15-21 percent) with compared to their width shrinkages (around 9-15 percent). Therefore, we can summarize than CO/SP rib fabrics show higher length and lower width shrinkages than CO rib fabrics. Reason for high length shrinkages of CO/SP rib fabrics may be due to good power of resiliency of spandex core (Mikucioniene, 2004; Quaynor et al., 1999). Then, length wise loop jamming also increased and resulted higher length shrinkages (Shanahan and Postle, 1973). The above mentioned length and width shrinkage behavior of CO/SP and CO rib fabrics can be justified by course and wale spacing values given in Table IV. In CO and CO/SP rib fabrics, length shrinkages reduce and width shrinkages increase, while increasing their structure tightness, at each relaxation stage (an anisotropic behavior with tightness factor). Thus, CO and CO/SP rib fabrics with medium and loose 40 TF-H

TF-M

TF-L

Length shrinkages (%)

35

30

25

20

Figure 5. Length shrinkages of CO/SP rib fabrics

15

Full

W1

W3 W5 Relaxation treatment

W8

W10

Note: TF- tightness factor (tex1/2 cm–1); L-low TF; M-medium TF and H-high TF

Characteristics of core spun cotton-spandex

16 TF-H

TF-M

TF-L

Width shrinkage (%)

12

51 8

4

0

Full

W1

W3 W5 Relaxation treatment

W8

W10

Note: TF- tightness factor (tex1/2 cm–1); L-low TF; M-medium TF and H-high TF

Figure 6. Width shrinkages of CO/SP rib fabrics

25 TF-H

TF-M

TF-L

Length shrinkages (%)

20

15

10

5

0

Full

W1

W3 W5 Relaxation treatment

W8

W10

Note: TF- tightness factor (tex1/2 cm–1); L-low TF; M-medium TF and H-high TF

Figure 7. Length shrinkages of CO rib fabrics

IJCST 19,1

16 TF-H

TF-M

TF-L

14 12 Width srinkage (%)

52

10 8 6 4 2 0

Figure 8. Width shrinkages of CO rib fabrics

Table IV. Density and spacing changes of ribs and courses of CO/SP and CO rib fabrics from full relaxation to 10th washing cycle

Full

W1

W3 W5 Relaxation treatment

W8

W10

Note: TF- tightness factor (tex1/2 cm–1); L-low TF; M-medium TF and H-high TF

Rib CO/SP CO

Spacing reduction (percent) Density increase (percent) Spacing reduction (percent) Density increase (percent)

3.29 3.39 1.18 1.16

H-TF Course 5.50 5.27 4.59 4.81

Rib 3.22 3.33 1.11 1.12

M-TF Course 6.47 6.92 4.80 5.04

Rib 3.04 3.14 0.99 0.93

L-TF Course 9.2 10.14 5.29 5.58

Notes: TF – tightness factor (tex1/2 cm2 1); L - low TF; M - medium TF and H - high TF

structure tightness do not show much difference in their length shrinkages (for CO, 4 percent average and for CO/SP, 2.8 percent average) compared to their tight fabrics (difference between tight and medium, for CO ,21.56 percent average and for CO/SP ,16.9 percent). When considering width shrinkages, CO/SP rib fabrics with loose and medium structure tightness have a difference of 27 percent average (higher than in the case of length shrinkages as mentioned above), but tight fabrics show more significant difference than this (the difference between tight-loose fabrics is ,80 percent average and between tight and medium is ,40 percent average). In the case of width shrinkages of CO rib fabrics, loose and medium fabrics show ,32 percent average difference and tight fabrics give more difference than this (between loose and tight fabrics ,56 percent average difference, and between medium and tight fabrics 18 percent average difference). Therefore, structure tightness gives significant influence on dimensional changes.

However, after full relaxation, laundering treatments give further dimensional changes due to attain to a more stable condition of rib stitches. At a glance, comparing of full relaxation and 10th cycle, for CO/SP rib fabrics, length shrinkages by 4.68, 5.4 and 7.2 percent and width shrinkages by 4.22, 21, 15.9 percent increases and for CO rib fabrics, length shrinkages by 14, 13, 7.6 percent and width shrinkages by 13, 12, 10 percent increases reported for loose, medium and tight fabrics, respectively. Table IV gives the course and wale density and their spacing changes of CO/SP and CO rib structures from full relaxation to 10th washing cycle. According to the data given in Table IV, following justifications for linear shrinkages can be made: . For CO/SP and CO rib fabrics, course spacing reduction percentages is higher than rib spacing reduction percentages. Therefore, length shrinkages are higher than width shrinkages and CO/SP show higher course spacing reduction, leading to higher length shrinkages than CO. . In CO and CO/SP material structures, fabrics with low tightness(L-TF) show higher course spacing reduction percentages, leading to higher length shrinkages in lengthwise compared to tight fabrics (H-TF). . For CO and CO/SP material structures, tight fabrics (H-TF) show higher rib spacing reduction percentages. This leads to higher width shrinkages in tight fabrics than loose fabrics (L-TF). However, most of literature suggest maximum allowable linear shrinkage for 1 £ 1 rib structure is length £ width: ^ 5 £ ^ 5 percent (most safer limits) (Knopten and Fong, 1970). However, our samples showed higher values than this. Area shrinkage variations To calculate area shrinkage (S), following formula was used: S ¼ {½LL0 £ LW 0 2 LL1 £ LW 1 =½LL0 £ LW 0 } £ 100 percent where LL0 and LW0 are the length and widthwise distance between data lines before relax treatment and LL1 and LW1 are the same distances, but after laundering treatments. Figures 9 and 10 give the variations of area shrinkages of CO and CO/SP rib fabrics during laundering treatments up to 10th cycle. CO/SP rib fabrics reported 38-43 percent average area shrinkages and CO rib fabrics showed 25-31 percent average area shrinkages (comparatively lesser). Thus, CO/SP rib tight fabrics show lower area shrinkages compared to medium and loose structures, because, its’ higher fabric density give restriction to change the area. Thus, comparatively lower length shrinkage of CO/SP H-TF fabrics may also affect to this result. CO/SP fabrics show 4.5 percent average reduction of area shrinkages, while structure tightness increases from loose to tight. However, CO rib fabrics show no significant difference in area shrinkages between structure tightness at each relaxation stage (from full relaxation to 10th cycle). Reason may be that their CPC and RPC are not much difference as CO/SP to give a significant area shrinkage difference. Average standard deviation of area shrinkage values of CO rib fabrics between three tightness factors is 0.72.

Characteristics of core spun cotton-spandex 53

IJCST 19,1

44 TF -H TF-H

TF -M TF-M

TF -L TF-L

54

Area shrinkage (%)

42

40

38

36

Figure 9. Area shrinkages of CO/SP rib fabrics

34 Full

W1

W3 W5 Relaxation treatment

W8

W10

Note: TF- tightness factor (tex1/2 cm–1); L-low TF; M-medium TF and H-high TF

35 TF-H

TF-M

TF-L

30

Area shrinkage (%)

25

20

15

10

5

Figure 10. Area shrinkages of CO rib fabrics

0 Full

W1

W3 W5 Relaxation treatment

W8

W10

Note: TF- tightness factor (tex1/2 cm–1); L-low TF; M-medium TF and H-high TF

At a glance, CO and CO/SP rib fabrics show further area shrinkages, after full relaxation towards 10th laundering cycle. Comparing full relaxation and 10th cycle, CO rib fabrics show 10.92, 16.57, 8 percent and CO/SP rib fabrics show 9.9, 7.55, 8.3 percent area shrinkage increases for loose, medium and tight fabrics, respectively. According to the literature, area shrinkages of most ensure machine washability of rib structures lie with in ^ 8 percent (Knopten and Fong, 1970). Similar to linear shrinkages, area shrinkages of our samples are also out of this limit.

Characteristics of core spun cotton-spandex 55

Spirality angle variations In theory, Spirality angle changes is restricted by higher structure tightness and the angle is increased during relaxation progress in laundering due to vigorous action of heat and tumbler action – wale distortion increased (Anand et al., 2002; De Araujo and Smith, 1989). Thus, spirality angle of rib structures show significantly lesser values and lesser changes than plain fabrics, due to balance structure of rib fabrics (Anand et al., 2002). Figures 11 and 12 show spirality angle variations up to 10th cycle for CO and CO/SP rib fabrics. It is clear that due to lesser tightness factor/lesser density of L-TF, wale distortion is higher than M-TF and H-TF. During wetting and vigorous tumbling action, these wale distortions further increase within washing regimes. Therefore, spirality angle increase during progressing of laundering and higher angles are shown with L-TF than M-TF and H-TF. Owing to higher CPC and RPC values of CO/SP rib structures (higher spacing reductions as in Table IV) results lesser spirality angles even after 10th washing cycle, compared to CO rib structures. Reason is that higher structure tightness impose higher opposition with providing less space to distort the wales during relaxation. 4 TF-H

TF-M

TF-L

Spirality angle (degrees)

3.5

3

2.5

2

1.5

1 Full

W1

W3 W5 W8 Relaxation treatment

W10

Figure 11. Spirality changes in CO rib fabrics

IJCST 19,1

3 TF-H

TF-M

TF-L

56

Spirality angle (degrees)

2.5

2

1.5

1

0.5

Figure 12. Spirality changes in CO/SP rib fabrics

0 Full

W1

W3 W5 Relaxation treatment

W8

W10

However, All spirality angle values of CO/SP and CO rib fabrics lie within the acceptable range (, 58). But, spirality angles of CO/SP lesser than that of CO rib fabrics. Reason for this characteristics of CO/SP rib fabric is the higher structure tightness, due to the resiliency power of spandex core. ANOVA analysis First, we have done a two factor ANOVA analysis for relaxation treatments from full relaxation to 10th cycle. By this, we observed the significance of effectiveness of tightness factor and relaxation progress. Table I and II gives the results for CO/SP and CO rib fabrics, respectively.. It has considered the significant level of 95 percent (a ¼ 0.025) for CO and CO/SP fabrics (Tables V and VI). From these two tables, for linear, area shrinkages (except for CO rib fabrics) and spirality angle changes, tightness factor and relaxation treatments effect significantly under 95 percent significant level. But, area shrinkage of CO has insignificantly effected by relaxation and structure tightness. Then, we did two factor ANOVA analysis for 3rd, 5th and 8th cycles to observe any effect by increasing washing cycle interval from two to three. But, there is no significant effect was noticed under 95 percent confidential level (a ¼ 0.025) for CO and CO/SP rib fabrics (Table VII). Conclusions Both CO and CO/SP rib fabrics show further shrinkage possibilities (consolidation), after their full relaxation. Same time, relaxation of knitted loops also take place. Higher

course and wale density values reported by CO/SP fabrics. CO and CO/SP structures have shown linear correlation to loop length 2 1 with an intercept. After 10th washing cycle, CO/SP fabrics show most stable state (by Kp values) than CO rib fabrics. Thus, K-values of CO/SP are significantly higher than CO rib fabrics. Higher length shrinkages and lower width shrinkages reported with CO/SP rib fabrics than CO rib fabrics. Tight fabrics show lower length shrinkage and higher width shrinkage for both CO and CO/SP rib fabrics. CO/SP fabrics show higher area shrinkages than CO fabrics. Their tight fabrics show lesser area shrinkages. But, CO/SP rib fabrics do not show significant area shrinkage differences with their structure tightness. CO/SP rib fabrics give lesser spiality angles (, 2.58) than CO fabrics (, 48). Tight fabrics show lesser spirality angle changes than loose and medium fabrics. Tightness factor and relaxation treatment affect on length-, width-, area- and spirality- changes of CO/SP and CO rib fabrics significantly, except area shrinkages of CO/SP rib fabrics. Thus, increasing washing cycle from two to three do not give any significant effect.

Dimensional characteristic changes

Effect of tightness factor

Effect of relaxation treatment

Length width Area Spirality

S S S S

S S S S

( p , 0.0001) ( p , 0.0001) ( p ¼ 0.002) ( p , 0.0001)

( p ¼ 0.004) ( p , 0.0001) ( p ¼ 0.0001) ( p ¼ 0.0001)

Dimensional characteristic changes

Effect of tightness factor

Effect of relaxation treatment

Length width Area Spirality

S ( p , 0.0001) S ( p , 0.0001) NS ( p ¼ 0.45) S ( p , 0.0001)

S ( p , 0.0001) S ( p , 0.0001) NS ( p ¼ 0.46) S ( p ¼ 0.0004)

Table VI. ANOVA for CO rib fabrics

Notes: S: significant (95 percent); NS: not significant (95 percent)

CO/SP NS p ¼ 0.50 CO NS p ¼ 0.25

Kc

Kw

Kp

NS p ¼ 0.56 NS p ¼ 0.33

NS p ¼ 0.36 NS P ¼ 0.99

NS p ¼ 0.42 NS p ¼ 0.42

Tested parameter Length Width Area shrinkage shrinkage shrinkage

Stitch density

Spirality

NS p ¼ 0.12 NS p ¼ 0.24

NS p ¼ 0.16 NS p ¼ 0.12

NS p ¼ 0.17 NS p ¼ 0.10

NS p ¼ 0.13 NS p ¼ 0.16

NS p ¼ 0.06 NS p ¼ 0.05

Notes: NS: not significant at 95 percent level; p: probability of existing

57

Table V. ANOVA for CO/SP rib fabrics

Notes: S: significant (95 percent); NS: not significant (95 percent)

Ks

Characteristics of core spun cotton-spandex

Table VII. ANOVA analysis for CO and CO/SP rib fabrics for 3rd, 5th and 8th

IJCST 19,1

58

References Anand, S.C., Brown, K.S.M., Higgins, L.G., Holmes, D.A., Hall, M.E. and Conrad, D. (2002), “Effect of laundering on the dimensional stability and distortion of knitted fabrics”, AUTEX Res. J., Vol. 2 No. 2, p. 85. De Araujo, M.D. and Smith, G.W. (1989), “Spirality of knitted fabrics, part 1: the nature of spirality”, Text. Res. J., Vol. 59 No. 5, p. 247. Heap, S.A., Greenwood, P.F., Leah, R.D., Eaton, J.T., Stevens, J.C. and Keher, P. (1983), “Prediction of finished weight and shrinkages of cotton knits- starfish project, part I: an introduction and general over view”, Text. Res. J., Vol. 53, p. 109. Heap, S.A., Greenwood, P.F., Leah, R.D., Eaton, J.T., Stevens, J.C. and Keher, P. (1985), “Prediction of finished weight and shrinkages of cotton knits- starfish project, part II: shrinkage and the reference state”, Text. Res. J., Vol. 55, p. 211. Knopten, J.J.F. and Fong, W. (1970), “Dimensional properties of knitted wool fabrics; part IV: 1 £ 1 rib and half-cardigen structures in machine-washing and tumbler drying”, Text. Res. J., Vol. 40 No. 12, p. 1095. Mikucioniene, D. (2001), “The reasons of shrinkage of cotton weft knitted fabrics”, Material Science, Vol. 7 No. 4, p. 304. Mikucioniene, D. (2004), “The dimensional change of used pure and compound cotton knit wear”, Material Science, Vol. 10 No. 1, p. 93. Quaynor, L., Nakajima, M. and Takashi, M. (1999), “Dimensional changes in knitted silk and cotton fabrics with laundering”, Text. Res. J., Vol. 69, p. 285. Shanahan, W.J. and Postle, R. (1973), “Jamming of knitted structures”, Text. Res. J., Vol. 43 No. 7, p. 532. Wool Science Review (1971), Vol. 41, p. 14. Corresponding author Bok Choon Kang can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Experimental analysis and modelling of textile transmission line for wearable applications Michel Chedid Saab Training Systems AB, Jo¨nko¨ping, Sweden, and

Ilja Belov and Peter Leisner School of Engineering, Jo¨nko¨ping, University, Jo¨nko¨ping, Sweden

Experimental analysis and modelling 59 Received July 2006 Revised August 2006 Accepted August 2006

Abstract Purpose – The paper seeks, by means of measurement and modelling, to evaluate frequency dependent per-unit-length parameters of conductive textile transmission line (CTTL) for wearable applications and to study deterioration of these parameters when CTTL is subjected to washing. Design/methodology/approach – The studied transmission line is made of Nickel/Copper (Ni/Cu) plated polyester ripstop fabric and is subjected to standard 608C cycle in a commercial off-the-shelf washing machine. The per-unit-length parameters (resistance and inductance) and characteristic impedance of the line are extracted from measurements before and after washing. Using the measurement data an equivalent circuit is created to model the degradation of the line. The circuit is then integrated in a three-dimensional transmission line matrix (TLM) model of the transmission line. Findings – Both an electrical equivalent circuit and a TLM model are developed describing the degradation of the conductive textile when washed. A severe deterioration of the electrical parameters of the line is noticed. Experimental and modelling results are in good agreement in the addressed frequency band. Research limitations/implications – Analysis is performed for frequencies up to 10 MHz. The developed TLM model can be used to conduct parametric studies of the CTTL. To counteract the degradation of the line, protective coating is to be considered in further studies. Originality/value – This paper extends knowledge of the subject by experimental and simulation-based characterization of the CTTL when subjected to washing cycles. Keywords Textiles, Electronics industry, Modelling, Impedance voltage Paper type Research paper

Introduction Both civilian and military markets of wearable applications are growing rapidly. Wearable electronics is designed to be body-worn while in use. It has to satisfy several design requirements: portability during operation, operation constancy, controllability, robustness and ergonomics (Chedid and Leisner, 2005). These requirements are crucial to guarantee good fitness of electronics to the human body. A wearable system has normally a modular architecture where different types of modules, e.g. sensors, radio and computer modules, are distributed on the body communicating to each other. A body area network based on conductive textile transmission lines (CTTL) that carries both direct current (DC) power and alternating current (AC) data signals, eliminates the The work was partly funded by the Swedish Knowledge Foundation through the Industrial Research School for Electronics Design, and Vinnova through the Summit Programme.

International Journal of Clothing Science and Technology Vol. 19 No. 1, 2007 pp. 59-71 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710717053

IJCST 19,1

60

need for traditional cables between the modules within the wearable system. This requires a low DC resistance to deliver power and known characteristic impedance in order to build matched transceivers. In order to get a wearable system to comply with the requirements on robustness and ergonomics, the conductive textile should exhibit the following characteristics: durability, washability and flexibility. Therefore, the CTTL for networking requires an analysis of the mechanical and electrical characteristics of such material and how it reacts to environmental stresses. Several studies have been done about conductive textile characteristics. In Cottet et al. (2003) conductive textile characteristics were studied and modelled for high frequency signal transmission. In Gorlick (1999) a separate power and data bus was developed where different modules could be connected to the bus through stingers. In Wade and Asada (2006) DC behaviour of conductive fabric networks was studied. In Jayoung et al. (2005) the effects of mechanical stresses on conductive textile were addressed at low frequencies. Robustness to washing induced stresses in combination with complex frequency dependent behaviour of the studied transmission line (TL) has not been investigated earlier. In this research the deterioration of the electrical properties of the CTTL is investigated by means of measurement of the per-unit-length parameters when CTTL is subjected to environmental stresses in form of washing cycles. Based on experiment, an equivalent circuit model is derived to predict the deterioration of the electrical properties. The circuit is then integrated in a three-dimensional transmission line matrix (TLM) model and the results are compared to experiment. Experimental setup In Figure 1, the flowchart of experimental and modelling steps is presented. The studied CTTL has a length of 1 m. It consists of two stripes of conductive textile with a low surface resistivity, sewed onto a non-conductive textile. Hardware modules are connected to the CTTL through conventional snap fasteners, see Figure 2. The conductive textile is a Nickel/Copper (Ni/Cu) polyester ripstop from Laird Technologies. Ripstop is made of very fine polyester filaments plainly woven with coarse fibres ribbed at intervals to stop tears, see Figure 3. To make it conductive, every filament is plated first with a thin layer of Copper followed by another layer of Nickel. The conductive layer has an approximate thickness of 1 mm. Electrical characteristics of the textile, taken from manufacturer’s datasheet, are shown in Table I. The snap fasteners are normally used in grounding products, e.g. electrostatic discharge (ESD) wrist straps. Before attaching the snaps to the textile, a highly conductive silicone based paste (volume resistivity: 0.01 V cm) is used to establish a reliable electrical connection between the snaps and the conductive textile. Silicone paste is chosen for its elastic properties compared to conductive silver epoxy adhesives, which were shown to be unreliable when exposed to washing. The width of the conductive stripes is 6 cm. They are folded to 2 cm before sewing onto the non-conductive textile which acts as a substrate, see Figure 3. The width of 6 cm is taken in order to achieve adequate DC resistance for power transmission. Taking the value of surface resistivity supplied by the manufacturer, the estimated value of the CTTL’s DC resistance is 2.3 V.

Experimental analysis and modelling

Conductive Textile TL 1-port Reflection Measurement Per Unit Length Parameter Extraction Equivalent Circuit Model

Conductive Textile Specification TLM Model Washing Cycle

61 Figure 1. Flowchart of the experimental and modelling steps

Figure 2. Experimental sample of the textile TL

Conductive textile

2cm

Nonconductive textile

Mag= 322x

100µm

EHT = 10.00 kv WD = 4mm

Signal A = SE2 Date : 26 Sep 2005 Photo No. = 37 Time : 12:09

Figure 3. Architecture of the CTTL (left) and a scanning electron microscope picture of the conductive textile (right)

IJCST 19,1

62

Using a network analyzer (NA) HP8714ET, measurements are performed on the CTTL stretched in the air 10 cm above a lab table. The experimental testing atmospheric conditions were measured to be 50 ^ 5 per cent for the relative humidity and 22 ^ 18C for the temperature. As a part of experimental study, the constructed CTTL is subjected to washing. A washing cycle consisted of a standard 150 min cycle in a commercial off-the-shelf washing machine (Elektro Helios TF 1002). The washing cycle included pre-washing at 408C, main washing at 608C, three rinsing sub-cycles and centrifuging. Washing was conducted without laundry detergent. After washing, the CTTL is dried at room temperature. Measurement procedure When designing a given communication system based on TL, the frequency dependent characteristic impedance of the TL determines the quality of communication. To ensure maximum power transfer between the transceivers, their input impedance must be matched to the TL characteristic impedance. The signal power at the receiver reaches a maximum when the transmitter, TL and receiver impedances are matched. A complete knowledge of the complex characteristic impedance of the TL is required in order to design an impedance matching network in the transceivers. Another reason for the measurement of the TL’s characteristics is to study the degradation of the conductive textile when it is washed. The parameters of main interest are R (resistance per-unit-length in V/m), L (inductance per-unit-length in H/m), G (conductance per-unit-length of the dielectric medium between conducting surfaces in S/m) and C (capacitance per-unit-length between conducting surfaces in F/m). Together, they form an RLGC model of the TL, see Figure 4. These parameters define propagation constant g and characteristic impedance ZC as equations 1 and 2, (Eisenstadt and Eo, 1992):

g¼

Table I. Ni/Cu polyester ripstop characteristics

Figure 4. RLGC model of a transmission line

pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ðR þ jvLÞðG þ jvCÞ

ð1Þ

Textile thickness (mm)

Filament thickness (mm)

Ni-Cu layer thickness (mm)

Surface resistivity of textile (Ohms/square)

Density of yarns (yarn/m)

Density of filaments (filament/yarn)

178

25

1

0.07

3,600

35

sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ R þ jvL ZC ¼ G þ jvC

ð2Þ

where v ¼ 2pf, and f is the frequency. In order to estimate unknown per-unit-length parameters of the CTTL, two reflection measurements are done for two different test setups: one end of the TL is always connected to the output port of the NA while the other end of the TL is short-circuit for the first measurement and open-circuit for the second measurement, see Figure 5. When a balanced TL is measured (as it is the case for the studied CTTL), it is necessary to connect a balance-to-unbalance balun (transformer) between the instrument and the TL. A balun is required for measuring balanced TL because the NA’s output terminal is unbalanced. Open/short/load calibration is done when the balun is connected to ensure accurate measurements. The measurements are done for the frequency range from 300 kHz to 10 MHz. The measured parameters, Gs and Go, are the voltage reflection coefficients for the short- and open-circuit measurement setup, respectively. Letting Z0 denote the NA’s impedance and l the length of the TL, ZC and g can be obtained from equations (3)-(6) (Agilent Technologies, 2002): 1 þ Go Z open ¼ Z 0 ð3Þ 1 2 Go Z short ¼ Z 0

1 þ Gs 1 2 Gs

Experimental analysis and modelling 63

ð4Þ

Z0 Textile Transmission Line

AC

Network Analyzer

Shortcircuit

Unbalancedto-balanced balun

Z0 Textile Transmission Line

AC

Network Analyzer

Unbalancedto-balanced balun

Opencircuit

Figure 5. Short- and open-circuit measurement setup for the textile TL

IJCST 19,1

ZC ¼

pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ Z open Z short

1 g ¼ tanh21 l

ð5Þ

sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ Z short Z open

ð6Þ

64 Combining equations (1), (2), (5) and (6), the four per-unit-length parameters can then be extracted as shown in equations (7)-(10) (Eisenstadt and Eo, 1992): R ¼ ReðZ C gÞ

ð7Þ

ImðZ C gÞ v

ð8Þ

L¼

g ZC

G ¼ Re

ð9Þ

g ZC

Im C¼

ð10Þ

v

where Re( ) and Im( ) denote the real part and imaginary part, respectively. The characteristic impedance can be approximated theoretically from the geometric configuration of the TL (Figure 6), (Gevorgian and Berg, 2001). Data provided in Table II gives a theoretical value of the characteristics impedance 194.7 V. Considering the fact that the CTTL is designed for installation on a human body, the TL will be bended in different directions and therefore the impedance will vary with time. However, studying the characteristics for the TL with a rigid geometry still gives a pointer on how conductive textile degrades when subjected to washing cycles. s w

Figure 6. Schematic of the CTTL cross-section

Table II. Dimensions of the CTTL

t h

Thickness of substrate (h) Width of conductive stripe (w) Actual thickness of conductive stripe (t) Separation between stripes (s) Relative permittivity of substrate (1r)

εr

0.5 mm 20 mm 0.45 mm 10 mm 3.6

Equivalent circuit of the CTTL Having performed the reflection measurements the resistance and inductance per-unit-length are extracted, see Figure 7. Rwc0( f ) and Lwc0( f ) are the extracted parameters from measurement before washing while effective parameters Rwc1( f ) and Lwc1( f ) are extracted from measurement made after washing. The large variation in the extracted inductance per-unit-length at low frequencies is due to variation from NA measurements. The capacitance per unit length (C) is found to be constant with a value of 22 pF/m. In addition, the conductance per unit length (G) at the frequencies considered approaches zero and is therefore neglected. The extracted data for Rwc0( f ) and Lwc0( f ) in Figure 7, is used to derive the RLGC model (Figure 4) of the CTTL before washing. By examining the measurement data for frequencies up to 10 MHz, Rwc0( f ) is fitted with pﬃﬃﬃ a linear function of frequency f and Lwc0( f ) is fitted with a nonlinear function of f . Since, the length of the studied TL is 1 m, the extracted parameters are considered as the per-unit-length parameters. The analytical expressions for per-unit-length parameters Rwc0( f ) and Lwc0( f ) are given in equations (11) and (12), respectively: Rwc0 ð f Þ ¼ 2:7 þ 3:3 · 1027 f ½V=m Lwc0 ð f Þ ¼ 656 þ

2:1 · 106 7:3 · 103 þ 2:5

Experimental analysis and modelling 65

ð11Þ

pﬃﬃﬃ ½nH=m f

ð12Þ

The CTTL changes characteristic behaviour due to washing. The new behaviour can be simulated by loading the RLGC model with an electrical equivalent circuit shown in Figure 8. This circuit consists of a resistor and two RC sub-circuits placed in series. 1200

200 Rwc1

L(nH/m)

R(Ohm/m)

180

0 washing cycle 1 washing cycle

1000

160 40 0 washing cycle 1 washing cycle

20

Lwc0

800

Lwc1

600 400 200

Rwc0 0

0 0

2

4

6

Frequency(MHz)

8

10

0

2

4

6

Frequency(MHz)

8

10

Figure 7. Extracted per-unit-length parameters Rwc0, Rwc1 (left) and Lwc0, Lwc1 (right) before and after washing

Figure 8. CTTL model after washing cycle (left) and a representative simplified circuit (right)

IJCST 19,1

66

The resistor R1 represents the fact that part of the conductive layer is washed out resulting in a reduction of effective conductivity and therefore higher resistance. The capacitances C2 and C3 represent the cracks that have built up within the conductive layer. These cracks force the current to take alternative paths and thus further increasing the resistance of the line (R2 þ R3). As the frequency increases the decreasing impedance of the capacitors C2 and C3 causes the current through R2 and R3 to decrease, respectively. At sufficiently high frequencies these RC sub-circuits are more or less shortcut and the only resistance seen is R þ R1. This explains the negative slope of the curve for parameter Rwc1 and the positive slope of the curve for parameter Lwc1 in Figure 7. The per-unit-length parameters corresponding to the original RLGC model are kept as they are to make it possible to load the line with the equivalent circuit in the subsequently created TLM model. The introduction of two RC sub-circuits proved to be sufficient to capture the CTTL’s behaviour when fitting the model to the measurement data. Furthermore, these two sub-circuits represent the fact that the displacement current through the various cracks is not attributed to a specific frequency. It rather increases within a frequency band. The loaded line can be represented as impedance X in series and admittance Y in parallel with the line conductors, see Figure 8. Admittance Y is not altered after washing due to the fact that the geometrical configuration of the CTTL and the material between the lines remain unchanged. Impedance X can be calculated from equation (13): 1 1 þ R3 = ð13Þ Xð f Þ ¼ Rð f Þ þ j2pfLð f Þ þ R1 þ R2 = j2pfC 2 j2pfC 3 The values of components R1, R2, R3, C2 and C3 can be extracted from the measurement by solving a system of nonlinear equation (14): 8 ReðXð f 1 ÞÞ ¼ Rwc1 ð f 1 Þ > > > > > > ReðXð f 2 ÞÞ ¼ Rwc1 ð f 2 Þ > > < ReðXð f 3 ÞÞ ¼ Rwc1 ð f 3 Þ ð14Þ > > > ImðXð f 1 ÞÞ ¼ Lwc1 ð f 1 Þ > > > > > : ImðXð f 3 ÞÞ ¼ Lwc1 ð f 3 Þ It is worth noting that parameters Rwc1 and Lwc1 do not stand for fundamental resistance and inductance, rather being the real part and the imaginary part of impedance X, respectively, with the imaginary part normalized to 2pf. Therefore, they are attributed as effective parameters. Taking f1 ¼ 1 MHz, f2 ¼ 5 MHz, f3 ¼ 10 MHz, the equation system is solved. The frequencies are chosen to capture boundary conditions and the dynamics in between. Using the values obtained from measurement results in the following values for the equivalent circuit in Figure 8: R1 ¼ 174.5 V R2 ¼ 5.8 V R3 ¼ 13.9 V C2 ¼ 17.1nF, C3 ¼ 1.1nF. TLM model of CTTL Constructing a TLM model requires beside the geometrical data, the electrical properties of materials used, i.e. conductivity, permittivity and permeability. Using the

method proposed in (Henn, 1998) to approximate surface resistivity of conductive textiles, the effective conductivity of the textile can be calculated as follows. Let df denote the diameter of the conductive filament and d the thickness of the Ni/Cu layer, then the area of the metal layer, Am, can be calculated according to equation (15): p 2 d f 2 ðdf 2 2dÞ2 Am ¼ ð15Þ 4 Every yarn is composed of n filaments. Let reff denote the effective resistivity of the Ni/Cu layer, then the resistance per-unit-length, r, of one yarn can be calculated according to equations (16) and (17): n 1 X 1 ¼ r reff =Am 1

r¼

reff nAm

ð16Þ

ð17Þ

Having the yarn density, N, which describes the number of yarn per-unit-length of textile, the surface resistivity of the textile, rs, can be approximated as in equation (18): " #21 N X 1 r reff rs ¼ ¼ ¼ ð18Þ r N nNA m 1 The effective conductivity, seff, can then be calculated according to equation (19): 1 1 seff ¼ ¼ ð19Þ reff nNAm r s Taking into account the values given in Table I, results in the effective conductivity of the studied conductive textile, 1.50 £ 106 S/m. A commercially available electromagnetic simulation tool, Flo/EMC from Flomerics is used in TLM simulations. The TLM tool solves Maxwell’s equations in time domain. The frequency domain data is then computed. The computational mesh is composed of cuboids of different size. Absorbing boundary conditions represent open faces of the computational domain. Convergence tests have been performed to determine the effect of the open boundary location on the simulation results. Convergence criterion is satisfied when initial system energy decays by at least 60 dB. The conductive textile in the CTTL is modelled as thin plates of isotropic dielectric material with fixed frequency independent relative permittivity and permeability and a fixed conductivity. These thin plates are penetrable to electromagnetic fields and the skin effect is simulated with acceptable accuracy. Wires are used to establish electrical connection between the two conducting plates. The NA is simulated by a wire port, including a voltage source and a 50 V impedance. For the wire port the TLM tool includes the wire current in the output, and associates with it the load impedance. A post-processing tool determines the signals entering and leaving the model through the port, resulting in the voltage reflection coefficient. In order to properly define thin plates in TLM simulation, their equivalent thickness teq has to be specified. It is calculated from the surface resistivity of the conductive

Experimental analysis and modelling 67

IJCST 19,1

textile provided by the manufacturer and from its effective conductivity seff. Every stripe in CTTL is made of three layers of conductive textile and therefore teq is calculated according to equation (20): teq ¼

68

Number of layers 3 ¼ 28:5 mm ¼ 1:5 £ 106 £ 0:07 seff r s

ð20Þ

Equation (21) (Chen and Fang, 2000) approximates the resistance of a rectangular cross-sectional conductor at a given frequency: s ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ﬃ 2 1 pf meff ð21Þ þ Rð f Þ ¼ 2 seff wt eq 4seff ðw þ t eq Þ2 where meff and w are the effective permeability and the width of the stripe, respectively. Although this formula takes only into account the skin effect of the TL conductors, it still gives an initial approximation of the permeability of the Ni/Cu layer. Using the measured value of resistance at 10 MHz, the effective relative permeability mr is predicted to be 350. In TLM simulations where other physical effects are included such as the proximity effect (Figure 9), mr is adjusted to 325 to get better agreement with measurement data. The input parameters used in the TLM simulations are summarized in Table III. As can be seen in Figure 9 the distribution of the surface current is frequency-dependent. Higher frequencies cause the surface current to flow on the innermost edges of the CTTL. This represents the proximity effect and together with the skin effect tends to cause the TL’s resistance to rise with increasing frequency. Modelling versus measurement Voltage reflection coefficients are obtained from measurements and from TLM simulations performed for the open- and short-circuit setups (Figure 5). Effective Surface Current (Amps/m) > 0.7 0.622 0.544 0.467 0.389 0.311

Figure 9. Surface current distribution in the CTTL TLM model at 300 kHz (left) and at 10 MHz (right)

0.233 0.156 0.0778 <0

Table III. TLM model input parameters

TL equivalent thickness (mm) 28.5

Effective conductivity (S/m)

Effective relative permeability of metallic layer

Relative permittivity of substrate

1.5 £ 10þ 6

325

3.6

per-unit-length parameters are then extracted according to equations (7)-(10). Results from measurement, equivalent circuit and TLM computation are shown in Figure 10. In Figure 10 (a), (c) and (e) the characteristic impedance and the extracted per-unit-length parameters (resistance and inductance) are presented for the TL before washing. Figure 10 (b), (d) and (f) presents the respective parameters after washing. Measured characteristic impedance ZC before and after washing is in good agreement with modelling results. ZC exhibits different frequency depending behaviour: in low-frequency region ZC is frequency dependent while in high-frequency region ZC is rather constant. After washing cycle, the high-frequency region is shifted up in

Experimental analysis and modelling 69

3000

340

Measurement TLM Equivalent circuit

320

Measurement TLM Equivalent circuit

2500

300 2000 |ZC|(Ohm)

|ZC|(Ohm)

280 260 240 220

1500 1000

200 500

180 160

0 0

2

4 6 Frequency (MHz)

8

10

0

2

4 6 Frequency (MHz)

(a)

10

(b) 202

8 Measurement TLM Equivalent circuit

7

Measurement TLM Equivalent circuit

200 198

6

196

5 R(Ohm/m)

R(Ohm/m)

8

4 3

194 192 190 188

2

186

1

184 182

0 0

2

4 6 Frequency (MHz)

8

0

10

4 6 Frequency (MHz)

(c)

1200

8

10

(d) 1200 Measurement TLM Equivalent circuit

1000

L(nH/m)

600

800 600

400

400

200

200 0

2

4 6 Frequency (MHz)

(e)

8

Measurement TLM Equivalent circuit

1000

800 L(nH/m)

2

10

0

2

4

6

Frequency (MHz)

(f)

8

10

Figure 10. Comparison between the measured and modelled characteristic impedance and per-unit length parameters for CTTL before and after washing

IJCST 19,1

70

frequency. Furthermore, the impedance plots reveal that the theoretically predicted value for the high frequency impedance (194.7 V) of the CTTL matches the experimental results when it is not subjected to any washing cycle. The resistance per-unit-length before and after washing is shown in Figure 10(c) and (d), respectively. The offset between the TLM model and measurement data before washing is due to the contact resistance that is not included in the TLM model. The curves however have the same slope. The resistance per-unit-length of the CTTL deteriorates dramatically after washing (from , 3 V/m to , 200 V/m at 300 kHz). The DC resistance of the TL is too high resulting in a TL that is no longer feasible for DC signals. Use of different types of protective top coatings should therefore be investigated to protect the conductive layer and enhance its robustness to washing. The inductance per-unit-length before and after washing is shown in Figure 10(e) and (f), respectively. The difference between the TLM model data and measurements seen in the results is due to the fact that permeability of the metallic layer is modelled as constant instead of being frequency dependent. The maximum error between measurement and TLM is bounded to approximately 15 per cent. The shape of the curves shows correlation between the TLM model and measurement results. There are several sources of uncertainty that contribute to measurement error. One major source is the ground plane that is not well defined in the measurement setup due to the fact that the CTTL is stretched 10 cm above the table. Another source of uncertainty is the snap buttons that are connected to the TL. The snap buttons add bias resistance to the measurement. While the error from the latter is small, the error from the former source is hard to predict and therefore sensitivity analysis based on TLM simulation is made and shown in Figure 11. The introduction of the ground plane leads to a relative change in extracted parameters of approximately 1 per cent, which is negligible. Conclusion Experimental and modelling-based study of frequency dependent (up to 10 MHz) electrical parameters of the TL made of Ni/Cu plated polyester ripstop fabric subjected to washing cycles led to the following conclusions: . Efficient combination of equivalent circuit modelling and TLM modelling was introduced in order to study deterioration of frequency dependent per-unit-length parameters of CTTL. The developed equivalent circuit was used to load the original TLM model of the non-washed TL and thus enabling the construction of

850

4

3.5

2

3.0 1 2.5

4

840 L(nH/m)

3

4.0

Relative change(%)

R(Ohm/m)

GP not included GP included Relative change

860

4.5

Figure 11. Sensitivity of simulation data to presence of ground plane

5

870

5 GP not included GP included Relative change

5.0

830

3

820 810

2

800 790

1

780

2.0

0 0

2

4

6

Frequency(MHz)

8

10

770

0 0

2

4

6

Frequency(MHz)

8

10

Relative change(%)

5.5

.

.

the TLM model of the washed Tl. The efficiency of such approach is confirmed by good agreement between measurement and simulation data. The developed approach can be of assistance during virtual prototyping in design phase of wearable applications. For example, computational experiments with different geometries that affect the characteristic impedance of the line can be conducted and thereby partly replacing real experiments. The studied conductive textile was found not to be appropriate in wearable application due to severe degradation of the Ni/Cu layer when washed. Different types of protective top coatings from the textile industry such as Polyurethane and Silicone coating should be investigated to improve robustness of the conductive textile to environmental stresses introduced by washing cycles.

References Agilent Technologies, Co. Ltd (2002), Agilent Technology Impedance Measurement Handbook, 2nd ed., Agilent Technologies, Co. Ltd, St Louis, MO. Chedid, M. and Leisner, P. (2005), “Roadmap: wearable computers”, Proceedings of IMAPS Nordic Conference, 11-14 September, Tønsberg, Norway, pp. 146-54. Chen, H. and Fang, J. (2000), “Modelling of impedance of rectangular cross-section conductors”, Proceedings of IEEE Conference on Electrical Performance of Electronic Packaging, 23-25 October, Scottsdale, USA, pp. 159-62. Cottet, D., Grzyb, J., Kirsten, T. and Tro¨ster, G. (2003), “Electrical characterization of textile transmission lines”, IEEE Transactions on Advanced Packaging, Vol. 26 No. 2, pp. 182-90. Eisenstadt, W.R. and Eo, Y. (1992), “S-parameter-based IC interconnect transmission line characterization”, IEEE Transactions on Components, Hybrids and Manufacturing Technology, Vol. 15, pp. 483-90. Gevorgian, S. and Berg, H. (2001), “Line capacitance and impedance of coplanar-strip waveguides on substrates with multiple dielectric layers”, Proceedings of European Microwave Conference, 25-27 September, London, UK, pp. 153-6. Gorlick, M. (1999), “Electric suspenders: a fabric power bus and data network for wearable digital devices”, Proceedings of the Third International Symposium on Wearable Computers, 18-19 October, San Francisco, CA, USA, pp. 114-21. Henn, A.R. (1998), “Calculating the surface resistivity of conductive fabric”, Interference Technology Magazine ITEM Update, available at: www.rbitem.com Jayoung, C., Jihye, M., Gilsoo, C. and Keesam, J. (2005), “An exploration of electrolessly Cu/Ni plated polyester fabrics as e-textiles”, Proceedings of the Ninth International Symposium on Wearable Computers, 18-21 October, Osaka, Japan, pp. 206-7. Wade, E. and Asada, H.H. (2006), “DC behavior of conductive fabric networks with application to wearable sensor nodes”, paper presented at International Workshop on Wearable and Implantable Body Sensor Networks, 3-5 April, Cambridge, pp. 27-30. Corresponding author Michel Chedid can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Experimental analysis and modelling 71

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,2

80 Received 15 September 2005 Revised 31 October 2006 Accepted 31 October 2006

Intelligent evaluation of fabrics’ extensibility from robotized tensile test Panagiotis N. Koustoumpardis, John S. Fourkiotis and Nikos A. Aspragathos Department of Mechanical and Aeronautics Engineering, University of Patras, Patras, Greece Abstract Purpose – The paper aims to propose an approach to intelligent evaluation of the tensile test. A robotized system is used that performs the fabrics tensile test and estimates the extensibility of the samples using a feed-forward neural network while trying to imitate the human expert estimation. Design/methodology/approach – The specifications of the tensile test are derived by an extensive observation of the respective experts’ estimation performance. The fabric sample size and the experimental conditions are specified. Linguistic values of the term “fabric extensibility” are extracted through a knowledge acquisition process. The tensile test is performed by a robot manipulator with a simple gripper and the experimental measurements (force, strain) are fed online into a neural network. The network is trained according to the extensibility estimations of the experts. The trained network is tested in estimating unknown fabric’s extensibility. Findings – The results demonstrate that the system is capable of estimating the extensibility of new fabrics. Originality/value – This work can be integrated in the robotized sewing process with intelligent control where the fabric’s extensibility in terms of linguistic values is necessary. The proposed system initiates a new approach, in which the fabric properties are expressed and used in a way that will facilitate the introduction of the artificial intelligence methods into the clothing industry. Keywords Fabric testing, Materials handling, Robotics, Neural nets Paper type Research paper

1. Introduction In the clothing industry, the experts traditionally estimate the fabric properties by hand actions, using their tactile feelings, which is a subjective measurement of fabric properties. Their evaluation is expressed through qualitative linguistic variables such as hardness, roughness, stiffness, etc. In the research community as well as in the clothing industry, there is a wide interest for objective measurements of the fabric properties. Therefore, a variant of testing methods and instruments have been developed. The fabric test instruments can be divided into two main categories: the “high-stress” and the “low-stress” instruments (Potluri and Porat, 1995; Saville, 1999). International Journal of Clothing Science and Technology Vol. 19 No. 2, 2007 pp. 80-98 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710725694

The authors would like to thank the knitting Greek industries J. Papafotis, S.A. and Sarris Jersey, S.A. for providing the most of the fabric samples. This work is carried out under the project “XPOMA-Handling of non-rigid materials with robots: application in robotic sewing” PENED 01 funded by the General Secretariat for Research and Technology of Greek Government. It is also integrated in the I *PROMS Network of Excellence.

The instruments based on high-stress tests are used to estimate the strength or the durability of a fabric, and usually it is a destructive test. The low-stress instruments are used for non-destructive tests by applying low stress close to the ones appeared in the normal use of a fabric, i.e. when it is handled or worn. The low-stress tests are used to estimate the fabric properties, which are essential for implementing an automated fabric handling system and/or are essential for the appearance of a cloth. The objective measurement of the fabric properties has been attracted the interest of researchers from the beginning of the previous century. Peirce (1930) presented the well-known cantilever-bending test, and a number of individual instruments for testing fabrics have been presented afterwards. In the clothing industry, these instruments have been occasionally used for determining the fabrics compression, tension, bending, shearing and frictional properties. The Fryma Fabric Extensiometer (Shirley Developments, 2003) is a representative tensile test instrument of this first-generation fabric test systems (Potluri and Porat, 1995), where the sample deformations are measured manually using rulers/micrometers while discrete loads are applied on the sample. These tests are performed by a human operator without any automation or control of the instruments or the entire experimental stage. Kawabata (1980) introduced the widespread Kawabata Evaluation System – KES available from Kato Tekko (Manuals, 1980), which is the first integrated system for low-stress fabric testing. The KES system can be considered as a real progress towards the objective measurements of fabric properties. In the KES-FB1 system, the shear and tensile tests are executed with the same instrument, where the sample is clamped among two chucks and the measured data are recorded automatically on an x-y plotter. Four parameters or tensile properties are extracted from the two curves representing the extension and the recovery fabric behavior, namely the linearity (LT), the tensile energy (WT), the tensile resilience (RT) and the extensibility (EMT). The first three of these parameters are used to estimate the primary hand values of the fabric and afterwards the total hand value of the fabric. The KES system is the most common low-stress testing system but it requires a considerable time to measure the fabric properties and it is composed of very high-cost instruments. A user friendly and less expensive system was presented by CSIRO Division of Wool Technology (1988). The developed set of instruments named SiroFAST (1994) was designed to meet the industrial needs for a simple and robust fabric test system, which is less flexible than the KES one. In the instrument for the tensile test (extension meter) discrete loads are used, its operating principle is similar with the first-generation instruments, however automatic data analysis is performed. According to many experts, KES and FAST systems are the main reliable equipments for fabric evaluation. A survey of experts’ views and discussions on this is given by Stylios (1991); a later survey on different methods is presented by Bishop (1996), while a recent comprehensive literature review is published by Matsudaira (2006). In both of KES and FAST systems, the necessary fabric handling is performed by a trained operator. The involvement of the human operator affects the repeatability and the accuracy of the fabric testing systems. The operational mechanisms of these instruments are based mainly on hard automation. There is no flexibility in terms of changes on the sample size or changes on the test parameters such as load, duration, etc. The request of flexibility was addressed by Potluri and Porat (1995, 1996) when they presented a flexible test system based on a robotic cell that performs the same

Evaluation from tensile test

81

IJCST 19,2

82

fabric tests with KES system. In this robotic system, several pre-test activities, such as the handling and clamping of the samples, as well as all the tests are automated reducing considerably the operator intervention. Nevertheless, the results of all the tests are just plotted and not analyzed. For the tensile test, a force-strain curve can be plotted automatically from the measured data and analyzed for calculating the LT, WT, RT and EMT fabric properties. Similarly, the same data management with KES system could be performed, in order to find the total hand value of the fabric. Furthermore, the results have to be evaluated by an operator/engineer in order to be utilized into an automated cloth making environment. Recently, artificial intelligent-based techniques have been introduced for the evaluation of the fabric properties. Gong and Chen (1999) used neural networks to predict the performance of fabrics based on fabric mechanical properties measured on the KES-FB system. Park et al. (2000, 2001) proved that the neural networks are more efficient, than the KES evaluation approach, for determining the total hand value of fabrics. Their approach is focused on the data processing after the test rather than on the test process. Selected fabric properties measured with the KES instruments are the inputs to the neural network, while the total hand value of fabrics is the output of the network. The supervised training of the network was based on the experts’ estimations about the fabrics total hand values. Matsudaira (2006), after a comprehensive literature review on fabric handle and mechanical properties, concluded that the accuracy of prediction with neural network models is higher than that of conventional regression methods using mechanical properties measured by KES system. Amano et al. (2000) tried to evaluate the fabric hand value by using qualitative data such as large/small, high/low, etc. which have been converted by the fabric mechanical properties measured on the KES system. Their qualitative data were analyzed by using the quantification theory I. They concluded that the objective evaluation of fabric hand using qualitative data can be in practical use at a similar level of the KES method. The introduction of the FAMOUS system (Stylios, 2001, 2003) responds to the industry demands for low-cost and user-friendly fabric testing instruments. This device establishes a new concept for effective measurements which is based on a single equipment. It performs the complete suite of measurements with only one sample and it reduces the human intervention and the required time is less than five minutes. The analysis of the measured data follows the same concept with the previous systems. The FAMOUS equipment is subjected to international patents and seems to be very promising. The majority of the industries producing or using fabrics have a separate laboratory for testing textile products. A testing laboratory is equipped with a variant of fabric testing instruments and operated by trained staff (Saville, 1999), which adds a noticeable cost to the product cost. The information acquired from these tests is used for improving the fabric’s quality but is not used or cannot be directly used, for the handling process of fabrics in clothing industry. Particularly for the sewing process, whether the fabric handling is performed by a human or by an automated system such as a robot, the fabric extensibility is necessary. This fabric property expresses the ability of a fabric to be extended under low stresses. The tensile properties (LT, WT, RT and EMT) derived from the tensile test using the above described instruments have to be interpreted by an expert in order to be used. The extensibility (EMT) is the strain at maximum load expressed as a percentage. Nevertheless, the extensibility of a fabric

is a non-linear property and cannot be expressed with only one parameter. Therefore, a combination of EMT with LT and WT would be necessary for estimating the fabric extensibility or a new evaluation approach of the tensile test should be considered. In this paper, a new approach is introduced for the estimation of the fabrics extensibility while the fabric tensile test is performed in a robotic cell that could be the same with that used for the sewing process. The acquired force-strain data from the tensile test are directly fed into a neural network, throughout the test. The neural network estimates the fabric extensibility according to its training which is based on experts’ estimations for the extensibility of a representative set of fabric samples. The introduced approach is part of a project for the development of an intelligent sewing system (Koustoumpardis and Aspragathos, 2003), where the extensibility was fed into a fuzzy logic system with a neuro-controller regulating the tension of a fabric in the robotized sewing. In the next sections, the introduced approach is described and the robotic cell for estimating the fabrics extensibility is presented. The requirements of the sewing process and the demand of knowing the fabric extensibility in this task are explained. The tensile test executed manually by the experts is investigated and the proposed system that tries to emulate the manual reasoning is described. The neural network topology and the training algorithm are analyzed. Finally, the performance of the proposed system is discussed and suggestions for further improvements are reported. 2. Problem definition and objective In the clothing industry traditionally, the expert seamstresses/tailors rely on their subjective estimations of the fabric properties and they do not use any testing equipments. This manual-oriented estimation is followed even nowadays in the majority of the clothing industry production lines and particularly in the fabric handling for sewing. The fabric mechanical properties are determined through hand actions, such as bending, stretching, shearing and rubbing. The experts express their evaluation through qualitative linguistic values, such as very low, low, medium, high, very high stiffness, extensibility, etc. The introduced approach has two targets. The first one is the automated estimation of the fabrics’ extensibility which could be used as an input to the intelligent robot controller for fabric handling in the sewing process presented by Koustoumpardis and Aspragathos (2003). The second one is a technique to measure online the extensibility of a fabric. Both targets require the development of a method for testing the fabrics and evaluating their extensibility with low-stress measurements. The fabric’s extensibility is a key fabric property for the majority of the fabric handling tasks. For the automated fabric handling during the sewing process, the fabric’s extensibility is an essential feature, which affects the design of the robot control scheme. Koustoumpardis and Aspragathos (2003) introduced an intelligent fabric handling approach focused on the control of tension applied on a fabric when it is guided to the sewing machine by a robot. When the sewing mechanism pulls the fabric from one edge, then the robot end-effector must guide the fabric like a seamstress/tailor. The determination of the appropriate fabric tension for sewing depends mainly on to the fabric’s extensibility. A fuzzy logic decision mechanism was proposed to determine the appropriate tension according to the fabric’s extensibility. The input of that fuzzy system was the expert’s estimation of the fabric’s extensibility.

Evaluation from tensile test

83

IJCST 19,2

84

Following the proposed approach, the estimation of the fabric’s extensibility could be performed automatically in a robotized sewing cell but the proposed extensibility estimation method could be integrated easily into any other extensibility measurement system. Particularly, this work contributes towards the automation of the fabrics’ extensibility estimation under the frame of a robotized sewing, as it is shown in Figure 1. The measured data of applied force and robot end-effector position are recorded in pairs. The corresponding curve, which represents the fabric’s extensibility, can be plotted automatically. Simultaneously, the measured data are evaluated “online” by the proposed neural network, which is shown in Figure 2. The network is trained with a target set which represents the experts’ estimation of the fabrics’ extensibility with linguistic values. Finally, the trained network is used for estimating the extensibility of any fabric with extensibility in the range of the fabrics used for training. This estimation could be used into an automated (robotic) fabric handling system with minimal or without human intervention. It must be noticed that the initial placement of the fabric sample on the robotic gripper is not automated and it is performed by an operator. The automation of this A Fabric Tensile Test

measured data

Figure 1. Tensile test and intelligent evaluation of measured data

B Estimation of fabric extensibility using Neural Network

C Intelligent robot controller for fabric handling

1 Fabric tensile curve

Tensile tests in a wide range of fabrics

Experts’ estimations of fabrics extensibility

Measured (force, strain) data for each fabric Feed-forward Neural Network

ET +

EN −

error = ET – EN training Training phase Testing phase

Figure 2. Measurement and intelligent evaluation of the fabric extensibility

Tensile test for unknown / known fabrics

Trained Neural Network

Fabric extensibility

preliminary stage of the test is out of the scope of this paper and has been faced successfully by Potluri and Porat (1995, 1996). The contribution of the present approach is focused on the post-processing of the tensile test measurements towards the mechatronic design and intelligent advancement of testing instruments. 3. Tensile test The tensile test measurements are performed under the following specifications.

Evaluation from tensile test

85

3.1 Specifications The sample size affects the fabric gripper dimensions as well as the dimensions of the overall test system. The determination of the sample size is deducted by the observation of the tensile test performed by the experts estimating the fabric’s extensibility. The experts are asked to evaluate the extensibility of a variety of fabric samples. The experts are seamstresses from small handicrafts of cloth manufacturing as well as some researchers having long research experience in textile handling. Six types of fabrics (three-woven with very low, medium and very high extensibility and three knitted with respective extensibility) and three sample sizes for each fabric type are used. The 18-samples are circulated in each one of the eight-experts in order to study the actions of the experts in the tensile test. The tensile test carried out by these experts is recorded and the main features of the procedure are extracted. In Figure 3, the initial grasping of a fabric sample is shown where can be observed that the two essential features for the sample size can be determined by measuring the distance, L between the hands of the expert and the width, D, along the gripping face of the thumb. The two dimensions L and D depend on the human body anatomy and size. The length, L depends on the distance between the two humans’ shoulders, and the width, D depends on the size of the hand. The value of L from several experts is found to be in the range between 5 and 15 cm but for the majority of them around 10 cm. Likewise, the value of D ranged between 3 and 8 cm but for most of them around the 5 cm. Hence, the dimensions of the active area of the sample are set equal to 10 £ 5 cm. The actual size of the samples is 20 £ 5 cm in order to grasp the two ends of the samples, as it is shown in Figure 4.

Figure 3. Tensile test as performed by an expert

IJCST 19,2

Figure 4. Fabric samples dimensions

5 cm

D = 5 cm

86

L = 10 cm

Tensile active area Gripping area

In the KES system, the size of the fabric samples for all measurements is 20 £ 20 cm; while for the tensile test the extension is applied in an active area of 20 £ 5 cm and along the 5 cm direction. The selection of this size is justified from the fact that it is a cloth typical area that supports the various deformations when the cloth is worn. This is a sound justification, because the scope of the KES system is the determination of the total hand value which concerns mostly the feeling of the fabric when it is worn. The same size of the sample is used in the FAMOUS system too. In the FAST system, and for the tensile test the samples must be 20 £ 5 cm (or 13 £ 6 cm for bias extension) while the tensile active area is 10 £ 5 cm (or 10 cm £ 6 cm for bias extension) and the load is applied along the 10 cm direction. In addition to the previous conclusions, the systematic investigation of the experts’ way of fabric handling for the tensile test revealed some remarkable features: . The distance between the two hands holding the fabric is not adjusted accurately. The experts do not care if their hands are placed out of plumb or not. The experts cannot repeat the test under the same active area of the sample, so small variations from test to test are presented. Nevertheless, this does not affect the repeatability of the test since the experts conclude to the same results. . The environmental conditions (humidity, temperature, etc.) of the test are not considered; and when a test is repeated by the same expert it is not performed under identical environmental conditions, especially, if the tests are executed into the industry. It should be noticed that the laboratory experiments were taken place from the summer to the winter but their judgments are not affected by the variation of the environmental conditions. . The experts express the extensibility of the fabric using linguistic values. Five linguistic values are recorded: “very low,” “low,” “medium,” “high” and “very high” extensibility. Despite the lack of testing specifications as described on the above features the cloth making people rely more on their subjective estimation for sewing than on the measured mechanical properties of the fabric derived by instrumental tests. This is a

common process followed in the clothing industry where the actual circumstances and conditions vary from the laboratory ones. Taking into account the conclusions of the afore-mentioned tests in the design of an automatic system, estimating the extensibility close to the human expert is a challenge and at the same time facilitates and reduces the cost of the implementation. It is really a challenge to design a system imitating the human behavior and the human way in estimating the characteristics of the fabric. The human expert has an overall approach in the evaluation and this is quite important for a material such as the fabric which is nonlinear and anisotropic. On the other hand, the hardware is simplified since it can be build up using low-accuracy and low-cost sensors and fixtures putting more effort in the intelligent system for the evaluation of the fabric properties, which is the main purpose of the present work.

Evaluation from tensile test

87

3.2 Development A sketch and the actual experimental stage are shown in Figure 5. The one edge of the fabric is fixed onto a horizontal table while a very simple robot gripper pulls the opposite edge of the sample. The entire active area of the fabric is slightly raised above the table surface, so as no friction force between the fabric and the table is presented. The robot end-effector pulls the sample and for each step of its movement equal to 0.01 mm the force and the robot position corresponding to the fabric elongation E are recorded and fed into the neural network. The tensile test stops when the force reaches the upper limit of 30 Nt or when the extension of the fabric reaches the 50 percent of the sample free length L. The robot maximum load imposes the restriction of 30 Nt, while the applied maximum force by a human expert lies between 10 and 15 Nt. In the KES system, the maximum applied force is 5 Nt/cm for woven fabrics and 2.5 Nt/cm for knitted fabrics. In the FAST system, the maximum force is 98.1 Nt/m and the maximum extension is 20 percent. In our case, the maximum extension is set to 50 percent which is close to the maximum extension for a test performed by hand in the cloth making industry. For the implementation of the tensile test the Mitsubishi RV-4 A robot is used, which is a 6-DOF vertical articulated manipulator. The robot’s specifications related to the tensile test are:

Fixed clamp

Fabric-cloth

Robot gripper

Force sensor

Figure 5. Fabric tensile test (extension)

IJCST 19,2

. . . .

88

Arm reachable radius (from the center point of robot’s base). 634 mm. Maximum resultant velocity. 5,800 mm/sec. Repeatability. ^ 0.03 mm. Maximum load. 4 kg.

The attached sensor is the g (65/5) F/T system from ATI Industrial Automation. The specifications of the sensor in the tensile direction are: . Sensing range. ^ 65 Nt. . Resolution. 0.05 Nt. . Accuracy. , 1 percent (at calibration temperature 228C). . Output data rate. 9,600 baud (< 126 Hz). The tensile test can be completed in less than 1 min for fabrics with a very high extensibility, but the required time is reduced to some seconds for a fabric with a very low extensibility since the upper bound of 30 Nt reached with a relatively small displacement of the robot end-effector. 3.3 Experimental procedure Fourteen fabric samples spanning a wide range of fabrics’ extensibility are chosen to form the training set for the neural network. The composition and construction of the 14 fabrics are shown in Figure 6. Six additional fabric samples shown in Figure 7 are chosen for testing the trained neural network. The extensibility of each fabric sample is estimated by the same team of experts presented in section 3.1. The experts are asked to characterize the samples using the linguistic values defined in section 3.1: “very low,” “low,” “medium,” “high” and “very high” extensibility which represented by the following numbers: 0.1, 0.3, 0.5, 0.7 and 0.9, respectively, in order to facilitate the neural network training. The experts’ estimations are shown in the last columns of Figures 6 and 7 where the value 0.6 represents cases of experts’ ambiguity and responding as follows “something between medium and high extensibility.” The force/strain curve for each of the samples shown in Figures 6 and 7, are shown in Figure 8. The measured force and strain for the first 14 samples are feed into the neural network for training which is described in section 4. As it is shown in Figure 8, the curves of the 14-fabrics cover the force-strain area adequately, starting with a fabric with a very low extensibility (No. 1) and come to a fabric with a very high extensibility (No. 14) while in a few cases two fabric curves cover nearby areas. Therefore, the choice of the 14-fabrics seams to be adequate for training a neural network that could estimate the fabrics extensibility into this area. The extensibility estimation is a classification problem, since the desired outputs (estimations of extensibility) for each input (data of tensile test) as well as the number of classes (five) are known from the available experts’ knowledge. The use of neural networks in classification requires supervised training in order to assign instances to predefined classes, which is followed in the proposed approach. The trained neural network is used to estimate the extensibility of the six fabric samples (No. 15-20), shown in Figure 7, which were not used for training and the network performance is presented in section 5.

Construction

Composition

1

Woven

viscose 65% polyester 35%

0.1

2

Woven

linen 100%

0.3

3

Woven

cotton 100%

0.3

4

Knitted inter rib

polyester 65% cotton 35%

0.6

5

Knitted

acrylic cotton

0.5

6

Knitted single-jersey

polyester 70% cotton 30%

0.6

7

Knitted

polyester 65% cotton 35%

0.5

8

Knitted rib

cotton 100%

0.7

9

Knitted interlock

polyester 100%

0.7

10

Knitted rib

11 12 13 14

Knitted single-jersey Knitted (No. 16 in warp direction) Knitted interlock Knitted

cotton 49% acrylic 49% lycra 2% polyester 94% lycra 6%

Fabric picture

Experts’ estimation: extensibility

Sample No.

Evaluation from tensile test

89

0.6 0.6

polyester 60% cotton 40%

0.7

nylon 100%

0.7

cotton 94% elastane 6%

0.9

It is evident that the curves shown in Figure 8 can be grouped in three categories A, B and C according to the type of fabrics. The A-curves are derived from the woven fabrics (Sample No. 1-3), the B-curves from the knitted synthetic fabrics with high percentage of polyester (Sample No. 4-7) and the C-curves from the knitted synthetic fabrics with high percentage of lycra or elastane or nylon (Sample No. 8-14). The fabric composition cannot be a basis for grouping the fabrics since the structure of the fabric and the structure of the yarns affect the extensibility of the fabric too. The direction of the applied force along the weft, warp or bias of a fabric is another factor that determines the extensibility of the fabric. Figure 8 shows that two (or more) curves could intersect in one (or more) points, which might cause problems to the performance of the neural network. The common point of the two intersecting curves a and b is (s, f ) where f is the measured force and s is the strain of the fabric, as it is shown in Figure 9.

Figure 6. Fabric samples used for training the neural network

IJCST 19,2

90

Construction

Composition

15

Woven

synthetic wool 50%-50%

0.3

16

Knitted (No. 12 in weft direction)

polyester 60% cotton 40%

0.5

17

Woven

18

Knitted

cotton 92% elastane 8% polyester 65% cotton 35%

Knitted rib Knitted rib

19

Figure 7. Fabric samples used for testing the neural network

20

Fabric picture

0.5 0.6

acrylic 100%

0.5

cotton 100%

0.9

14-fabrics for training 6-fabrics for testing

35 1 30

A

B

15

5

2

4

C 9 8 10

7

19 18 6

17

25 Force (Nt) f

Experts’ estimation: extensibility

Sample No.

3

11

20

12 16

15 10

13

5

20 14

Figure 8. Fabrics’ force/strain curves

0

0

5

10

15

Figure 9. Improving the network performance by feeding with current and previous measurements

force

b

20

30

35

40

a

(s, f ) (asp, afp) (bsp, bfp) strain

25 Strain (%) s

s f sp fp

Neural Network

Extensibility (EN)

45

50

Since, the point (s, f ) is one of the input pair for training the network and this input set corresponds to two desired outputs, the extensibility corresponding to each curve, the training of the network can be obstructed. This problem is faced by feeding the network with the local gradient of the curves, but to avoid differentiation, in each instance the neural network is fed with the current point (s, f ) and the previous one (sp, fp).

Evaluation from tensile test

91 4. The structure and the training of the neural network The tensile test data (pairs of strain and force) are feed in a feed-forward neural network. Each input set is composed by four elements: j j si ; f i ;j si21 ;j f i21 ð1Þ where j is the number of the curve corresponding to the jth fabric sample j ¼ 1 2 14;j si ;j si21 are the ith measured strain and the previous one, j f i ;j f i21 are the ith measured force and the previous one. The range of i varies for each curve j since the duration of the tensile test and hence the number of the measurements varies from fabric to fabric. For a fabric with very low extensibility a low number of measurements are recorded while for a fabric with very high extensibility a large number of measurements are recorded. For each input set a network output (EN) is calculated which is compared with the desired output or target set (ET) and produces the error in order to train the network. The target j ET; is the experts’ estimation about the extensibility for the jth fabric sample. The error used for the supervised training of the network is: error ¼ j ET 2 j ENi

ð2Þ

where, j ENi is the output of the network for the ith pair of measurements of the jth fabric sample. Although various types of neural networks and training algorithms are available (Matlab Documentation, 2001), there is no currently a standard method regarding the determination of the appropriate number of hidden layers and neurons. Therefore, in order to determine the appropriate number of the hidden neurons a large number of experiments with a variation of numbers of hidden neurons were performed. For this commonly used trial and error approach, the quality of the response rate, the response smoothness and especially the overshooting were the criteria for the selection of the appropriate number of the hidden layers and neurons. In order to simplify and speed up the training procedure, it was decided to train a number of feed-forward neural networks with a single hidden layer and an increasing number of hidden neurons, until a satisfactory performance was met. Eventually, a number of seven hidden neurons proved to be enough for the network satisfactory performance. Networks with more than seven neurons in the hidden layer presented faster response and better precision to the desired values. Nevertheless, the minimum number of seven neurons was chosen since the nature and the specifications of the problem do not require extremely precise response but it requires low computational time if it would be incorporated into the cloth

IJCST 19,2

making process as an online test. Neural networks with two or more hidden layers were not tested, since such a structure increases the computational time considerably. The neurons in the input layer are linear, while for each of the neurons in the hidden and output layers a sigmoid activation function is assigned, given by:

92

output ¼

1 1 þ e2input

ð3Þ

The neural network structure is shown in Figure 10. The input layer of the network consists of four neurons, where the current force (Nt) and strain are fed in the first two, while the previously measured force (Nt) and strain are fed into the third and fourth.

B

B

S K11

W11

jf

i

L S

js i

S

L

j

ENi

S

S

jth

fabric extensibility S

jf

i–1

Output Layer

L

S js

i–1

L

K71 W47

Figure 10. The 4-7-1 neural network topology

Input Layer with 4 neurons

S

Hidden Layer with 7 neurons

S: sigmoid neuron L: linear neuron B: bias Wij: weight array for the hidden layer Kj1: weight array for the output layer

4.1 Data preprocessing The training of the network was performed without preprocessing any of the data in order to reduce the computational time. However, to make the training easier and the network’s response “realistic” parts of the available data were used. All training patterns with force components less than 1.5 Nt and strain components less than 0.05 were ignored because this data area, shown with a gray box in Figure 8, contains indistinguishable curves due to the very intense overlapping of the measured data, which could cause problems in the training of the neural network.

4.2 The supervised training of the network The standard back-propagation algorithm (with or without momentum) is a well-known and certainly the most common algorithm for training feed-forward neural networks. Nevertheless, due to its slow convergence, it is ill-suited for practical and online applications. Many high performance algorithms have been proposed (Haykin, 1999) with much higher convergence rate than the standard back-propagation algorithm. The Levenberg-Marquardt algorithm (Hagan and Menhaj, 1994) is the fastest algorithm for moderately sized networks and function approximation problems. However, as the number of weights increases the algorithm’s advantages decrease. All the training algorithms provided by the neural network Toolbox of MatLab (Matlab Documentation, 2001) have been tested in order to find the most appropriate. Finally, the Levenberg-Marquardt algorithm is used for the training. The starting value of the m coefficient (Hagan and Menhaj, 1994) is set equal to 0.001. The m coefficient is changed according the performance function which is the sum of squares of errors over all inputs. After each successful iteration (in which the performance function is decreased), the m coefficient is multiplied by 0.1 to decrease its value. If the iteration is unsuccessful, the m coefficient is multiplied by ten so as to increase its value, with a maximum equal to 1 £ 10þ 10. The training procedure terminates when m reaches the maximum value. The neural network is trained for 1,000 epochs, which means that the input set (all the measurements for all fabrics) is fed 1,000 times in the network. The mean square error (MSE) (sum of squares of errors over all inputs) is about 0.00041 at the end of the training, as shown in Figure 11. For our case, this MSE is very satisfactory since the trained network can estimate the extensibility for each of the 14 samples with accuracy about ^ 0.001. The extensibility of a fabric estimated by the experts is represented by one decimal digit number (0.1, 0.3, 0.5, 0.7, 0.9) while the network estimates the extensibility of the same fabric approximating this number with accuracy ^ 0.001 after ten epochs, while the MSE is about ^ 0.1 at the beginning of the training process, as shown in Figure 11. The training of the network splits the force-strain region into extensibility classes. For example, the force-strain region composed by the fabrics with the C-curves (Figure 8) has been classified by the trained network as a region with fabrics’ extensibility between 0.6 and 0.7. Small misalignments in sample grasping do not affect the introduced classification, since the network is robust enough to reject such disturbances.

Evaluation from tensile test

93

IJCST 19,2

MSE is 0.00040846 100

10−1

Training Error

94

10−2

10−3

Figure 11. The MSE of the training process

10−4 0 10

101

102

103

Epochs

5. Testing the network performance – results The trained neural network is tested with the set of the six unknown fabrics shown in Figure 7, which are not used for the training of the network. Tensile test is performed for each sample and the measured forces and strains are fed online into the trained network. The output of the network is the estimation of the extensibility for the tested sample. The response of the network for each of the six fabrics is shown in Figure 12. For each of the samples the horizontal solid line is the extensibility estimated by the experts, and the dashed line is the output of the network. The network is capable to make a rough estimation from the very beginning of the tensile test, and its estimation is improved as the test is progressed. It is important to remind that the training of the network was performed with data greater than 0.05 for the strain and greater than 1.5 Nt for the force. Nevertheless, the network makes satisfactory estimations even for this ignored data range, while quite far before the end of the test (approximately in the 20-30 percent of the total test time) the desired estimation is closely approximated. Each pair of the measured force and strain is fed immediately into the trained neural network, during the tensile test; consequently it is not necessary to complete the tensile test in order to estimate the fabric extensibility. As it is shown in Figure 12, the responses for the samples 15 and 20 are not as satisfactory as the other ones since their force-strain curves are close to the limits of the training set, as shown in Figure 8. It is well-known that when the fed sample is close to the limits of the training set the neural network performance is deteriorated. Moreover, the samples 15 and 20 are fabrics with a very low and a very high extensibility,

Sample 15

Evaluation from tensile test

Sample 16

0.35 0.3

0.5 Extensibility

Extensibility

0.25 0.2 0.15 0.1

0.4 0.3

95

0.2 0.1

0.05 0 0

0.01

0.02

0.03 Strain

0.04

0.05

0

0.06

0

0.02 0.04 0.06 0.08 Strain

0.7

0.6

0.6

0.5

0.5

0.4 0.3

0.4 0.3

0.2

0.2

0.1

0.1

00

0.05

0.1 Strain

0.12 0.14

Sample 18

0.7

Extensibility

Extensibility

Sample 17

0.1

0.15

0 0

0.2

0.05

Sample 19

0.1 Strain

0.15

0.2

Sample 20 1

0.6 0.8 Extensibility

Extensibility

0.5 0.4 0.3

0.6 0.4

0.2 0.2

0.1

0

0 0

0.05

0.1 Strain

0.15

0.2

0

0.1

0.2

0.3 Strain

0.4

0.5

respectively, where a small number of training samples was used compared to the other regions. This problem could be faced if new samples of fabrics with tensile curves outside these limits are used to train the network. The proposed system initiates a new approach, in which the fabric properties are expressed and used in a way close to the experts’ conception. The ability of neural networks, to approximate non-linear functions, is used to estimate and express the extensibility of a fabric with one parameter as close as possible to the expert

Figure 12. The neural network estimations of the extensibility for samples 15-20

IJCST 19,2

96

estimation. The innovation of this approach is based on the intelligent evaluation of the tensile test measurements rather than on the experimental procedure. The abilities of neural networks for generalization and abstraction (Wasserman, 1989) can facilitate towards this direction. This work is an attempt to introduce into the clothing industry and particularly in the robotic sewing process the use of linguistic variables (qualitative data) for expressing the fabric properties. These linguistic variables are already used by a seamstress/tailor during the sewing process as well as throughout the majority of the fabric handling tasks in the clothing industry. This subjective way of estimating the fabric properties is a significant experience which by tradition is passed from the expert to the trainee. The optimistic perspective of this work is the scientific establishment of a common language between the staff of a clothing industry and the utilization of this language in the automated sewing systems which can be extended to other flexible automation systems based on neural networks, fuzzy logic and genetic algorithms. 6. Conclusions The proposed robotic system performs the tensile test of fabrics and estimates their extensibility using a simple feed-forward neural network. This system will be integrated in the intelligent robotic sewing workcell where the linguistic estimation of the fabric’s extensibility is necessary but it can be integrated in intelligent fabric testing systems too. This approach is close to the experts’ way of testing and evaluating the fabric extensibility. The specifications of the tensile test derived by the observation of the respective tests carried out by experts using their hands, senses and brain. The intelligent information processing, through neural networks obtains results close to the experts estimations. The proposed system is flexible and the additional cost for hardware is low if it will be incorporated in a robotic cell for automatic sewing process. Compared with conventional test instruments and their fabric properties evaluations, the proposed method examines the extensibility of a fabric as an overall approach, as an expert does. The evaluation of a force-strain curve as a whole allows a better comparison of fabrics as it contains more information about their extensibility than the simple figures for EMT, LT and WT. The neural networks approach for the final estimation of the extensibility takes into account the global shape of the curve and not only the final point or some average characteristics. The future directions of the proposed approach are focused on the bending and shearing fabric tests in relation with the fabric handling automation. The observation and the attempt to imitate the human way of testing and evaluating the fabric characteristics should be the guidance for that work. References Amano, T., Maenaka, Y., Kitaura, T. and Obara, K. (2000), “Fabric hand evaluation using qualitative data”, Sen-I Gakkaishi, Vol. 56 No. 1, pp. 41-5. Bishop, D.P. (1996), “Fabrics: sensory and mechanical properties”, Textile Progress, Vol. 26 No. 3, ISBN 1870812751. CSIRO Division of Wool Technology (1988), Fabric Assurance by Simple Testing, CSIRO Division of Wool Technology, Melbourne.

Gong, R.H. and Chen, Y. (1999), “Predicting the performance of fabrics in garment manufacturing with artificial neural networks”, Textile Research Journal, Vol. 69 No. 7, pp. 477-82. Hagan, M.T. and Menhaj, M. (1994), “Training feedforward networks with the Marquardt algorithm”, IEEE Transactions on Neural Networks, Vol. 5 No. 6, pp. 989-93. Haykin, S. (1999), Neural Networks: A Comprehensive Foundation, Prentice-Hall, Englewood Cliffs, NJ, ISBN: 0-13-273350-1. Kawabata, S. (1980), The Standardization and Analysis of Hand Evaluation, 2nd ed., The Textile Machinery Society of Japan, Osaka. Koustoumpardis, P.N. and Aspragathos, N.A. (2003), “Fuzzy logic decision mechanism combined with a neuro-controller for fabric tension in robotized sewing process”, Journal of Intelligent and Robotics Systems, Vol. 36 No. 1, pp. 65-88. Manuals (1980), Manuals for the Kawabata Evaluation System, Kato Tekko Ltd, Kyoto. Matlab Documentation (2001), Neural Network Toolbox, Ver. 4.0.1., The MathWorks, Inc., Natick, MA. Matsudaira, M. (2006), “Fabric handle and its basic mechanical properties”, Journal of Textile Engineering, Vol. 52 No. 1, pp. 1-8. Park, S.W., Hwang, Y.G., Kang, B.C. and Yeo, S.W. (2000), “Applying fuzzy logic and neural networks to total hand evaluation of knitted fabrics”, Textile Research Journal, Vol. 70 No. 8, pp. 675-81. Park, S.W., Hwang, Y.G., Kang, B.C. and Yeo, S.W. (2001), “Total handle evaluation from selected mechanical properties of knitted fabrics using neural network”, International Journal of Clothing Science & Technology, Vol. 13 No. 2, pp. 106-14. Peirce, F.T. (1930), “The handle of cloth as a measurable quantity”, Journal of the Textile Institute, Vol. 55, pp. T377-T416. Potluri, P. and Porat, I. (1995), “Towards automated testing of fabrics”, International Journal of Clothing Science & Technology, Vol. 7 Nos 2/3, pp. 11-23. Potluri, P. and Porat, I. (1996), “Low-stress fabric testing for process control in garment assembly application of robotics”, International Journal of Clothing Science & Technology, Vol. 8 Nos 1/2, pp. 12-23. Saville, B.P. (1999), Physical Testing of Textiles, Woodhead Publishing Ltd, Cambridge, ISBN 1 85573 367 6. Shirley Developments (2003), Shirley Developments Catalogue No. 12, SDL International Group, Manchester, NH. SiroFAST (1994), Report No. WT92.02, CSIRO Division of Wool Technology, Melbourne, ISBN 0 643 06025 1. Stylios, G.K. (1991), Textile Objective Measurement and Automation in Garment Manufacture, Ellis Horwood Ltd, Chichester, ISBN 0139127429. Stylios, G.K. (2001), “The FAMOUS concept in the textile industries”, paper presented at International Textile Congress, 18-20 June. Stylios, G.K. (2003), “The concept of fabric automated measurement and optimisation system for textiles and clothing”, paper presented at 4th IMCEP, Maribor, pp. 1-6. Wasserman, P.D. (1989), Neural Computing Theory and Practice, Van Nostrand Reinhold, New York, NY, ISBN 0-442-20743-3.

Evaluation from tensile test

97

IJCST 19,2

98

Further reading Potluri, P., Atkinson, J. and Porat, I. (1995), “A robotic flexible test system (FTS) for fabrics”, Mechatronics, Vol. 5 Nos 2/3, pp. 245-78. About the authors Panagiotis N. Koustoumpardis is a Research Assistant in the Robotics Group of the Mechanical and Aeronautics Engineering Dept., University of Patras, where he received a Mechanical Eng. Diploma (1998). Since, 2004, he is an Instructor at the Dept. of Mechanical Engineering, Technological Educational Institute of Patras. His main research interests are intelligent control, neural networks, machine vision for robotics applications and intelligent robotic handling of non-rigid materials. In these fields, he has published and announced his work in national and international conferences and journal. He has been involved in several research projects and he is currently working on one national and two EU research project. He is a member of Greek Society of Mechanical and Electrical Engineers, and of Technical Chamber of Greece. Panagiotis N. Koustoumpardis is the corresponding author and can be contacted at: [email protected] John S. Fourkiotis is a Mechanical Engineer. He received the Mechanical Eng. Diploma (2004) from the Mechanical and Aeronautics Engineering Dept., University of Patras, Greece. During his final diploma he worked on robotized handling of non-rigid materials as well as on the experimental development of tensile tests. He is a Postgraduate student in the School of Mechanical Engineering of the National Technical University of Athens. E-mail: [email protected] Nikos A. Aspragathos leads the Robotics Group of Mechanical and Aeronautics Engineering Department, University of Patras. He received the Electrical Engineering Diploma (1975) and the PhD (1981) both from School of Engineering, University of Patras, Greece. He joined Greek industrial companies as a construction and maintenance Engineer (1976-1980). In 1982, he was appointed as a Lecturer in Mechanical and Aeronautics Engineering Department, University of Patras. He was a visiting member of the staff in UWIST, Cardiff from 1987 to 1988. His main research interests are robotics, intelligent planning, control and design, industrial automation, CAD/CAM, and simulation. He is reviewer in several Journals and member of the editorial board of the Mechatronics Journal. He has been and is currently involved in research projects funded by Greek and European Union sources. He is a member of Greek Society of Mechanical and Electrical Engineers, and of Technical Chamber of Greece. E-mail: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Salt-free reactive dyeing of cotton

Salt-free reactive dyeing of cotton

D.P. Chattopadhyay Department of Textile Chemistry, Faculty of Technology and Engineering, M. S. University of Baroda, Vadodara, India

R.B. Chavan Department of Textile Technology, Indian Institute of Technology, New Delhi, India, and

J.K. Sharma

99 Received 10 January 2006 Revised 22 September 2006 Accepted 22 September 2006

Department of Chemistry, G. J. University, Hissar, India Abstract Purpose – Fibre reactive dyes are very popular for cellulosic garments as they are environmentally safe and having good overall fastness properties. But application of these dyes requires a very high concentration of salt. The salt released from garment dyeing increases salinity in drain water stream which has a negative impact on environmental ecology. The present work aims to eliminate the usage of salt during dyeing of cotton goods with reactive dyes. Design/methodology/approach – The methodology adopted here, for the elimination of salt in cotton dyeing, was based on the principle of cationisation (to develop a positive charge) of cotton. The same was achieved by subjecting the caustic pretreated cotton fabric samples to a treatment of 1, 2 dichloroethane followed by methylamine to introduce amino groups in the cellulose structure. The treated cotton when dyed from slightly acidic bath generates positive sites due to protonation in the amino group. The reactive dyes being anionic (negatively charged) in solution get attracted to the positive charges on the fibre which eliminates the salt requirements for satisfactory dye exhaustion. Findings – The investigation was conducted for cold brand, hot brand and highly exhaustive reactive dyes. The modified cotton showed excellent dye exhaustion for all the dyes in the absence of salt. The treatment was found to improve the dye fixation also. The modification was assessed through elemental analysis. Research limitations/implications – This study may be further extended to viscose material after suitably modifying the treatment conditions. Practical implications – A pretreatment to cotton which could eliminate the usage of salt in its dyeing with reactive dyes is revealed. Originality/value – The study explored a newer technique of cotton dyeing without salt usage. Both garment dyeing units and fabric/yarn finishing industries would thus be helpful. Keywords Dyes, Cotton, Salt, Chemical reactions Paper type Research paper

Introduction Garments made of cellulosic fibres are mostly dyed with reactive dyes. These dyes are much brighter, longer-lasting, and easier-to-use. Unlike other dyes, they actually form a covalent bond with the cellulose molecule. No wonder one can safely wash a garment that has been dyed in bright fibre reactive colours with white clothing, a hundred times, without endangering the white. Salt performs two key functions during dyeing of cotton garments with reactive dyes. Sodium chloride or sodium sulphate is fully dissociated in water to loose ion pairs

International Journal of Clothing Science and Technology Vol. 19 No. 2, 2007 pp. 99-108 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710725702

IJCST 19,2

100

2 of sodium (Naþ ) and chloride (Cl2 ) or Sulphate (SO2 4 ) ions. The positive sodium ion has the capability of traveling to the fiber/water interface and effectively negating the zeta potential, destroying the barrier to the initial dye/fiber interaction. Picturing the water as a lattice structure that has only limited sites for accommodating charged species, and considering the fact that dissociated sodium salt is more compatible with the water lattice than a large molecular weight, dye molecule with a few sulphonate groups hanging off, the salt is the preferred species to occupy the limited lattice sites. Dye is thus effectively “salted out” of the bath, with the distribution coefficient of the dye shifted from solution towards the fiber. Both these actions of the salt lead to the reactive dye exhaustions achieved during dyeing of cotton with reactive dyes. In the absence of adequate salt concentration, reactive dye bath exhaustion is poor. Therefore, the major part of the dye remains in the bath and subsequently gets hydrolysed in the dye bath in the presence of alkali. Part of these hydrolysed dye gets exhausted on the fabric, but remains unfixed. To remove the hydrolysed/unfixed dye, time consuming, energy intensive and expensive washing off procedures are required. A number of processes have therefore been proposed from early 1930s, till date to improve the substantivity of anionic dyes towards cellulosic fibres by introducing cationic sites on the fibres. Schlack (1938) was the first to report improved affinity of acid dyes towards cellulose modified through the introduction of aminated epoxy groups. Rupin (1976) and Rupin et al. (1970) studied the dyeability of cellulose towards direct and reactive dyes after pretreatment with glycidyl trimethyl ammonium chloride (Glytac). It was reported that the Glytac pretreatment improved the dyeability of cotton with reactive dyes in the presence of alkali and salt. Pretreatment of cotton with polyamide epichlorohydrin type polymer, e.g. Hercoset125 (Hercules Powder Corporation) offers the opportunity for increasing both the substantivity and reactivity of cotton towards reactive dyes under neutrtal to acidic condition due to the introduction of cationic azetidinium group (Earle et al., 1971). Wu and Chen (1992) treated cotton with polyepichlorohydrin (PECH) dimethylamine which was manufactured by initial polymerization of epichlorohydrin, followed by amination with dimethylamine. The epichlorohydrin was polymerized in carbon tetrachloride using boron trifluride etherate as catalyst. The dyeability of treated cotton towards direct dyes was investigated and it was found that PECH-amine could improve the direct dyeability of modified cotton. In another work they have reported (Wu and Chen, 1993) the effect of PECH-amine treatment on the reactive dyeability of cotton. It was found that the modified cotton can be dyed with selected low reactivity dyes under neutral condition using limited salt concentrations or with selected high reactivity dyes without salt. The effect of modification of cotton using various N-methylolacrylamide derivatives, viz. bis(N-methylol-2-cabamoylethyl)butylamine, N-(N,N0 -dimethylol2-cabamoylethyl) diethylamine and N-(N,N 0 -dimethylol-2-cabamoylethyl) dimethylamine on acid dyeability has been investigated by El-kharadly et al. (1996). These agents were applied to cotton fabric by pad-dry-cure procedure. Such treatments were found to improve the crease recovery angle and acid dyeability. But these agents release hazardous formaldehyde during curing as well as on storage.

Scientists of German Wool Research Institute reported (Junger et al., 1999) the modification of lyocell fibres using epichlorohydrin catalyzed with Zn(BF4)2 followed by amination with ethlenediamine dihydrochloride at alkaline conditions. The modification was aimed at improving the acid deability of lyocell for one-bath dyeing of blends with wool. Treatment of cotton with 2,4-dichloro-6-(2-pyridineethylamino)-s-triazine was reported to improve its reactivity towards amino alkyl dyes (Lewis and Lei, 1992). Application of fibre reactive chitosan derivative, O-acrylamidomethyl-N-[(2-hydroxy-3-trimethylammonium)propyl] chitosan chloride to cotton is reported (Lim and Hudson, 2004) to provide zero-salt dyeing of cotton with direct and reactive dyes. Most of the techniques developed so far for reducing salt concentration in cotton dyeing are complicated in nature. Some of these methods are successful only for highly reactive dyes. Under this background, efforts were made to find out easier method for salt free reactive dyeing of cotton. In this study, scoured and bleached cotton cellulose (C0) was first converted to soda cellulose by treatment with 20 per cent (w/v) aqueous sodium hydroxide solution. This cotton sample was then reacted with 1, 2 dichloroethane (DCE) to get a product (C1) as shown in Figure 1, which was further aminated with various amines. The best aminating agent found was methylamine which resulted in the formation of Cell-O-CH2CH2NHCH3 (C2). Pretreatment with dilute acetic acid before dyeing could generate cationic site in this modified cotton.

Salt-free reactive dyeing of cotton

101

Experimental Materials Cotton. Bleached and unmercerised cotton fabric having following specifications was used. Ends/in.: 143, picks/in.: 71, warp count: 45s, weft count: 40s. Cell-OH

(C0)

NaOH

Cell-ONa

CICH2CH2CI (1,2-dichloro ethane)

Cell-O-CH2CH2CI

(C1)

CH3NH2 Cell-O-CH2CH2NHCH3 H+ +

Cell-O-CH2CH2NH2CH3

(C2)

Figure 1. Chemical reactions undergone by cotton cellulose during the course of treatments

IJCST 19,2

102

Dyes and auxiliaries. All dyes (Table I) and auxiliaries used were of commercial grade. The chemicals were: . sodium hydroxide, laboratory regent; . DCE, analytical reagent; . methylamine (40 per cent solution), laboratory reagent; and . acetic acid, laboratory reagent. Modification of cotton The cotton fabric was padded with 20 per cent (w/v) aqueous sodium hydroxide solution using a laboratory padding mangle by two dip and two nip sequence. This fabric was then treated with an emulsion of water and DCE in a closed vessel of laboratory beaker dyeing machine. The resultant fabric was rinsed thoroughly with hot and cold water to remove excess alkali and unreacted DCE. This fabric was finally aminated using a mixture of an aminating agent and water. The fabric was finally washed, pretreated with a dilute acetic acid solution and taken for dyeing. Dyeing of untreated and treated cotton The dyeings of treated and untreated cotton were carried out using cold brand, hot brand and highly exhaustive (HE) reactive dyes (Table I) on Rota Dyer laboratory dyeing machine keeping material-to-liquor ratio at 1:50 for three shade percentages, viz. light (0.25 per cent), medium (1.5 per cent) and dark (4.5 per cent). All dyeings were performed as per the standard method prescribed by the dye manufacturers. The untreated (control) sample was dyed both in the absence and presence of recommended quantities of salt and compared with the no-salt dyeing results of treated cotton. The concentrations of salt used in dyeing are given in Table II. The pH of the exhaustion

Table I. Dyes used in the experiments

Table II. The concentration of salt and soda used in dyeing

Dye

Manufacturer

C.I. Reactive

Reactofix Yellow M4G Procion Red M5B Procion Blue MR Reactofix Yellow H4G Reactofix Red H8B Reactofix Yellow HE-6GI Reactofix Red HE8B Reactofix Blue HE-2R

JAY Atul Atul JAY JAY JAY JAY JAY

Yellow 22 Red 2 Blue 4 Yellow 18 Red 31 Yellow 135 Red 152 Blue 172

Chemicals

L

Cold M

Common salt (g/l) Soda ash (g/l)

30 3

40 4

D 60 10

L

Dye type Hot M

D

L

HE M

D

40 20

60 20

100 20

30 10

60 15

90 20

Note: L, M and D stands for light, medium and dark shades, respectively

dye bath was maintained between 6 and 7 for treated cotton whereas control samples were dyed at neutral pH. The fixation of all dyeings was carried out using alkali as per Table II. Measurements Exhaustion. The dye bath exhaustions were calculated from the bath absorbance values at lmax before and after dyeing using equation (1): E¼

A1 2 A2 £ 100 A1

ð1Þ

where A1 ¼ absorbance of dye solution before the commencement of dyeing; A2 ¼ absorbance of dye solution after the dyeing process. Dye fixation. Measurement of the extent of dye fixation on both treated and untreated cotton was performed by stripping the unfixed dye from a part of the dyed cotton using a 25 per cent aqueous pyridine solution at 1008C and a liquor ratio of 10:1. Such stripping treatment continued until no colour was visible in the stripped liquor. The K/S value of the stripped and unstripped dyed sample were measured on Macbeth spectrophotometer (model 2020þ ) integrated with a computer colour matching system. The percentage dye fixation was calculated using equation (2): F¼

f1 £ 100 f2

ð2Þ

Where F ¼ percentage fixation; f1 ¼ K/S value of dyed sample before stripping; f2 ¼ K/S value of dyed sample after stripping. Elemental analysis. The carbon hydrogen and nitrogen content of the treated and untreated cotton were estimated using a CHNS analyzer (Vario EL CHNS Elemental Analyser, Elementar Analysen Systeme GmbH, Germany). In general, cotton fibres show a crystallinity index of 55-65 per cent depending on the maturity of the fibres (Hequet and Abidi, 2001; Gamble, 2005; Kane and study, 1953; Gohl and Vilensky, 1983). The per cent of carbon, hydrogen and nitrogen which should be present theoretically in the reaction products and weight of a particular element on the same were calculated assuming cotton has 60 per cent crystallinity and 40 per cent accessible region for reaction and considering the reaction to occur per glucose unit. The unit weight of C0, C1, and C2 were calculated as follows: Unit weight of C0 ðC6 H10 O5 Þ ¼ 162 Unit weight of C1 ðCell – O 2 CH2 CH2 ClÞ ¼ ½162 þ 0:4{63:5 2 1 ðless substituted hydrogenÞ} ¼ 187 Unit weight of C2 ðCell – O 2 CH2 CH2 NHCH3 Þ ¼ ½162 þ 0:4{58 2 1 ðless substituted hydrogenÞ} ¼ 184:8 < 185 Assessment of washing fastness. Washing fastness was tested according to ISO 3 and the colour change was studied using the grey scale for colour change.

Salt-free reactive dyeing of cotton

103

IJCST 19,2

104

Results and discussion Optimisation of treatment conditions To find out a suitable condition for introducing positive site in cotton the scoured and bleached fabric was first padded with 20 per cent (w/v) aqueous caustic soda solution. The fabric was then treated with an emulsion of DCE and water using varying concentrations of DCE for different time and temperature. After the DCE treatment, the fabric samples were subjected to amination with various aminating agents like ammonia, methylamine, dimethylamine, trimethylamine and monoethanolamine using a fixed concentration of amine for a particular time and temperature. The effectiveness of these aminating agents were evaluated by comparing the improvement in K/S values of the treated samples to that of the control when dyed with Procion Red H8B (C.I. Reactive Red 31) in absence of salt. The maximum improvement was noticed with methylamine treated sample. The various parameters for DCE and amination treatment were then optimized and the optimized conditions were as follows: (1) Treatment with DCE: . DCE ¼ 16 per cent (v/v); . material-to-liquor ratio ¼ 1:50; . time ¼ 90 min; and . temperature ¼ 708C. (2) Amination with methylamine: . methylamine (40 per cent) ¼ 10 per cent (v/v); . material to liquor ratio ¼ 1:50; . time ¼ 90 min; and . temperature ¼ 708C. Elemental analysis study Table III shows the elemental analysis results of the control cotton cellulose (C0), after reaction with DCE (C1) and aminated cotton (C2). Information on per cent of element which should be present theoretically in each of the above compounds was also provided in the table. Theoretical calculations were based on the assumption that Sample Total avg. weight per Weight of element per Elementala reference Element glucose unit or its derivative glucose unit or its derivative analysis (per cent) C0 C1 C2

Table III. Elemental analysis of cotton and its derivatives

C H N C H N C H N

162 187 185

Note: aThe values given in parentheses

72 10 – 81.6 11.6 – 86.4 13.2 5.6

43.03 (44.4) 6.01 (6.2) 0.02 42.6 (43.6) 6.09 (6.2) 0.03 44.6 (46.7) 6.24 (7.13) 0.27(3.02)

40 per cent non crystalline part of cotton cellulose is accessible for such reaction and reaction to occur per glucose unit. The presence of nitrogen in sample C2 supports the reaction shown in Figure 1. Sodium fusion test confirmed the presence of chlorine in sample C1 by the formation of a very faint silver chloride precipitate, when AgNO3 was added to the sodium fusion extract of the sample.

Salt-free reactive dyeing of cotton

105 Dyeing properties of modified cotton Tables IV-VI show the effect of modification on the exhaustion and fixation behaviour of cotton dyed with three types of reactive dyes, viz. cold brand, hot brand and HE for three different shades per cent, i.e. light, medium and dark. Table IV shows that the cotton samples when dyed in the absence of salt with cold brand dyes like C.I. Reactive Yellow 22, C.I. Reactive Red 2 and C.I. Reactive Blue 4 exhibited very poor exhaustion for all the three concentrations of the dyes. But in the presence of recommended concentrations of salt the exhaustions were approximately double or more. For example, the exhaustion of Reactive Red 2 was only 25.1 per cent in the absence of salt which was increased to 58.7 per cent when dyed in the presence of

Sample C.I. Reactive Yellow 22 Control (without salt) Treated (without salt) Control (with salt) C.I. Reactive Red 2 Control (without salt) Treated (without salt) Control (with salt) C.I. Reactive Blue 4 Control (without salt) Treated (without salt) Control (with salt)

Sample C.I. Reactive Yellow 18 Control (without salt) Treated (without salt) Control (with salt) C.I. Reactive Red 31 Control (without salt) Treated (without salt) Control (with salt)

Shade (per cent) 0.25 per cent 1.5 per cent 4.5 per cent E (per cent) F (per cent) E (per cent) F (per cent) E (per cent) F (per cent) 31.2 79.4 60.6

68.2 95.0 76.7

29.4 62.9 62.1

82.8 90.4 82.1

25.6 63.8 65.2

84.3 94.5 87.7

25.1 66.2 58.7

80.4 97.3 80.0

23.1 59.6 60.9

82.0 84.3 81.0

21.5 57.3 62.7

79.7 90.6 94.5

26.4 77.4 52.7

70.7 99.4 79.1

25.1 76.0 56.8

85.8 90.6 78.5

23.3 75.2 59.1

71.3 84.1 74.5

Table IV. Effect of modification of cotton on exhaustion and fixation of cold brand reactive dyes

Shade (per cent) 0.25 per cent 1.5 per cent 4.5 per cent E (per cent) F (per cent) E (per cent) F (per cent) E (per cent) F (per cent) 66.9 88.9 76.2

64.3 87.8 71.4

63.0 85.3 77.5

66.6 93.7 78.4

61.5 79.3 80.6

62.7 80.4 70.4

54.3 81.6 78.7

55.5 84.7 58.2

50.1 77.3 81.0

72.0 77.5 62.7

49.2 75.5 83.2

62.7 80.4 70.4

Table V. Effect of modification of cotton on exhaustion and fixation of hot brand reactive dyes

IJCST 19,2 Sample

106

Table VI. Effect of modification of cotton on exhaustion and fixation of HE reactive dyes

Shade (per cent) 0.25 per cent 1.5 per cent 4.5 per cent E (per cent) F (per cent) E (per cent) F (per cent) E (per cent) F (per cent)

C.I. Reactive Yellow 135 Control (without salt) Treated (without salt) Control (with salt) C.I. Reactive Red 152 Control (without salt) Treated (without salt) Control (with salt) C.I. Reactive Blue 172 Control (without salt) Treated (without salt) Control (with salt)

81.0 92.1 85.1

85.4 94.2 88.0

78.7 90.3 91.7

87.6 93.7 94.0

77.1 89.1 92.5

86.2 95.1 91.2

80.6 90.5 88.2

92.0 98.4 91.4

79.0 88.2 91.3

82.1 94.0 86.7

78.1 84.6 94.0

80.6 90.6 83.9

82.1 95.0 86.7

88.3 99.9 91.6

80.0 94.5 89.9

86.2 96.7 90.1

79.6 92.7 91.8

84.0 91.1 90.2

30 g/l salt. An addition of 60 g/l salt increased the exhaustion from 21.5 to 62.7 per cent for 4.5 per cent depth. Salt addition is thus inevitable in conventional dyeing. The treated samples, on the other hand, caused remarkable improvements in dye uptake in the absence of salt. For Yellow 22 and Red 2 the exhaustions of the treated samples for 0.25 and 1.5 per cent depth were equivalent to the control samples dyed using recommended salt concentrations. In case of dark shades like 4.5 per cent the treated samples were slightly lighter to the control samples dyed with salt. But in case of Blue 4 the treated samples for all the concentrations were quite darker than their control counterparts dyed using salt. The pretreatment could improve the dye fixation also, to a significant extent. Salt-free dyeing of cotton materials with cold brand dyes was thus made possible with such pretreatment. The improved dye uptake of treated materials may be attributed to the generation of positive charge as shown in Figure 1. The results of hot brand dyes are presented in Table V which indicates that the dye exhaustions of control samples were better than the former classes even in the absence of salt. Addition of salt caused further improvement. The Yellow 18 was exhausted more on the modified cotton for both 0.25 and 1.5 per cent depth whereas for darker shade like 4.5 per cent it was exhausted almost to the same extent as that on the untreated sample in presence of salt. The exhaustion of Red 31 was also improved on modified cotton but the extents of improvement for the medium and dark shades were found to be lower than that achieved on the untreated control samples in presence of salt. The salt requirements of this class of dyes were also very high. The fixation of dyes was also found to be improved greatly on the modified cotton. The exhaustions of HE dyes on control sample were reasonably good (Table VI) even in the absence of salt. But a high concentration of salt is generally recommended to avail the excellent exhaustion characteristic with this class of dyes. The Blue 172 was exhausted very well on the modified cotton for all the concentrations and secured higher exhaustion values compared to the untreated sample dyed using recommended salt concentration. For lighter shade like 0.25 per cent, the exhaustions of Yellow 135 and Red 152 on modified cotton were equivalent to the control dyed with salt whereas for higher depths control showed little better exhaustion.

Alkali addition, during fixation, encourages the reaction of the dye with the fibre or with the water present in the system. The rate of the reaction depends on the reactivity of the dye. The dyes exhausted by the fibre preferentially react with the fibre, therefore higher exhaustion results in higher degree of fixation also. The improvement in fixation was also reflected in the washing fastness ratings shown in Table VII. Such treatment was, therefore, adequate in screening the negative surface charge of the fibre and thus enabling appreciable exhaustion and fixation of reactive dye even in the absence of salt.

Salt-free reactive dyeing of cotton

107

Conclusions Cotton cellulose was treated with DCE followed by methyl amine to introduce free amino groups in the cellulose skeleton. The free amino groups get protonated in presence of acids. Therefore, dye molecules are attracted towards the fibre when dyed in slightly acidic bath which reduces the requirement of salt during dyeing. The dyeing behaviour of the treated cotton especially in the replacement of salt in dyeing with various fibre reactive dyes, was studied and compared with the untreated control sample. The investigation was done with three types of reactive dyes, viz. cold brand, hot brand and HE reactive dyes for three depths of shade namely light (0.5 per cent), medium (2 per cent) and dark (4.5 per cent). The modified cotton showed excellent dye exhaustion results for all the dyes in the absence of salt. The effect was more prominent in case of cold and hot brand dyes. In the absence of salt untreated cotton sample, in general, exhibited very low exhaustion compared to the corresponding treated samples. The difference in such exhaustion values was in the following order cold brand . hot brand . HE dyes. For all the depths, exhaustion of the treated sample dyed with C.I. Reactive Blue 4 and C.I. Reactive Blue 172, in absence of salt, was found higher to the corresponding exhaustion values of the control samples dyed in presence of salt. Other dyes also showed a similar effect for light and medium shades. In case of HE dyes the difference in the dye exhaustion values, of untreated and treated samples, in absence of salt, was found to be relatively less compared to that found with cold and hot brand dyes. The modification caused substantial improvement in dye fixation also for all the dyes used in this study. The results of this work thus revealed that cotton garments after such pre-treatment can be dyed with environmentally friendly and most popular fibre reactive dyes without any use of salt during its dyeing and thus can cut down the dye effluent load.

Dye (C.I. Reactive) Yellow 22 Red 2 Blue 4 Yellow 18 Red 31 Yellow 135 Red 152 Blue 172

Control (with salt) 4 3-4 3-4 4 3-4 3-4 4 3-4

Washing fastness Treated (without salt) 5 4 4-5 4-5 4 4 5 4-5

Table VII. Effect of modification on washing fastness of cotton dyed with reactive dyes (1.5 per cent shade)

IJCST 19,2

108

References Earle, R.H., Saunders, R.H. and Kangas, R.L. (1971) paper presented at Applied Polymer Symposium No. 18, p. 707. El-kharadly, E.A., Shakra, S.S., Youssef, B.M. and Youssef, Y.A. (1996), J Fibre Text Res, Vol. 21, p. 35. Gamble, G.R. (2005), Journal of Cotton Science, Vol. 9 No. 1, pp. 56-64. Gohl, E.P.G. and Vilensky, L.D. (1983), Textile Science, 2nd ed., Longman, Cheshire. Hequet, E. and Abidi, N. (2001), Proceedings of the Beltwide Cotton Conference,Vol. 2, pp. 1247-50. Junger, O., Schafer, H. and Hocker, H. (1999), Melliand Int, Vol. 1, p. 78. Kane, D.E. (1953), “A study of the relationship between dielectric constant and accessibility of cellulose”, PhD thesis, The Institute of Paper Chemistry, Appleton, Wisconsin, June. Lewis, D.M. and Lei, X.P. (1992) paper presented at AATCC Internat. Conf. and Exhib. Book of Papers, Atlanta, p. 259. Lim, S.H. and Hudson, S.M. (2004), Color. Technol., Vol. 120, p. 108. Rupin, M. (1976), Text. Chem. Colorist, Vol. 8, p. 139. Rupin, M., Veaute, G. and Balland, J. (1970), Textilveredlung, Vol. 5, p. 829. Schlack, P. (1938), USP 2131120. Wu, T.S. and Chen, K.M. (1992), J Soc Dyers Colour, Vol. 108, p. 388. Wu, T.S. and Chen, K.M. (1993), J Soc Dyers Colour, Vol. 109, p. 153. Corresponding author D.P. Chattopadhyay can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Mechanical simulation of the plain weft knitted fabrics

Plain weft knitted fabrics

Savvas G. Vassiliadis Department of Electronics, TEI Piraeus, Athens, Greece, and

109

Argyro E. Kallivretaki and Christopher G. Provatidis School of Mechanical Engineering, National Technical University of Athens, Athens, Greece

Received 27 February 2006 Accepted 22 October 2006

Abstract Purpose – The present work aims to focus on the simulation of tensile, shear and bending deformation of the plain-weft knitted fabrics in an analogous manner to the tests performed on the Kawabata Evaluation System for Fabrics. Design/methodology/approach – The simulation of the tests is based on the modelling of the fabric microstructure and the application of the boundary conditions and the equivalent loading that correspond to each mechanical test, with a respect to the contact phenomena. A three-dimensional model consisting of three bodies in contact represents the unit cell of the fabric microstructure. Finite element analysis is used for the prediction of fabric performance since the complexity of the structure, the anisotropic properties of the yarns and the interaction phenomena between the yarns at the contact areas preclude the use of analytical methods. Findings – The proper definition of the boundary conditions and the appropriate load is of great significance for the realistic simulation of the mechanical tests under examination. The results of the simulated deformations compared to the respective measurements of the laboratory tests are correlated very well and this enables the consideration of the computational analysis as a powerful textile design tool. Originality/value – The prediction of the mechanical properties of the knitted fabrics based on the computational modelling supports the estimation of the fabric hand during the design stage and before its manufacturing. Keywords Knitwear, Modelling, Finite element analysis Paper type Research paper

Introduction The low-stress mechanical tests of fabrics are performed on the Kawabata Evaluation System for Fabrics (KES-F) for the objective measurement of the mechanical properties and the estimation of the fabric hand. The tests are uniaxial tensile, shear and bending test along the course and wale direction and compression. For the final hand estimation are also necessary the measurement of surface roughness and friction coefficient as well as the measurement of fabric thickness and weight per unit area (Kawabata, 1980). The deformation process of the yarns during the low-stress mechanical tests of the knitted fabrics is a superposition of the bending, the elongation and the interaction between the yarns. The simulation of the mentioned tests is an attempt supporting the The present work has been partially supported by the “Archimedes1” EPEAEK Research Project in TEI Piraeus, co-financed by the EU (ESF) and the Greek Ministry of Education.

International Journal of Clothing Science and Technology Vol. 19 No. 2, 2007 pp. 109-130 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710725711

IJCST 19,2

110

prediction of fabric hand and in the same time mainly contributes to the understanding of the deformation process. The study of the mechanical analysis of the knitted fabrics will be based on the modelling of the fabric microstructure. The geometrical complexity of the knitted fabrics’ unit cell, the coexistence of elongation and bending of the yarns in the deformation process, the yarn profile flattening and the interaction phenomena between yarns in the contact areas have a significant influence in the mechanical analysis of the knitted fabrics. Thus, a simplified model considering the fabric as continuous homogenous membrane-like material cannot describe precisely its mechanical behaviour. The same reasons impose the use of finite element analysis (FEA) as a suitable tool for the mechanical analysis of the fabrics (Araujo et al., 2004). Another point that supports the use of FEA for the mechanical modelling of fabrics is the anisotropic properties of the yarns. The microscopic observation of the cross-section of common yarns concludes that usually approximately 50 per cent of it is covered by fibres’ cross-sections and the rest corresponds to the empty space between the fibres. Thus, an objective and realistic modelling of the fabric structure should use as first structural unit the fibres of the yarn. But a model like this would increase extremely the complexity of the structure and the calculation time for the mechanical analysis. The modelling of the yarns as homogenous cylinders simplifies the representation of the yarn structure. However, it produces realistic results only when the material of the cylinders is considered as transverse isotropic. In this case, the different elastic properties along the yarn axis direction and the radius direction can describe the non-homogenous behaviour of the yarn. This assumption imposes the use of finite elements having the capability of anisotropic properties. The simulation of the tests because of the symmetry of the structure is based on the mechanical analysis of one half of a loop, the structural unit cell. The use of the minimum possible structural unit refers to the fastest computation of the mechanical parameters, without any distortion of the final results and without limiting the generalization of them. The precise and successful modelling of even one unit cell requires a relative big number of finite elements. The nodes on the external surfaces of the yarns are responsible for the maintenance of the round shape of the yarn. If these nodes are not relatively close, sharp edges will appear and thus undesirable penetrations of the one body inside the other in the contact area will appear. The finer mesh resulting in a bigger number of finite elements is also important especially in the regions close to the contact areas to ensure the normal transfer of the contact pressure between the yarns. The use, of course, of one loop for the simulation of the real mechanical tests imposes the application of proportional loads and the respective boundary conditions. The precise geometrical modelling of fabric microstructure has a great significance for the accuracy of the simulation of the mechanical tests. The accuracy of the used geometrical model has been estimated comparing the theoretically calculated loop length of the model to the measured one (Vassiliadis et al., 2005a, b). The mean and the maximum estimated error resulted lower than the respective values that correspond to the prevailing existing models.

Method Representation of the model The main structural parameters of a single jersey fabric are: the course-spacing (c) the wale-spacing (w) and the thickness of the yarn used (D). The rest of the geometrical parameters required for the complete description of the structure derive analytically from them. The estimation of the geometrical parameters has been based on the assumption of the ideal elastic yarn. Thus, the yarns are represented as homogenous cylinders of constant diameter, with initial restricted contact area between them. Figure 1 shows the top view (a), front view (b) and side view (c) of the used geometrical model. We consider initially the independent parameters c, w, D and in addition the: distance t as it is noticed in Figure 1. It is important to mention that for a realistic design the values of parameters D, w must fulfill the constraint D # w=4. So in the case that D . w=4 it is assumed D ¼ w=4. The rest of the parameters derive from the geometric relationships mentioned below: c t D R¼ 2 2 ; 2 2 2

r¼

tan w ¼

where 0 # t # c 2 2D

n 2 2 o c 2 D2 2 2t 2 D2 þ 2t ð2DÞ

ðON Þ 2 ðODÞ ¼ R

ð2Þ

qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2 2 2 2 r þ D2 2 2c 2 r 2 D2 2 Dþt 2 R

tan v ¼

ð3Þ

0:5c r þ 0:5D

ð4Þ

c 2 D sin g D

ð5Þ

sin g ¼

k¼

ð1Þ

p w c 2 2 R · tan 2v 4 2 2

ð6Þ

The value of the parameter t, which was temporarily considered as independent, can be defined from a repeated calculation process. For every value of the parameter t, belonging in the field [0, (c-2D)], the central axis of the loop is generated and the loop length l is calculated. By considering the elastic behaviour of the yarn while forming the loop, the selected value for the parameter t is the one that corresponds to the minimum loop length l min . For the calculation purposes, the values of the distance t start from zero and in steps of magnitude depending on the accuracy required increases to its final value. Figure 2 shows an example of the loop length values l as a function of parameter t for a single jersey. The method of modelling presented applies optimisation criteria by the assumption of the elastic energy minimization of the yarns composing the relaxed fabric.

Plain weft knitted fabrics

111

IJCST 19,2

(a) k Λ

K

112 M

(b)

k

w/2

1/2

h

Λ

K ω c

M

D

y

Σ

x

(c) K ϕ

τ

Figure 1. Geometrical model of single jersey knitted fabric structure

c/2

M

y H g Σ

z

N

∆

O

The relaxation of the yarns from the remaining stresses is accompanied by the minimization of the existing deformations. Thus, the elongation of the yarns imposed by the knitting machine during the fabric production is minimized. During relaxation the adjacent yarns slide over each other until the minimum elastic energy structure will appear.

Plain weft knitted fabrics

loop length - parameter t

loop length (mm)

2.92 2.88 2.84

113

2.80 (tselected, lmin)

2.76 2.72 0.000

0.030

0.060

0.090 t (mm)

0.120

0.150

0.180 tmax

Yarn properties The yarn is considered in general as a transverse isotropic material. Its main axis of symmetry is the yarn axis. The local coordinate system xyz of the yarn is used. The z axis corresponds to the longitudinal yarn axis and the plane xy corresponds to the yarn cross-section. The elastic parameters of the yarn are assumed as having the normal values: E x ¼ 10 N=mm2

vxy ¼ 0:30

Gxy ¼ 3:85 N=mm2

E y ¼ 10 N=mm2

vyz ¼ 0:30 Gyz ¼ 5 N=mm2

E z ¼ 800 N=mm2

vxz ¼ 0:30

Gxz ¼ 5 N=mm2

The mesh The modelling and simulation of the mechanical tests of the single jersey unit cell is based on the ANSYS software. After the geometrical design a meshing phase is following. The generation of the mesh, i.e. the segmentation of the virtual bodies into small parts, is a stage of great significance for the accuracy of the simulation since it presides over the success of the load transfer, the stress and strain distribution. Usually the definition of the type of finite elements, the mesh density and the pattern of the mesh arises from the experience obtained by a series of tests. Practically the role of the mesh is to distribute evenly the stresses and strains among the finite elements. A fine mesh provides obviously a more even distribution than a coarser one. On the other hand, the finer the mesh the higher the computational time is required for the solution. The proposed approach is based on the creation of a mesh with approximately 10,000 finite elements of different sizes. The smaller elements are placed in the high stress areas. The ANSYS meshing tool permits the creation of a free or mapped mesh. Choosing the mapped option the created mesh is restricted in terms of the element shape that contains and the pattern of the mesh. The mapped mesh typically has a regular form, with obvious rows and columns of elements. The used mesh consists of hexahedral solid elements. The elements are defined by eight nodes with three degrees of freedom

Figure 2. Selection method for the parameter t for a single jersey (c ¼ 0.50800, w ¼ 0.84667, D ¼ 0.17003)

IJCST 19,2

114

at each node: the translations in the nodal x, y, and z directions. The orthotropic properties of the element are activated in order to approach the real mechanical behaviour of the yarn. The element coordinate system is shown in Figure 3 is used for the definition of the orthotropic properties. The geometrical model of the unit cell is built as a group of volumes as shown in Figure 4 in order one edge of each volume to be lying at the central axis of the yarn. The element size is defined using the meshing tool of ANSYS and the geometrical model is meshed with the mapped option. The volume of the produced finite elements belong in a range of 1.2 £ 102 6-1.2 £ 102 5 mm3. The z axis of the element coordinate system is parallel to the central axis of the yarn. Thus, the orthotropic elastic properties of the material can be used (Figure 5).

O

P M N

Figure 3. Element geometry

Figure 4. Creation of volumes

ELEMENT COORDINATE SYSTEM

K

L Z I

Y X

J

Plain weft knitted fabrics

115

Figure 5. Meshed unit cell

Contact analysis A special care has been taken for the support of the contact analysis. The contact often presents significant difficulties, such as high localised nonlinearity, inaccurate information about the contact regions, penetration between the bodies in contact and sliding of one body over the other with simultaneous transfer of the initial contact region. If the mentioned phenomena will not be prevented, they will cause convergence problems during the solution process. The first step for the contact analysis as using the ANSYS code is the definition of the surfaces of the bodies where it is expected that the contact will occur. The two surfaces, one of each body, specify a contact pair. The proposed model includes at least two contact pairs, since three bodies are in contact. The exact size of the surfaces where the contact phenomena will appear is initially unknown and sometimes is moved due to the sliding of the bodies. Thus, initially wider contact surfaces must be declared. The exact surfaces will be known after the preliminary simulation of the mechanical test. Also in addition to the predicted contact areas, some others can be observed during the mechanical test. For example, the bodies 2, 3 (Figure 6) will come into contact during the bending test along course direction when the fabric is bent with the technical face body 2

body 3

Figure 6. The mesh close to contact area body 1

IJCST 19,2

116

inside. The successful definition of the contact pairs is necessary in order the convergence conflicts to be avoided. The second step is the placement of the contact and target elements on the surfaces constituting the contact pair. These elements are responsible for the load transfer between the contact areas, the friction forces application and the definition of the contact state (sticking or sliding). In the current case of deformable bodies in contact the surface of one body is arbitrarily taken as a contact surface and the surface of the other body as a target surface. The “contact-target” pair concept has been widely used in finite element simulations. The contact and target elements are considered as an overlay on the solid elements of the contact and target surface, respectively. Hence, the target elements are the geometric entities that sense and respond to the movements of the adjacent contact elements. Corresponding conjugated nodes have been foreseen in the contact areas, on the two bodies surfaces (Figure 6). The mesh generation methodology including the previously described two steps has been found to be the optimal for the convergence of the computational FEM algorithm and the higher precision of the results. The penetration problems of the solution process result from two main factors. Considerable initial penetration between the bodies can be introduced after the meshing by numerical round-off, even if the solid model is built in a just-touching state. The definition of exact and precise initial contact conditions is the most important aspect of building a contact analysis model. Penetration between the bodies can also be introduced during the computational analysis, when the mesh in the contact area is not fine enough for the even distribution of the contact pressure. In such cases, the contact elements may overestimate the contact forces, resulting in non-convergence of the algorithm or in breaking-away of the components in contact. Although the current modelling concerns low-stress mechanical tests, the sliding of one yarn towards to the other in contact is a quite intense phenomenon. Even in the case that the sliding of the yarns is not obvious, their elongation is followed by the nodal movements which in the contact region correspond to the sliding of the yarns. The ANSYS code provides the options of node to node, node to surface and surface to surface contact analysis. The mentioned different types express the kind of the collaboration between the contact and target areas. The experience resulted from a series of tests showed the surface to surface contact analysis as the optimal one especially when sliding is included in the contact phenomena. Boundary conditions and loading The current models simulate the conditions of the low-stress mechanical tests of the fabrics as they are performed on the KES-F. The realistic definition of the loads applied on the model and the boundary conditions concerned support the success of the simulation of the mechanical tests. The linear distribution of the load f (gf/cm) in the cases of the computational simulation and the corresponding experimental test must be of the same value. Thus, the load applied in the unit cell for the simulation of a mechanical test results from the multiplication of the linear distribution of the load by the length of the loaded edge of the unit cell. The difficulty in the proper definition of boundary conditions arises from the fact that the simulation of the tests is using as model a structural entity smaller than the

total sample used for the real tests. Generally, the boundary conditions derive from the observation of the deformation process of the mechanical test. The realistic boundary conditions are a series of restrictions in the nodal degrees of freedom so that the unit cell will be deformed according to the simulated test. The infinite displacement of the unit cell in the vertical direction to the plane of loading is avoided by the boundary conditions. The modelling of only one unit cell of the sample imposes the introduction of boundary conditions for the maintenance of the continuum of the adjacent unit cells. Thus, a virtual placement of the adjacent deformed loops produces the total deformed sample. The description of the applied boundary conditions and load is referenced to the global coordinate system and the areas numbering of Figure 7. The nodal displacements are represented as ux, uy, uz along the x, y, z axis, respectively.

Plain weft knitted fabrics

117

Tensile test The tensile test using the KES-F is performed on an orthogonal sample of dimensions 20 £ 2.5 cm. The sample is mounted between two parallel chucks of 20 cm length and is pretensioned by applying a weight of 20 gf. The one chuck is clamped and the other is moving kept parallel to the clamped chuck in order to impart the tensile deformation, i.e. the elongation to the sample. The deformation is considered as strip biaxial and affects the small dimension of the sample. The elongation rate is constant and equal to 0.1 mm/s. When the tensile load reaches the value of fm ¼ 50 gf/cm the imposed deformation stops and the reverse procedure begins. Load and boundary conditions of the tensile test along the wale direction. The load F is applied along the y direction on the areas A1 and A2 . The opposite load ( –F) is applied on the areas A3 , A4 . The value of load is calculated from the linear load f, FðgfÞ ¼ f ðgf=cmÞ £ wðcmÞ. The symmetry in terms of geometry and load imposes the concurrent displacement of the nodes lying on the areas A1 and A2 along y direction: uyðA1 Þ ¼ uyðA2 Þ

A1

A2

y

x

z A5

A6

A3

A4

Figure 7. Numbering of areas and global coordinate system

IJCST 19,2

The same is considered for the nodes of A3 and A4 areas: uyðA3 Þ ¼ uyðA4 Þ Owing to the symmetry, the displacements also fulfill the equation:

118

uyðA1 ; A2 Þ ¼ 2uyðA3 ; A4 Þ The nodes lying on the pair of areas (A1, A3) or (A2, A4) located by the same coordinates x, z present the same value of displacements ux and uz. The last consideration supports the maintenance of the continuum of the deformed unit cell with the adjacent ones and it can be described by node-to-node constraint equations (node-node CE). The displacements of the nodes belonging to the A5 ; A6 areas along the x axis are set equal to zero, since the tensile test is considered as strip biaxial. The indefinite displacement of the unit cell along z direction is avoided by the restriction uz ¼ 0 for the nodes that are shown in Figure 8. Load and boundary conditions of the tensile test along the course direction. The load F is applied along the x direction on the area A6 and the opposite load ( –F) is applied on the area A5 . The value of load is calculated from the linear load f, FðgfÞ ¼ f ðgf=cmÞ £ cðcmÞ. The symmetry of the geometry and the load imposes the reverse displacement of the nodes lying on the areas A5 and A6 along x direction: uxðA5 Þ ¼ 2uxðA6 Þ The nodes lying on the pair of areas ðA1 ; A3 Þor ðA2 ; A4 Þ located by the same coordinates x, z present the same value of displacements ux and uz (node-node CE). The last consideration supports the maintenance of the continuum of the deformed unit cell with the adjacent ones. The displacements of the nodes belonging to the areas A1 , A2 , A3 and A4 along the y axis are set equal to zero, since the tensile test is considered as strip biaxial. The indefinite displacement of the unit cell along z direction is avoided by the restriction uz ¼ 0 for the nodes that are shown in Figure 9.

Figure 8. uz ¼ 0

Plain weft knitted fabrics

119

Figure 9. uz ¼ 0

Shear test In the shear test the fabric sample subjected to deformation has dimensions of 20 £ 5 cm. The shearing is executed along the course and wale direction. The sample is mounted exactly like in the tensile test. The one chuck is restricted to move in the longitudinal sense and the other is moving parallel to it. A pretension of lw ¼ 10 gf/cm is applied along the wale direction for the coursewise shearing and of lc ¼ 10 gf/cm along the course direction for the walewise shearing. When the shear angle achieves the value of 88 the shearing deformation is completed and the chuck moving along the shearing direction starts the reverse movement. The moving chuck introduces a shearing angle of 88 in the reverse direction and thereafter the comeback starts. Thus, the shearing test is executed along the two directions. Since, the loop configuration is considered symmetric for the modelling (the yarns are ideally elastic) the simulated test is executed in one direction. Load and boundary conditions of the shear test along the wale direction. The simulation of shearing is constituted by two load steps. The first load step corresponds to the imposition of the load applied along the course direction before the starting of shearing. The second load step introduces the shearing in the deformed by the first load step model. The first load step corresponds to the application of the load: F x ðgfÞ ¼ l c ðgf=cmÞ · cðcmÞ on the area A6 along the course direction. Moreover, the displacements ux of the nodes lying on the area A5 are set equal to zero. The rest boundary conditions are applied as the ones of the tensile test along the course direction. The second load step retains the load F x and the resulted deformation. A load F y applied on the area A6 corresponds to the shearing force. The displacements ux and uy along x and y directions for the nodes of the area A5 are set equal to zero. The nodes lying on the areas ðA1 ; A3 Þ and ðA2 ; A4 Þ located by the same coordinates x, z present the same value of displacements ux, uz, uy (node-node CE). The last consideration supports the maintenance of the continuum of the deformed unit cell with the adjacent

IJCST 19,2

ones. The indefinite displacement of the unit cell in z direction is avoided by the restriction uz ¼ 0 for the nodes that are shown in Figure 9. Load and boundary conditions of the shear test along the course direction. For the first load step the load: F y ðgfÞ ¼ 2l w ðgf=cmÞ · wðcmÞ

120

is applied on the areas A3 and A4 along the y direction while the displacements uy of the nodes lying on the areas A1 and A2 are set equal to zero. The rest boundary conditions are applied as the ones of the tensile test along the wale direction. The load F y and the resulted deformation remain in the second load step. A load F x is also applied on the areas A3 , A4 along x direction and the opposite load 2F x is applied on the areas A1 and A2 . The zero displacement uy along y direction also remains for the nodes of the areas A1 and A2 . The nodes lying on the two areas ðA5 ; A6 Þ located by the same coordinates y, z present the same value of displacements ux, uy, uz (node-node CE). The nodes lying on the pairs of areas ðA1 ; A3 Þ or ðA2 ; A4 Þ located by the same coordinates x and z present the same value of displacements uz and the opposite value of displacements ux (node-node CE). The indefinite displacement of the unit cell along z direction is avoided by the restriction uz ¼ 0 for the nodes that are shown in Figure 8. Bending test The bending test is performed on an orthogonal sample of dimensions 20 £ 1 cm. The sample is mounted between two parallel chucks of 20 cm length. The one chuck is clamped (O position) and the other is moving in the locus A-B as it is shown in Figure 10. We consider the Cartesian coordinate system of the fabric with the plane xy lying on the plane of the fabric. The x, y axes correspond to the course and wale direction, respectively. Then the motion of the second chuck is described by the functions of curvature, K: A 1 l1

j

y or x (cm)

l0

Figure 10. Locus of moving end of the specimen and tangential angle of KES-F2 for bending deformation (sample length ¼ 1 cm)

l10

O

z (cm)

B

z¼ y¼

1 2 cos K K

ð7Þ

sin K sin K or x ¼ K K p f¼ 2K 2

Plain weft knitted fabrics

ð8Þ ð9Þ

121

The lines l 0 ; l 1 . . .l 10 correspond to the projection of the bent sample in the vertical plane to the plane of the fabric for K ¼ 0, 0.25 . . . 2.5 cm2 1, respectively. The function y(K) of equation (8) is used when the fabric is mounted with the courses parallel to the chucks, thus the fabric is bent along the wale direction. Contrarily the function x(K) is used for the bending test along the course direction, mounting the sample with the wales parallel to the chucks. The curvature varies between the values K ¼ 2 2.5 and 2.5 cm – 1 with a constant rate of 0.50 cm2 1/s. The curvature is considered as positive when the fabric is bent with the technical face inside, according to the placement of fabric during the experiment. Load and boundary conditions of the bending test along the wale direction. Two unit cells are placed along the wale direction for the configuration of a model (Figure 11) of dimensions ð2cÞ · w for the simulation of the bending test along the wale direction. Since, the projection of the sample during the bending test in the yz plane is known, the displacement of the model can be calculated. Lines of length 2c are lying on the lines of projection (l 0 ; l 1 . . .l 10 ) with starting point the position O in order the central axis of the bent sample for different values of curvature K to be produced. Thus, the ending point of each produced line defines the displacement of the bent model. The model is bent considering the areas B3 , B4 as clamped and the application of displacements on the central nodes of the areas B1 , B2 . Thus, the displacements along y B1

B2

B6

B5

B8

B7

B3

B4

Figure 11. Structural model for the bending test along the wale direction

IJCST 19,2

122

and z axes of the nodes lying on the areas B3 , B4 are set equal to zero. The displacements of the nodes lying on the areas B5 , B6 , B7 , B8 along the x axis are considered equal to zero since the bending is executed along the small dimension of the sample. The bending moment is calculated on the clamped areas (B3 , B4 ). The use of two unit cells along the wale direction for the walewise bending test aims to the prediction of the pressure developed by the contact of the adjacent courses. Load and boundary conditions of the bending test along the course direction. Two unit cells are placed along the course direction for the configuration of a model (Figure 12) of dimensions c · ð2wÞ for the simulation of the bending test along the course direction. The displacement of the model during the bending test is calculated considering a line of length 2w lying on the lines of projection of the bent sample at the xz plane with starting point the position O. The model is bent considering the area C 9 as clamped and applying the displacements on the central node of the area C 10 . Thus, the displacements along the x and z axes of the nodes of the areas C 9 are set equal to zero. The displacements along the y axis of the nodes lying on the areas C 1 , C 2 , C 3 , C 4 , C 5 , C 6 , C 7 , C 8 are also set equal to zero. The bending moment is calculated on the clamped area ðC 9 Þ. The use of two unit cells along course direction for the coursewise bending test aims to the prediction of the pressure developed by the contact of the adjacent wales. Results The accuracy of the simulated tests is estimated by comparing the curves produced from the KES-F during the tests to the predicted ones by the mechanical simulation. In the current paper the evaluation is based on the modelling of four fabrics of different structural parameters. The elastic yarn properties are the same in the four samples. The parameters C (courses/in.) and W (wales/in.) and yarn count (tex) of the four samples are measured in laboratory. The yarn thickness D (mm) is calculated from the yarn count using experimental data. The values are given in Table I. C1

C3

C4

C9

Figure 12. Structural model for the bending test along the course direction

C10

C5

Sample Table I. Structural parameters of used samples

C2

1 2 3 4

C6

C8

C7

C (courses/in.)

W (wales/in.)

Stitch density C· W

Yarn count (tex)

D (mm)

48.4 52.3 41.4 52.3

37.5 30.5 27.6 37.0

1815.0 1595.2 1142.6 1935.1

19.7 19.7 35.4 20.0

0.1845 0.1845 0.2657 0.1858

Figures 13-18 show a typical model in its initial and deformed state for each test. Figures 19-24 show the predicted (dashed line) and experimental curves (continuous line for Figures 19-23 and open circles for Figure 24). Conclusions For the quantitative comparison of the experimental curves produced from the KES-F during the tests to the predicted ones are calculated the physical parameters resulting from both curves. Thus, the tensile energy produced by the maximum applied load is calculated for the estimation of the accuracy of the simulated tensile test. The tensile energy corresponds to the area formed by the curve, the axis of elongation and the

Plain weft knitted fabrics

123

Figure 13. Typical deformed model by the walewise tensile test

Figure 14. Typical deformed model by the coursewise tensile test

IJCST 19,2

124

Figure 15. Typical deformed model by the walewise shear test

Figure 16. Typical deformed model by the coursewise shear test

straight line vertical to the axis of elongation for the maximum value of elongation. The shear stiffness G is calculated for the evaluation of the shear test simulation. The value of shear stiffness corresponds to the slope of the curve between w ¼ 0.5 and 2.58. The bending rigidity B, is calculated from the mean slope of the curve in the range K ¼ 0.5 , 1.5 cm2 1 and K ¼ 2 0.5 , 2 1.5 cm2 1. The mean value of the bending rigidity is calculated for the evaluation of the simulation of the bending test. The predicted and the experimental values of the constants are presented in Tables II-IV.

Plain weft knitted fabrics

125

Figure 17. Typical deformed model by the walewise bending test

Figure 18. Typical deformed model by the coursewise bending test

IJCST 19,2

Sample 2 60 50 40 30 20 10 0

load (gf/cm)

load (gf/cm)

126

Sample 1 60 50 40 30 20 10 0 0

4

8

0

12

4

60 50 40 30 20 10 0 0

4

8

0

12

4

Sample 2 load (gf/cm)

load (gf/cm)

Sample 1

10 strain (%)

15

20

60 50 40 30 20 10 0 0

5

0

5

10 strain (%)

10 strain (%)

15

20

15

20

Sample 4

60 50 40 30 20 10 0

load (gf/cm)

load (gf/cm)

Sample 3

Figure 20. Load – elongation curves of coursewise tensile test for the predicted and the experimental data

12

strain (%)

60 50 40 30 20 10 0 5

8

60 50 40 30 20 10 0

strain (%)

0

12

Sample 4

load (gf/cm)

load (gf/cm)

Sample 3

Figure 19. Load – elongation curves of walewise tensile test for the predicted and the experimental data

8 strain (%)

strain (%)

15

20

60 50 40 30 20 10 0 0

5

10 strain (%)

8

8

6 4 2

6 4

3

6

127

2

0 0

0

9

0

3

angle (deg.)

8

8

load (gf/cm)

10

load (gf/cm)

9

Sample 4

10

6 4

6 C

4

Figure 21. Load – shear angle curves of walewise shear test for the predicted and the experimental data

2

2 0

0 0

3

6

9

0

angle (deg.)

3 6 angle (deg.)

Sample 1

9

Sample 2

10

10

8

8

load (gf/cm)

load (gf/cm)

6 angle (deg.)

Sample 3

6 4 2

6 4 2 0

0 0

3

6

0

9

3

6

9

6

Figure 22. Load – shear angle curves of coursewise shear test for the predicted and the 9 experimental data

angle (deg.)

angle (deg.) Sample 3

Sample 4

10

10

8

8

load (gf/cm)

load (gf/cm)

Plain weft knitted fabrics

Sample 2 10 load (gf/cm)

load (gf/cm)

Sample 1 10

6 4 2

6 4 2 0

0 0

3

6 angle (deg.)

9

0

3 angle (deg.)

IJCST 19,2

Sample 1

Sample 2 0.15 M (gf.cm/cm)

128

M (gf.cm/cm)

0.10 0.05 −3

−2

0.00 0 −1 −0.05

1

2

3

0.10 0.05 −3

−2

Sample 3

M (gf.cm/cm)

2

3

Sample 2

0.06

0.06

0.04

0.04

–1 0 –0.02

1

2

3

–3

–2

–0.06 K (1/cm)

Sample 3

Sample 4

0.00 –1 0 –0.05 –0.10 K (1/cm)

1

2

3

M (gf.cm/cm)

M (gf.cm/cm)

–2

2

3

1

2

3

1

2

–0.04

0.10

–3

0.00 –1–0.02 0

–0.04 K (1/cm)

0.05

1

0.02

0.15

Figure 24. Bending moment – curvature curves of coursewise bending test for the predicted and the experimental data

, 0 −1−0.05 −0.10

Sample 1

0.00 –2

−2

−0.15 −0.20 K (1/cm)

0.02 –3

0.00 −3

−0.30 K (1/cm)

M (gf.cm/cm)

M (gf.cm/cm)

1

−0.20

M (gf.cm/cm)

0.10 0.05

0.10 0.00 0 −1 −0.10

3

Sample 4

0.20

−2

2

−0.15 K (1/cm)

−0.15 K (1/cm)

−3

1

−0.10

−0.10

Figure 23. Bending moment – curvature curves of walewise bending test for the predicted and the experimental data

−0.00 0 −1 −0.05

–3

–2

0.09 0.07 0.05 0.03 0.01 –0.01 –1 0 –0.03 –0.05 K (1/cm)

3

Tensile test

Sample

Coursewise

1 2 3 4 1 2 3 4

Walewise

Shearing test Coursewise

Walewise

Sample 1 2 3 4 1 2 3 4

Bending test

Sample

Coursewise

1 2 3 4 1 2 3 4

Walewise

Tensile energy (gf/cm) Experimental Predicted 384.1 320.5 290.8 295.4 201.4 218.3 166.7 180.2

402.6 321.4 300.5 346.1 202.7 212.0 162.6 191.7

Plain weft knitted fabrics

129 Table II. Evaluation of simulation of tensile test

Shear stiffness [gf/(cm · deg)] Experimental Predicted 0.930 0.820 1.100 1.300 0.750 0.600 1.000 0.900

1.056 0.864 1.197 0.978 0.626 0.653 1.233 1.008

Table III. Evaluation of simulation of shear test

Bending rigidity (gf.cm2/cm) Experimental Predicted 0.0163 0.0132 0.0315 0.0207 0.0273 0.0342 0.0512 0.0346

0.0137 0.0171 0.0350 0.0225 0.0263 0.0391 0.0603 0.0427

References Araujo, M., Fangueiro, R. and Hong, H. (2004), “Modelling and simulation of the mechanical behaviour of weft-knitted fabrics for technical applications, Part III: 2-D hexagonal FEA model with non-linear truss elements”, AUTEX Research Journal, Vol. 4 No. 1, pp. 25-32. Kawabata, S. (1980), The Standardization and Analysis of Hand Evaluation, The Textile Machinery Society of Japan, Osaka. Vassiliadis, S., Kallivretaki, A. and Provatidis, Ch. (2005a), “Estimation of the geometrical structure of the plain weft knitted fabrics for use in numerical modelling”, Proceedings of 5th World Textile Conference AUTEX 2005, Portorozˇ, Slovenia, 27-29 June, pp. 412-7.

Table IV. Evaluation of simulation of bending test

IJCST 19,2

130

Vassiliadis, S., Gurkan, P., Kallivretaki, A. and Provatidis, Ch. (2005b), “Fabric hand: an important quality parameter of the textile fabrics”, paper presented at the 2nd International Scientific Conference in Information Technology and Quality, Spetses, 4-5 June. Further reading Araujo, M., Fangueiro, R. and Hong, H. (2003), “Modelling and simulation of the mechanical behaviour of weft-knitted fabrics for technical applications, Part II: 3-D model based on the elastica theory”, AUTEX Research Journal, Vol. 3 No. 4, pp. 166-72. Kawabata, S. (1989), “Knitted fabrics”, in Chou, T.W. and Ko, F.K. (Eds), Textile Structural Composites, Composite Materials Series 3, Elsevier Science Publishers, Amsterdam, pp. 99-114. Munden, D.L. (1959), “The geometry and dimensional properties of plain-knit fabrics”, Journal of the Textile Institute, Vol. 50 No. 3, pp. T448-51. Corresponding author Savvas G. Vassiliadis can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Fit evaluation within the made-to-measure process

Fit evaluation

Chin-Man Chen Department of Fashion Merchandising, Shih Chien University, Taiwan, Republic of China Abstract

131 Received 31 March 2006 Revised 21 September 2006 Accepted 21 September 2006

Purpose – The purpose of this study is to evaluate fit of the basic garments made for Taiwanese female students with various figure characteristics. The basic garments are produced according to patterns derived from the PDS 2000 and APDS-3D systems. Design/methodology/approach – This study recruited ten Taiwanese female subjects who represented various figure characteristics. After scanning each subject, the body measurements with additional functional ease were manually entered into the APDS-3D system accompanied with the PDS 2000 system to generate the block patterns. These patterns were used to make basic garments worn on the subjects for fit evaluations. T-test and one-way ANOVA were employed to investigate if any statistically significant differences between figure characteristics of subjects exist. Findings – After statistical analysis, results showed that the percentage of tolerance allowed by the system in preventing incorrect measurements has to be revised and more measurements have to be included into the APDS-3D system. Furthermore, female students who exhibit multiple figure variations complicate fitting problems. For example, sloped-shoulder subjects with narrow shoulders and forward stance generate the problem of extra fabric gathering at the shoulder tips as well as looseness at the upper chest. Therefore, figure variations have to be analyzed in a future study. Research limitations/implications – The convenient sample with limited size does not allow generalization of figure variations associated with fit problems in all colleges or universities located in Taiwan. Originality/value – Few researchers have analyzed fit problems on garments made for females with figure variations, but none of them use 3D body scanners in combination with computer-aided design systems to test fit on basic garments for females with various physical characteristics. Keywords Clothing, Women, Computer aided design Paper type Research paper

Introduction Fit problems have plagued the women’s apparel industry for a long time. Kurt Salmon Associates, a consultant company, has found that more than 50 percent of consumers are unable to find well-fitting garments. Also, research shows that fit problems are the primary factors causing customers to return apparel products they purchase from catalog to web site retailers (Anderson, 2004). The fit issue has been so critical in influencing sales that some companies apply technological tools to improve apparel fit. Manufacturers and retailers install 3D body scanners and computer-aided design (CAD) systems to assist a made-to-measure process. The made-to-measure process is aided with a 3D body scanner that extracts body measurements without physically touching human bodies. Then, the extracted measurements are entered into alteration systems of CAD where existing patterns are stored. Finally, existing patterns are selected and modified to be individual patterns (Istook, 2002). Using a 3D scanning system accurately transfers body shape and measurements to CAD systems and

International Journal of Clothing Science and Technology Vol. 19 No. 2, 2007 pp. 131-144 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710725720

IJCST 19,2

132

efficiently helps improve the quality of fit, according to the ASTM (2005). Indeed, using a 3D scanner is more effective in analyzing body shape and evaluating fit than the photographs or videotapes that have been used to for decades (Ashdown et al., 2004). Although research indicates that the 3D scanner is an efficient tool to help improve fit, apparel firms, such as Levi-Strauss and Brooks Brothers, have conducted the customized process but not successfully turned a profit. The high rate of customer returns could be a result of fit problems (Lee et al., 2002). Most fit problems occur due to figure variations in body contour, posture, and proportion (Kwong, 2004). Few researchers have analyzed fit problems on garments made for females with figure variations, but none of them use 3D body scanners in combination with CAD systems to test fit on basic garments for females with various figure characteristics. This study applies a 3D body scanner and a CAD system to produce block patterns for ten Taiwanese female students, as diverse apparel technologies have been promoted in Taiwan. Specifically, these female students who are scanned exhibit figure variations related to proportion between shoulder and hip, shoulder and chest shape, curve of back, and posture. Block patterns altered to fit the females are cut and used to produce basic garments. Fit evaluation on the basic garments is conducted, and results and suggestions are provided. Purpose of the study The purpose of this study is to evaluate fit of the basic garments made for Taiwanese female students with various figure characteristics. The basic garments are produced according to patterns derived from the PDS 2000 and APDS-3D systems. Review of literature Apparel fit Apparel fit is defined as the relationship between the size and contour of the garment and those of the human body. Fit problems arise as an incongruent relationship between the garment and the human body occurs (Huckabay, 1992). Whether the garment conforms to the body depends on five criteria; they are ease, line, grain, balance, and set. Ease, including functional ease and styling ease, is the dimensional difference between the garment and the individual wearing the garment. Functional ease is considered as the amount of fabric allowed for body movement, and styling ease is defined as the amount of fabric needed to demonstrate the design of the garment. The second criterion is line associated with seams of garments. A well-fitting garment presents vertical seams perpendicular to the floor and parallel to the body center when it is worn on the body. The third criterion is grain focusing on the relationships among fabric, patterns, and wearers. Grain lines of garments should be parallel or perpendicular to the floor as well. The fourth criterion is balance. A balanced jacket evenly hangs on the body, so the distance from the right side of the body to the center is the same as the distance from the left side to the center. The last criterion is set which refers to the absence of wrinkles on garments (Erwin and Kinchen, 1974). Fit evaluation is a complicated process in which the relationship of the garment to the body is analyzed based upon certain criteria. Fit evaluation determined by individuals wearing garments tends to be subjective. Two individuals who have the same body measurements may perceive clothing fit differently (Alexander et al., 2005). On the other hand, fit evaluation conducted by expert panels is more objective.

Trained judges follow criteria to analyze fit, and they discuss and negotiate rating scales to reach consensus before final decisions are made. Therefore, the fit analysis by trained judges is believed to provide reliable and valid data (Ashdown et al., 2004). Self-reported body types and fit problems Fit evaluation can be achieved by either general individuals or expert panels. Few researchers have investigated fit problems perceived by individuals with various body types. Alexander et al. (2005) studied fit problems of females with four body shapes: pear, hourglass, rectangular, and inverted triangular. Their study acquired 223 questionnaires from female ages 18 to 29 years. They found that the participants who identified their body shapes as rectangular, pear and hourglass were more likely to express fit problems at the bust area than those who perceived themselves as the inverted triangular shape (Alexander et al., 2005). In contrast to Alexander et al. (2005) who studied females aged between 18 and 29, Kind (1995) focused on fit preference of college females on ready-to-wear garments. Kind (1995) used the height measurements of 358 college females to classify their body types into petite, average, tall, and large sizes. Kind (1995) surveyed their fit perceptions and found that the college females, with exception of the average group, were dissatisfied with pant length (Kind, 1995). Rather than focusing on fit perceptions of general college females, Feather et al. (1997) collected data from 503 female collegiate basketball players from two southeastern states. These researchers investigated self-perception of body type and fit problems of uniforms and street wear. They reported that almost 50 percent of the players identified their body types as ecto-mesomorph. More than 25 percent of the participants perceived themselves as ectomorphs. The players rated the sites of hips, crotch of pants and buttocks on both uniforms and street wear the lowest satisfaction scores (Feather et al., 1997). Body-type classification and fit evaluation The body-type classification of ecto-, meso- and endo-morphs is originated from the study of Sheldon (1954). Sheldon (1954) and his colleagues excluded subjects with extreme figure variations from their well-known study. Human subjects who had irregular development of bone and muscle and imbalanced deposits of fat in their bodies were not recruited. By applying a somatotyping technique, the body types of the male college students were categorized as meso-, ecto-, and endo-morphs (Sheldon, 1954). Although the body-type classification developed by Sheldon and his colleagues has been applied by many researchers in various fields, Johnson (1990) suggested that two more categories, ecto- and endo-mesomorphs, needed to be added into the body classification, because the two additional forms in a revised body classification system would more accurately reflect the physical characteristics of current individuals (Johnson, 1990). Based on the somatotyping technique developed by Sheldon, Douty (1968) established a method of “somatometry” to analyze the body shape. She photographed 300 college females as they stood silhouetted against a grided screen. The photographs were called somatographs which were used to evaluate postures, body masses, proportions, and body shape. Three experienced judges evaluated the somatographs, and the body shapes were classified into five categories of the body-build and posture.

Fit evaluation

133

IJCST 19,2

134

Douty (1968) concluded that the back shape, shoulder shape, buttocks shape, posture, and body build were important elements that researchers needed to pay more attention to while conducting a fitting test on female bodies. Directed by Douty, Brinson (1977) incorporated the somatometry method to alter patterns and to make basic garments of bodices and skirts for ten female students. Among the ten subjects, six were in the experimental group and four were in the control group. Brinson (1977) measured the angles on the somatographs of the experimental group, and she used angle measurements to determine how darts and side seams on basic patterns could be altered. The traditional pattern alteration method used for her study applied only length and circumferences to produce basic garments for the group members. Results showed that there was no significant difference on the fit between the two groups of subjects. However, the experimental method allowed a subject with figure variations to easily obtain a good fit without numerous fittings and alterations (Brinson, 1977). Unlike Brinson (1977) who tested the fit of bodice and skirt garments, Pouliot (1980) incorporated the somatometry method to develop pant patterns and test the fit of altered pants. She used a computer program, designed to correct the distortion of measurements which resulted from the somatometry method, to acquire body shapes, proportions, and angle measurements of 36 female subjects between the ages of 12 and 55. Finally, she chose five models representing hip shapes of round hips, pear-shaped hips, weight-in-front, weight-in-back, and average-figures for a fit evaluation. Each subject wore basic pants altered by both the unit method and the experimental method. The unit method referred to the application of length and circumference measurements to alter patterns, while the experimental method combined length and circumference measurements with angle measurements taken from somatographs. Pouliot (1980) determined that using the experimental method was better in creating pants that fit well on front waist position, front waist darts, back crotch, and horizontal grain orientation. Lesko (1982) further used the somatometry method and same computer software to produce bodice garments. She photographed 30 subjects and selected eight subjects who exhibited one or more of these figural variations: erect vs slumping posture, large bust vs small bust, high bust vs low bust, square shoulders vs sloping shoulders, and overweight vs underweight. Two garments were made according to patterns drafted using the unit and the experimental methods for each subject. She concluded that there was no adequate evidence proving that one method was better than the other (Lesko, 1982). Because the experimental alteration method that incorporated angle measurements did not significantly improve the fit of garments, Heisey et al. (1986) tried to improve the method. They realized that the transformation of 2D patterns from the somatometry method could be derived from a conical relationship between the 3D body form and the 2D patterns. Thus, they established a conical theory to formulate equations and to apply equations to certain areas in patterns that can be modeled as cones (Heisey et al., 1986). Their conical theory was not tested on the human body until 1993. Shen and Huck (1993) tested the conical theory by producing basic garments to be worn on the human body. They used somatographs in combination with four body measurements to develop bodice patterns for 12 female subjects with a variety of upper torso configurations. Bodice patterns for each subject were drafted using both the experimental method incorporating angles with four body surface measurements and the traditional method requiring 27 body surface measurements. They particularly

treated the bust area as a conical shape in the experimental group, and the cone was transformed to a 2D wedge form as the basis to adjust darts in patterns. Garments made using the experimental method were judged to have significantly better fit on most areas (Shen and Huck, 1993). In summary, previous research relied on photographs to extract body angles, and the angle measurements are incorporated with length and circumference in an attempt to create well-fitting garments. The attempt to improve apparel fit for female customers with a variety of body shape has been a challenge. As new technologies such as body scanners and CAD systems have been updated, the tests of fit on garments made by applying these technologies are practical and profound. However, none of research has used these technologies to create garments and test fit on females with various figure characteristics. This study extracts body measurements from a body scanner, deprives patterns from a CAD system, produces basic garments for fit evaluation, and analyzes female body shapes associated with fit problems. Methodology In the apparel industry, pattern construction is a necessary step to produce garments. Most apparel firms construct block patterns to develop patterns for new styles or design. The shape and size of block patterns should accurately represent the human body (Burns and Bryant, 1997). The block patterns used in this study were derived from the APDS-3D system. Asahi Chemical Industry Co. Ltd introduced the APDS-3D system to the market in 1995. The APDS-3D system starts with the dress form which contains 88 measurement points. The standard dress form can be viewed from different angles on the screen, and the dress form can be modified using measurements that are extracted from 3D body scanner (Tait, 1995). The system incorporates with the Gerber Pattern Design Systems (PDS) 2000 into a complete CAD system enabling the function of design, pattern making, and production. In addition to the CAD system, the VITUS 3D body scanner developed by Vitronic, a Germen company is also used in this study. The VITUS body scanner is accompanied by a software system equipped with VITUS View, RAMSIS Individual, and VITUS Measure to extract anthropometric measurements (Figures 1 and 2). After taking the six body measurements on each subject (Table I), the scanning measurements with additional functional ease were manually entered into a size-9 dress form measurement table of the APDS-3D system. The size-9 dress form was selected because it represents the average figure of Taiwanese women. The six scanning measurements were added to additional ease measurements of 8 cm for bust circumferences and 4 cm for waist circumferences. The values were entered into the measurement table, shown on Table II. When entering the values of measurements, the APDS-3D system only accepts values that are within a specific range. If measurement input is outside the tolerance of a selected form’s measurements, a white frame is marked around the measurement and the form modification screen becomes non-functional. Therefore, the measurements of waist circumference 81.3 cm for subject 4 and 82.5 cm for subject 8 were rejected, and they were limited to 79.3 cm and 78.5 cm instead. Once the measurements were accepted by the system, they were transformed into 3D visualization. The 3D visualization was then converted into 2D patterns of bodice front and bodice back. Both block patterns of bodice front and back with shoulder darts and waist darts were imported to the PDS 2000 which allows users to make changes on the 2D patterns. The completed block

Fit evaluation

135

IJCST 19,2

136

Figure 1.

Figure 2.

patterns were cut out and used to make basic garments for the ten subjects who had different figure characteristics (Table III). The body forms of the ten subjects were categorized by a panel of three professional judges who were faculty teaching clothing construction and apparel fit. The ten female subjects, who enrolled in the advanced fashion design class, were voluntary for this study. The judges followed the illustrations taken from the book Pattern Making for

Fashion Design written by Armstrong (1987) to classify the subjects’ body characteristics (Figures 3-8). The physical characteristics of the subjects were categorized as “body type,” “shoulder type A,” “shoulder type B,” “chest type,” “back type,” and “posture.” Eventually, the judges evaluated the basic garments worn

Measurements (cm) Bust circumference Waist circumference Center length (back) Bust span Neck Across shoulder (back)

2

3

4

83.3 68.5 40.0 15.5 35.6 40.6

87.3 70.8 35.0 16.0 37.4 38.5

85.5 69.3 35.0 19.0 36.0 39.4

89.0 77.3 36.0 17.5 39.9 41.0

Measurements (cm)

Size-9

Bust circumference Waist circumference Back center length Bust span Neck Back across shoulder

87.7 63.5 37.9 17.4 38.8 38.2

86.9 71.9 35.0 16.7 34.9 42.7

1

2

3

4

91.2 72.5 40.0 15.5 35.8 40.6

95.3 74.8 35.0 16.0 37.5 38.5

93.5 73.3 35.0 19.0 36.3 39.5

97.0 79.3 36.0 17.5 39.8 40.0

1 Body type

Shoulder type A Shoulder type B Chest type Back type Posture

Ideal Hourglass Straight Broad shoulders and small hips Narrow shoulders and large hips Ideal Sloped Square Muscular Bony Ideal Hollow chest Pigeon breast Ideal Flat Round Perfect stance Forward stance Upright stance

137

Subjects 5 6

1

91.1 64.0 36.0 16.4 37.1 39. 5

Subjects 5 6 94.9 75.9 35.0 16.7 35.1 42.0

2

99.1 68.0 36.0 16.4 37.3 39. 5

3

Fit evaluation

4

7

8

9

10

79.0 65.5 36.0 15.4 30.1 39.1

92.5 78.5 38.0 16.2 43.7 45.5

86.7 61.7 38.0 15.2 38.4 38.5

82.6 67.2 37.0 15.1 30.6 38.5

7

8

9

10

86.9 69.5 36.0 15.4 30.6 39.0

100.5 78.5 38.0 16.2 43.4 45.5

94.7 65.7 38.0 15.2 38.1 38.0

90.6 71.2 37.0 15.1 31.1 32.1

Subjects 5 6 7

8

9

V

Table I. The scanning measurements of subjects

Table II. The measurements of size-9 and scanning measurements with ease of subjects

10 V

V V

V V

V

V

V

V

V

V

V

V V V V

V V

V

V

V

V V

V V

V V V

V V

V V

V

V V

V V

V

V V V

V

V

V V

V V

V

V

V

V

V V V

V V

V

V V

Table III. The figure characteristics of subjects

IJCST 19,2

138

Figure 3. Body type

Figure 4. Shoulder type A

Ideal

Ideal

Hourglass

Straight

Sloped

Broad shoulders Small hips

Narrow shoulders Large hips

Square

by the subjects whose ages between 20 and 25. These subjects wore the basic garments over undergarments for the fit evaluation. The judges received copies of the rating scale created by Shen (1991) to evaluate fit based on the appearance of the basic garments. They marked a score ranging from 2 4 to þ 4 on the 25 items of the rating scale. The closer the score was to zero, the better fitting the basic garments were. T-test was employed to test differences on the means of 25 items between the physical characteristics of the shoulder type B category. One-way ANOVA was run to investigate the potential presence of significant differences on the means of 25 items between characteristics of each category: body type, shoulder type A, chest type, back type, and posture. When the ANOVA tests showed significant difference, the Tukey multiple comparison test was used to examine differences between possible pairs. Results and discussion The mean scores show that the basic garments have adequate ease and proper seam placement. However, a particular concern is the mean of 1.73 on the back

Fit evaluation

139

Muscular

Bony

Figure 5. Shoulder type B

Figure 6. Chest type Ideal

Hollow chest

Pigeon breast

armhole shape. This score indicates that most of the basic garments have the armhole shape much outside the natural body curves. Because the size-9 dress form has a wider across back measurement than that of the subjects’ bodies, the form used in the Taiwanese apparel industry has to be remolded. Otherwise, the measurement across the back must be included in the base measurement table of the APDS-3D system. The same problem occurs in the front armhole shape as well. Although the mean of the front armhole shape, 1.30, is not as high as that for the back armhole shape, the basic

IJCST 19,2

140

Ideal

Flat

Round

Figure 7. Back type

garments for subjects two, four, and five have front armhole shapes much outside the curves of their bodies. Along with inaccurate measurements of the dress form, the APDS-3D system’s tolerance limits and figure variations also cause fit problems. The APDS-3D system limits the measurement input of the waist circumferences if the values exceed the selected dress form’s waist circumference ^ 25 percent; waist circumferences outside this range are rejected by the system. Since, the waistline measurement inputs for subjects 4 and 8 went beyond the system’s tolerance, their waistline measurements were limited to the allowance and the tight waistline on the basic garments generated fit problems. Indeed, software designers need to consider eliminating the waist darts on the block patterns or develop a formula to adjust the waist darts. The waist darts can be eliminated because the block patterns with shoulder darts still portray the 3D shape of female bodies. Furthermore, software designers might create a formula to calculate the differences between the block pattern’s waistline circumference and the person’s waistline circumference containing basic ease. The intake of waist darts is the difference resulted from the block pattern’s waistline circumference minus the person’s waistline circumference. Either using patterns with only shoulder darts or adjusting the intake of the waist darts would solve the problem of an exceedingly tight waistline. In fact, the software designers should revise the tolerance limits of the APDS-3D system in terms of waistline circumferences as well as neck circumferences. Particularly, the tolerance of ^ 25 percent is too big to prevent incorrect input when used for neck measurements. The unusual neck measurements, 30.1 and 30.6 cm for subjects 7 and 10, respectively,

Fit evaluation

141

Tilting waistline

Tilting hemline

PERFECT STANCE

FORWARD STANCE

UPRIGHT STANCE

were resulted from an inaccurate measure during the scanning process. However, these measurements were accepted by the APDS-3D system, so the basic garments appeared poor fit on the necklines. The necklines of the basic garments were so small that they became stuck before they reached the bottom of the neck; the necklines did not fit and could not be smoothly positioned on the subjects’ necks. Therefore, the percentage of tolerance allowed for the neck measurement should decrease to assure proper fit.

Figure 8. Posture

IJCST 19,2

142

The fit problems occur due to the APDS-3D system as well as figure variations. After a series of ANOVA tests, significant differences exist between the physical characteristics within the categories of body types, shoulder type A, chest type, back type, and posture. First, the findings indicate that the subjects with the hourglass characteristic are significantly different from other groups on the items of bustline circumference and strain/looseness at the bust level; p-values for bustline circumference and strain/looseness at the bust level were 0.023 and 0.0001 (, 0.05), respectively. The hourglass type exhibit the fit problem – tightness around the bust area. The subjects with the hourglass characteristic appear to have more prominent bust than those who are not the hourglass type. This finding seems to be consistent with the study of Alexander et al. (2005) who report that the hourglass type expresses fit problems at the bust area. Findings also show that those with narrow shoulders and large hips are significantly different from the groups of straight and broad shoulders and small hips on the item called “strain/excess at shoulder tip” (0.009 , 0.05). The subjects with narrow shoulders and large hips accompanied with sloped shoulders have excess fabric around the shoulder tips compared with the groups of straight and broad shoulders and small hips. Moreover, the shoulder seams of the narrow shoulders and large hips group move much toward the front bodice. A possible reason for the occurrence of improper shoulder seam position could be the subjects’ narrow length measurement across the chests. In combination with narrow lengths across chests, the subjects who exhibit narrow shoulders and large hips stand in a relaxed posture. Their arms drop down and the shoulder seams are pulled toward the front bodice. Within the category of shoulder type A, there are significant differences on the items of gapping/strain at back armhole, strain/excess ease at shoulder tip, and shoulder seam position. The p-values for gapping/strain at back armhole, strain/excess ease at shoulder tip, and shoulder seam position were 0.014, 0.046, and 0.001 (, 0.05), respectively. The group of subjects with the physical characteristics of square shoulders is significantly different from the others on these three items. The square-shoulder subjects have more strains at the back armhole of the basic garments compared with those whose shoulders are ideal. The subjects with square shoulders tend to have flat backs that lead diagonal strains toward the shoulder tips. Furthermore, the group with square shoulders is significantly different from the group with sloped shoulders on strain/excess ease at the shoulder tip. Square-shoulder subjects showed much more strains at the shoulder tips compared with the subjects of sloped shoulders. Moreover, the square-shoulder group of subjects significantly differs from the ideal- and sloped- shoulder groups on the shoulder seam position; the shoulder seams of basic garments worn on the square-shoulder subjects move more toward the back bodice. The possible reason could be that the subjects’ shoulder points tend to be closer to front bodices, so their shoulder points do not match the shoulder points of the basic garments. Within the category of the chest type, significant differences exist for the items of gapping/strain at back armhole and side seam position. The p-values for gapping/strain at back armhole and side seam position were 0.001 and 0.046, respectively, among the three characteristics (ideal-, hollow-, and pigeon-chests). Subjects with hollow chest have more gapping at the back armhole compared with the group of ideal chest because subjects with hollow chest also have round back

generating gapping at the back armhole. Meanwhile, the protruding back creates another fit problem on the side seams in which the hollow-chest subjects’ side seams move toward the back bodice. Owing to prominent back and smaller bust, the side seams move more toward the back than the side seams of the ideal-chest group. The group of round back is significantly different from the group of flat back on the items of gapping/strain at back armhole, strain/looseness at bust level, and bustline circumference; p-values were 0.009, 0.037, and 0.037, respectively. It is a common occurrence for the group of round back to have more gapping at the back armhole, more strains at the bustline circumference, and more strains at the bust level. The round back produces gapping at the back armhole, and the subjects with round back accompanied with big bust causing tighter bustline circumference and strains at the bust level. There are significant differences within the category of posture on the mean of gapping/strain at back armhole, bustline circumference, and strain/looseness at bust level; p-values were 0.049, 0.039, and 0.021, respectively. Subjects with the characteristic of forward stance have more gapping at the back armhole, strains at the bustline circumference, and strains at the bust level than the subjects with upright stance. The possible cause is that subjects with the forward stance could also have round back. Limitation and conclusion This study recruited subjects who enrolled in the Department of Fashion Design at Shih Chien University in Taiwan that represent various physical characteristics. The convenient sample does not represent a generalization of figure variations in Taiwanese college women. However, the findings provide some information valuable for software designers. This research suggests software designers need to add measurements such as across back, across chest, and front center length that will enhance well fitting. Also, the tolerance for neck circumference in the APDS-3D system has to be decreased in order to ensure accurate measurements. Meanwhile, the intake of the waist darts has to be adjustable, depending on the difference between the females’ waist circumferences with additional ease and block patterns’ waist dimensions. Furthermore, this research reveals that fit problems become complicated when one figure variation combines with another and when the physical characteristics are not all ideal. Future research can focus on investigating commonly occurring figure variations among females and combination of figure variations. References Alexander, M., Connell, L.J. and Presley, A.B. (2005), “Communications: clothing fit preferences of young female adult consumers”, International Journal of Clothing Science & Technology, Vol. 17 No. 1, pp. 52-64. Anderson, G. (2004), If The Clothes Fit, Buy ’Em, available at: www.retailwire.com/Print/ PrintDocument.cfm?DOC_ID ¼ 10058 (2005, December). Armstrong, H.J. (1987), Pattern Making for Fashion Design, Harper & Row Publishers, New York, NY. Ashdown, S.P., Schoenfelder, L.K. and Lyman-Clarke, L. (2004), “Visual fit analysis from 3D scans”, Abstracts of the Fiber Society 2004 Annual Meeting and Technical Conference, p. 111.

Fit evaluation

143

IJCST 19,2

144

ASTM (2005), Global Notebook, American Society for Testing and Materials, West Conshohocken, PA, available at: www.astm.org/cgi-bin/SoftCart.exe/SNEWS/ NOVEMBER_2002/gn_nov02.html?L ^ mystore ^ druk4676 (accessed December 2005). Brinson, E.G. (1977), “Pattern alterations predicted and quantified using angle measurements”, unpublished master’s thesis, Auburn University, Auburn, AL. Burns, L.D. and Bryant, N. (1997), The Business of Fashion, Fairchild Publications, New York, NY. Douty, H.I. (1968), “Visual somatometry in health related research”, Journal of Alabama Academy of Science, Vol. 39, pp. 21-34. Erwin, M.D. and Kinchen, L.A. (1974), Clothing for Moderns, 5th ed., Macmillan, New York, NY. Feather, B.L., Herr, D.G. and Ford, S. (1997), “Black and white female athletes’ perceptions of their bodies and garment fit”, Clothing and Textiles Research Journal, Vol. 15 No. 2, pp. 125-8. Heisey, F.L., Brown, P. and Johnson, R.F. (1986), “A mathematical analysis of the graphic somatometry method of pattern alteration”, Home Economics Research Journal, Vol. 15 No. 2, pp. 115-23. Huckabay, D.A. (1992), “Perceived body cathexis and garment fit and style proportion problems of petite women”, unpublished master’s thesis, Virginia Polytechnic Institute and State University, Blacksburg, VA. Istook, C.L. (2002), “Enabling mass customization: computer-driven alternation methods”, Journal of Clothing Science and Technology, Vol. 14 No. 1, pp. 61-76. Johnson, K.K.P. (1990), “Impressions of personality based on body forms: an application of Hillestad’s Model of appearance”, Clothing & Textiles Research Journal, Vol. 8 No. 4, p. 34. Kind, K.R.O. (1995), “Specialty-size college-age females: satisfaction with apparel fit, body cathexis, and retail attributes”, unpublished master’s thesis, University of Georgia, Athens, GA. Kwong, M.Y. (2004), “Garment design for individual fit”, in Fan, J., Yu, W. and Hunter, L. (Eds), Clothing Appearance and Fit: Science and Technology, Textile Institute; CRC, Cambridge, pp. 196-233. Lee, S.E., Kunz, G.I., Fiore, A.M. and Campell, J.R. (2002), “Acceptance of mass customization of apparel: merchandising issues associated with preference for product, process, and place”, Clothing and Textiles Research Journal, Vol. 20 No. 3, pp. 138-46. Lesko, L.J. (1982), “Bodice fit compared: conventional alterations with and without graphic somatometry”, unpublished master’s thesis, Iowa State University, Ames, IA. Pouliot, C.J.T. (1980), “Pants alteration by graphic somatometry techniques”, unpublished master’s thesis, Iowa State University, Ames, IA. Sheldon, W.H. (1954), Atlas of Men, Harper, New York, NY. Shen, L. (1991), “Bodice sloper development using somatometic and physical data: an exploratory study”, unpublished master’s thesis, Kansas State University, Manhattan, KS. Shen, L. and Huck, J. (1993), “Bodice pattern development using somatographic and physical data”, International Journal of Clothing Science & Technology, Vol. 5 No. 1, pp. 6-16. Tait, N. (1995), “The Asahi apparel CAD 3D-PDS system”, Apparel International, Vol. 26 No. 11, pp. 35-7. Corresponding author Chin-Man Chen can be contacted at: [email protected] To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

COMMUNICATIONS

The analysis of personal and delay allowances using work sampling technique in the sewing room of a clothing manufacturer Sinem Gunesoglu and Binnaz Meric

Analysis of personal and delay allowances 145 Received 12 January 2006 Revised 1 August 2006 Accepted 1 August 2006

Textile Engineering Department, Uludag University, Bursa, Turkey Abstract Purpose – The aim of this paper is to study the operator activities in garment industry and the percentages of distribution of operations and to analyze the personal and delay allowances by observing the operations and deriving the ratios within a manufacturing period. Design/methodology/approach – A work sampling technique is used. Relevant reports (1978-2004) are studied to give the basis and methodology of the technique. In accordiance with work sampling techique, the operations to be observed in a sewing room are defined, the number of observations and observers required for each day and the procedure for making observations are determined and the distributions of work flows are calculated. Findings – It is found that 72.7 per cent of working time in an general sewing room was spent for productive activities and 23.2 per cent for personal and unavoidable delay allowances. Practical implications – Work sampling technique gives information about personal and delay allowances in a work flow of any sewing room. When the distributions of activites are determined, it is possible to find which activities are most responsible for low efficiency. For this purpose, standard operations time in a sewing room should be determined by time measurement studies and work flow should be organized. Originality/value – This paper deals with an actual sewing room and gives general information about the distributions of activites in work flow which should be used for organization of any sewing room. Keywords Optimization techniques, Work sampling, Garment industry Paper type Research paper

1. Introduction Garment manufacturing in nature is more complicated than many other industries. It involves a number of machines, hundreds of employees and thousands of bundles of sub-assemblies producing different styles simultaneously. In clothing production, garment components are assembled through a sub-assembly process until they are gathered into a finished garment. The production process involves a set of work stations in each of which a specific task in a restricted sequence is carried out (Hui and Frency Ng, 1999). Many factors such as the properties of fabrics and human emotions will affect the performance of operatives that ultimately will cause variance on the task time. Personal and delay allowances for the apparel industry are very important. Delays can be broken down into work elements which can be readily measured as fixed or variable and these measurements are then combined into work standards.

International Journal of Clothing Science and Technology Vol. 19 No. 2, 2007 pp. 145-150 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710725739

IJCST 19,2

146

Work study is used in the development and control of work situations. Work study encompasses all those procedures concerned with work measurement and motion study. Motion study is qualitative analysis of a work situation leading to the design or improvement of an operation. Work measurement techniques such as time study and work sampling are used in measuring or forecasting the rate of output of an existing or newly designed operation, as well as in determining how much time is consumed for various productive and non-productive activities of a process or operation. Also involved is the determination of standard times which represent the allowable time for the performance of work (Chuter, 1990; Jenkins and Orth, 2004). Time study is used to determine a standard time for an operation by direct time measurement. Work Sampling Technique can be used to determine the required data necessary for the application of a percentage allowance of personal needs, fatigue and unavoidable delays (Brock, 1978, Pape, 1991). The main difference between time study and work sampling is that time study requires a definite period of uninterrupted time to observe and considerable time to work up the study. Work sampling, however, will provide essentially the same results in 1/3-1/6 of the time and cost. Other advantages of work sampling studies can be made over a period of days or weeks thus decreasing the change of day-to-day or week-to-week variations (Barnes, 1980; Brisley, 1992; Brock, 1978, Chuter, 1990). In this study, the operator activities in garment industry and the percentages of distribution of operations are studied. The operations are observed, the ratios within manufacturing period are derived. 2. Experimental This study was carried out in sewing section of a garment manufacturer company and work sampling technique was used to determine the allowances (Kiremitci, 1999) and it was chosen due to mentioned advantages in Introduction. Work sampling technique is simply the act of observing an operator or operations at random time and then noting whatever was going on at the time the operator or operations was observed. For example, a particular machine being down could be noticed while walking through the sewing plant. If the operator is working (sewing, for example), the tally sheet should record any activity of the operator during that time, whether he/she is working, or is in any other state or activity. In setting up any work sampling study, there are six important steps to be considered. These steps are: (1) Accurately define the project or problem(s). (2) Determine the economic accuracy required. (3) Define the operation to be observed. (4) Determine the number of observations and observers required for each day. (5) Determine the procedure for making observations (starting place of tour, length of tour, manner of recording observations and when questions of operators or of supervisors should be asked). (6) Explanation of the work sampling procedure to all concerned (Barnes, 1980; Pape, 1991; Stohlman, 1978).

The operations are determined to be observed for distribution centre activities. The groups of determined operations are presented in Table I. In Table I: . G, the productive activities which belong to flow intervals in the measurement period. . Vsv, unavoidable delays that shows additional assignment or breakdown times which belong to flow intervals in the measurement period. . Vp, personal allowances that is caused by individual purposes or self requirements which belong to flow intervals in the measurement period. . N, avoidable delays. It is additional operations caused by the operator which belong to flow intervals in the measurement period.

Analysis of personal and delay allowances 147

Number of observations and tour periods are calculated after the flow types are determined. We assumed that a confidence level of 99 per cent and an accuracy of ^ 1 per cent are satisfactory. Also assuming that the binomial distribution is used as the basis for determining error and then the formula for determining the number of observations required (n) is Equation (1) (Barnes, 1980): Nr

Operations

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

To take a bundle To sew To tie the bundle To take pieces for sewing To leave sewn piece To control bundle and write on card To regulate a piece before sewing To control and check work To bring material To tie a broken thread To change apparatus To discuss about work To change thread bobbin To organize work place To correct faults To correct his/her faults To break of electricity To wait for work To wait for cleaning To change machine Needle break Machine is out of work and operator has additional work Machine break down Machine maintenance and adjustment Individual purposes or self requirements To speak privately To leave the work early To leave the factory To start working late Unknown

G

Vsv

Vsv

Vp N

X

Table I. The groups of determined operation types according to flow types

IJCST 19,2

148

ðZ a=2 Þ2 pð100 2 pÞ ð1Þ f 02 Where p: percentage occurrence of an activity or delay being measured; f 0 : degree of confidence; Za/2: value taken from normal distribution table. In preliminary (Brisley, 1992; Pape, 1991) studies for the sewing room, p ¼ 30 per cent and f 0 ¼ 1.0 are assumed at the measurement of distribution time ratios. Thus, Za/2 ¼ 2.575 is derived from the normal distribution table and then n is found 14,000 from the Equation (1). The work sampling period is defined as 8 weeks by taking 2 days in a week. According to this, daily number of observation nT is calculated using Equation (2): 14; 000 ¼ 875 observations=day ð2Þ nT ¼ ð2 days=weekÞ £ 8 weeks About 30 work systems consisting of 20 lock-stitch machines and ten over-edge stitch machines are chosen for the observation in the sewing room. Thus, tour number to be realised in a day RT is calculated using Equation (3): nT RT ¼ < 30 tours=day ð3Þ 30 Random time tables are used to determine the tour periods. Working hours, resting periods and noon hour are taken into account for determination the time periods. Tour plan for the sewing room has been determined as shown in Table II. n¼

3. Results Observations were continued until distribution time ratios gave measured degree of confidence f smaller than f 0 ¼ 1 in this study. The distribution of work flows within f is in Table III.

Table II.

Weeks

Days

1 2 3 4 5 6 7 8

Tuesday-Thursday Monday-Wednesday Thursday-Friday Wednesday-Thursday Tuesday-Wednesday Tuesday-Thursday Tuesday-Friday Monday-Thursday

Flow type

Table III. The distribution of the flows

1. Productive activities 2. Unavoidable delays 3. Personal activities (2 þ 3) 4. Avoidable delays Total

Symbol G Vsv Vp V N

Number of observation (n)

p (per cent)

f (per cent)

9,815 2,011 1,123 3,134 551 13,500

72.7037 14.8963 8.3185 23.2148 4.0814 100

0.9873 0.7891 0.6120 0.9357 0.4385

It is found that 72.7 per cent of working time was spent for productive activities and 23.2 per cent for personal and unavoidable delay allowances at the end of work sampling study. In early similar studies (Barnes, 1980; Pohl, 1979; Brock, 1978; Stohlman, 1978), personal and unavoidable delays allowances which we found as 8.3 and 14.9 per cent, respectively, were both recorded as 5 per cent. The percentages of distribution of operations consisting of delays are shown in Figure 1. As seen from Figure 1, the operation caused by individual purposes or self requirements (Operation Nr: 25) covers 36 per cent of total distribution time periods. Work control operation (Operation Nr: 8) has the second place with 11 per cent percentage of non-productive activities. This operation is simply the operator’s eye control of the quality of patch after sewing. Regulating before sewing (Operation Nr: 7) has the share of 10 per cent of non-productive activities. This time period is spent for regulation of small pieces like collars, sleeve cuffs, etc. This operation is caused especially by imprecise cuttings and would be eliminated by using modern automated cutting systems. Waiting for the piece to be sewn (Operation Nr: 18) takes 6 per cent of non-productive activities. The main reason for this is not to balance the production planning and production line well. Another reason for this is the lack of communication between production units. The improper material planning and late delivery of cutted pieces from cutting department can be given as examples. The percentages of personal delay and unavoidable allowances according to the productive activities were calculated 11.4 and 20.5 per cent, respectively.

Analysis of personal and delay allowances 149

4. Conclusions This study investigates personal and unavoidable delay allowances of the sewing room of a garment manufacturer. Operation types are classified and the work flow is

40 35 30

p (%)

25 20 15 10 5 0 25

8

7

16

6

12 18 22 13 non-productive activities

10

15

17 Other

Figure 1. The distribution of non-productive activities

IJCST 19,2

150

composed; then each operation activity time is measured and allowances are calculated. Distribution of operations within non-productive activities are also determined. We found that personal based operations or intervals have the greatest amount of non-productive activities. Controlling and checking the work, cutting action before sewing and waiting the pieces have also remarkable percentages. To increase the efficiency of a sewing room, distribution of these actions should be reduced since wrongly production line and missions determined cause delays during the execution of a work. All materials should be in required place at correct time to prevent delays. For this purpose, standard time should be determined by time measurement studies and work flow should be organized. References Barnes, R. (1980), Motion Time and Time Study Design and Measurement of Work, 7th ed., Wiley, New York, NY, pp. 406-40. Brisley, L.C. (1992), “Work sampling and group time technique”, in Hudson, W.K. (Ed.), Maynard’s Industrial Engineering Handbook, 4th ed., McGraw-Hill, New York, NY, pp. 4.39-68. Brock, H. (1978), Work Measurement Technique in the Apparel Distribution Center, JSN International, Agra, November, pp. 42-9. Chuter, A.J. (1990), Introduction to Clothing Production Management, BSP Professional Books, London, pp. 43-59. Hui, C.L.P. and Frency Ng, S.F. (1999), “A study of the effect of time variations for assembly line balancing in the clothing industry”, International Journal of Clothing Science & Technology, Vol. 11 No. 4, pp. 181-8. Jenkins, J.L. and Orth, D.L. (2004), “Productivity improvement through work sampling”, Cost Engineering, Vol. 46 No. 3, pp. 27-32. Kiremitci, S. (1999), “The investigation of reasons for additional activities and their ratios in production time at an apparel plant”, MSc thesis, Uludag University, Bursa. Pape, E.S. (1991), “Work sampling”, in Salvendy, G. (Ed.), Handbook of Industrial Engineering, 2th ed., McGraw-Hill, New York, NY, pp. 1699-721. Pohl, H. (1979), “Aufbau und Ho¨he von Verteil und Erholunszeiten fu¨r Arbeitsverrichtungen der Bekleidungsfertigung”, Bekleidung und Waesche, pp. 2-12. Stohlman, D. (1978), Work Sampling – What is and How to Use it in the Sewn Products Industry, JSN International, Agra, December, pp. 32-40. Corresponding authors Sinem Gunesoglu can be contacted at: [email protected]; Binnaz Meric can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

A minimisation algorithm with application to optimal design of reinforcements in textiles and garments

A minimisation algorithm

159

Zˇeljko Sˇomodi, Anica Hursa and Dubravko Rogale Faculty of Textile Technology, University of Zagreb, Zagreb, Croatia Abstract Purpose – This paper aims to develop an efficient two-variable minimisation algorithm and to apply it in engineering optimisation of buttonhole reinforcements. Design/methodology/approach – An iterative extreme search is based on quadratic approximation of objective function, locating approximate solution at the minimum of corresponding elliptic paraboloid. Stress analysis is performed using plane stress finite element model. Optimal selection of geometrical parameters of buttonhole-type reinforcement is done in terms of balance between maximum stress and material consumption. Findings – Adopted minimisation algorithm is assessed in a selected test example and proven to perform well in comparison with methods available in literature. In the selected case two local minima have been found within the predefined optimisation domain, with slight difference in objective function values. Research limitations/implications – Research is limited to homogeneous isotropic elastic model and a single representative load case. Objective function is restricted to two influence factors and predefined optimisation domain. Practical implications – The method can be useful for engineers/practitioners in the branch of clothing technology as a tool for computational estimate of optimal design of structural reinforcements. Originality/value – The main quality of the paper is in software fusion of finite element analysis and advanced optimisation algorithm. The proposed theoretical-numerical model is discussed in terms of applicability in parameter setting on buttonholer in garment production. Keywords Clothing, Technology led strategy, Optimization techniques, Numerical analysis Paper type Research paper

1. Introduction The application of sophisticated computer aided methods of computational analysis is becoming ever more present in various fields of technology and engineering. The procedures of engineering analysis and design are strongly supported by the software packages based on finite element method or other techniques of discrete analysis. The range of applications reaches well beyond the basic linear elastic static analysis and covers dynamics, nonlinearities of material and geometric nature, as well as specialist areas such as fracture mechanics, micro structured materials and automatic optimisation. The issues of optimal solutions in engineering design have received much attention, of course primarily in the fields other than textile and clothing engineering. In our recent paper (Sˇomodi et al., 2005), we have dealt with the subject in the context of

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 159-166 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741623

IJCST 19,3/4

160

“buttonhole type” reinforcements, seeking minimum of an objective function incorporating the opposed criteria of maximum stress and amount of built-in material. The aim of the present work is to continue and improve on this in terms of dimensionality (objective function of more than one variable) and programming fusion of finite element analysis and automatic optimisation procedure. For this purpose, in Section 2 a novel technique for minimum search is proposed. We refer to it as “quadratic search”. The efficiency of the algorithm is assessed on the classical test function introduced in 1960 by Rosenbrock (“banana valley”) (Rosenbrock, 1960; Donald, 1986). In Section 3, the algorithm is then applied in the optimisation of reinforcements in a situation of “buttonhole type” – tension strip with transverse central cut reinforced in the zones of stress concentration. Optimum search is done in the space of two variables, namely the size and the thickness of the reinforced zone of prescribed rectangular shape. The finite element computations incorporated in the programme are based on our previous development and applications (Sˇomodi et al., 2003a, b). 2. Minimisation algorithm: quadratic search There exist a number of algorithms for extreme search, from simplex method of linear programming and gradient-based methods such as Cauchy’s steepest ascent/descent method, to more recent and more sophisticated methods such as Davidon-Fletcher-Powell and Broyden-Fletcher-Goldfarb-Shanno (Donald, 1986). Still, there is no such thing as the universally best method, and from case to case different methods may appear more or less effective, and sometimes some of them even fail to converge. So, in the context of minimisation problem related to optimal design of structural reinforcements, it appeared inviting to invent our own extreme search algorithm. In the paper (Sˇomodi et al., 2005), the one-dimensional minimum search is performed by approximating the function with the parabola in the vicinity of the trial point – the procedure clearly inspired by Newton’s iterative method. Now let the similar procedure be stated in the case of more than one dimension: obviously, the approximation of the function shall be the quadratic polynomial. To evaluate the constants in such a quadratic approximation, it is necessary to know the value of the function and its first and second derivatives at the trial point – these can be computed by the usual finite difference expressions from the values of the function in a certain number of points in the vicinity of the trial point. For simplicity, the details are elaborated here for the case of function of two variables f(x1, x2). Let the values of the variables x1 and x2 at the trial point (starting point of the step k in the iterative search for the extreme of f(x1, x2)) be denoted by xk1 ; xk2 . In search for the kþ1 of the point of minimum let us proceed as follows (for approximation xkþ1 1 ; x2 notational simplicity the superscript k is omitted): the gradient of the scalar function f(x1, x2) is defined by the first derivatives and can be evaluated by the finite difference schemes:

›f f ðx1 þ Dx1 ; x2 Þ 2 f ðx1 2 Dx1 ; x2 Þ < ; 2Dx1 ›x1 ›f f ðx1 ; x2 þ Dx2 Þ 2 f ðx1 ; x2 2 Dx2 Þ < 2Dx2 ›x2

ð1Þ

A minimisation algorithm

The second derivatives are:

›2 f f ðx1 þ Dx1 ; x2 Þ 2 2f ðx1 ; x2 Þ þ f ðx1 2 Dx1 ; x2 Þ < ›x21 Dx21 ›2 f f ðx1 ; x2 þ Dx2 Þ 2 2f ðx1 ; x2 Þ þ f ðx1 ; x2 2 Dx2 Þ < ›x22 Dx22

ð2Þ

161

›2 f f ðx1 þ Dx1 ; x2 þ Dx2 Þ þ f ðx1 ; x2 Þ 2 f ðx1 þ Dx1 ; x2 Þ 2 f ðx1 ; x2 þ Dx2 Þ < Dx1 Dx2 ›x 1 ›x 2 Note that central difference scheme is used, except for mixed second derivative in equation (2) – this scheme might be called “upper-right”. Note also that data in equations (1) and (2) are computed from trial point plus five surrounding points. The second derivatives (2) form the Hessian matrix: 72 f ¼

›2 f ; ›x i ›x j

i; j ¼ 1; 2

ð3Þ

It can easily be shown that in the case of co-ordinate axes rotation, i.e. if x1, x2 are replaced by the rotated linear combination: x 1 ¼ x1 cos w þ x2 sin w;

x 2 ¼ x2 cos w 2 x1 sin w;

ð4Þ

the transformation law for equation (3) is equivalent to familiar behaviour of tensor quantities such as stress – therefore, let us refer to equation (3) as “second derivative tensor”. Our extreme search algorithm relies on the eigen values of equation (3) – these eigen values can be viewed as principal curvatures of the surface f(x1, x2) at the trial point. If both principal second derivatives of f at the trial point are non-zero and positive (i.e. the Hessian matrix is positive – definite), the surface f(x1, x2) is fully concave and we approximate it with a quadratic polynomial (elliptic paraboloid) of the form: pðx1 ; x2 Þ ¼ ax21 þ bx22 þ cx1 x2 þ dx1 þ ex2 þ f

ð5Þ

The constants a to f in equation (5) follow from the conditions of equality of p(x1, x2) and f(x1, x2) and the respective first and second derivatives at xk1 ; xk2 . Finally, supposing similar behaviour of the function f(x1, x2) and its approximation (5), let us set the next approximation for the point of minimum of f(x1, x2) to the point of minimum (apex) of the paraboloid equation (5): x1kþ1 ¼

ce 2 2bd ; 4ab 2 c 2

xkþ1 ¼ 2

cd 2 2ae 4ab 2 c 2

ð6Þ

kþ1 Approximation (6) may produce quite a substantial jump from xk1 ; xk2 to xkþ1 1 ; x2 , especially in the case when at least one eigen value of equation (3) is close to zero. Therefore, it can be useful to check it out in the following two control steps: or xkþ1 (or both) may turn out to be outside the admissible (1) Control step 1. xkþ1 1 2 (physically meaningful) domain of the variables x1, x2. For example, geometric parameters in a case like the one we deal with later in this paper (e.g. size or

IJCST 19,3/4

162

thickness of the reinforced zone) may become negative – this cannot be kþ1 in accepted. In such a situation let us reduce the jump from xk1 ; xk2 to xkþ1 1 ; x2 some way – here we simply propose to cut it in half until both variables are back in the admissible domain. kþ1 are admissible, we may want to check out (2) Control step 2. Even if xkþ1 1 ; x2 weather the iteration step has led to lower value of f(x1, x2), as expected from the minimisation procedure. If the evaluation shows that on the contrary kþ1 kþ1 kþ1 k k k k the f ðxkþ1 1 ; x2 Þ . f ðx1 ; x2 Þ, meaning that between x1 ; x2 and x1 ; x2 function f(x1, x2) has deviated considerably from the presumed quadratic nature and got higher instead of lower, we may want to correct the iteration step. Once again, we propose without further elaboration simply to cut the jump from kþ1 kþ1 k k in half until the condition f ðxkþ1 xk1 ; xk2 to xkþ1 1 ; x2 1 ; x2 Þ , f ðx1 ; x2 Þ is satisfied. This control step is optional and the algorithm may, depending on the case, perform better with or without it. The described procedure, as mentioned before, works in the “regular” case when both eigen second derivatives are positive, i.e. surface f(x1, x2) is concave in both principal curvatures. Otherwise the function is at xk1 ; xk2 either fully convex or of saddle type (concave-convex) and there is no point in using the approximations (5) and (6). In that case, let us use the simple “steepest descent” strategy: from the components (1) find the direction of gradient and move on the surface f(x1, x2) in downward direction by the step of an adopted length, for example, five per cent of the typical range of variables x1 and x2. If the function f(x1, x2) has a minimum, this temporary strategy should lead us to the fully concave region, where we proceed with approximations (5) and (6) as described before. The generalisation for more than two variables seems straightforward, although computationally more demanding, and is beyond the scope of this paper. Note that the algorithm requires six function evaluations per iteration step, plus one for control step 2, but this one can be used in the next step. If correction takes place in control step 2, number of function evaluations is increased by one for each correction step. In the following, the assessment of the performance of the proposed minimisation algorithm is demonstrated in the example of a classical test function. The function was first proposed by Rosenbrock (1960) and is referred to as “Rosie” or “banana valley” (Donald, 1986). It reads: 2 f ðx1 ; x2 Þ ¼ 100 x21 2 x2 þð1 2 x1 Þ2

ð7Þ

It has the shape of a parabolic valley (Figure 1(a)) with the minimum min f ¼ 0 at the point (1, 1). A typical benchmark test for a minimisation algorithm is to minimise equation (7) to below 102 3 starting from the point (21.2, 1). It is quoted in Donald (1986) that Powell’s method requires 125 function evaluations for that task, while some other algorithms available at the time required from 150 to about 500 evaluations. In Figure 1(b) and (c), the performance of our “quadratic search” algorithm is demonstrated. In Figure 1, co-ordinate axes are set to x1 ¼ x2 ¼ 2 2 as the left and bottom margins of the domain of interest. In Figure 1(b) and (c), iteration is shown in the aerial view with points and adjacent numbers indicating the end of the respective steps. White colour represents the region of full concavity, while in the parabolic shaped area the function is saddle shaped (concave-convex). Figure 1(b) shows the iteration with

corrections in control step 2 as described above. The task was fulfilled in 17 iteration steps, and corrections were necessary in steps two (three correction steps), 12 and 16 (one correction step each). Simple calculation gives the number of function evaluations: 17 £ 6 þ 3 þ 1 þ 1 þ 1 ¼ 108. Somewhat surprisingly at first glance, even better performance was obtained when control step 2 was turned off (Figure 1(c)). The iteration took 11 steps and 11 £ 6 þ 1 ¼ 67 function evaluations. As we see from this example, the algorithm shows potential for good performance. It remains to be verified on more test functions and in practical problems related to minimisation. In the next section, we apply it in a problem of optimal design of structural reinforcements. The problem can be viewed as simplified theoretical example related to reinforcements of buttonholes and similar cases where geometric irregularities cause considerable stress concentration.

A minimisation algorithm

163

3. Example of application: optimisation of reinforcements Consider the case of a thin rectangular body with central transverse cut (Figure 2(a)). When the body is exposed to tension, stress concentration takes place in the vicinity of the edges of the cut. Let this critical regions be reinforced by increasing the thickness in the depicted zones of rectangular shape. We chose to describe the optimisation space by two variables: x1 related to the size and x2 related to the thickness of the reinforced zone. The value of these dimensionless variables shall roughly correspond to the attributes of the geometrical parameters in the following way: x ¼ 2 2, very low; x ¼ 2 1, low; x ¼ 0, medium; x ¼ 1, high; x ¼ 2, very high. Some limitations on the geometrical parameters must be given in advance, not only in order to have physically meaningful domain (for example, size of reinforced zone cannot be negative or larger than the body itself), but also due to possible technological requirements or limitations of the computational model. So let the medium value of the base of reinforced zone bz be set to 0.4b and the range from “very low” to “very high” to 0.4b. This means that it varies from 0.2b as very low to 0.6b as very high. Similarly, in this example we choose for the height of the reinforced zone: medium at 0.3h, range 0.2h; and for the thickness of reinforcement (added thickness in the reinforced zone): medium at 0.8t0, range 0.8t0, where t0 denotes basic thickness of the body. Adopted geometry of the body implies h ¼ 2b, with the length of the cut also exactly equal to 2b. Stress is computed for a typical tensile load case. For symmetry reason one quarter, upper-right, of the body is modelled by a mesh of triangular (CST) finite elements. In Figure 2(b), the coarse finite element mesh is visible for the case X2

X2

Y

2

2

1

1

1

X2

17 16 15 14 13 12

START

2

3 4

0

5

1

10

910 8 7 6

11

6 7 89

11

START

3

0

5 4

2

−1

−1

X1

(a)

−2 −2

−1

0

(b)

1

2 X1 −2 −2

−1

0

(c)

1

2X1

Figure 1. Minimisation of “Rosie” by quadratic search algorithm: (a) shape of the function; (b) iteration with corrections in control step 2; and (c) iteration without corrections

IJCST 19,3/4

hz

2bz

2h

164

Figure 2. Thin body with cut and reinforcements: (a) geometry of tension strip, transverse cut and reinforced zones; and (b) finite element discretisation of one quarter of the strip

2b

(a)

(b)

x1 ¼ 0 (medium size of reinforced zone), as well as boundary conditions (nodal supports) and loading by nodal forces. We declare certain solution to be optimal if it corresponds to minimum of the objective function – a quantity incorporating several criteria and expressing whatever is unwanted. In this case, like in our previous work on the subject (Sˇomodi et al., 2005), let the governing opposed criteria be the maximum stress and the amount of built-in material. In the dimensionless objective function these are expressed with respect to corresponding values at a reference point. We choose the reference point in the middle of the expected optimisation domain, i.e. at x1 ¼ x2 ¼ 0. Additionally, let us complete the objective function with correction members expressing the criteria that the solution should stay within pre-defined limit values of optimisation parameters. In accordance with the contents of the first paragraph of this section, we set these limit values at x1,2 ¼ ^ 3. The corresponding correction member can have a form: ci ¼

1 12

xi 2 3

;

i ¼ 1; 2

ð8Þ

Like other contributions in the objective function, equation (8) reduces to unity at the reference point x1,2 ¼ 0. In order to not disturb the governing criteria of stress and material consumption (except near the domain limits), correction members (8) are given very low weight factors of up to three per cent. So, the objective function has the form: f ðx1 ; x2 Þ ¼ ws

smax smax 0

2

þwV

V V0

2 þwc1 c1 þ wc2 c2

ð9Þ

Sum of weight factors in equation (9) is ws þ wV þ wc1 þ wc2 ¼ 1, so that at the reference point function (9) takes the value of unity. The optimisation problem reduces to finding the pair (x1, x2) for which equation (9) is minimal. In our programming solution it is done in two steps. In the first step, a rough domain scan is done with integer values of 2 2 # xi # 2, i ¼ 1, 2. This produces the picture of the domain in which regions of possible minima are roughly visible, and then in the second step iterative minimum search by the quadratic search method is done, starting from the trial point specified by the user. Let us illustrate the described procedure in the example of ws ¼ 0.54, wV ¼ 0.4, wc1 ¼ wc2 ¼ 0.03 (Figure 3). As indicated by the rough domain scan, there are two regions of low values of objective function (white squares in the picture). Iteration has produced two local minima of the objective function (Figure 3(a) and (b)): fmin1 ¼ 0.92946 at x1 ¼ 2 0.909, x2 ¼ 1.265 (low size, high thickness) and fmin2 ¼ 0.93184 at x1 ¼ 0.7701, x2 ¼ 2 2.1056 (medium to high size, very low thickness). Obviously, the first minimum is better but just slightly. Numerical results of the iterative minimum search of Figure 3(a) are also given in Table I. Values in columns entitled x1, x2 and f are the two optimisation variables and X2

+

+

+

4 5 67 +START 3 2

1

+

+

+

+

+

+

+

+

−2

−1

2

1

+

+

+

+

−1

+

+

−1

+

+

+

+

−1

+

+

−2

+

+

+

+

START + −2 94

0

165

X2

−2

Iteration step 1 2 3 4 5 6 7 8

2

A minimisation algorithm

1

2

X1

0

1

2

+

+

1+ 2 8 1037 6 5

+

X1

Size of reinforcement, x1

Thickness of reinforcement, x2

Objective function, f

20.918878 21.146610 20.759221 20.987508 20.587962 20.972129 20.917779 20.909291

1.30572 0.97688 1.07651 1.48335 1.46428 1.35286 1.28703 1.26482

0.929482 0.959524 0.932764 0.930155 0.974931 0.930082 0.929470 0.929456

Figure 3. Minimisation of equation (9) with ws ¼ 0.54, wV ¼ 0.4, wc1 ¼ wc2 ¼ 0.03: (a) start of iteration at x1 ¼ 2 1, x2 ¼ 1; and (b) start of iteration at x1 ¼ 0, x2 ¼ 22

Table I. Numerical results of iteration of Figure 3(a)

IJCST 19,3/4

166

objective function at the end of the respective step. Note that the very first step starting from (2 1, 1) has almost reached the minimum, but then some oscillations took place (“steepest descent” step has led to actually higher value of objective function in steps two and five) before the iteration finally converged. 4. Conclusions Minimisation algorithm proposed in this paper as “quadratic search method” offers solid possibilities in applications related to optimisation or other extreme search situations. Its applicability is proved in the standard test example of Rosenbrock’s “banana valley” function, as well as in “practical” case of optimal design of buttonhole reinforcements. Still, it remains to be thoroughly tested on a number of other test functions and compared with existing algorithms. There also seems to be some space for refinement in the “steepest descent” step. We hope that our programme solution incorporating software fusion of finite element analysis and extreme search algorithm can be useful for engineers practitioners in the branch of clothing technology (and maybe some other branches as well) as a tool for computational estimate of “optimal” design of structural reinforcements. Although some limitations of the model restrict it to simplified theoretical picture of actual situation, it can be viewed as an attempt to introduce rational engineering analysis procedures in decision-making process, so far based more on ad hoc estimates or instinct. Objective function definition in terms of criteria and weight factors remains to be elaborated by experts, and in this sense our method is quite open. As far as the practical example is concerned, the distribution of the objective function showed two local minima. It leads to a conclusion that sometimes there can be a thin line between more than one “optimal” solution. References Donald, A.P. (1986), Optimisation Theory with Applications, Dover Publications Inc., New York, NY. Rosenbrock, H.H. (1960), “An automatic method for finding the greatest or the least value of a function”, The Computer Journal, Vol. 3 No. 3, pp. 175-84. Sˇomodi, Zˇ., Hursa, A. and Rogale, D. (2003a), “Finite element analysis of textile in plane stress”, in Matejicˇek, F. (Ed.), Proceedings of 4th International Conference Innovation and Modelling of Clothing Engineering Processes – IMCEP 2003, Faculty of Mechanical Engineering, Maribor, October, pp. 108-14. Sˇomodi, Zˇ., Hursa, A. and Rogale, D. (2003b), “Nonlinear modelling of flexible materials in plane stress with application to stress concentration in textile”, in Matejicˇek, F. (Ed.), Proceedings of 4th International Congress of Croatian Society of Mechanics, Croatian Society of Mechanics, Zagreb, September, pp. 517-22. Sˇomodi, Zˇ., Hursa, A. and Rogale, D. (2005), “Numerical optimisation of mechanical reinforcements of weak spots in textiles and garment”, International Journal of Clothing Science and Technology, Vol. 17 Nos 3/4, pp. 200-8. Corresponding author Zˇeljko Sˇomodi can be contacted at: [email protected] To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Investigating the development of digital patterns for customized apparel Yunchu Yang and Weiyuan Zhang

The development of digital patterns 167

Fashion Institute, Donghua University, Shanghai, People’s Republic of China, and

Cong Shan Fashion Institute, Donghua University, Shanghai, People’s Republic of China and Shanghai University of Engineering Science, Shanghai, People’s Republic of China Abstract Purpose – The paper aims to provide an overview of the area of digital pattern developing for customized apparel. Design/methodology/approach – The paper outlines several methods of digital pattern developing for customized apparel, and discusses the principles, characters and applications. Digital pattern developing process has two paths. One path develops apparel according to traditional 2D pattern-making technology. There are three methods: parametric design, traditional grading technique, and pattern generating based on artificial intelligence (AI). Another path develops pattern through surface flattening directly from individual 3D apparel model. Findings – For parametric method, it can improve greatly the efficiency of pattern design or pattern alteration. However, the development and application of parametric Computer-Aided-Design (CAD) systems in apparel industry are difficult, because apparel pattern has fewer laws in graphical structure. For grading technique, it is the most practical method because of its simple theory, with which pattern masters are familiar. But these methods require users with higher experience. Creating expert pattern system based on AI can reduce the experience requirements. Meanwhile, a great deal of experiments should be conducted for each garment with different style to create their knowledge databases. For 3D CAD technology, two methods of surface flattening have been outlined, namely geometry flattening and physical flattening. But many improvements should be done if the 3D CAD systems are applied in apparel mass customization. Originality/value – The paper provides information of value to the future research on developing a practical made-to-measure apparel pattern system. Keywords Customization, Clothing, Parametric measures, Artificial intelligence, Flatness measurement Paper type Viewpoint

Introduction In today’s apparel market, most of consumers desire to personalize the style, fit and color of the clothes. They require high-quality customized products at low prices with faster delivery. With this sort of consumer interest in mind, the concept of mass customization emerged in the late 1980s (Seung-Eun and Chen, 2000). Mass customization is a hybrid of mass production and customization and is a new manufacturing trend. It is an effective competing strategy for maximizing customers’ satisfaction and minimizing inventory costs. In the book, Mass Customization, Pine (1993) defined mass customization as “the mass production of individually customized goods and services”. Pine stated that information technology and automation are prerequisite of implementing mass customization because they constitute

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 167-177 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741632

IJCST 19,3/4

168

the connection between the consumer’s wants and needs and the ability of a manufacturer to create the products accordingly. In the apparel industry, several new technologies have helped mass customization operations. Taking advantage of technological developments in combination with proximity to the target market is the differential advantage that domestically manufactured products have over foreign produced goods. However, while technology may enable mass customization efforts, the processes involved are far from automatic. A significant amount of “behind the scenes” effort is still required in order to provide fit of each garment that might be requested by individual customers. Made-to-measure (MTM) apparel pattern systems are a vital part of mass customization because it can solve how to produce attractive and accurately fitting apparel according to customers’ needs (Istook, 2002). These systems will enable the creation of garments, customized for fit, in a very quick and accurate manner and can be inserted into normal production lines as an additional size and produced like every other garment of the same style (Turner, 1994). They also allow styles to be customized, repeatedly, without time consuming preparing activities. This means that successful companies with huge libraries of garment styles would be able to implement mass customization strategies with relatively little effort. The purpose of this paper is to explore the technologies of digital pattern developing for MTM pattern systems. The authors outline several methods and present their advantage and disadvantage. Digital pattern developing process has two paths. One path develops apparel according to traditional 2D pattern-making technology. These are based upon relationships among apparel topography, critical body scan measurements and 2D flat pattern alteration heuristics. Details on this path are presented in the second section. Another path, discussed in the third section, develops pattern through flatting directly from individual 3D apparel model. Pattern-making technique for customized apparel in 2D CAD system An essential key to realize mass customization for apparel industry is the ability of Computer-Aided-Design (CAD) systems to integrate measurement information and make changes to patterns, as necessary, without permanently changing the basic, original garment pattern (Istook, 2002). Many companies have developed MTM pattern systems based on 2D CAD technology to satisfy the needs of mass customization. There are several typical approaches, including: . individual pattern auto-making based on parametric design; . individual pattern altering base on grading rules; and . individual pattern generating based on artificial intelligence (AI). Individual pattern auto-making based on parametric design Parametric design means, based on the known constraint condition and topology structure, generating the new design drawings according to the modified parameter value. It was presented in the practical application process of CAD technology. The approach brings not only the interactive function for CAD system, but also the auto-design function. In the data list of parametric CAD, not only the entity objects but also their topology structure and constraint relationship among them are store (Yang and Ping, 2002; Zhang Hong et al., 2005). From the view of geometry, apparel pattern is regarded as a set of geometric elements, such as point, line, arc, and curve. The graphical descriptions of the pattern include three parts: topology structure,

geometric parameters and the relationship between the geometric parameters and structure parameters (Zhang Hong et al., 2005). The topology structure of apparel pattern is a set of rules of pattern making. The geometric parameters just are the geometric information, like points’ coordinates of pattern. The structure parameters are critical part of the parametric design approach. In the MTM apparel pattern system, it is through changing the values of these structure parameters to generate the new individual pattern when we develop apparel in the parametric way. For apparel pattern, these structure parameters can be classified as the following category: . Measurement parameters, involving the key measurements of body or garment. . Design parameters defined by pattern maker according to apparel style, such as ease value, pleat value, shrinking value. . Graphical variables, defined by pattern maker too, often represent the length of geometric elements (line or curve). For example, pattern-masters often determine the length of back shoulder (BS) of the garment according to the length font shoulder (FS). So the length of FS can be defined as a graphical variable. . Compound parameters, which is a synthesis of the parameters above through a formula. The formula was defined by pattern maker according to specialty knowledge and experience on pattern engineering. Consequently, through defining these parameters, pattern masters can enlarge the free space for pattern designing or pattern alteration.

The development of digital patterns 169

The principle of pattern-making based on parametric design can be outlined as follows: firstly, record the process of pattern making (graphics topology structure) and define it as a pattern template; secondly, define a set of parameters to make a constraint in the pattern; finally, carry out the process of dimension driven through modifying the set of parameters and generate new pattern from the pattern template. Figure 1 shows the process of pattern auto-making based on parametric design. From the figure, there are two critical parts: dimension constraint and dimension driven (Yang and Ping, 2002). Dimension constraint in the process of parametric design. Traditional CAD systems define geometric elements with a set of fixed dimension values and only the coordinates of the elements are stored in the file format of these systems. As the graphics as concerned, the products only represent one simple connection among all kinds of elements, like point, line, circle, arc or curve. It does not involve the internal relation. The concept of “constraint” is to establish the internal relation of the product utilizing some rules or limiting conditions. The process of the parametric design is just the process of constraint satisfied, and is also the process of creating and solving the constraint-based model. All kinds of parameters Knowledge database

Constraint Pattern template Driven Individual pattern

Figure 1. The parametric design process of the apparel pattern

IJCST 19,3/4

170

The constraint can be classified geometric topology constraint and dimension constraint (Yang and Ping, 2002). The geometric topology constraint represents the special relationship among the elements of the graphics, including connecting relation and positional relation. This kind of constraint is the unchangeable object for parametric design. It limits the changing of the position or connecting relation among all elements in order to maintain the old shape of graphics when the graphics is modified. The dimension constraint is used for controlling the relative position of all elements and represents their intrinsic properties. It is the changeable object for parametric design. The parametric graphics objects are generated through altering the dimension constraint. Dimension driven in the process of parametric design. How to control the outline of the graphics precisely according to the constraint is a critical technology of parametric design. Dimension driven is just a mechanism where the outline of the graphics will be altered automatically as the value of the constraint is changed. This mechanism is a kind of operation on graphics data. It can modify the geometric data of the graphics. Meanwhile, it should satisfy the conditions of the geometric topology constraint at the same time. Apparel pattern CAD system, based upon parametric design, can improve greatly the efficiency of pattern design or pattern alteration. Users are required to design the topology structure of the pattern template only one time and a series of individual patterns will be generated conveniently by altering the geometric parameters. However, there are some problems in parametric design of apparel pattern. Apparel pattern is different with other industries’ products, like machine, because pattern has more flexible. It is well known that pattern making often depend on pattern-masters’ experience. That is why the development and application of parametric CAD systems in apparel are more difficult than other industry. Generally, parametric design systems are used for customizing the apparel pattern with relatively fixed topology structure, such as suit, jean and shirt. Individual pattern altering base on grading technique Alteration of garment patterns is an essential step in producing attractive and accurately fitting clothing from patterns that already exist. Commercial CAD alteration systems are based on one of the following two approaches. One approach uses basic patterns and apply a mathematical formula to change them to fit specific body measurements. Another approach uses graded patterns in conjunction with sizing and alteration tables to alter patterns to fit a person with specific body measurements (Istook, 2002; Turner, 1994). Both approaches have common feature of reference to size charts for either standard size grading or calculation of the alterations to be applied. Figure 2 shows the common workflow of pattern alteration in MTM CAD system. The altering process Develop point numbering

Develop alteration rules

Verify alteration rules

Knowledge database (Including numbering convention and alteration rules)

Individual pattern

Standard pattern

Figure 2. The workflow of pattern alteration based on traditional grading technique

Determine critical alteration points

This set of activities requires a strong knowledge of garment design, grading, and garment construction, as well as an understanding of how computer programs “think.” Generally, pattern altering technique base on grading rules adopted by some CAD system involves three main steps (Istook, 2002). Determine critical alteration points. Garments are generally altered at grade points or cardinal points. Since, a great deal of effort has been made previously to allow garment patterns to grow as the human body does, this makes sense. Therefore, critical alteration points are usually related to the human circumferential measurements of the neck, chest/bust, waist, hips, biceps, thigh, knee, and ankle, and the length measurements of arms, back of neck to waist, waist to hip, waist to knee, and waist to floor. However, since no one person is exactly average, normal grading methods leave much to be desired in relation to fit. Fortunately, CAD systems have the ability to alter garment patterns exactly where they need to be altered, if set up correctly and determined in advance. For this to occur the tailor must determine where fit problems might occur in any specific garment, and may be required to add points to these locations. Develop point numbering. In order for a CAD system to perform alterations on a pattern, it must be able to identify each point by number or name and it must be able to differentiate between points so that the requested action does not cause confusion or occur inappropriately. One of the most time consuming activities in the CAD pattern alteration process is the development of a numbering convention and the application of that convention to the pattern pieces and in alteration rules. If a numbering convention is developed in advance, the tailor can take advantage of that process in all similar garment situations. The first step in creating the numbering convention is to identify all of the garment categories that would have similar alterations performed on them. The second step is to get a layout of all of the pattern pieces used in the most complicated garment. The last step is to identify all of the points on any specific part of a garment that would be involved in the CAD alteration process. Develop alteration rules. For the fit of a specific garment, alteration rules are generally created for each key body measurement, such as the bust, waist, hips, back waist length, and waist to knee length. For women with significant fit problems, however, additional alteration rules might also be needed, such as across chest, across back, high hip, and combined thigh girth. Within each alteration rule, instructions must be developed to control every point on a garment related to a specific location where an alteration should occur. Instructions for alteration rules are based on types of movements that can be used to generate an alteration. The five types are as follows (Istook, 2002): (1) move along the counter-clockwise without extension; (2) move along the counter-clockwise with extension; (3) move along the clockwise without extension; (4) move along the clockwise with extension; and (5) X and Y move. To provide control for these alteration movements, hold points and move points must generally be defined. Beginning with one alteration rule at a time, identify all points that should be adjusted and determine the kind of move that would be most

The development of digital patterns 171

IJCST 19,3/4

172

appropriate for each. It is possible that more than one kind of move operation could cause the same change in the pattern piece. Once all of the alteration rules have been determined and the descriptions defined by the point number, movement type, and mount, an alteration rule table will be created in the CAD system. Verify alteration rules. Every alteration rule that is created must be checked for accuracy before it is implemented. Most of the CAD systems have a feature allowing the operator to test the alteration rule. In the Gerber system, the “display piece” function enables each pattern piece to be viewed with an alteration rule applied. The operator should test each rule using an exaggerated measure. Once each rule has been tested on its own, and then begins testing combinations of rules. When all rules have been tested using a variety of measurements and combinations, as might occur in real life, you can feel fairly comfortable that the rules have been created accurately, at least for this specific garment silhouette. Individual pattern generating based on artificial intelligence Individual pattern generating can basically be implemented with the two methods mentioned above. However, the parameters or the values of modification have to be inputted by pattern experts with a significant level of knowledge and practical experience. Pattern masters should take many years to gain experiences in flat pattern technique by intensive practice. Therefore, many researchers began to adopt AI to create expert pattern system in order to reduce the experience requirements. In relative research works (Ng, 2004), the authors have tried two different approaches to the simulation of the pattern masters’ expertise using artificial intelligent methods, namely artificial neural network and fuzzy logic (FL). As a pilot study, a total of 735 trousers patterns of various body sizes were generated. They were used to train the feed-forward back-propagation neural network and the FL control unit. Another research was carried out to predict shirt patterns from body measurements, fabric properties and fitting requirements using AI (Chen et al., 2003a, b). The application of artificial neural network. Artificial neural network has proven to be feasible and useful in developing models for nonlinear systems (Granino, 1995). It likes the proverbial black box. That is to say, the users have not got to understand how does the network learn, just to know how to use the network. From the view of math, apparel pattern is a set including point, line, curve and sizing. Pattern making in MTM system is just to decide the set according to individual body measurements and fabric properties. Meanwhile, pattern-making technique is a very complex process involving the nonlinear deformation of a 3D surface (garment) to the corresponding 2D pattern. So the problem of pattern making can be solved through constructing one ANN model, in which the input data is body measurements and fabric properties and the output data is the coordinate data of pattern. Figure 3 shows the three-layer BP neural network model for pattern-making system. Four phases are included as follows (Ng, 2004; Granino, 1995): (1) Data preparing. In the phase, sizing table and material properties should be determined. A large quantity of garment patterns with the same style needs to be made with CAD system. The accuracy of these patterns should be validated and the unfit patterns must be eliminated. (2) Data normalizing. In order to satisfied the conditions of ANN, the data of the experiment, including sizing table, material properties and the coordinate data of the patterns, should be transformed to the values within the interval [2 1,1].

hidden layer

The development of digital patterns

output layer Y1

X2

Y2 ...

...

Xn

Pattern data

X1

...

Body measurements Material properties

Input layer

Yn

173 Figure 3. Three-layer BP neural network model for pattern-making system

(3) The training phase of the model. The purpose of the phase is to control the prediction accuracy of the pattern within the acceptable range for apparel enterprise by carrying the training data set. (4) The validating phase of the model. In the last phase, the set of validating data will be used for evaluating the prediction accuracy of the ANN model. The results of previous works indicate that the accuracy is within the objective error range (Chen et al., 2003a; Granino, 1995). The application of fuzzy logic. There are many problems, which were not described precisely in pattern making. For instance, how fit is a garment? That belongs to vague concept. In pattern making, determining the ease value is also a vague problem and often depends on pattern experts’ experience. FL provides a simple way to arrive at a definite conclusion based upon vague, ambiguous, imprecise, noisy, or missing input information. FL’s approach to control problems mimics how a person would make decisions, only much faster. It uses an imprecise but very descriptive language to deal with input data more like a human operator. FL incorporates a simple, rule-based IF X AND Y THEN Z approach to a solving control problem rather than attempting to model a system mathematically. The FL model is empirically based, relying on an operator’s experience rather than their technical understanding of the system (Granino, 1995). Previous experiments (Ng, 2004) had been carried out aiming at create fuzzy expert pattern system and the results showed that it was feasible. Figure 4 shows the FL controller for pattern making. Like the ANN experiment, the FL linguistic variables should be normalized within the interval [0,1]. There are three problems should be solved in the FL controller:

Fuzzy reasoning

Rule table

Defuzzyfication

Material properties

Fuzzyfication

body Measurements

Static ease Compressive ease

+

Total ease

Figure 4. The FL controller for pattern making

IJCST 19,3/4

174

(1) Quantize the input-variables. The input-variables, including body measurements and material properties, should be transformed into the membership functions of the fuzzy sets for the reason that the FL controller can recognize the input values. (2) Generation of the rule-table and fuzzy reasoning. Firstly, a series of IF X AND Y THEN Z rules should be created according to the experience of apparel designers and pattern experts. Then FL produces the joint membership functions for all combinations of the inputs and outputs which are admitted by the given rule table. Finally, FL computes the membership function of the fuzzy logical union to all fuzzy sets. (3) Defuzzify the output-variables. A defuzzication algorithm generates a specific crisp output as a sort of weighted average over this computed output membership function. Pattern flattening technique based on 3D apparel model Most of commercial apparel CAD systems currently used by designers are 2D-based pattern design systems that do not include ways to visualize garments in 3D graphics. With the developing of 3D apparel CAD systems, considerable attention has been paid to surface flattening technique based on 3D apparel model. Surface flattening is applied in many applications, such as in the aircraft, shoe, and garment industries. Contrary to the problem of constructing garments, flattening the designed 3D surfaces to 2D plane is 3D to 2D problem. Fan et al. (1998) first introduced techniques of 3D-2D transformation. The problem definition can be described as follows (Mccartney et al., 2005; Wang et al., 2005): given a 3D freeform surface and the material properties, find its counterpart pattern in the plane and a mapping relationship between the two so that, when the 2D pattern is folded into the 3D surface, the amount of distortion – wrinkles and stretches – is minimized. 3D garment surface is a complex surface, which cannot be flattened 2D plane without distortion and error. According to the latest theoretic study of the surface flattening, two typical methods for surface fattening are outlined, namely geometry flattening and physical flattening: (1) Geometrical flattening. The geometrically-based flattening techniques do not consider the physical properties of materials. The method maps all 3D triangles of the surface in a 2D plane with some geometrical constraint conditions. Parida and Mudur (1993) presented the flattening approach of complex surface based on constraint satisfying. Shimada and Tada (1991) stated an approach to transform approximately an arbitrary curved surface into a plane using dynamic programming. Yang et al. (2001) introduced a method for making complex surface developable and developing it. A developable surface is obtained between two curves on the complex surface and approximates to complex surface. Using this method, 3D front garment basic pattern is made developable and developed into plane pattern. (2) Physical flattening. The physical-based technique usually represents cloth models as triangular or rectangular grids with points of finite mass at the intersections. The various forces and energies are considered in surface fattening. In Parida and Mudur (1993), a surface flattening algorithm based on the energy-minimization on the 3D surface is presented by Benjamin Yeung et al., which calculates the optimal 2D pattern that folds to the 3D surface with minimum deformation. Both isotropic

and woven-like anisotropic materials are supported by the algorithm. Darts insertion is also incorporated in the algorithm to further enhance the flattening result. In Mccartney et al. (2005), an orthotropic strain model is adopted to convert the strain values to energy values. Energy minimization analysis is performed during the flattening process to minimize distortion. Wang et al. (2005) presented a garment design system to design 3D garment around a 3D human model through 2D sketch strokes. The resulting garment meshes are defined as seam lines to split the garment mesh into pieces. A flattening procedure is performed to generate flat garment pieces that can be used for manufacturing the garment.

The development of digital patterns 175

In general, the physical flattening process from 3D garment surface can be summarized by following procedures: . Create the energy model for fabric distortion. During the flattening process, fabric distortion will be embodied. The energy model will calculate the strains implicit in distorting the 2D form of a triangle to its 3D representation. Figure 5 shows the three types of individual strains that contribute to the total distortion for a typical fabric. The detailed algorithm of the energy model can be found in previous works (Wang et al., 2005; Parida and Mudur, 1993; Shimada and Tada, 1991). . Map all 3D triangle into 2D plane with constraint. The appropriate starting point should be determined to avoid the energy’s building up dramatically by the time. Firstly, the first triangle is arbitrarily laid down in 2D in a completely undistorted manner. Then the three neighboring triangles can be mapped with constraint in order to assure the integrated topology relation of these triangles. According to the sequence, all 3D triangles are mapping into 2D plane. During this iterative process, if the 3D surface is not developable, it will soon emerge that a triangle must deform if it is to be flattened. . Minimize the energy by diffusion. Owing to the choice of the starting point and the mapping sequence of the triangles, the strain energy of the fabric model is usually not at the minimum state. Therefore, the strain energy must be minimized to obtain a better mapping by applying a diffusion-like process. The flattening technology has been adopted to develop pattern by several commercial 3D garment CAD systems, such as Gerber, PAD and DressingSim. v

v

v

v φ Sv

Su u Original woven element (a)

u

u

Weft strain

Warp strain

(b)

(c)

Notes: Su and Sv are the train force for the weft and warp directions respectively; φ is the angle of the shear transformation

u Shear stain (d)

Figure 5. Affine transformations of woven element

IJCST 19,3/4

176

However, the practicability of these systems is not good and there are always some distortions in the 2D pattern flattened from 3D apparel model. It is difficult to construct the real garment model in these systems, especially for the apparel with complex style. Therefore, many improvements should be done if the 3D CAD systems are applied in apparel mass customization. Conclusion MTM apparel pattern system is a vital part of mass customization. The paper outlines two paths of digital pattern making for customized apparel, namely 2D pattern making and 3D surface flattening. Three typical methods are discussed in the former path, including parametric design, traditional grading technique and pattern generating based on AI. For the first method, it can improve greatly the efficiency of pattern design or pattern alteration. However, the development and application of parametric CAD systems in apparel industry are more difficult than other industry, because apparel pattern has fewer laws in graphical structure. For the second method, it is the most practical method because of its simple theory, which pattern masters are familiar with. Currently, several companies have developed the MTM pattern systems based the method. But these systems required users with higher experience lever. Creating expert pattern system based on AI can reduce the experience requirements. Meanwhile, a great deal of experiments should be conducted for each garment with different style to create their knowledge databases. In another path, two methods for surface fattening are outlined, namely geometry flattening and physical flattening. The flattening technology has been adopted to develop pattern by several commercial 3D garment CAD systems. But many improvements should be done if the 3D CAD systems are applied in apparel mass customization. References Chen, A., Fan, J. and Yu, W. (2003a), “A study of shirt pattern drafting methods, part 1: experimental evaluation of existing methods”, Sen-I Gakkaishi, Vol. 59 No. 8, ISSN 0037-9875, pp. 319-32. Chen, A., Fan, J. and Yu, W. (2003b), “A study of shirt pattern drafting methods, part 2: prediction of shirt patterns using human body anthropometrical data”, Sen-I Gakkaishi, Vol. 59 No. 8, ISSN 0037-9875, pp. 328-33. Fan, J. et al. (1998), “A spring-mass model-based approach for warping cloth patterns on 3D objects”, The Journal of Visualization and Computer Animation, Vol. 9, ISSN 1099-1778, pp. 215-27. Granino, A.K. (1995), Neural Networks and Fuzzy-Logic Control on Personal Computers and Workstations, The MIT Press, London, ISBN 0262112051. Istook, C.L. (2002), “Enabling mass customization: computer-driven alteration methods”, International Journal of Clothing Science and Technology, Vol. 14 No. 1, ISSN 0955-6222. Mccartney, J., Hinds, B.K. and Chong, K.W. (2005), “Pattern flattening for orthotropic materials”, Computer-Aided Design, Vol. 37, ISSN 0010-4485, pp. 631-44. Ng, R. (2004), “Garment pattern design with artificial neural network and fuzzy logic”, paper presented at International Fashion Culture Festival, Shanghai. Parida, L. and Mudur, S.I.P. (1993), “Constraint-satisfying planar flattening of complex surface”, Computer Aided Design, Vol. 25 No. 4, ISSN 0010-4485, pp. 225-32.

Pine, B.J. (1993), Mass Customization, Harvard Business School Press, Boston, MA, ISBN 0875849466. Seung-Eun, L. and Chen, J.C. (2000), “Mass-customization methodology for an apparel industry with a future”, Journal of Industrial Technology, Vol. 16 No. 1, ISSN 08826404. Shimada, T. and Tada, Y. (1991), “Approximate transformation of an arbitrary curved surfaces into a plane using dynamic programming”, Computer Aided Design, Vol. 23 No. 2, ISSN 0010-4485, pp. 153-9. Turner, J.P. (1994), “Development of a commercial made-to-measure garment pattern system”, International Journal of Clothing Science and Technology, Vol. 6 No. 4, ISSN 0955-6222, pp. 28-33. Wang, C., Tang, K. and Yeung, B. (2005), “Freeform surface flattening based on fitting a woven mesh model”, Computer-Aided Design, Vol. 37, ISSN 0010-4485, pp. 799-814. Yang, J.X. et al. (2001), “A new method for marking complex surface developable and its development”, Mechanical Science and Technology, Vol. 20 No. 4, ISSN 1003-8728, pp. 520-2. Yang, Y. and Ping, L.Y. (2002), The Principle and Practice of CAD/CAM, China Railway Publishing House, Beijing. Zhang Hong, Z. et al. (2005), The Principle and Application of Apparel CAD, China Textile Press, Beijing. Corresponding author Yunchu Yang can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The development of digital patterns 177

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,3/4

Investigating the positioning edge accuracy of sensors in textile and clothing manufacture

178

Gojko Nikolic´ and Goran Cˇubric´ Department of Clothing Technology, Faculty of Textile Technology, University of Zagreb, Zagreb, Croatia Abstract Purpose – This paper seeks to pursue the research of different types of sensors suitable for positioning edge accuracy of textile material. Design/methodology/approach – A measuring device is used to install different types of sensors on the slide holders and to change their interspace as well as the space between the sensors and textile fabric. Findings – The new measuring equipment has been established. Research limitations/implications – Only the results of woven fabric measurement are analyzed in this paper, while the results of knitted and nonwoven fabric measurement will be elaborated in the future papers. Originality/value – The measuring equipment is original. Keywords Automation, Clothing, Sensors Paper type Research paper

1. Introduction The problem of accurate product positioning on machines for clothing production is more complex than similar problems in other fields of production and installation. The main reason for this is the type of material that is used. Namely, it is easier to position accurate over the edge of the product the parts produced from rigid materials. The problem is even more complicated in the positioning of materials like woven and nonwoven fabrics that are soft, pliable, elastic, porous, with uneven edges, etc. There are many different sensor types that could be used for this purpose on the market, but the systematic analysis or comparison of their real possibilities has not been carried out up to now. The aim of the experiment is described research. For this purpose has been manufactured a new measuring device which moves the textile material with the 0.01 mm shifting what enables the precise determination of the sensor activation spot or its detection characteristic. For the research have been used: capacitive, optical (light barrier and reflective) and pneumatic sensor (pneumatic barrier).The research has been long so only the results of woven fabric measurement are analyzed in this paper, while the results of knitted and nonwoven fabric measurement will be elaborated in the future papers. International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 178-185 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741641

2. Sensors The characteristic of capacitive sensors is its reaction on any material that reaches its proximity, either it is electrical conductive or non-conductive. They sense metal, wood, plastic, fabric, liquid, etc. as well as live body. They are based on the capacitor capacity

change in RC circle when some material approaches (Figure 1) (Nikolic´ 2001; Ebel and Nestel, 1992). The electric field exists between one “active” plate capacitor and other that represents the ground. When the target approaches, the dielectric changes. The capacity change depends on the gap of target, its mass (thickness) and dielectric constant. Capacitive sensors react at range from 5 to 20 mm. Optical sensors are the most extended sensors which are used in industry, as well in textile and clothing industry. They are sensors which use different light wave length for the target sense. In the praxis are mostly used the sensors that use red (visible) light ray in wave length of 660 nm and invisible infrared ray wave length around 880 nm. These sensors consist of emitter who emits a ray and receiver that collects the ray. Depending if emitter and receiver are in the same housing or they are separated, we differentiate some types of optical sensors. By the light barrier emitter and receiver are one against another. When the target arrives between them the light ray is interrupted and it senses the presence of target. By reflective optical sensors, emitter and receiver are in the same housing (Figure 2) (Herold, 1993). The ray reflection from the target catches receiver and output signal appears. When the optical fibers are added to sensors they make narrower light beam and it is possible to note very small objects or edges. Using these fibers is also possible to get the light barrier or reflective sensors (Figure 3).

Investigating the positioning edge accuracy 179

OUTDOOR SUPPLY INTERNAL SUPPLY CAPACITOR

OUTPUT SIGNAL

G

OSCILLATOR DEMODULATOR TRIGGER

LED

AMPLIFIER

Figure 1. Capacitive sensor diagram

OUTDOOR SUPPLY INTERNAL SUPPLY OSCILLATOR AMPLIFIER OUTPUT SIGNAL & FOTOELECTRIC EMITTER FOTOELECTRIC RECEIVER PREAMPLIFIER

SIGNAL CONVERTER LOGICAL OPERATION

LED

Figure 2. Reflective optical sensor diagram

IJCST 19,3/4

180

Pneumatic non-contact sensors use compressed airflow (pressure 2.5-20 kPa) to sense the object. Those are pneumatic barrier and pneumatic reflective nozzle (Figure 4(a) and (b)). When the target by pneumatic barrier comes into airflow work field, between transmitter and receiver nozzle it causes disarrangement of pressure or the pressure fall in receiver nozzle. By annular nozzle using reflected airflows (reflective nozzle), where the transmitter and receiver nozzle are in same housing, a target arrives near the sensor causing the increase of pressure in receiver nozzle. These disarrangements are used as output signal and they have low pressure (round $ 5 kPa) so it is necessary to reinforce them. 3. Experiment 3.1 Measuring device A special measuring device has been projected for the investigation (Figure 5(b)). It is possible to install different types of sensors (optical, capacitive, pneumatic, and in different constructions barrier or reflective) on the slide holders and to change their interspaced as well as the space between the sensors and textile fabric. The sample in size 70 £ 70 mm (8) is placed on the carrier (6) up to the anterior boundary (10) and fixed by the sample holder (4). The carrier with the sample is placed on the wiper (3) that is moved by micrometer in the direction of x-axis (5) with the movement readings of 0.01 mm. The boundary of the carrier insures the identical placement of the edges for all the samples. The sensors (9) placed on the holders can be arbitrary moved and fixed in the direction of y-axis. If one of the holders is not needed (reflective sensor type), it could be simple rotated for 908 and disengaged form work. The whole measuring device is placed on the massive pedestal (7).

EMITTER EMITTER

Figure 3. Optical sensors with optical fibers: (a) light barrier; and (b) reflective sensor

RECEIVER

RECEIVER (a)

(b)

2

2

1

1

1

1(P) 1(P)

1(P)

2(A)

2(A)

Figure 4. Non-contact pneumatic sensors: (a) pneumatic barrier; and (b) reflective nozzle

1(P) 2(A)

1(P)

1(P) 2(A)

TRANSMITTER NOZZLE

(a)

RECEIVER NOZZLE

(b)

9

Y 4

8

5

10 X 9

1

Investigating the positioning edge accuracy 181

6 3 2

Figure 5. Measuring device: (a) photography; and (b) scheme

7

(a)

(b)

3.2 Measured samples For the investigation has been chosen nine woven fabrics that differ in weave, thickness, the mass per unit area and the weave density (Table I). The samples were chosen intentionally to investigate the influence of different factors on the results and measuring possibilities as well as to investigate the hysteresis of each sensor type. Hysteresis is a difference between the moments of sensors turning on and turning off, when the edge of textile material approaches the sensor and moves away. From the each woven fabric has been cut a sample in dimensions 70 £ 70 mm. 3.3 Measurement For each sample has been done 30 measurements on each individual sensor distance from woven fabric and the average value of obtained results, has been calculated. It is well known from practical work that sometimes it is not possible to put the sensor near the place in which it must to sense something. In that case we must put sensor further from the measurement place. That is why the research was founded on how each sensor reacts on different distances from fabric. Distances have been determined according to experimental measurements which have been done earlier. For the research with optical sensors the optical fibers attached on optical connector have been used. Obtained sensor signal is forwarded to distribution unit which besides LED has the sound signal too. Optical sensor works with visible (red) light ray in wave length of 660 nm. By the research with pneumatic sensors has been used low pressure amplifier unit that

Sample

Color

TB1 TB2 TB3 TS1 TS2 TS3 TC1 TC2 TC3

White Light brown Light brown Grey Green Brown Black Dark blue Black

Mass per unit area g/m 181 253 353 128 188 365 138 284 319

2

The weave Cloth Cloth Cord Twill 3/1 Cloth Cord Cloth Cloth Cord

Weave density threads/cm Warp Weft 24 12 21 50 27 21 30 18 21

24 12 17 23 25 17 27 6 17

Table I. The sample designation and description

IJCST 19,3/4

182

reinforces signal from $ 0.05 kPa to 5 £ 105 Pa (Nikolic´ 1990). Such a signal is converted into electrical signal by pneumatic-electrical converter (Figure 6). 4. Results In the Table II are shown the results of measurements using the optical sensor (light barrier) and in Table III the results of measurement using the reflex optical sensor. The hysteresis is in mm. The Figures 7 and 8 shows graphically the results from the tables. 24V

2

EMITTER

AMPLIFIER

1

PNEUMATIC ELECTRICAL CONVERTER

BN WH BU

Figure 6. Schematic illustration of connection: (a) light barrier; and (b) pneumatic barrier

500 kPa COMPRESOR

0V

1

5 kPa RECEIVER

(a)

(b)

The hysteresis for each sample in mm Distance between the sensor and fabric in mm TB1 TB2 TB3 TS1 TS2 TS3 TC1 TC2 TC3

Table II. The hysteresis obtained by light barrier

1 3 5 7 9 11 13 15 18 21 24 27 30 33 36 39 42 45 48 51 55 60 65 70 80 90 100

0.38 0.23 0.17 0.13 0.36 0.07 0.18 0.06 0.15 0.05 0.11 0.05 0.07 0.05 0.17 0.06 0.06 0.05 0.08 0.05 0.05 0.04

0.06 0.10 0.15 0.20 0.23 0.26 0.28 0.30 0.34 0.32 0.30 0.21 0.16 0.12 0.08 0.07 0.06 0.06 0.05 0.05 0.05 0.05 0.04 0.04 0.04 0.04 0.04

0.61 0.26 0.36 0.24 0.17 0.18 0.13 0.12 0.08 0.07 0.06 0.05 0.05 0.05 0.05 0.05 0.04 0.04

0.07 0.10 0.14 0.18 0.19 0.18 0.20 0.19 0.18 0.18 0.17 0.14 0.10 0.09 0.07 0.06 0.05 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03

0.05 0.13 0.14 0.10 0.17 0.13 0.13 0.10 0.07 0.07 0.07 0.07 0.08 0.07 0.05 0.06 0.05 0.04 0.04 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03

The hysteresis for each sample in mm Distance between the sensor and fabric in mm TB1 TB2 TB3 TS1 TS2 TS3 TC1 TC2 TC3 1 3 5 7 9 11 13 15 18 21

0.12 0.10 0.12 0.13 0.15 0.20 0.31 0.35 0.52 0.99

0.09 0.09 0.14 0.20 0.25 0.31 0.41 0.47 0.60 1.30

0.09 0.11 0.14 0.16 0.22 0.27 0.37 0.48 0.53 0.69

0.11 0.11 0.14 0.18 0.24 0.29 0.36 0.44 0.66 0.89

0.14 0.11 0.18 0.24 0.31 0.45 0.60 0.86 1.17 1.20

0.07 0.10 0.17 0.22 0.28 0.36 0.46 0.57 0.65 0.74

0.05 0.13 0.20 0.30 0.42 0.70 0.99 0.99

0.09 0.15 0.23 0.29 0.37 0.48 0.79 1.27 1.39 1.50

0.08 0.14 0.21 0.32 0.38 0.51 0.72 0.80 0.92 1.19

Investigating the positioning edge accuracy 183 Table III. The hysteresis obtained by reflex optical sensor

LIGHT BARRIER 0.70 TB2

HYSTERESIS in mm

0.60

TB3 0.50

TS1

0.40

TS2

0.30

TS3

0.20

TC1 TC2

0.10

TC3

Figure 7. Hysteresis with light barrier

0.00 0

10

20

30

40

50

60

70

80

90

100

DISTANCE FROM THE SAMPLE in mm

REFLECTIVE OPTICAL SENSOR

HYSTERESIS in mm

1.60

TB1

1.40

TB2

1.20

TB3

1.00

TS1 TS2

0.80

TS3 0.60

TC1

0.40

TC2

0.20

TC3

0.00 1

3

5

7

9

11

13

15

DISTANCE FROM THE SAMPLE in mm

17

19

21

Figure 8. Hysteresis with reflective optical sensor

IJCST 19,3/4

184

5. Discussion 5.1 Research with optical sensors The results of research with optical sensors (light barrier) are shown in Table II and Figure 7. It is seen that sensor did not sense fabric TB1. The cause is the passing of one part of light through the fabric that has low weave density. The other fabric sensor did not sense bright color until the large distance. Sensor has sense fabrics between bright and dark color (TS1, TS2, TS3), that he sense fabric TS1 not until larger distance. Dark color fabrics (TC1, TC2, TC3) have been sensed too. Only the fabric TC1 sensor sensed at distance of 21 mm. The cause of it is like in previous case, one part of light pass through the fabric. From Figure 7, it could be seen that the hysteresis at beginning increase of distance between sensor and fabric and later it decrease. At larger distances hysteresis is approximately the same. That could be explained with dissipation of light with larger distance and lower light intensity that comes to receiver. It could be also seen from Figure 7 that hysteresis vary depending on distance of sensor from fabric. Fibers (or yarns) turned one part of light to receiver and caused the bringing of large amount of light to sensor. Therefore, histeresis value varies. Every fabric has start at different sensor distance from woven fabric because of different weave density. The results of research with reflective optical sensors are shown in Table III and Figure 8. It is seen that hysteresis increases as the sensor moves from woven fabric. For darker fabric the hysteresis is larger then for other. That was expected because darker materials reflect smaller amount of light then brighter fabrics. For the sample TC1 with lower weave density, the part of light did not reflect, but has passed through material. Therefore, reflective sensor did not sense TC1 at distances larger then 15 mm. Hysteresis is at distance of 21 mm larger then 1 mm and the research did not continue at larger distances. 5.2 Research with pneumatic sensors The results of the research with pneumatic sensor (pneumatic barrier) have caused a problems. The woven fabric vibrates and moves because of the compressed airflow. The range of vibration depends on fabric itself (the mass per unit area), compressed air pressure and sensor distance from fabric. When the pressure is higher, the amplitudes of vibrating fabric, the amplitudes of fabric vibration are higher too. The fabric vibration appears already at the pressure of 1 kPa (round 1 mm). When the fabric approaches the sensor, at one moment the edge of the fabric moves back (Figure 9). That occurs because of the air turbulence which appears under the fabric. The air lifts the fabric and moves edge of the fabric inwardly. Therefore, it is not possible to determine when the sensor has sensed the fabric. When the air pressure (105 Pa) is higher the problems are better expressed. The researches with pneumatic reflective sensor had not been performed because the problems are the same as those by the research with air barrier (vibrates and moves of fabric). 5.3 Research with capacitive sensor The research with capacitive sensor (standard construction) has shown that its sensibility is not high enough to react at only one layer of fabric. Sensor did not sense any of researched fabrics. The further research has shown that capacitive sensor senses fabrics with higher mass per unit area. The measurement have shown that sensor reacts on the closing of two or more connected parts of same fabric with total mass per unit area higher than 550 g/m2.

Investigating the positioning edge accuracy 185

Figure 9. The edge moving by the research with pneumatic sensor (pneumatic barrier)

6. Conclusion The research has shown that pneumatic sensors are not appropriate for positioning edge accuracy of woven fabric. If used, appear the problem of fabric vibration and movement of fabric edge inside because of turbulence. Also the used capacitive sensors are not appropriate for positioning because they sense fabric with mass per unit area higher than 550 g/m2. They are more appropriate for determination of whole clothing or parts of cloth. By the used of optical sensors we can very strictly determine of presence of fabric edge. The positioning can be done with reflective sensors but the distance between sensor and fabric should not be higher than 11 mm. At larger distances hysteresis is too high for accuracy positioning. For larger distances light barrier could be used. Optical sensor hysteresis is within the limit of acceptability for textile and clothing industry and can be used for positioning edge accuracy of textile material. References Ebel, F. and Nestel, S. (1992), “Sensors for handling and processing technology”, Proximity Sensors, Festo Didactic KG, Esslingen, ISBN 3-8127-3046-4. Herold, H. (1993), Sensortechnik, Hu¨thig Buch Verlag, Heidelberg, ISBN 3-7785-2138-1. Nikolic´, G. (1990), Pneumatsko upravljanje, Sveucˇilisˇna naklada Liber, Zagreb, III. izdanje, ISBN 86-329-0242-3. Nikolic´, G. (2001), Osnove automatizacije strojeva za proizvodnju odjec´e, TTF and Zrinski d.d., Cˇakovec, ISBN 953-155-056-5. Corresponding author Gojko Nikolic´ can be contacted at: [email protected] To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,3/4

Image analysis method of surface roughness evaluation Jirˇ´ı Militky´ and Miroslav Mazal

186

Department of Textile Materials, Technical University of Liberec, Liberec, Czech Republic Abstract Purpose – The main aim of this paper is description of new apparatus and approach for contact less evaluation of surface roughness. For characterization of surface roughness, the procedures based on classical and non-classical (complexity) parameters are proposed. Design/methodology/approach – For obtaining the roughness profile in the selected direction (on the line transect of the surface), the special arrangements of textile bend around sharp edge is used. The image analysis is used for extraction of surface profile. The system of controlled movement allows one to obtain surface roughness profile in two dimensions. Findings – By using aggregation (cut length principle), the roughness resolution is decreased and roughness profile is created without local roughness variation. After application of cut length principle, the direct combination of slices leads to the creation of roughness surface. Research limitations/implications – There exists plenty of roughness characteristics based on standard statistics or analysis of spatial processes. For evaluation of suitability of these characteristics, it will be necessary to compare results from sets of textile surfaces. Practical implications – The measurement of fabric roughness by an RCM device is useful as simple tool for description of roughness in individual slices and in the whole rough plane. This method replaces the traditional contact stylus profiling methods Originality/value – The reconstruction of surface roughness from individual slices. The utilization of aggregation principle for creation of micro and macro roughness. The evaluation of roughness parameters based on the geometrical characteristics, harmonic analysis and complexity indices. Keywords Surface-roughness measurement, Profile measurement, Image scanners Paper type Research paper

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 186-193 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741650

1. Introduction The wearing comfort is important for all kinds of clothing type fabrics. Important parts of mechanical comfort are tactile properties including roughness. Roughness of engineering surfaces has been traditionally measured by the stylus profiling method creating of surface profile called surface height variation (SHV) trace (Greenwood, 1984; Kawabata, 1980). Modern methods are based on the image processing of surface images or images of properly bend fabric. Surface irregularity of plain textiles has been identified by friction, contact blade, lateral air flow, step thickness meter or subjective assessment (Greenwood, 1984; Kawabata, 1980; Militky´ and Bajzı´k, 2001; Ajayi, 1988; Ajayi, 1992; Ajayi, 1994; Stockbridge et al., 1957). Standard characteristics of surface profile are based on the relative variability characterized by the variation coefficient (analogy with evaluation of yarns mass unevenness) or simply by the standard deviation. Standardized parameters describing roughness of technical surfaces are given in the ISO 4287 standard ISO 4287 (1997). For characterization of roughness of textiles surfaces, the mean absolute deviation (MAD) (denoted by Kawabata as SMD) is usually used.

The main aim of this contribution is description of new apparatus and approach for contact less evaluation of surface roughness. This approach is based on the image analysis of especially prepared fabric images. For obtaining the roughness profile in the selected direction (on the line transect of the surface), the special arrangements of textile bend around sharp edge is used. The image analysis is used for extraction of surface profile. The system of controlled movement allows obtaining surface roughness profile in two dimensions (surface heights in the whole plane).

Image analysis method

187

2. Non-contact roughness evaluation Roughness profile of textile surfaces at given position along machine direction can be obtained from the analysis of especially prepared fabric images. For good image creation, it is necessary to select suitable lighting and fabric arrangement (Figure 1). The original system RCM designed for non-contact surface roughness evaluation uses the special arrangements of textile bend around sharp edge (Figure 2) and laser lighting from the top. Result after image treatment is so-called “slice” which is the roughness profile in the cross direction at selected position in machine direction (the line transect of the fabric surface). The system RCM allows reconstruction of surface roughness plane in two dimensions (Figures 1 and 2). For this purpose, the sample holder is moved (step by step) in controlled manner. From set of these slices, it is possible to reconstruct the roughness plane as well. Power Pc+software

Control unit

Motor Sample holder

Light source

Measuring edge

Laser

CCD camera

Optical system

Figure 1. RCM system components

Power

(a)

(b)

(c)

Figure 2. Details of RCM apparatus: (a) fabric arrangements; (b) CCD camera position; (c) lighting system

IJCST 19,3/4

188

3. Surface roughness characterization There are two reasons for measuring surface roughness. First is to control manufacture and is to help to ensure that the products perform well. In the textile branch, the former is the case of special finishing (e.g. pressing or ironing) but the later is connected with comfort appearance and hand. From a general point of view, the rough surface display process which have two basic geometrical features: (1) Random aspect. The rough surface can vary considerably in space in a random manner, and subsequently, there is no spatial function being able to describe the geometrical form. (2) Structural aspect. The variances of roughness are not completely independent with respect to their spatial positions, but, their correlation depends on the distance. Especially, surface of weaves is characterized by nearly repeating patterns and therefore some periodicities are often identified. Characterization of roughness based on the MAD is the classical descriptive statistical approach. This statistical characteristic is useful for random SHV traces where elements of SHV trace are statistically independent each other. The SHV profile of a lot of fabrics has been identified as irregular and more structured. From the SHV or SFV, trace is possible to evaluate a lot of roughness parameters. Let us define roughness characteristics for SHV (the same equations are also valid for SFV). Classical roughness parameters are based on the set of points z(dj) j ¼ 1. . . M defined in the sample length interval L. The measurement points dj are obviously selected as equidistant and then z(dj) can be replaced by the variable zj. The standard roughness parameters used frequently in practice are: . MAD. This parameter is equal to the mean absolute difference of surface heights from average value (Ra). For a surface profile this is given by: MAD ¼

.

1 X zj 2 R a M j

ð1Þ

This parameter is often useful for quality control. However, it does not distinguish between profiles of different shapes. Its properties are known for the case when zj’s are independent identically distributed (IID) random variables (Meloun et al., 1994). Standard deviation (Root Mean Square) Value SD. This is given by: sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 X ðzj 2 Ra Þ2 SD ¼ ð2Þ M j Its properties are known for the case when zj’s are IID random variables. One advantage of SD over MAD is that for normally distributed data can be simple to derive confidence interval and to realize statistical tests. SD is always higher than MAD and for normal data is SD ¼ 1.25 MAD. It does not distinguish between profiles of different shapes as well. The parameter SD is less suitable than MAD for monitoring certain surfaces having large deviations (corresponding distribution has heavy tail) (Meloun et al., 1994).

.

Mean height of peaks (MP). This is calculated as the average of the profile deviations above the reference value R (often is R ¼ Ra). It is given as mean value of peaks Pi, i ¼ Np, where: P i ¼ zi 2 R

.

.

.

.

.

for zi 2 R . 0 and P i ¼ 0 elsewhere

Mean height of valleys (MV). This is calculated as the average of the profile deviations below the reference value R (often is R ¼ Ra.). It is given as mean value of valleys Vi, i ¼ Nv, where: V i ¼ R 2 zi

Image analysis method

for zi 2 R , 0 and V i ¼ 0 elsewhere

The parameters MP and MV give information on the profile complexity. Exceptional peaks or valleys are not considered but are useful in tribological applications. (v) The standard deviation of profile slope (PS). This is given by: vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ u u 1 X dzðxÞ 2 PS ¼ t ð3Þ M j dx j The standard deviation of profile curvature (PC). This quantity called often as waviness is defined by the similar way: vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ u u 1 X d2 zðxÞ 2 ð4Þ PC ¼ t M j dx 2 j The slope and curvature are characteristics of a profile shape. The PS parameter is useful in tribological applications. The lower the slope the smaller will be the friction and wear. Also, the reflectance property of a surface increases in the case of small PS or PD. Mean slope of theprofile (MS). This is given by: 1 XdzðxÞ ð5Þ MS ¼ M j dx j Mean slope is an important parameter in several applications such as in the estimation of sliding friction and in the study of the reflectance of light from surfaces. (viii) Ten point average (TP). This characteristic is defined as the average difference between the five highest peaks and five deepest valleys within a surface profile. The parameter TP is sensitive to the presence of high peaks or deep scratches in the surface and is preferred for quality control purposes.

These parameters are useful in the case of functional surfaces or for characterizing surface bearing and fluid retention and other relevant properties. For the characterization of hand will be probably best to use waviness PC. The characteristics of slope and

189

IJCST 19,3/4

190

curvature can be computed for the case of fractal surfaces from power spectral density, autocorrelation function or variogram (Zhang and Gopalakrishnan, 1996). 4. Experimental part A multifilament filament fabric with relatively good structural homogeneity was selected for demonstration of RCM system capability. The fabric surface is shown on the Figure 3(a). Individual slices roughness profile were created by using of threshold and set of morphological operations. Result of these operations is vector of surface contours in cross direction at specified machine direction. By using aggregation, the resolution is decreased and roughness profile is created without local roughness variation. This aggregation has the same function as cut length at unevenness measurement and therefore we use word “cut length X” for the case of aggregation of X subsequent values. After application of cut length principle, the direct combination of slices leads to the creation of roughness surface (Figure 4). The principle of aggregation can be used for smoothing in the machine direction as well (it is equal to the local averaging of slices). Output from data pretreatment phase is array of slices, i.e. array of vectors Rj(i ) where index i corresponds to the position in jth slice. The simplest way to roughness characterization is to use characteristics for individual slices and averaging or plotting of characteristics for all slices. For computation of these characteristics, the program ROSQ in MATLAB has been created. The following characteristics are computed: . Mean absolute deviation (MAD). . Mean profile slope (MS).

Figure 3. Tested fabric: (a) surface view; (b) raw image of one slice

0.1 mm

2 1 0 200

2 1 0 200 150

150 100

Figure 4. Roughness surface: (a) cut length 1; (b) cut length 10

100 50 0

50

100

150

200

50 50

150

100

0 (a)

(b)

200

. . . . . .

Image analysis method

Standard deviation of profile slope (PS). Standard deviation of profile curvature (PC). Ten point average (TP). Variation coefficient CV ¼ SD/Ra. Mean fractal dimension DF. Initial fractal dimension DFp.

191

For characterization of surface spatial roughness, the variograms and fractal dimensions in selected directions are computed. The overall roughness variation is divided into parts corresponding to the selected directions 5. Results and discussion For illustration of ROSQ capability, the PC for cut length 1 and cut length 10 are shown on the Figure 5. In Table I, the mean values of variation coefficients for all slices are given. For evaluation of periodicities, the frequency corresponding to the maximum amplitude of slices were computed. Results are given in Table I. It is visible that for cut length 1 (resolution 2 mm) corresponds the maximum amplitude to the filament diameter and for cut length 10 (resolution 20 mm) is maximum amplitude roughly equal to the yarn diameter. For characterization of roughness in plane, the CV in various directions were computed. Results are given in Table I. It is visible that in the main directions is CV for cut length 1 higher that CV for cut length 10. The excess of CV, i.e. dCV2 ¼ CV2 ðl ¼ 1Þ 2 CV2 ðl ¼ 10Þ can be used as characteristics of micro-roughness. curvature 0.023 0.022 0.021 0.02 0.019 0.018 0.017 0.016 0.015 0.014 0.013

curvature 0.135 0.13 0.125 0.12 0.115 0.11 0.105

0

50

100

150

200

250

300

350

400

0.1

0

10

(a)

Characteristics Variation coefficient (percent) Fourier frequency (mm) CV machine (percent) Variance cross (percent) Variance longitudinal (percent) Variance transversal (percent) Overall dispersion index

20

30

40

50

60

70

80

Figure 5. Standard deviation of profile curvature PC of individual slices: (a) cut length 1; (b) cut length 10

(b)

Cut length 1

Cut length 10

29.81 8.52 0.0239 0.48 4.77 5.21 0.00257

31.06 900.5 0.0114 0.175 4.95 5.05 0.00149

Table I. Some characteristics of mean slice roughness and surface roughness

IJCST 19,3/4

192

For deeper investigation of anomalous regions, the moving windows were used. Principle is division the study area to the several local neighborhoods of equal size (moving windows) and within each local window the mean and variance are computed. The dimension of moving windows can be gradually changed to obtain good identification of local anomalies. The plot of local means and variances are shown in Figure 6. The slice means and variances are shown as well. There are visible some departures from constancy of mean and variance (stationarity). 6. Conclusion The contact less measurement of fabric roughness by using of RCM device is useful for description of roughness in individual slices and in the whole rough plane. There exists a plenty of roughness characteristics based on standard statistics or analysis of spatial processes. For evaluation of suitability of these characteristics, it will be necessary to compare results from sets of textile surfaces. The program ROSQ can compute variograms for selected directions (directions 0, 45 and 908). Averaging of variograms calculated in all directions leads to the omnidirectional variogram. The omidirectional variogram can be used for computation of fractal dimension as well. Another possibility is to use Fourier transformation for computation of power spectral density estimation. The system RCM can be used for characterization of other peculiarities of planar textile surfaces as surface hairiness, pilling and abrasion. Window local means

Window local variances

1.5

0.08 0.06

1 0.04 0.5 0.02 0

0.01 0

1

2

3

0

1

2

3

× 10 4

× 10 4

Row local mean

Figure 6. Windows local characteristics (first row) and slice local characteristics (second row) A

Row variances

0.8

0.05

0.7

0.04

0.6

0.03

0.5

0.02

0.4 0

50

100

150

200

0.01 0

50

100

150

200

References Ajayi, J.O. (1988), “Some studies of frictional properties of fabrics”, doctoral thesis, University of Strathclyde, Glasgow. Ajayi, J.O. (1992), “Fabric smoothness, friction and handle”, Text. Res. J., Vol. 62, pp. 87-93. Ajayi, J.O. (1994), “An attachment to the constant rate of elongation tester for estimating surface irregularity of fabric”, Text. Res. J., Vol. 64, pp. 475-6. Greenwood, J.A. (1984), “A unified theory of surface roughness”, Proc. Roy. Soc. London, Vol. A393, p. 133. ISO 4287 (1997), Geometrical Product Specification, GPS-surface Texture, Profile Method – Terms, Definitions, and Surface Texture Parameters, Beuth Verlag, Berlin. Kawabata, S. (1980), The Standardization and Analysis of Hand Evaluation, 2nd ed., HESC, Text. Mach. Soc., Osaka. Meloun, M., Militky´, J. and Forina, M. (1994), Chemometrics for Analytic Chemistry Vol. I and II, Statistical Models Building, Ellis Horwood, Chichester, ISBN 0-13-123788-7. Militky´, J. and Bajzı´k, V. (2001), “Characterization of textiles surface roughness”, Proc. 7th Int. Asian Textile Conference, Hong Kong August 2001. Stockbridge, H.C. et al., (1957), “The subjective assessment of the roughness of fabrics”, J. Text. Inst., Vol. 48, pp. T26-T34. Zhang, C. and Gopalakrishnan, S. (1996), “Fractal geometry applied to online monitoring of surface finish”, Int. J. Mach. Tools Manufact., Vol. 36, pp. 1137-50. Corresponding author Jirˇ´ı Militky´ can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Image analysis method

193

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,3/4

Garment abrasion strength evaluation: a comparative methods study

194

Rui Miguel, Jose´ Lucas, Lurdes Carvalho and Manuel Santos Silva Textile and Paper Materials Research Unit, Department of Textiles Science and Technology, University of Beira Interior, Covilha˜, Portugal, and

Albert Manich CSIC – Research and Development Centre, Barcelona, Spain Abstract Purpose – The breaking strength compromises fabric wearing out during wear of garments and is a determining parameter in their useful life. Thus, it is intended to compare the efficacy of each method concerning the understanding of results, which is, the explanation of the phenomenon, namely through statistical models which characterize abrasion strength, measured by each method, as a function of fabric assurance by simple testing (FAST) parameters. Design/methodology/approach – The simulation of abrasion mechanism was done on Martindale wear and abrasion tester, following two ways: the weight loss and two yarns breakage methods. The average weight loss of a fabric was determined among four specimens (in mg/5,000 cy). Fabric abrasion was done against a standard wool fabric under a 12 KPa pressure. In the two yarns breakage method, the number of abrasion cycles required to break two yarns is determined according to Woolmark Company TM 112 test standard. Findings – In general, no better models to explain abrasion strength by the breakage of two yarns method are achieved when FAST variables are considered respecting those based on classical ones. However, for woollen fabrics, there is an interesting model, which gathers classical and FAST variables with higher explanation: shear rigidity and polyamide composition. This allows one to conclude that this FAST parameter performs an important role in the abrasion behaviour of fabrics. Originality/value – As the abrasion consequences cause fabric degradation, a better understanding is obtained of the different abrasion stage mechanisms, which explain both evaluation methods: breakage of two yarns and weight loss. It is intended to gather conditions for a future definition of a single and comprehensive evaluation methodology of abrasion. Keywords Textile manufacturing processes, Fabric testing, Clothing, Abrasion, Statistical testing, Quality control Paper type Research paper

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 194-203 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741669

1. Introduction The breaking strength compromises fabric wearing out during wear of garments and is a determining parameter in their useful life. The wearing out is the change of initial fabric when subjected to multiple functions during wear. At the beginning of abrasion, fabric becomes to suffer a surface appearance change resulting from the alteration of pile distribution and from fibre ends cut. After, all fibre ends have been removed, fabric The authors wish to thank to Ms. Margarida Rodrigues at the Textile Department Fabrics Laboratory of University of Beira Interior for her contribution to the experimental work.

surface becomes completely clean (weight loss method). Continuing abrasion, fabrics suffer a yarn structure degradation ending by yarn breakage (yarn breakage method). After Allen et al. (1980), during abrasion testing of wool fabrics using the Martindale apparatus, the first fibre breakages are produced by fibrillation as a result of bending, twisting and tensile forces caused on fibres during abrasion. The preferential cutting planes start at intermolecular bonds and intermicrofibrilar connections. The effect of yarn and fabric structures, as well as of finish, on abrasion strength is explained by the structure ability to dissipate the abrasion energy. Among the most important structural characteristics, fabric weight per unit area and yarn twist may be considered. The wearing out phenomenon caused by abrasion and its laboratory evaluation has not found a satisfactory solution yet. The simulation of abrasion mechanism is carried out on Martindale wear and abrasion tester, following the weight loss and two yarns breakage methods. The Martindale apparatus comprises four testing heads. The fabric sample under study slides, under a given pressure, against the abrasive surface according to two simple harmonic movements in the horizontal plane. In the weight loss method, the abrasion stage before the structural degradation of yarns is analysed and the testing methodology followed evaluates the weight loss of fabric after 5,000 abrasion cycles (Miguel, 2000). In the two yarns breakage method, the number of abrasion cycles needed to break two yarns is determined, this meaning a structural degradation of yarns. Manich et al. (2001) estimated the effect of abrasion on the surface and the structural degradation of fabric. To this purpose, they have studied the abrasion resistance of wool fabrics using the Martindale apparatus according to the British Standard method, and then used a fast method to determine the initial weight loss rate and the mean abrasion gradient (MAG) by fitting the weight loss over the first 5,000 abrasion cycles, using Ratkowsky’s regression models. They fitted the model: X ð1Þ Y¼ A£X þB to the weight loss results, where Y is the weight loss in mg and X is the number of abrasion cycles (cy). All the fits were highly significant. The initial weight loss rate (1/B) in mg/cy seems to be related to the surface degradation of the fabric. In early stages, the higher the values of B, the lower will be the rate of surface degradation. They obtained the mean value of the weight loss rate for 5,000 abrasion cycles, this is, the MAG in mg/cy, by the definite integral of the fabric weight loss distribution divided by the number of abrasion cycles between 0 and 5,000 cy: 1 5; 000 þ ðB=AÞ ð2Þ MAG ¼ log A B=A They used the multiple regression analysis to relate both the initial weight loss rate (1/B) and the MAG with the fibre, yarn and fabric characteristics. Equation (3) shows that the initial weight loss rate (1/B) in mg/1,000 cy increases with the wool composition (La) and weight per square meter of the fabric (PM2) and decreases with fineness (Fi) and the number of ply threads per yarn (NC): 1 ¼ 12:12 þ 0:09La þ 0:08PM2 2 1:19Fi 2 3:27NC B

ð3Þ

Garment abrasion strength 195

IJCST 19,3/4

Equation (4) shows that the MAG increases with the PM2, the weave coefficient (CL) and the yarn count (Tex) and decreases with the NC: MAG ¼ 270:21 þ 0:33PM2 þ 23:41CL 2 9:24NC þ 0:68Tex

196

ð4Þ

Equations (3) and (4) show that there is a significant relationship between fibre, yarn and fabric characteristics with the surface and structural abrasion of the wool and blended fabrics for the 5,000 first cycles of abrasion. The higher the PM2 and the lower the NC, the higher the surface and structural abrasion of the fabric. This means that the abrasion resistance of woollen fabrics is lower than that of worsted fabrics. It seems that fibre characteristics like wool ratio and fineness play a significant role in surface abrasion, while yarn and fabric characteristics like Tex and CL are significantly related with structural abrasion. During the last years, it has been a subject of research the objective evaluation of subjective properties of fabrics. The most common example is the objective evaluation of fabric handle, for which several mechanical and physical parameters contribute, that also influences the fabric behaviour during garment making and wear, mainly those that are measured by the Fabric Assurance by Simple Testing (FAST) system. Reference is made to bending rigidity (FB), extensibility to 100 g/cm (F100), formability (FF), hygroscopic expansion (FHE), relaxation shrinkage (FRS), shear rigidity (FG), surface thickness (FST) and released surface thickness (FSTR). With this study, it is intended to compare the efficacy of each method concerning the understanding of results, this is, the explanation of the phenomenon, namely through statistical models which characterize abrasion strength, measured by each method, as a function of fabric FAST parameters. 2. Materials and methods 2.1 Selected fabrics The study is done with 82 wool and blended fabrics, 51 worsted and 31 woollen fabrics, having different structural characteristics. The only one, which is common to all fabrics, is the existence of wool in their compositions. Thus, selected fabrics have different compositions, PM2, yarn types, opacity (CoT), weave and finish, so that they are representative of industrial fabrics of the wool industry. The evaluation of fabric structural characteristics was done by laboratorial testing and according to current standards. Following are presented the tested characteristics and their corresponding units: . Composition (La, PE, PA, EA) – in per cent – standard NP 2248. . Average fibre fineness (Fi) – in m – standard NP 3160. . Average fibre length (Co) – in mm – standard BS 6176:1981(1991) – WIRA apparatus. . Yarn count and number of yarn plies (Tex, NC) – in Tex – standard NP 4105. . Twist of yarns and plies – turns/m – standard NP 4104. . Weight per square meter (PM2) – in g/m2 – standard NP 1701. . Thickness (ESh) – in mm – Standard UNE 40-224-73 – Louis Schopper apparatus. . Yarn density (D) – in yarns/cm – standard NP EN 1049-2. . Weave – standard NP 4114/1700.

Based on the tested structural characteristics and mathematically relating some of them, we find parameters that contribute for the complete definition of fabric construction: twist coefficient (CTt), CL, weave type (TL), average float (Amed), cover factor (FaC), CoT and porosity (Po). The kind of finish (TA) was also determined according to a 1-5 scale as a function of the fabric felting degree. Considering that FAST system developed at CSIRO Textile and Fibre Technology, is important and easy to determine parameters to evaluate fabric behaviour during cloth manufacturing and wear, they have been included in the set of variables to be studied. Reference is made to FB, extensibility to 100 g/cm (F100), FF, FHE, FRS, FG, FST and FSTR. 2.2 Applied measurement techniques The simulation of abrasion mechanism was done on Martindale wear and abrasion tester, following two ways: the weight loss and two yarns breakage methods. The Martindale apparatus comprises four testing heads. The fabric sample under study slides, under a given pressure, against the abrasive surface according to two simple harmonic movements in the horizontal plane. The testing methodology was as follows: for each fabric sample, four round specimens with a 32 mm diameter were taken from representative areas of the fabric sample. These were kept in the testing laboratory atmosphere, weighted right before and after a 5,000 cy abrasion testing. The average weight loss was determined among the four specimens (in mg/5,000 cy) (Miguel, 2000). During testing, a 12 KPa pressure was applied to each specimen. The abrasive fabric used was the Shirley Developments Ltd 100 per cent wool standard fabric. In the two yarns breakage method, the number of abrasion cycles required to break two yarns is determined according to Woolmark Company TM 112 test standard. 3. Results and discussion The best equations found to explain abrasion strength – weight loss (RAM) and breakage of two yarns (RAR) methods, the respective determination coefficients (R 2) and significance levels (P) are shown in the Table I (Miguel, 2000). The relative weight of equation variables, which most explain fabric breaking strength – stripe and grab methods, is shown in Table II. 3.1 Discussion of the equation of worsted and woollen fabrics – weight loss method The abrasion strength (RAM) increases as thickness at 100 gf/cm2 (EF100) decreases. It has a maximum for FG values near 45 N/m and a minimum for thickness values at 2 gf/cm2 near 1.2 mm (Miguel, 2000). The thickness at 100 gf/cm2 represents the fabric that one without taking into account the FST, this is, it comes form the structural fabric parameters. Thus, what contributes to its decrease is the weight/m2 and yarn count in Tex decrease, with the consequent increase of yarn twist and number of plies. By technical construction reasons, fabrics with these characteristics have a tendency to increase its polyester composition. Thus, conditions are met to a decrease of weight loss caused by abrasion (less material, more cohesion and more polyester). The abrasion strength increases with FG up to 45 N/m and decreases as FG goes up above this value (Figure 1). In the first case and respecting to the second one,

Garment abrasion strength 197

Worsted Woollen Woollen

Weight loss FAST Breakage of two yarns FAST Breakage of two yarns Classical þ FAST Woollen

Fabric family

Table I. Equations that explain abrasion strength – weight loss (RAM) and breakage of two yarns (RAR) methods

Property and variables RAM ¼ 17.65 2 0.95FBm þ 0.01FGm2 2 8.80EF2m2 þ 45.16EF10m RAR ¼ 45552 þ 7557FHEm þ 641FGm 2 31241F5m 2 43722FSTRm2 RAR ¼ 20611 þ 871PA 2 12725TL þ 610FG

Equation

0.1

0.1

78.5 82.6

0.1

P (per cent) 83.2

R2 (per cent)

Good

Good

Good

Accuracy level

IJCST 19,3/4

198

fabrics having lower TA, higher polyester composition and higher TL variation range are found (0.5-4 versus 0.33-2). In both cases, there are fabrics having 1, 1 and 2, and 2 ply yarns. In the first group of fabrics, due to low TAs, the increase of FG depends on the increase of fabric FaC and CoT. This increase becomes easier by an increase of TL, which can have a large variability from a weft majority to warp in the fabric face. Considering fabrics having polyester, an increase of abrasion strength may be due to an increase of the number of two ply threads (with higher cohesion), caused by an increase of TL and FaC. In the second group of fabrics, the increase of FG is dependent from the increase of TA and from the increase of weight/m2 and polyamide composition (PA), with a consequent decrease of polyester. As it is known, these conditions lead to an increase of weight loss during abrasion and to a decrease of abrasion strength. The abrasion strength decreases as thickness at 2 gf/cm2 increases up to 1.2 mm and increases with it above this value (Figure 1). In the first case, there are worsted and woollen fabrics and, in the second case, woollen fabrics only, most of them having polyamide in their composition. In the first group, the increase of thickness is due to the increase of TA, yarn count in Tex and Po, since there is a decrease of spinning type, this meaning a change from worsted to woollen fabrics. These conditions, followed by a decrease of polyester, allow a decrease of abrasion strength.

Method and variables

Fabric family

Variables (per cent)

Weight loss – FAST Breakage of two yarns – FAST Breakage of two yarns Classical þ FAST

Worsted/woollen Woollen Woollen

EF10-62 FG-61 FG-63

FG-9 FHE-8

Garment abrasion strength 199

Not explained (per cent) EF2-8 TL-8

17 22 17

Table II. Relative weight of the most important equations variables

Abrasion Strength Weigth Loss Method (mg/5000 cycles)

EF2 - Thickness at 2gf/cm2(mm)

2.4

.0

4

8

12 16 20

2

24 28 32

1.6

36

1.2 0.8 40

0.4 0 0

20

40

60

80

FG-Shear Rigidity (N/m)

100

120

Figure 1. Influence of FG and thickness at 2 g/cm2 on abrasion strength – weight loss method

IJCST 19,3/4

200

In the second group of fabrics, the increase of thickness is dependent from the increase of FST. It is important to mention that this last one is included in the thickness at 2 gf/cm2. Actually, although thickness increases with FaC, CoT and TL of fabrics, it also increases as yarn count in Tex decreases. These reverse effects lead to the conclusion that the increase of thickness is not due to these structural parameters. It is believed that the increase of FST has to do with the decrease of twist factor and NC plies, as well as with the slight increase of TA. As a matter of fact, this fabric group presents high finish levels, its variation being small but important as it corresponds to reach the raised finish. In this finish, the raising machine pulls equally all fibres towards fabric surface, as opposed to felted fabrics whose fibres at surface are mostly wool. This leads to a situation in which an increase of FST and, thus, of thickness at 2 gf/cm2 is followed by an increase of the amount of different fibres in the FST and fabric composition. This means an increase of polyamide amount and, mainly, of polyester on fabric surface with the consequent effect on the decrease of weight loss during abrasion and the increase of the corresponding strength, but also in the existence of high levels of pilling formation at fabric surface. 3.2 Discussion of the equation of woollen fabrics – breakage of two yarns method The abrasion strength (RAR) increases with FG and FHE (Miguel, 2000). An increase of FG may be due to the increase of fabric CoT, TA and weight/m2. As a matter of fact, increasing these fabric characteristics difficult the relative mobility of warp and weft yarns. On the other hand, in this group of studied fabrics, the increase of TA (felting and raising intensity) is associated to an increase of PA due to the finishing characteristics of fabrics having this type of fibre. The FG has distinct influences on abrasion strength concerning worsted and woollen fabrics. In the first ones, abrasion strength increases as FG decreases, and, considering the second ones, the reverse takes place. It is believed that the explanation lies on the different technical construction between the two fabric families. Concerning worsted fabrics, CoT is higher as in woollen ones, and, thus, closer to the acceptable maximum value. This means that yarn positions in worsted fabrics are more rigid, being important a decrease of FG to allow yarns to better adapt to new positions caused by the movements of abrasion process, avoiding, in this way, a premature wearing out by abrasion. In the generality of woollen fabrics, CoT values are far from the acceptable maximum. For this reason, yarns are not subjected to a high positional rigidity. Thus, the movement flexibility given by FG to yarns in worsted fabrics takes place in a natural manner in woollen fabrics. In this way, the influence of FG on abrasion strength of these fabrics has a different explanation as compared to worsted fabrics. Considering woollen fabrics, one can say that yarns have higher positional stability for fabrics with higher opacities and lower positional and the reverse. Often, in this last case, it is the TA that complements the required stability concerning the garment making and wear behaviours. Thus, for low CoT fabrics, and, consequently with low FG, yarns are quite vulnerable to abrasion, since they have positional flexibility high enough to be moved by friction and allowing twisting themselves, suffering higher damages as compared to yarns from higher CoT fabrics. In these, the reduced yarn mobility allows lower abrasion strength, as they can only have a better adaptation to abrasion movement. On the other hand, low CoT fabrics let a distribution of abrasion efforts to a smaller NCs as compared to high CoT ones and, thus, for those fabrics, each

yarn has to suffer a higher abrasion effort. This is a possible explanation for the fact that, for woollen fabrics, abrasion strength increases with FG. Concerning the influence of FHE on abrasion strength, this has mainly to do with polyester and PAs. Thus, the FHE is strongly related to polyester composition, increasing as this goes down, because of a large difference of water absorption capacity verified between polyester and wool fibres. However, the polyester composition in woollen fabrics, by itself, has not a marked influence on abrasion strength, being superimposed by other fabric characteristics. But, in this group of fabrics, a decrease of polyester is followed by an increase of PA. As explained before, this probably is the possible cause to the abrasion strength improvement. 3.3 Discussion of the equation of woollen fabrics with classical and FAST variables – breakage of two yarns method The abrasion strength (RAR) increases with FG and PA and with the decrease of TL (Miguel, 2000). It is important to notice that an interesting model has been achieved using classical and FAST variables all together. This model has a R 2 ¼ 0.826 with three variables, the one based only on classical variables has an R 2 ¼ 0.839 including five variables and that based on FAST variables has an R 2 ¼ 0.785 and four variables. As a matter of fact, concerning the explanation of the phenomenon, the three variables that compose the first model are the most relevant ones in the other two models (classical and FAST variables). 4. Conclusions The evaluation of abrasion strength by the weight loss method is useful to determine the surface appearance degradation of fabrics during the abrasion process associated to wear. For the generality of fabrics, in the group of FAST variables which most influence abrasion strength are thickness, resulting from the structural fabric parameters and FG. In this way, these FAST variables reflect the influence of classical variables weight/m2, yarn count in Tex, yarn twist, NC plies and polyester composition, to which Po and TA are associated. In general, no better models to explain abrasion strength by the breakage of two yarns method are achieved when FAST variables are considered respecting to those based on classical ones. However, for woollen fabrics, there is an interesting model, which gathers classical and FAST variables with higher explanation: FG and PA. This allows concluding that this FAST parameter performs an important role in the abrasion behaviour of fabrics. Thus, the yarn adaptation to cyclic abrasion movements may influence breakage. Considering woollen fabrics, their surface characteristics (high amount of surface pile, often resulting from raising) work as a protection layer for yarns, retarding wearing out and breakage. On the other hand, this effect is improved on fabrics having polyamide in their composition, since this fibre has good abrasive characteristics. 4.1 Terms and definitions . Coefficient of determination, R 2. A statistic that is widely used to determine how well a regression fits. It represents the fraction of variability in y that can be explained by the variability in x.

Garment abrasion strength 201

IJCST 19,3/4

.

.

202

.

. .

.

.

.

. .

.

. .

. .

.

. .

.

.

.

Significance level, P. In hypothesis testing, it is the criterion used for rejecting the null hypothesis. The lower the significance level, the more the data must diverge from the null hypothesis to be significant. Accuracy level. A quality scale where: low ¼ R 2 , 40 per cent; acceptable ¼ R 2 $ 40 per cent and , 60 per cent; good ¼ R 2 $ 60 per cent and , 90 per cent; high ¼ R 2 $ 90 per cent. Yarn count, Tex. Indicates fineness of yarns. It is the number of grams of a 1,000 m yarn length. pﬃﬃﬃﬃﬃﬃﬃﬃ Twist coefficient, CTt. Yarn twist (turns/meter)/ Tex £ 100 Weave coefficient, CL. Number of yarn inflections between fabric face and reverse in a weave repeat. Weave type, TL. Relationship between the amount of warp and weft in fabric face. Finish type, TA. Pile intensity at fabric surface expressed in a 1-5 qualitative scale. One means a clear finish fabric and five a fabric completely covered with pile. Average float, Amed. Average length of yarn portions between two consecutive inflection points in a weave repeat. Yarn density, D. Number of warp (or weft) yarns per 1 cm fabric length. Cover factor, FaC. Relationship between yarn and linear densities. pﬃﬃﬃﬃﬃﬃﬃﬃ FaC ¼ D £ Tex/10. Opacity, CoT. Ratio between cover factor of the fabric and that of a fabric with maximum yarn densities. Average fibre fineness, Fi. Average fibre diameter expressed in microns. Average number of ply yarns, NC. Number of individual plies twisted to form a yarn. Porosity, Po. Volume of air entrapped in a unit volume of fabric. Relaxation shrinkage, FRS. Relationship between sizes of a dry fabric sample and a dried one, after being wet. Hygroscopic expansion, FHE. Relationship between sizes of a wet fabric sample and a dried one, after being wet. Formability, FF. A FAST parameter given by: FF ¼ (F20 2 F5) £ FB/14.7. Extensibility, F100, F20, F5. Extensibility of a 5 cm wide fabric sample under a load of 100, 20 or 5 g/cm, respectively. Bending rigidity, FB. Bending rigidity of a 5 cm wide fabric sample caused by its own weight. It is given by FB ¼ fabric weight/m2 £ c 3 £ 9.81 £ 102 6, where c is the fabric bending length. Shear rigidity, FG. Strength of a 5 cm wide fabric sample under a load in bias direction. FG ¼ 123 £ EB5, where EB5 is the fabric extensibility at 5 g/cm, also in bias direction. Surface thickness, FST. Difference between thicknesses measured at 2 and 100 g/cm2.

.

Released surface thickness, FSTR. Difference between thicknesses measured at 2 and 100 g/cm2, after fabric being steamed.

References Allen, A.G. et al. (1980), CSIRO Division of Textile Technology, Geelong, pp. 185-97. Manich, A. and Miguel, R.A.L. et al. (2001), “Abrasion kinetics of wool and blended fabrics”, Textile Research Journal, Vol. 71 No. 6, pp. 469-74. Miguel, R.A.L. (2000), “Modelling the influence of structural characteristics on wear properties of wool and blended fabrics”, PhD thesis, University of Beira Interior, Covilha˜, pp. 51, 141, 148, 149, 152 and 159. Corresponding author Rui Miguel can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Garment abrasion strength 203

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,3/4

Investigation of the strength of ultrasonically welded sails Edita Vujasinovic´

204

Faculty of Textile Technology, University of Zagreb, Zagreb, Croatia

Zeljka Jankovic´ Amadeus M.A.J., Zagreb, Croatia, and

Zvonko Dragcˇevic´, Igor Petrunic´ and Dubravko Rogale Faculty of Textile Technology, University of Zagreb, Zagreb, Croatia Abstract Purpose – Today when the newest high-tech fibers and sophisticated material constructions are used for the production of sails, forming 3D sail shape from 2D sailcloth has still remained very primitive because classic sewing techniques are mostly used. Since, the clothing and technical textile industry has been recently using some of contemporary joining techniques (ultrasonic, thermal, high-frequency) replacing classic sewing, this paper seeks to investigate the possibility of ultrasonic welding in the production of sails, and the strength of obtained bonds. Design/methodology/approach – Concerning the aim, sails were made employing the classic and modern (ultrasonic) joining technique whereby bonding parameters such as amplitude and welding speed, geometry of anvil wheels were varied. Objective quality evaluation of the bond made in such a way, was performed in order to be more exact about its strength. Findings – Based on the obtained results it has been concluded that ultrasonic welding may successfully replace the classic sewing of sails, selecting an anvil wheel with suitable engraving and optimal parameters of welding (speed and amplitude). Practical implications – Selection of optimal welding parameters not only increases the sail’s bond strength in comparison with classic seam, but also provides sail air impermeability, being one of the basic aerodynamic requirements for sail making. Originality/value – This paper has presented the novel and successful approach in 3D sail shape forming from 2D sailcloth. Keywords Technology led strategy, Textiles, Naval vessels, Ultrasonic welding, Seam welding Paper type Research paper

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 204-214 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741678

Introduction Since, the early beginnings of sailing or nearly for 6,000 years past when animal leathers and papyrus were used for sail making and when sailing ships discovered new paths, worlds and civilizations (Doyle, 1997; Veljic and Govedic, 1989; Whidden and Levit, 1990) until today, when the most modern materials are used for sail making, and sails are mostly used for sports and fun, sail-makers have tried to make the best possible sail, i.e. the sail which is elastic when there is no wind, but strong enough as not to be weak when the wind is strong (Bond et al., 2000). Based on experience and knowledge, sail makers can design an optimal aerodynamic shape of the desirable sail, Presented work is a part of investigations that are carried out within joint Slovenian-Croatian scientific project entitled “Implementation of new materials objective measurement and evaluation system in the processes of technical and intelligent textiles design”.

but how it will retain its shape under different winds and different weather conditions depends on the applied material (its construction, porosity, UV stability, resistance to elongation, strength (Doyle, 1997; Whidden and Levit, 1990; Wallace, 1987) strength and compactness of the finished sail. Hence, sail making represents a process of constant optimization which does not only refer to the design of the appropriate material for sail making, but also to the engineering and final design of the sail itself. Although today the newest high-tech fibers (Vujasinovic and Cunko, 2002; Vujasinovic, 2003) and sophisticated material constructions are used for the production of sails (Figure 1) (Dimension Polyant, 2005), making a 3D sail shape from 2D sailcloth has still remained very primitive. Namely, joining individual sail parts is mostly carried out on sewing machines by drawing through the thread between material layers, making various sewing stitches with necessary characteristics, whereby seam elasticity and strength are to be taken into account. To join sails, lockstitch and chain stitch sewing machines are mostly used that are in contrast to clothing machines considerably sturdier, larger and stronger, operating with smaller nominal speeds (Figure 2) (Halsey Ligard Sailmakers, 2005). Besides universal sewing machines, special sewing machines (first of all zigzag stitch machines), automatic sewing machines (primarily for bar tacking) and sewing robots as well as CNC sewing machines play an increasing role in sail sealing. Sewing threads also play a decisive role because they need to meet high quality requirements, increasingly becoming threads for targeted purposes and narrow application ranges, i.e. for specific materials, machines or production method. To make up sails, extremely strong, UV and weather resistant sewing threads are required. They should be resistant to increased temperatures in sewing and use. Besides the sewing thread, sewing machine needles play a decisive role in making up sails so that increasing demands are made on needles in regard to the hardness of the needle surface

Strength of ultrasonically welded sails 205

Film

Tafeta

Skrim

Tafeta

Figure 1. New sailcloth design

IJCST 19,3/4

206

Figure 2. Sail stitching

as well as concerning needle point and sewing needle geometry, depending primarily on the material type to be sewn (Rogale and Dragcevic, 2002). Taking account of the above-said and the fact that the technical textile industry, primarily mobiletex, has been recently successfully replacing the classic sewing process with some of contemporary joining techniques such as thermal, high-frequency or ultrasonic sealing (Ujevic and Rogale, 2004), the goal of this work has been to investigate whether it is possible to apply ultrasonic welding in sail making. Methodology Concerning the aim and in line with the foreseen investigation plan (Figure 3), sails were made employing the classic and modern (ultrasonic) joining technique whereby bonding parameters such as: amplitude and welding speed, geometry of anvil wheels were varied. Objective quality evaluation of the bond made in such a way, to be more exact its strength was done. US welding Anvil wheel Amplitude A1

Can ultrasonic bond on sails replace classic seam?

A2

B2

A3

B3 Seam strength

IDEA

Figure 3. Investigation plan

Speed B1

BOND

OBJECTIVE EVALUATION

Sail demands (air impermeability)

CLASSIC SEAM

Quality evaluation of ultrasonic bond GOAL Quality evaluation of seam

Is the US better?

Methods and devices for joining sailcloth into the sail For the purposes of this work sailcloth samples were joined on: . The Pfaff 8310 Seamsonic ultrasonic welding machine with a 400 W ultrasonic generator and a frequency of 30 kHz (Pfaff, 2006). Sonotrodes are made of titanium with a 104 mm diameter and a maximum weld width 10 mm. Welding speed ranges from 6 to 136 dm min2 1, and welding pressure ranges from 0 to 800 N (5 bar). Amplitude and gap between the sonotrode and anvil wheel may be regulated. The appearance of the ultrasonically welded seam, i.e. the impression of an engraved pattern depends on the shape of anvil wheels. Four anvil wheels were used in order to select the best joint in sails (Figure 4). Welding parameters were varied for each anvil wheel so that samples were joined at different welding speeds as follows: 6, 18 and 30 dm min2 1, and amplitudes 50, 76 and 100 percent. The gap between the sonotrode and anvil wheel and pressure are constant, amounting to 0.18 mm and 384 N (2.4 bar). . A special sewing machine Minerva-Boskovice class 525 with designation ID ZZ TD especially designed for sail making (lengthened long-arm). The seam was made with a zigzag stitch, and a NM 100 needle and a polyester thread DABOND 2000 VVR (Heminway Bartlett; V 92 Shade White) were used for sewing. Test methods The usual characterization of the seam joining two textile parts includes the description of aesthetic appearance and the determination of seam strength with special emphasis on thread shear. But the characterization of bond or seam requires a different approach to the evaluation of the joint itself on account of specific requirements that are to be met by the sail in order to be effective. Since, sails are subjected to considerable tensile stresses during use, bond strength represents the first and most important quality parameter, whereas bond impermeability is a requirement providing an efficient aerodynamic utilization of the sail. Hence, two standardized or especially adjusted methods and testers were used for the purposes of this work (Table I) in order to make an objective evaluation of sailcloth bond (seam). Test specimen For the purposes of the investigations within the scope of this work a polyester sample of the material for sailmaking designated with Dacron 3.7 us oz (, 42.83 gm2 2) by Polyant USA was used. The following designations of samples were used in this paper: (1) The designations relating to the method of joining: . (A) – anvil wheel with point engraving for ultrasonic welding (Figure 4(a)). . (B) – anvil wheel with zigzag engraving for ultrasonic welding (Figure 4(b)). . (C) – anvil wheel with one-line engraving for ultrasonic welding (Figure 4(c)). . (D) – anvil wheel with three-line engraving for ultrasonic welding (Figure 4(d)) and K-sample joined with the threefold zigzag stitch. (2) These designations refer to the speeds of ultrasonic sealing: 1 – designates velocity v ¼ 6 dm min2 1, 2 – designates velocity v ¼ 18 dm min2 1 and 3 – designates velocity v ¼ 30 dm min2 1.

Strength of ultrasonically welded sails 207

IJCST 19,3/4

208

(a)

(b)

(c)

Figure 4. Geometry of the anvil wheels used: (a) anvil wheel with point engraving (A); (b) anvil wheel with zigzag engraving (B); (c) anvil wheel with one-line engraving (C) and (d) anvil wheel with three-line engraving (D)

(d)

Determination of seam strength

Determination of bond/seam permeability

According to the standard: HRN F.S2.017 & HRN F.S2.024 Instrument: TensoLab STRENGTH TESTER

Instrument: Universal research microscope with transmitted light Olympus BX51; equipped with analySIS software package for digital analysis of micrographs and microphotometry Microscoping the seam in transmitted light makes possible to determine material damage caused by bonding that can be seen as holes through which light passes

Three parallel measurements of air-conditioned samples with dimensions of 50 £ 400 mm with a gap between the clamps of 200 mm, clamping speed of 500 mm min2 1 and preloading of 5 N

Strength of ultrasonically welded sails 209 Table I. Applied methods and procedures

(3) The designations of the amplitudes used for ultrasonic welding: a – designates amplitude A ¼ 50 percent, b – designates amplitude A ¼ 76 percent and c – designates amplitude A ¼ 100 percent. For example, A1a – represents the sample (PES Dacronw) ultrasonically welded by the anvil wheel with point engraving at a speed of 6 dm min2 1 and 50 percent amplitude, and K-represents the sample (PES Dacronw) made by sewing with the classic threefold zigzag stitch. Results and discussion Table II and Figures 5-8 show the results of determining the resistance of the sail seams to tensile force. Figure 5 shows that the tensile strength of the ultrasonic bonds obtained by use of the anvil wheel with point engraving (A) ranges from 0.44 to 31.77 N mm2 2, whereby the tensile strength of the bond obtained at the medium amplitude (76 percent) and welding speed (18 dm min2 1) is the highest, and the lowest one of the bond obtained at the lowest amplitude (50 percent) and the highest speed (30 dm min2 1). In the sample bonded by use of the lowest welding speed (6 dm min2 1) the tensile strength decreases with the rising amplitude, whereas in the samples bonded by use of other speeds (18 and 30 dm min2 1) the tensile strength of the bonds made by use of the medium amplitude (76 percent) is the highest, and with the 50 percent amplitude is the lowest. The speed affects the bond strength mostly when bonding at the lowest amplitude where the difference between the highest and the lowest value of the tensile strength obtained by use of the highest and lowest welding speed amounts to approximately 98.5 percent. With other applied amplitudes the impact of the speed on bond strength is not so high. Figure 6 shows that the tensile strength of the ultrasonic bonds obtained by use of the anvil wheel with zigzag engraving (B) ranges from 2.31 to 13.48 N mm2 2, whereby Sample K Breaking force, FB (N) Elongation, 1B (percent) Strength, sB (N mm2 2) Elasticity module, 10 (N mm2 2)

xa

xmin

xmax

CV (percent)

PE

629.59 21.90 13.66 1.15

625.66 21.30 13.58 0.75

636.45 22.80 13.80 1.49

0.95 3.62 0.87 32.79

^ 0.20 ^ 0.03 ^ 0.00 ^ 0.01

Table II. Resistance of the classic sail seam to tensile force

IJCST 19,3/4

31.77 35 30 Strength [Nmm-2]

210

25 20 15 0.44

10

6 5

Figure 5. Tensile strength of the ultrasonic bonds obtained by use of the anvil wheel with point engraving (A)

0 30

50

18 Welding speed [dmmin-1]

76 100 Amplitude [%]

13.84 15

Strength [Nmm-2]

12

2.21

9 6 6

3

Figure 6. Tensile strength of the ultrasonic bonds obtained by use of the anvil wheel with zigzag engraving (B)

18

0 50

30 76 100 Amplitude [%]

Welding speed [dmmin-1]

Strength of ultrasonically welded sails

6.74

8

211

0.71 Strength [Nmm-2]

6

4

6

2

Welding speed [dmmin-1]

18 0 30

50 76 100 Amplitude [%]

Figure 7. Tensile strength of the ultrasonic bonds obtained by use of the anvil wheel with one-line engraving (C)

17.34 18

Strength [Nmm–2]

15 12 9 6

0.52

3 6 0 18 50 30

76 Amplitude [%]

100

Welding speed [dmmin–1]

Figure 8. Tensile strength of the ultrasonic bonds obtained by use of the anvil wheel with three-line engraving (D)

IJCST 19,3/4

212

the tensile strength of the bond obtained at the medium amplitude (76 percent) and the highest welding speed (30 dm min2 1) is the highest, and the lowest of the bond obtained at the highest amplitude (100 percent) and the lowest speed (6 dm min2 1). At the speeds of 6 and 18 dm min2 1 the inversely proportional dependence of the tensile strength of the amplitude is recognizable, whereas this is not the case at a speed of 30 dm min2 1. In the samples bonded at the 76 percent amplitude a proportional dependence of the tensile strength on the welding speed is recognizable, which is not the case for the lowest and highest amplitude applied (50 and 100 percent). According to the above-said, it may be concluded that there is no general regularity, i.e. dependence of tensile strength on amplitude and speed. Figure 7 shows a parallel representation of the strengths of all samples obtained by ultrasonic welding using the anvil wheel with one-line engraving (C). The figure displays that the tensile strength of the ultrasonic bonds obtained by use of the anvil wheel with one-line engraving ranges from 0.71 to 6.47 N mm2 2, whereby the tensile strength of the bond obtained at the lowest welding speed and amplitude (6 dm min2 1 and 50 percent) is the highest, and the lowest of the bond obtained at the medium speed and amplitude (18 dm min2 1 and 76 percent). In the samples bonded at the lowest welding speed (6 dm min2 1) the tensile strength decreases with the rising amplitude. If the values of the tensile strength of the bond at the lowest amplitude of welding are analyzed, it becomes evident that at rising speed the bond strength drops to a large extent. These regularities are not valid for the other samples, i.e. welding speeds and amplitudes. Figure 8 shows that the tensile strength of the ultrasonic bonds obtained by use of the anvil wheel with three-line engraving (D) ranges from 0.52 to 17.34 N mm2 2 which corresponds to the range of values found out at different amplitudes at a speed of 30 dm min2 1. For other speeds (6 and 18 dm min2 1) the regularity holds true that bond strength drops with rising amplitude. If the values of bond strength are observed at constant amplitude, only at 76 percent amplitude it holds true that bond strength increases with rising speed of welding. According to the results shown in Figure 9 it may be concluded that ultrasonic welding together with the selection of optimal parameters (speed and amplitude) can successfully replace classic sail sewing. Namely, the strength of ultrasonic bond is as a rule higher than the strength of classic sail seam (for anvil wheel A up to 130 percent), which makes a contribution to safe sailing. Initial modulus is also higher as well as the limit of aerodynamic efficiency of the designed sail shape to which a lower elongation at break makes its considerable contribution. An ultrasonic sail bond is by approximately 2/3 thinner than a classic seam, it is less rough, and in the case of optimally selected welding conditions it is completely air impermeable. Namely, although sail strength represents a very important element for the objective evaluation of sails from the point of view of safety and application (wind strength), the sail should also have certain aerodynamic characteristics to fulfill its basic function. The sail as a whole as well as all its parts, and also sailcloth bond (seam), should have as small a thickness and roughness as possible. It should be air impermeable, representing the second element for an objective evaluation of the sail as a whole (Figure 10).

35

Strength of ultrasonically welded sails

σB[Nmm–2] 31,77

30 25

213

20

17,34 13,48

13,66

15 10

6,47

5 0 K

(a)

US A

(b)

US B

(c)

US C

(d)

US D

(e)

Notes: (a) Ultrasonic bond with an anvil wheel with point engraving (A); (b) ultrasonic bond with an anvil wheel with zigzag engraving (B); (c) ultrasonic bond with an anvil wheel with one-line engraving (C); (d) ultrasonic bond with an anvil wheel with three-line engraving (D); and (e) classic seam

Conclusion Based on the results of investigation conducted within the scope of this work, it has been concluded that ultrasonic sailcloth bonding can successfully replace classic sewing if an anvil wheel with appropriate engraving and optimal welding parameters (speed and amplitude) is selected. It does not only increase bond strength in comparison to classic seam, but it also provides sail air impermeability, being one of the basic aerodynamic requirements put for sail making. References Bond, B. et al. (2000), The Complete Book of Sailing, Chancellor Press, London. Dimension Polyant (2005), The Sail, The Cloth, The Company, Catalog of New Product TT, Dimension Polyant, Putnam, CT.

Figure 9. Parallel representation of the strengths of all sail samples

Figure 10. Micrograph of a sailcloth bond ( £ 40)

IJCST 19,3/4

214

Doyle, B.E. (1997), “Strong fabrics for fast sails”, Scientific American, Vol. 277 No. 7, pp. 46-53. Halsey Ligard Sailmakers (2005), “Loft tour – from design to delivery of finished sails”, available at: www.halseyligard.com/LoftTour/tour.html (accessed May 2005). Pfaff (2006), “Welding machines”, available at: www.sewingperfection.com.an (accessed January 2006). Rogale, D. and Dragcevic, Z. (2002), “Tehnike konfekcioniranja tehnickog tekstila”, Tekstil, Vol. 51 No. 2, pp. 64-77. Ujevic, D. and Rogale, D. (2004), “Pregled razvojnih dostignuca na podrucju odjevnih tehnologija prikazanih na IBM 2003”, Tekstil, Vol. 53 No. 1, pp. 18-24. Veljic, A. and Govedic, S. (1989), Brodovi na jadranu, RO Tiskara Varteks, Varazdin. Vujasinovic, E. (2003), “Investigations on the possibility of objective characterization of sailcloth”, PhD thesis, Faculty of Textile Technology University of Zagreb, Zagreb. Vujasinovic, E. and Cunko, R. (2002), “Sailcloth – past, present and future”, in Dragcevic, Z. (Ed.), Proceedings of the 1st International Textile, Clothing & Design Conferences ‘Magic World of Textiles’, University of Zagreb Faculty of Textile Technology, Zagreb, pp. 130-5. Wallace, R. (1987), Sail Power – The Complete Guide to Sail and Sail Handing, Alfreda Knopf Pub. Inc., New York, NY. Whidden, T. and Levit, M. (1990), The Art and Science of Sails, St Martin’s Press, New York, NY. Corresponding author Edita Vujasinovic´ can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

New measuring device for estimating the pressure under compression garments Elzbieta Maklewska and Andrzej Nawrocki

Device for estimating the pressure 215

Clothing Research Department, Institute of Knitting Techniques and Technologies TRICOTEXTIL, Ło´dz´, Poland

Krzysztof Kowalski Faculty of Textile Engineering and Marketing, Technical University of Łodz´, Ło´dz´, Poland, and

Ewa Andrzejewska and Wojciech Kuzan´ski Department of Paediatrics’ Surgery and Oncology, Medical University of Łodz´, Ło´dz´, Poland Abstract Purpose – This paper aims to describe new measuring device designed for measuring the pressure exerted by textile products used in healing therapy of hypertrophic scars. The testing device called “Textilpress” has been used for verification of the usually used method of designing and manufacturing ready-made compression garment products. Design/methodology/approach – The pressure measurement, realized by use of the “Textilpress” device, is an indirect measuring method, which is based on the Laplace Law. The investigations described in this paper concern the pressure measuring under textile bands placed on the models representing selected parts of the human body with pre-set circumferences. For this purpose, rigid cylinders were prepared, covered by a layer of neoprene which simulated the susceptibility of human skin. Findings – The investigations described in this paper indicate that the “Textilpress” test-device may be used for pressure measuring exerted by compression bands on the cylinder surface. In order to estimate the pressure exerted on a particular body part with the shape close to a cylinder, a measurement should be carried out on a cylinder with a circumference similar to that of the selected part of the human body. Research limitations/implications – The “Textilpress” test-device may be used by the manufacturers for measuring pressures exerted by compression bands (manufactured from knitted fabrics) on the cylinder surface of a pre-set diameter. In order to estimate the pressure exerted on a particular body part with the shape close to a cylinder, a measurement should be carried out on a cylinder with a circumference similar to that of the selected part of the human body. Originality/value – As hospitals do not possess appropriate measuring devices, which would enable one to measure the pressure on the scar exerted by the textile garment, the pressure efficiency and the appropriate fittings are estimated subjectively for the particular case. The “Textilpress” device enables monitoring such pressure. Keywords Compressive strength, Pressure, Garment industry Paper type Research paper

The investigation presented in this paper was carried out as part of the research project No. 4T08E05425 financially supported by the Polish Ministry of Scientific Research and Information Technology.

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 215-221 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741687

IJCST 19,3/4

216

1. Introduction Pressure therapy, also known as compression therapy, for scar management is an established component of a recovering burn patient’s rehabilitation program. Pressure can be achieved through special tight-fitting compression garments (CG) or dressings made of elasticized fabrics. Elastic bandages or garments used to provide constant and equal pressure provide compression over the healed burn. This compression minimizes development of scarring and deformity caused by serious burn injury (Figure 1). These scars are warm, red in colour due to increased vascularity, raised due to increased collagen disposition. These type of hypertrophic scars (HS) are shown on Figure 1. CG help to reduce bulky, thick, hard scars by pressing and flattening the scars. The tightness of a compression garment also helps to stop the itching associated with a healing burn. feeling of skin tension and prevent of limitations of limb movements, improve their physical fitness (Lamberty and Whitaker, 1981; Linares et al., 1993; Reno et al., 2001; Williams et al., 1998; Giele et al., 1997; Macintyre et al., 2003; Finnie, 2000). The photos in Figure 2 show examples of using CG. All garment-products of this type should be worn 23 h per day, from 6 to 2 months, until flattening, color fading and tissue softening would be observed. CG, should be designed so that the pressure on the protected body area with the scar would be within the range of 20-25 mm Hg (26.7-33.3 hPa). Too high, and not checked pressure during the healing period, may cause disadvantageous changes especially at small children. As hospitals do not possess appropriate measuring devices, which would enable to measure the pressure on the scar exerted by the textile garment, the pressure efficiency and the appropriate fittings are estimated subjectively for the particular case. The pressure measuring methods used hitherto are evaluated critically, and a broad literature review devoted to the methods, which allow measuring pressure

Figure 1. Samples of hypertrophic scars (HS) caused by serious burn injury

Figure 2. Examples of using “burn-garments”

Vest with sleeves

Glove with open finger tips

Face mask

exerted by garments, is presented in a paper of Finnie (2000). Most often, the measurements are performed by means of the electro-pneumatic method. The research work presented herein, presents selected results of tests carried out with the use of a new measuring stand designed for measuring the pressure exerted by textile products used in healing therapy of hypertropic scars. The testing device was designed and built at the “Tricotextil” Institute of Knitting Techniques and Technologies’ (Poland) and was named “Textilpress”. To underline the difficulties of the measurements realized, should be emphasized that under real conditions the pressure gauge must measure pressure values within the range of atmospheric pressure changes.

Device for estimating the pressure 217

2. Aim of investigation The aim of the paper, presented herein, is a description of the new test device named “Textilpress” designed for measuring the pressure exerted by textile products used in the healing therapy of hypertropic scars. The “Textilpress” device has been used for verification of the often used method of designing and manufacturing ready-made compression garment products. This method is based on a constant reduction of the circumferences of knitting elements of garment by 10-20 per cent in relation to the body circumferences (Macintyre et al., 2003). 3. Investigation methods 3.1 Test stand description The pressure measurement, realized by use of the “Textilpress” device, is an indirect measuring method, which is based on the Laplace Law. Te Laplace Law has been widely used to calculate the pressure delivered to a cylinder of known radius by a fabric under known Tension. However, this Law was originally developed by Laplace in 1806 to explain the surface tension phenomenon in liquids and their ability to form droplets or soap bubbles (Macintyre et al., 2003) The Laplace Law is shown on the Figure 3 and by Equation (1). P¼

T R

ð1Þ

where: P – pressure, the normal force per area unit; T – tension expressed as force per width unit of the pressure band (T ¼ F/w); and R – radius of cylinder curvature. The basic element of the test stand Textilpress, where are placed sensors, is a measuring matrix (Figure 4). Textile band Wall tension T

External pressure P.

R

P

w

Cylinder

Figure 3. Scheme illustrating the Laplace Law

IJCST 19,3/4

Figure 4. Measuring matrix; (a) scheme; (b) general view

(a)

Additional gauges

Y Longitudinal direction (main measuring direction)

218

Measuring field dimensions: XxY=70 50 mm

1

Open gauge

Closed gauge

Open gauge

Open gauge

X Transverse direction (auxiliary)

2

2 2

1

(b)

1

3

Notes: 1 - "open" gauges for circumference force measuring; 2 - "closed' gauges for curvature measuring; 3 - additional curvature gauges

The measuring matrix is equipped with two kinds of measuring tensometric sensors called “closed” and “open”. The “closed” gauges are for determining the radius curvature R of the circumference tested. The “open” gauges measure the tension T. The measuring matrix includes totally 3 sets £ 6 (4 “closed” þ 2 “open”) ¼ 18 gauges which enable determining the tension and the curvature in three independent circuits of the measuring field. Additional curvature gauges are placed parallel to the axis of the compression band and allow us correcting the pressure measurement if the tested body is not an ideal cylinder. The pressure distribution at the basic points, and the approximated pressure distribution between these points are accepted as the measuring result, together at 25 points. An example of recorded test results for testing with use of Textilpress is shown in Table I. The dimensions of the measuring field are 70 £ 50 mm. Technical parameters of the “Textilpress” device are listed below: . pressure measuring range: 0.70 mm Hg ^ 10 per cent; . measuring resolution: ^ 1 mm Hg;

Pressure (hPa)

Table I. An example of recorded test results report for testing with use of Textilpress

Information: test III DT 15:26:29 Curvature diameter (mm)

16 24 32 27 22 16 24 32 27 23 16 24 32 27 23 16 24 32 27 23 15 24 32 27 23 Average pressure value ¼ 24.3 S ¼ 5.4 Max ¼ 7.9 Min ¼ 2 9.0 V ¼ 22.2 per cent

114 114 114 114 114 114 115 114 113 114 115 114 115 115 115 114 118 117 116 115 Average diameter value ¼ 114.4 S ¼ 1.0 Max ¼ 3.1 Min ¼ 21.4 V ¼ 0.9 per cent

114 114 113 113 114

. . .

range of the tension T: 0 4 200 N/m; range of the curvature radius 25 4 200 mm; and minimum radius of bending the matrix: 20 mm (exceeding this value may cause a damage of the matrix.

A view of the “Textilpress” device is shown on the photos in Figure 5.

Device for estimating the pressure 219

3.2 Investigation procedure The investigations described in this paper concern the pressure measuring under textile bands placed on the models representing selected parts of the human body with pre-set circumferences. For this purpose, rigid cylinders were prepared, covered by a layer of neoprene which simulated the susceptibility of human skin. Before performing the tests should be done a calibration of the measuring stand. The calibration of the curvature radius and the pressure is carried out with the use of cylinders with determined dimensions, for tensions of 28 N/m and 170 N/m. In order to perform a measurement, the tested band should be placed on the cylinder with the matrix fastened to it, and then the test could be carried out (Figure 5). A protocol example of recorded test results for a compression band placed on a 0.1 m diameter cylinder is presented in Table I. There are shown the values of pressure and curvature diameter recorded in 25 places-points. There are also given average pressure and diameter values for tested band.

1

2

3

4

5

2

Note: 1 – textile compression band; 2 – measuring matrix placed under the compression band; 3 – cylinder; 4 – cylinder holder; 5 – strain gauges amplifier; 6 – computer

6

Figure 5. View of the “Textilpress” measuring stand during measuring

IJCST 19,3/4

4. Test material As a testing material, it was prepared the series compression bands, from a warp-knitted fabric, manufactured by a professional producer of garments used in hypertropic scars therapy. The dimension of circumference G0 of the bands were calculated according to equation:

220 G0 ¼

G1 1þ1

ð2Þ

G0 – circumference of the band, in free state (cm); G1 – circumference of the cylinder (cm); and 1 – relative elongation of the bands and is equal 0.1. The dimensions of the bands are listed in Table II. 5. Test results and discussion The results of pressure measurement under the tested compression bands, carried out with the use of the Textilpress’ test-device are presented in Table II and on the Figure 6. The maximum and minimum pressure recommended in compressive therapy was marked, on the Figure 6, as Pmax ¼ 33.3 hPa and Pmin ¼ 26.7 hPa, respectively.

Table II. Dimensions of the bands and test results of the pressure exerted by compression bands

LP

Cylinder denotationa

1 2 3 4 5

Cylinder circumference G1 (m) Band denotation Band circumference G0 (m) Relative elongation Pressure (hPa)

B

C

D

E

F

G

H

I

0.19 BT 0.17 0.1 27.5

0.28 CT 0.25 0.1 27.1

0.35 DT 0.32 0.1 24.4

0.41 ET 0.35 0.1 22.1

0.50 FT 0.45 0.1 19.9

0.66 GT 0.60 0.1 18.4

0.82 HT 0.73 0.1 16.2

0.85 IT 0.75 0.1 16.2

Note: aThe cylinders C, E, and I are pattern cylinders, for calibrations

35 Pmax

Figure 6. Graphical illustration of the test results of pressure measuring under the bands of both series I and I; the maximum and minimum values of the pressure recommended in pressing therapy have been marked in the figure

Pressure [hPa]

30 Pmin 25 20 15 10 0.19

0.28

0. 35

0.41

0.5

0.66

0.82

0.85

Cylinder circumference G1 [m] Pressure under bands

Pmax

Pmin

An analysis of the results of pressure measuring under compression bands manufactured by the professional producer of textile compression products and placed on appropriate cylinders, indicates that the pressure value exerted by those bands, decreases with the increasing the cylinder diameter. Almost all of pressure values are below of recommended pressure minimum Pmin ¼ 26.7 hPa. It should be underlined, that the relative difference between band circumferences of this series and the corresponding cylinder circumferences has a constant character, and is equal 0.1. 6. Conclusion The investigations described in this paper, indicate that the “Textilpress” test-device may be used for pressure measuring exerted by compression bands on the cylinder surface. In order to estimate the pressure exerted on a particular body part with the shape close to a cylinder, a measurement should be carried out on a cylinder with a circumference similar to that of the selected part of the human body. Theoretically, the matrix may be placed also directly on the body, but in reality the realization of such a measuring test would be objectively difficult. The object with the matrix placed on it, should be motionless, that for example by testing children would be rather impossible. The investigations carried out, demonstrate that designing method based on using a constant difference value between the compression band circumference in free state and the circumference of the model part of the body is erroneous and do not provide the expected level of pressure value. It seems, that this difference should be depended on the circumference of the model part of the body. Additionally, the pressure under garment should be checked by testing with measurement device. References Finnie, A. (2000), “Interface pressure measurements in leg ulcer management”, British Journal of Nursing, Vol. 9 No. 6, pp. 8-10. Giele, H.P., Liddiard, K. and Wood, F.M. (1997), “Direct measurement of cutaneous pressures generated by pressure garments”, Burns, Vol. 23 No. 2, pp. 137-41. Lamberty, B. and Whitaker, J. (1981), “Prevention and correction of hypertrophic scarring in posburns deformity”, Physiotherapy, Vol. 67 No. 1, pp. 2-4. Linares, H. et al. (1993), “Historical notes on the use of pressure in the treatment of hypertrophic scars or keloids”, Burns, Vol. 19 No. 1, pp. 17-21. Macintyre, L. et al. (2003), “The study of pressure delivery for hypertrophic scar treatment”, paper presented at International Conference on Healthcare Medical Textiles, Bolton, July. Reno, F., Grazianetti, P. and Cannas, M. (2001), “Effects of mechanical compression on hypertrophic scars: prostaglandin E2 release”, Burns, Vol. 27 No. 3, pp. 215-17. Williams, W., Knapp, D. and Wallen, M. (1998), “Comparison of the characteristics and features of pressure garments used in the management of burn scars”, Burns, Vol. 24, pp. 329-35. Corresponding author Elzbieta Maklewska can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Device for estimating the pressure 221

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,3/4

Technical systems in intelligent clothing with active thermal protection

222

Snjezˇana Firsˇt Rogale, Dubravko Rogale, Zvonko Dragcˇevic´ and Gojko Nikolic´ Department of Clothing Technology, Faculty of Textile Technology, University of Zagreb, Zagreb, Croatia, and

Milivoj Bartosˇ Pelcom d.o.o., Zagreb, Croatia Abstract Purpose – This paper seeks to examine the practical construction of a patented intelligent article of clothing with active thermal protection, outer shell with a variable thickness, system of thermoinsulating chambers, sensors and measuring systems, microcontroller and control system, actuator system and power supply system based on hardware aspect. Design/methodology/approach – Based on practical construction and software aspect, a measuring and control program with algorithm of intelligent behavior is presented. Findings – An intelligent article of clothing with thermal protection like a complex technical system consisting of a very complex architecture connected by different bus types to constitute a complex system acting synchronously and effectively controlled by the microcontroller containing the program which is based on the algorithm of the intelligent behavior. Research limitations/implications – The technical systems described represent a suitable basis for experiments and scientific research during the introduction of intelligent clothing with active thermal protection into human life. Practical implications – Introduction of intelligent clothing into human life. Originality/value – The prototype of an intelligent article of clothing with active thermal protection was designed and constructed. At the moment it is being subjected to the first experiments of functional investigations of intelligent clothing with active thermal protection. Keywords Clothing, Intelligent manufacturing systems, Thermal protection Paper type Research paper

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 222-233 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741696

1. Introduction Clothing belongs to one of the oldest human articles of daily use which have always provided a protective function against climatic influences (primarily protection against cold, moisture, wind, heat, rain, sun’s UV radiation) and then against environmental influences (dust, mud). Later, the clothing assumed a mechanical protective function to protect craftsman, farmers and warriors against mechanical injuries and blows. Afterwards, it assumed additional functions to mark the social class of its wearer or his rank in hierarchy within organization groups (army, clergy, and sovereigns). Only later, it assumed attributes of embellishment (adornment, self-assertion and seduction). One of the latest functions of clothing, developed during the twentieth century, was the function to express opinions (moral, sociological and religious) which was called language of clothing.

At the meeting of the “Thematic Expert Group, TEG No. 6 Smart Textiles & Clothing” within the European Technology Platform for the future of textiles and clothing organized by European Apparel and Textile Organization (EURATEX) and held in January 20, 2006, 37 experts coming from all the European countries accepted the definition and characteristics of the term intelligent clothing. The experts agreed that three sets of instruments should be integrated into an article of clothing: sensors for measuring and information input which collect input information, processing unit for interpreting input information and making decisions (microcomputers, microprocessors or microcontrollers with accompanying programs) and output actuators for adapting an article of clothing and provide output information. This definition is in accordance with investigations performed in the sector of developing intelligent clothing and publications in the course of two previous years at the Department of Clothing Technology of the Faculty of Textile Technology of the University of Zagreb. 2. Architecture of intelligent article of clothing with thermal active protection Based on the investigations performed, the whole architecture of an article of clothing with thermal protection is clear, consisting of the following technical subsystems: . system of the outer shell with a programmable variable thickness with an outer and inner protective fabric layer; . system of thermoinsulating chambers with possible control of conduction and convection of body heat; . sensors and systems of measuring input variables; . subsystem of measuring the temperature of environment and microclimate of an article of clothing; . subsystem of pressure measuring in thermoinsulating chambers; . microcontroller measuring and control system for an intelligent article of clothing; . actuator system for an intelligent article of clothing with active thermal protection with elements of micropneumatics for controlling output variables; . power supply; and . measuring and controlling microcontroller program with algorithm of intelligent behavior of an article of clothing. Hereinafter, the architecture elements with integrated technical systems are described. 3. System of outer shell with a programmable variable thickness with an outer and inner fabric layer To manufacture the outer shell and the lining of an intelligent article of clothing with active thermal protection, the model of a men’s sports jacket, Figure 1(a), was chosen for whose construction and modeling a special construction of the basic cut was designed. It is acceptable for modeling all models from which wear comfort and freedom of movement is expected (Ujevic´ et al., 2004). A garment size of 54 was selected for the construction of the basic cut of the outer shell, Figure 1(b), of an intelligent article of clothing with active thermal protection.

Technical systems in clothing 223

IJCST 19,3/4

224

On the basis of main body measurements construction measurements were calculated, and then the construction of the basic cut of the outer shell and lining followed. 4. System of thermoinsulating chambers with possible control of conduction and convection of body heat By courtesy of Bayer Epurex Films GmbH, Germany, a few types of high-elastic polyurethane foils were delivered for the manufacture of thermoinsulating chambers. All foils were subjected to extreme stresses and pressures. The foil designated as Walopur 4201AU showed the best results. It is characterized by a material density of 1.15 g/cm3, a softening point from 1408C to 1508C and a very high elongation caused by breaking force, amounting to 550 percent. Moreover, the material is highly UV resistant, hydrolytically stable, has good properties in joining by thermal and ultrasonic methods, and a good microbiological stability which is important for the incorporation into the clothes. The selected high-elastic polyurethane foil showed better properties in ultrasonic sealing than hot wedge or hot air sealing. The construction of the thermoinsulating chamber is shown in Figure 2. The Pfaff 8310-003 Seamsonic ultrasonic welding machine was used for sealing. Little pipes are used to inject compressed air into thermoinsulating chambers. 6

2

2 Svi

Svi + 1

Dr ½ Odr

¼ Do

3

8

Do

4

Figure 1. Model of a men’s sports jacket: (a) model sketch; and (b) construction of the basic cut

Figure 2. Representation of a sample of the thermoinsulating chamber

Dl

Db (a)

Ultrasonic joint of a airtight seam

(b)

Measuring sample of the thermoinsulating chamber

Pneumatic connection element

Little inflation pipe

5. Sensors and systems of measuring input variables Intelligent article of clothing with active thermal protection should contain sensors of so-called input variables in order to get information about the condition in its environment and the microclimate space between the intelligent article of clothing and human body as well as the information about pressures in the thermoinsulating chambers. In the case of an intelligent article of clothing with thermal protection five input variables are included: (1) environmental temperature of the intelligent article of clothing with active thermal protection (so-called external or ambient temperature) – tout; (2) temperature inside the intelligent article of clothing with active thermal protection (so-called internal or microclimate temperature) – tin; (3) pressure of the shoulder thermoinsulating chamber – pr; (4) pressure of the chest thermoinsulating chamber – pp; and (5) pressure of the waist thermoinsulating chamber – pb.

Technical systems in clothing 225

In accordance with the representation shown in Figure 3, the system of measuring input variable is composed of: . temperature measurement subsystem; . pressure measurement subsystem; and . subsystem for the stabilization of program voltage.

tout

1

TEMPERATURE MEASUREMENT SUBSYSTEM

tin

Shoulder thermoinsulating chamber

2 Chest thermoinsulating chamber Waist thermoinsulating chamber

1

2

SUB SYSTEM FOR THE STABILIZATION OF OPERATING VOLTAGE

pr PRESSURE

pp MEASUREMENT SUBSYSTEM

pb

MEASUREMENT SYSTEM OF INPUT VARIABLES

2 SENSOR SYSTEM

SBUS OF AIR-CONDITIONED MEASURING SIGNALS MEASURING VOLTAGE FOR MEASURIN MICROCONTROLLER PORTS

SENSOR SIGNAL BUS

Air-conditioned and increased measuring voltages of input variables of environment temperatures and microclimate of the intelligent article of clothing and pressures of the

POWER SUPPLY OF THE INTELLIGENT ARTICLE OF CLOTHING

Microclimate inside the intelligent article of clothing

1 - Temperature sensors 2

- Pressure sensors

Figure 3. Representation of the sensor and measurement system of the input variables of an intelligent article of clothing

IJCST 19,3/4

226

air thermoinsulating chambers are led by a measuring bus to the microcontroller ports. The subsystem for the stabilization of program voltage, which is necessary for the operation of the measuring system of input variables, is also used to supply the microcontroller system. 5.1 Subsystem for measuring the pressure in thermoinsulating chambers A RS235-5784 pressure converter by RS Computers was used to measure pressure in the thermoinsulating chambers. The pressure converter dimensions are comparatively small. It was selected because of its good properties and a comparatively small thickness of only 8 mm, which made it possible to integrate the converter successfully into the intelligent article of clothing. According to the production specification (Instruction Leaflet, 1998), the pressure converter in question can measure pressures up to 1 psi according to the English system of measurement, i.e. up to 68.95 bar. It permits a 20-fold overloading up to about 1.4 bar. The pressure sensor in the converter is a thin diaphragm which is deformed with increasing pressure. Four piezo-resistors, which change their electric resistance according to diaphragm deformation, are attached to it. Piezo-resistors are connected to a Wheatstone bridge. Inlet bridge resistance amounts to 5 kV and the recommended input terminal voltage amounts to 10 V. A higher input terminal voltage would cause the heating up of measuring piezo-resistive elements and measuring faults. Therefore, a voltage stabilizer of the input terminal voltage with Whitestone bridge of 5 V was selected for experiments whereby, the supply current amounted to 1 mA. The highest declared voltage output from the converter is 45 mV at the highest pressure of 68.95 mbar when supplying the measuring bridge with a voltage of 10 V. Therefore, the highest sensitivity of the converter amounts to 0.653 mV/mbar. Owing to a reduced input terminal voltage in the experiments conducted, a sensitivity of 0.327 mV/mbar was reached and the highest output voltage from the measuring bridge of 22.55 mV, respectively, at the highest pressure connected to the measuring converter. 5.2 Subsystem for measuring the environment temperature and the microclimate of the article of clothing An intelligent article of clothing with active thermal protection is devised in such a way that it can measure two significant temperature parameters: external temperature (or environment temperature) or microclimate temperature inside the garment. A measuring converter of the third generation with digital processing measurement data designed DS18B20 by Dallas Maxim Semiconductors was selected for temperature measurements. Its measurement accuracy is better than 0.28C, and the measuring range lies between 558C and þ 1258C. The resolution of an incorporated A/D converter can be adjusted between 9 and 12 bit resolution, depending on the user needs. Each DS18B20 temperature converter has a unique and unchangeable mark integrated into a 64-bit serial number in the ROM memory of the converter which serves as an address mark in the bus to which a practically unlimited number of measuring converters can be connected, each with its own address. They all can be connected to the bus which represents only one data conductor and mass (so-called “1-Wire bus”). The supply of the converter may be realized from 3 to 5.5 V local DC source or via a data conductor (so-called “parasite power”). Conversion of temperature

data in the highest resolution of a 12-bit digital word lasts mostly for 750 ms, and a 9-bit one (lowest resolution) lasts for about 94 ms. Thanks to the serial data flow, the DS18B20 temperature converter needs only one data connection, and due to integrating all the components into the basic silicium plate within the converter no external additional component is necessary for its operation. This is very significant information since a remarkable complexity of the measuring sensor at internal level is reflected in a simple construction and realization of an intelligent article of clothing with active thermal protection. 6. Microcontroller measuring and control system of intelligent article of clothing Two microcontrollers are used to control the operation of an intelligent article of clothing with active thermal protection. Most of the functions, including the performance of algorithm of intelligent behavior, are controlled by a very powerful PIC 16F877 manufactured by Microchip, and rational electric power management is controlled by a smaller PIC 16F628 microcontroller of the same manufacturer (Microchip PIC 16F87XA Data Sheet, 2003). The PIC 16F877 microcontroller has 40 connections on the box where a powerful central processing unit with reduced instruction set computer (RISC) architecture is located so that it can recognize and perform 35 different computer instructions necessary for program processing. Its operation speed can be adjusted by changing the control frequency which can amount to 20 MHz whereby an instruction can be performed in only 200 ns. A FLASH program memory with 8 k capacity, RAM data memory with 368 £ 8 byte capacity and EEPROM data memory with 256 £ 8 byte capacity were installed into the same box. Processing requirements can be addressed from 14 possible sources; it can operate in the so-called sleep mode or in a state of nonoperation with reduced electric power consumption. It can operate with supply voltages ranging from 2 to 5.5 V, and at individual connections output current can reach up to 25 mA. Typical consumption amounts to 0.6 mA with a supply of 3 V and a work cycle of 4 MHz. If the work frequency is reduced to 32 kHz, with the same supply voltage its current consumption amounts to only 20 mA. In the sleep mode, its consumption is less than 1 mA. The described microcontroller has three installed timers, two comparators, multichannel analog-to-digital converters with a ten-bit resolution and a possibility of serial and parallel communication with its environment. The smaller microcontroller responsible for a rational electric current consumption is located in the box with 20 leads into which the RISC processing unit as well as program and data memories are built in. The FLASH program memory has a capacity of 20 kilobyte, RAM data memory has a capacity of 224 £ 8 byte and EEPROM data memory has a capacity of 128 £ 8 byte. The work frequency can also reach up to 20 MHz. Two analog comparators and one source of reference voltage and a communication assembly are also built in. The consumption of the microcontroller is less than 2 mA with a supply voltage of 5 V and a work frequency of 4 MHz. At a reduced work frequency of 32 kHz, it consumes approximately 15 mA, and in the sleep mode the bias-current amounted to less than 1 mA. It recognizes and performs 35 program instructions whereby all instructions are performed in one cycle of 1 ns, apart from program branching which are performed in two cycles.

Technical systems in clothing 227

IJCST 19,3/4

228

7. Actuator system for an intelligent article of clothing with active thermal protection with elements of micropneumatics for controlling output variables The actuator system uses micropneumatic elements shown in Figure 4. Compressed air is generated by the DR-4X2PN microcompressor by the American company Clark (11). The microcompressor operates on diaphragm principle and is supplied with a direct current of 9-14 V consuming 200 mA. Maximum work pressure of the microcompressor may be 0.75 bar. Its dimensions are: width 25 mm, length 50 mm, height 66 mm, mass 197 g. Compressed air is led through small tubes (12) to the stop valve (14) behind which there is the pressure regulator (15). The value of the regulated pressure is measured by the sensor (16). Electrovalves (18), (19) and (20) are used to inject compressed air into the thermoinsulating chambers. The pressure in the thermoinsulating chambers is measured via sensor (16). Electrovalves (21), (22) and (23) are used to discharge the air from the thermoinsulating chambers. The microcontroller system used in line with the decisions based on the algorithm of intelligent behavior controls the operation of the electrovalves. The microprocessor is activated when it is necessary to inflate the thermoinsulating chambers, and it is supplied directly from the electric power supply system. 8. Electric power supply system A pack of rechargeable NiCd batteries with a total voltage of 24 V is used to supply sensors, microcontroller system and actuators. Battery capacity is 1,200 mAh.

13

13

13 11 EM 16 21

22

12

14 15 18

Figure 4. Actuator system of the intelligent article of clothing

19 17

16

20

23

The system is supplemented by an adapter which is connected to the power supply system. The microcontroller monitors the battery state-of-charge, and measurement results are displayed. 9. Measuring and control microcontroller program with the algorithm of the intelligent behavior of the article of clothing The operation of an intelligent article of clothing with active thermal protection is based on an abridged conceptual flow diagram shown in Figure 5. After activating the system, a self diagnosis of the state of the technical system of the garment is performed. In the mentioned procedure, the proper operation of all temperature and pressure sensors in the chambers, microcompressors, all for inflating and deflating valves and the battery state are examined. After finishing the self-diagnosis procedure of the state, the data are displayed, and the external and internal temperature is measured. Afterwards, it is possible to select a control mode for the thermoinsulating chambers which can be manual or automatic. Provided that the manual control mode has been selected, it is necessary to make a subjective evaluation of thermal comfort. If a feeling of thermal comfort is expressed, the system is directed to the first subprogram (SUB 1) in which the system deflates all the thermoinsulating chambers in sequence. If the subjective feeling of a severe cold is expressed, the second subprogram (SUB 2) is activated which activates the sequential inflation of all thermoinsulating chambers. By a subjective evaluation of medium cold the third program (SUB 3) is activated which starts the inflation of the shoulder and waist thermoinsulating chamber. Subjective feeling of coolness can be evaluated through two states: cooler and slightly cool. In case of a higher coolness, the fourth program (SUB 4) is activated and this inflates the shoulder chamber. In case of feeling slightly cool, the subprogram 5 (SUB 5) is activated to start the waist chamber. After finishing any of these five subprograms, the external and internal temperature is measured, and the program is reset, i.e. to select a control mode microclimate temperature. Thereupon the lower microclimate temperature limit and the upper microclimate temperature limit are determined. As the initial condition of the lower and upper limit a temperature difference of ^ 28C is selected. After this adjustment and determination of the initial conditions, the internal and external temperature is measured in order to collect parameters for determining the algorithm of intelligent behavior. The first step in making a decision which can lead to branching in the program flow is the comparison between the internal temperature and the upper limit temperature. In the event that the measurement shows that the internal temperature is higher than the upper limit temperature, a conclusion is made that the wearer of the garment is too hot, and it is necessary to activate the SUB 1 and to deflate all the chambers in sequence. In the event that the internal temperature is not higher than the upper limit temperature, the algorithm of behavior will continue according to its basic tree. Thereupon, a comparison of the value of the internal temperature and the lower limit temperature of the microclimate is made. In the event that the internal temperature is higher than the temperature of the lower microclimate limit, the system makes a conclusion that it is within the limit

Technical systems in clothing 229

IJCST 19,3/4 ACTIVATION OF THE SYSTEM

PROCEDURE OF THE SELF-DIAGNOSIS OF THE STATE

230

PRINTOUT OF THE STATE AND TEMPERATURE tv,tu

MANUAL

SUB 1: DEFLATE ALL THE CHAMBERS IN SEQUENCE

NO YES VERY COLD?

SUB 2: INFLATE ALL CHAMBERS IN SEQUENCE

NO YES

SUB 3: INFLATE THE SHOULDER AND WAIST CHAMBER

NO YES COOLER?

SUB 4: INFLATE THE SHOULDER CHAMBER

NO

SLIGHTLY COOL?

NO

YES

MICROCLIMATE TEMPERATURE

MICROCLIMATE TEMPERATURE

tm – 2 = td

td - lower limit of the microclimate

tm + 2 = t g

tg - upper limit of the microclimate

t , tu

MEASURE v

YES

tu > tg

(Meaning: it is too hot)

NO

YES

SUB 1: DEFLATE ALL THE CHAMBERS IN SEQUENCE

tu > t d

(Meaning: temperature is inside ± 2 ºC od tm.Do not react.)

SUB 5: INFLATE THE WAIST CHAMBER

tm = DESIRABLE

tm - INPUT OF THE

THERMAL PROTECTION IS NOT NECESSARY (environment temperature is higher or within limits of the adjusted microclimate temperature)

YES COMFORTABLE?

MEDIUM-COLD?

AUTOMATIC

SELECTION OF THE CONTROL MODE

(Meaning : wearer is too cold. Protection is to be activated)

NO

CALCULATE ∆ t

MEASURE tv, tu

tu > tg tu > td tu = tm ± 2˚C

IT IS TOO HOT OR COMFORTABLE WITHIN LIMITS

∆t > 15˚C

∆t > 10˚C

∆t > 5˚C

∆t < 5˚C

VERY COLD

MEDIUM-COLD

COOLER

SLIGHTLY COOL

SUB 2

SUB 3

SUB 4

SUB 5

THERMAL PROTECTION IS NECESSARY (environment temperature is lower than the adjusted microclimate temperature, and depending on ∆t 4 degrees of thermal protection are selected)

∆t

= tm – tv NO

∆t

> 15˚C NO

(Meaning: medium-cold) YES

∆t

YES

(Meaning: it is very cold)

SUB 2: INFLATE ALL CHAMBERS IN SEQUENCE

> 10˚C

SUB 3: INFLATE THE SHOULDER AND WAIST CHAMBER

NO

(Meaning: cooler) ∆t

> 5˚C

(Meaning: slightly cool)

YES

NO

SUB 5: INFLATE THE WAIST CHAMBER U

SUB 4: INFLATE THE SHOULDER CHAMBER

tm ± 2˚C SUB 1

Figure 5. Operation flow diagram of the intelligent article of clothing with active thermal protection

YES

INTERRUPTION OF THE AUTOMATIC CONTROL

NO

tolerances and hysteresis of ^ 28C of the desirable microclimate temperature, and it makes a decision that the system response is not necessary so that the program is reset to the starting point of the decision-making tree. In the event that the internal temperature is not higher than the temperature of the lower microclimate limit, the system makes a conclusion that the wearer of the garment is cold, and the thermal protection of the garment is to be activated. In the event that the automatic control mode is selected, it is necessary to enter data on the desirable. For the system to make a decision on response intensity of thermal protection, it is necessary to calculate the difference between the desirable microclimate temperature and the measured external temperature. The response intensity of the technical system of the intelligent article of clothing and the decision on combinations of activated thermoinsulating chambers will depend on the above-mentioned difference. If the temperature difference (Dt) between the desirable microclimate temperature and the measured external temperature higher than 158C, the system concludes that it is very cold and activates SUB 2 in which it inflates all the thermoinsulating chambers in sequence. If the temperature difference is lower than 158C, the system finds out whether the temperature difference is higher than 108C. If the difference is higher than 108C, the system concludes that it is medium cold and makes a decision to inflate the shoulder and waist chamber. If the temperature difference is lower than 108C, a comparison is made whether the temperature difference (Dt) is higher than 58C. If the decision made is positive, the system concludes that it is cooler and activates the shoulder chamber, and if the decision made is negative, it concludes that it is slightly cool and activates the waist chamber by subprogram 5 (SUB 5). After finishing any of five subprograms, an enquiry is made of interrupting the automatic control. If there is no desire to interrupt, the program is reset to the starting point of the branching tree; if there is a desire to interrupt the automatic control, the system is switched to manual control.

10. Conclusion An intelligent article of clothing with thermal protection is a complex technical system consisting of a very complex architecture. The whole architecture includes a system of outer shell, thermoinsulating chambers, sensors, microcontroller system, actuator system and power supply system. All the mentioned architectural elements are connected by different bus types to constitute a complex system acting synchronously and effectively (Figure 6). The operation of the system is controlled by the microcontroller containing the program which is based on the algorithm of the behavior of the intelligent article of clothing. The architecture described and the technical systems in the intelligent article of clothing with thermal protection presented in this paper are protected by patent (Rogale et al., 2004, 2005). The technical systems described represent a suitable basis for experiments and scientific research during the introduction of intelligent clothing into human life.

Technical systems in clothing 231

IJCST 19,3/4

232

Figure 6. The prototype of an intelligent article of clothing with active thermal protection

References Instruction Leaflet (1998), Uncompensated Pressure Transducers RS235-5784, tt. RS Computers, pp. 1-4. Microchip PIC 16F87XA Data Sheet (2003), 28/40/44-pin Enhanced Flash Microcontrolers, DS39582B, Dallas Maxima, TX, pp. 1-234. Rogale, D., Firsˇt Rogale, S., Dragcevic, Z. and Nikolic, G. (2004), Intelligent Article of Clothing with an Active Thermal Protection, State Intellectual Property Office of Croatia, Zagreb, P20030727A.

Rogale, D., Firsˇt Rogale, S., Dragcevic, Z. and Nikolic, G. (2005), Intelligent Article of Clothing with an Active Thermal Protection, European Patent Office, Munich, International Application No.PCT/HR2004/000026. Ujevic´, D., Rogale, D. and Hrastinski, M. (2004), Technics of Garment Construction and Modelling, Faculty of Textile Technology University of Zagreb, Zagreb, pp. 235-9. Corresponding author Snjezˇana Firsˇt Rogale can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Technical systems in clothing 233

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,3/4

234

Development of a mathematical model for the heat transfer of the system man – clothing – environment Jelka Gersˇak and Milan Marcˇicˇ Faculty of Mechanical Engineering, University of Maribor, Maribor, Slovenia Abstract Purpose – To develop a mathematical model of the heat transfer from human body to environmental air that will be numerically solved by a personal computer. Design/methodology/approach – Development of the heat transfer model for the system man – clothing – environment is based on studying all forms of dry heat transfer (radiation, convection and conduction). Findings – The advantage of the model for heat transfer is the ability of the fast evaluation of heat transfer for various textile materials incorporated into the clothing system under different ambient conditions. Originality/value – Modelling the heat transfer is original and has an important contribution to thermal comfort. Keywords Clothing, Textile structure, Heat transfer, Modelling Paper type Research paper

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 234-241 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741704

1. Introduction Human thermal sensation is mainly related to the thermal balance of his or her body as a whole. This balance is influenced by physical activity and clothing, as well as by the environmental parameters. Several studies have been conducted in order to quantify and qualify the variables that affect thermal comfort (Fanger, 1970; Mecheels, 1991; Holmer and Elna¨s, 1981; Gagge et al., 1986). They show that comfort can be evaluated from three different classes of variables, this are ambient variables (air temperature, mean radiant temperature, air humidity and air velocity), physical activity and clothing. The first practical system of units for the description of the heat exchange of man and environment was proposed by Gagge et al. (1941) as early as 1941. Clothing insulation was measured in units of “Clo”. Later, Fanger (1970) developed a model of the whole body thermal comfort; known as the predicted mean vote (PMV) model. Fanger’s PMV model was development in the 1970s from the laboratory and climate chamber studies. It was based on thermoregulation and heat balance theories. The PMV model combines four physical variables (air temperature, air velocity, mean radiant temperature, and relative humidity) and two personal variables (clothing insulation and activity level) into an index that can be used to predict thermal comfort. The index provides a score that corresponds to the SIST EN ISO 7730 (2006) seven-point thermal sensation scale as well as ASHRAE thermal sensation scale (ASHRAE Standard 55-1992, 1992). The work presented offers a short survey of some models developed, as well as a presentation of our original model of heat transfer from the body, through the clothing system, into the environment.

2. Models overview In the last 40 years, numerous studies have focused on the heat transfer from human body to surrounding air. A considerable amount of papers have considered modelling heat transfer from human body to the surrounding air (Stolwijk, 1971; Farnworth, 1986; Mecheels, 1998; Negra˜o et al., 1999; Fiala et al., 1999; Song et al., 2004). J.A.J. Stolwijk developed a multinode comfort model, based on six body segments: head, torso, arms, hands, lens and feet (Stolwijk, 1971). Stolwijk’s 25-node model of thermoregulation (Stolwijk and Hardy, 1966) set out the fundamental concept, algorithm, physical constants and physiological control subsystems for many contemporary multimode models. Stolwijk considered clothing as an insulation without mass. Later, B. Farnworth developed a numerical model of the combined diffusion of heat and water vapour through clothing (Farnworth, 1986). He considered the combination of the tree heat flow mechanisms, conduction by air, radiation, and diffusion of water vapour, in a numerical model of multi-layered clothing systems. The calculations were performed in a time-dependent mode and compared to experiments performed on a sweating hot plate in a non-steady state mode. Berkeley multinode comfort model can be also mentioned. The Berkeley multinode comfort model was based on the Stolwijk model as well as on the work by Tanabe in Japan (Tanabe et al., 1995), but included several significant improvements over the Stolwijk model. The Berkeley model used 16 body segments corresponding to the Berkeley segmented thermal manikin (Huizenga et al., 1999; Huizenga et al., 2001). The model is capable of predicting human physiologic response to transient, non-uniform thermal environment. Fiala et al. (1999) presented a computer model of human thermoregulation for a wide range of environmental conditions. The numerical formulation of the passive system was verified by comparison with the known analytic solutions for conduction in cylinders. Further, he presented the active system and compared the model predictions with the experimental data and the predictions by other models. The active system known as a dynamic model is capable of predicting human thermal responses in cold, cool, neutral, warm, and hot environments. On the other hand, Song et al. (2004) developed a numerical model to predict skin burn injury resulting from heat transfer through a protective garment worn by an instrumental manikin exposed to laboratory-controlled flash exposures. This model incorporated characteristics of the simulated flash fire generated in the chamber and the heat-inducted changes in fabric thermophysical properties. The model proved to be a powerful tool for engineering thermally protective garments. Although the models mentioned are fairly complex, a greater part of them have not yet gained widespread use. Reason for this may include lack of confidence in the predictive abilities, limited rang applicability and many times deficient modelling of the heat exchange with the environment. 3. Development of a model for heat transfer from human body to environment From the point of view of thermophysiology, clothing can be seen as a “quasi physiological system” which impacts thermoregulation of the body of its wearer in such a manner that he/she feels comfortable under various weather conditions and at various levels of physical activity (Mecheels, 1998; Gersˇak and Grujic´ 2000).

Development of a mathematical model 235

IJCST 19,3/4

236

The feeling of comfort is in this context a subjective notion, e.g. a psychological condition, while comfort in wearing clothing is expressed as a result of a balanced process of heat transfer among the body, clothing and environment. Since, the evaluation of comfort has the above two aspects, the problem of thermal comfort and objective evaluation of heat exchange between the human body and the environment is not only highly complex, but also quite undetermined. The heat transfer from the human body, through the air gap between the body and the clothing system and the clothing system itself to the environment is a complex process. The logic diagram of the design model of the heat transfer in the system man – clothing – environment is shown in Figure 1. The starting points for an objective evaluation of the heat exchange between the human body and the environment and the development of a suitable model for the heat HEAT TRANSFER OF THE SYSTEM MAN – CLOTHING – ENVIRONMENT

Design human body

Human body

Clothing system

Posture Activity level Environment Air temperature Air velocity Relative humidity

Metabolic rate M Body surface area ADu Clothing area factor fcl

Figure 1. Diagram of the design model of the heat transfer in the system man – clothing – environment

Computation software

Input data

Numerical model Clothing system Air gab Skin heat transfer Skin temperature distribution

Air gab size & distribution Skin resistance to heat transfer Heat transfer coefficients Fabric matter properties

Numerical model Clothing system Vapour transfer

Sweating rate Skin resistance to moisture transfer Fabric matter properties

Prediction & Simulation Heat transfer of the system Man – Clothing – Environment

transfer from the human body to the environment are the study of the relationship between the thermophysical properties of different textile materials incorporated into the clothing systems and the level of thermophysiological wearing comfort achieved at various weather conditions, as well as at different physical activities (Gersˇak and Grujic´ 2000; Plazl, 2004; Gersˇak, 2005). According to the diagram of the design model, Figure 1, the development of the model for heat transfer from the human body to the environment is subdivided into five parts: (1) Modelling a human body as the thermal physiological model. (2) Determining clothing area factor as a basis for the calculation and simulation of heat exchange between the human body and the environment through the layers of textile materials incorporated into the clothing system. (3) Designing a mathematical model for the heat transfer in the system man – clothing – environment. (4) Calculation and simulation of the heat transfer from the human body to the environment. (5) Testing the developed model and its implementation. A model of the human body will be defined by the processed data of the photos and serves as the basis of the virtual mannequin. Parts of the body will be modelled in an object-oriented way parameterised by significant dimensions. From the point of view of thermal physiological model, the three-dimensional model of the body will consist of two kinds of interactive systems: a human tissue system and a thermoregulatory system. Human thermal model will contain a detailed clothing system model as well, characterised by temperature, humidity, and airflow in the air gap between the skin surface and clothing system. The key idea of a model based in the above manner is to construct a real model for an objective evaluation of human thermal comfort, based on the data concerning thermal physical parameters of the material, the data concerning heat and humidity transfer from the body, through the clothing system, and into the environment, as well as on the data concerning the changes of testing person physiological parameters on wearing the clothing. To achieve these goals, extensive theoretical and experimental research work has been carried out, which will enable us to develop the model for the objective evaluation of the heat exchange between the human body and the environment, based on analytical methods and results of the experimental work. Having in mind the complexity of the research, we shall present a mathematical model used to describe heat transfer from the body, through a double-layered clothing system, and into the environment. Body surface area covered by a single or multiple clothing layers can be expressed as a percentage of the total body area according to oSIST prEN ISO 9920 (2005), Figure 2. Clothing area factor fcl is defined as the surface area of the clothed body Acl divided by the surface area of the nude body AD: f cl ¼

Acl : AD

ð1Þ

Total body surface area is traditionally estimated from the simplified equation of D. DuBois and E. F. DuBois (DuBois and DuBois, 1916; Formulas, 2006), that is:

Development of a mathematical model 237

IJCST 19,3/4

1

1

2

7

7

6

238 8

9

4

11

10

Figure 2. Template for the estimation of the body surface area covered by one or more layers of clothing and the nude area

6 3

12

13

14

15

9

8

5

10

11

16 17

13

12

15 14

17 16

Designation 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Total

Segment Head + neck Chest Back Abdomen Buttocks Right upper arm Left upper arm Right lower arm Left lower arm Right hand Left hand Right thigh Left thigh Right calf Left calf Right foot Left foot

% of total 8.7 10.2 9.2 6.1 6.6 4.7 5.2 3.2 3.0 2.5 2.5 9.1 9.3 5.1 6.2 3.6 3.8 100.0

Source: ISI/DIS 9920 (2005)

ADu ¼ 0:20247H 0:725 W 0:425 ; ð2Þ where ADu – DuBois surface area, in m2; W – body weight; kg; H – body height, m. The parameters involved in the calculation are defined for standardised (static body, wind still; i.e. speed , 0.2 m s2 1) condition. When the air moves, or when the body moves, this will affect the heat transfer (typically lower it). The heat exchange between the human body and the environment exhibits several modes. Figure 3 shows how thermal energy is transferred from the human body through the air gap and the clothing system to the environment when the skin temperature Ts is higher than the ambient temperature Ta(Ts . Ta). Heat energy is conducted through the fabric incorporated into the clothing system by conduction and radiation simultaneously. Heat transfer exhibits the following modes: (1) A radiation, convection and conduction mode from skin to clothing system. A radiation mode can be described as follows: ð3Þ qs ¼ s1s T 4s 2 T 4a 2 s1cl T 48 2 T 4a ; where s is the Stefan-Boltzmann constant, 1s and 1cl are the emissivities of the skin and clothing system, Ts is skin temperature and Ta is the ambient temperature. Heat transfer by conduction/convection through the trapped air located between the internal surfaces of the clothing system and human skin can be modelled as a resistance to heat flow between the two surfaces. The heat transfer coefficient agap between the clothing system and the skin is modelled as a function of the size of the air gap and the temperature of the trapped air as: lair agap ¼ Nu; ð4Þ dgap where Nu is Nusselt number, lair is the thermal conductivity of the air and dgap is the size of the air gap.

Development of a mathematical model

Ts T1 T2 ∆T

T4

T3

T5

T7 T6

Temperature drop

239

T8 Ta

ps psat p1

p2

Condensation p3 p4 p

5

∆ p Water vapour pressure drop

p6

psat p7

Skin

Air

d1

d2

Textile I

Textile II

p8

Figure 3. Temperature and water vapour pressure during heat transfer

pa

Air

Heat transfer is: qa ¼ agap ðT s 2 T 2 Þ:

ð5Þ

(2) The conductive mode through the clothing system. Figure 3, shows a two-layered clothing system. The conductive mode can be described as: lavg qcl ¼ ðT 8 2 T 2 Þ; ð6Þ d1 þ d2 where: 1

lavg

¼

1 1 þ ; l1 l2

ð7Þ

where lavg is the average value of the clothing system thermal conductivity. l1 and l2 are thermal conductivities of the clothing textile layers (Figure 3). Heat transfer from the clothing system to the environmental air can be described as: qcl2a ¼ acl ðT 8 2 T a Þ;

ð8Þ

where acl is the heat transfer coefficient from clothing system surface to the ambient air.

IJCST 19,3/4

240

(3) At certain environmental temperature human body starts to secrete sweat which evaporates on the skin surface and starts cooling the body. The evaporated sweat moves from the skin to the clothing system, where it partly condenses or moves through the clothing system (Figure 3). The evaporation and condensation processes are hard to model. The presented model of heat transfer does not contain the evaporation and condensation processes of the sweat. 4. Conclusion The paper presents a development concept of the mathematical model enabling an objective evaluation of the human thermal comfort, respectively, simulation of the heat transfer from the human body to the ambient air. The presented mathematical model is solved using a numerical method by a personal computer. Based on the simulation results, it can be concluded that the numerical solution is able to predict heat transfer from the human body to the ambient air. The simulation model does not include the sweat evaporation on the skin surface and condensation in textile materials incorporated into the clothing system, since these processes are highly complex. The next step in the improvement of the existing model will be to include sweat evaporation on the skin surfaces and sweat condensation in the layers of textile materials into the designing model, as well as the development of a numerical model for an objective evaluation of the human thermal comfort and predicting human physiological response to a transient thermal environment. The greatest advantage of the heat transfer computer simulation is the ability of a fast evaluation of heat transfer for various textile materials incorporated into the clothing system under different ambient conditions. References ASHRAE Standard 55-1992 (1992), Thermal Environment Condition for Human Occupancy, ASHRAE, Atlanta, GA. DuBois, D. and DuBois, E.F. (1916), “A formula to estimate surface area, if height and weight are known”, Archives of International Medicine, Vol. 17, p. 863. Fanger, P.O. (1970), Thermal Comfort. Analysis and Application in Environmental Engineering, McGraw-Hill Book Company, New York, NY. Farnworth, B. (1986), “A numerical model of the combined diffusion of heat and water vapour through clothing”, Textile Research Journal, Vol. 56 No. 11, pp. 653-65. Fiala, D., Lomas, K.J. and Stohrer, M. (1999), “A computer model of human thermoregulation for a wide range of environmental conditions: the passive system”, Journal of Applied Physiology, Vol. 87 No. 5, pp. 1957-72. Formulas (2006), “References and formulas used by the body surface area calculator, formulas for body surface area”, available at: www.halls.md/body-surface-area/refs.htm (accessed 20 June 2006). Gagge, A.P., Burton, A.C. and Bazett, H.C. (1941), “A practical system of units for the description of the heat exchange of man with his environment”, Science, Vol. 94, p. 428. Gagge, A.P., Forbelets, A.P. and Berglund, P.E. (1986), “A standard predictive Index of human response to thermal environment”, ASHRAE Transactions, Part 2, Vol. 92, pp. 709-31. ˇ Gersak, J. (2005), “Influence of kind of materials, for car seat on thermophysiological comfort of driver”, Proceedings: [IN-TECH-ED ’05 inovation-technics-education]. Budapest: Budapest

Tech, Rejto˝ Sa´ndor Faculty of Light Industry Engineering, paper presented at V: 5th International Conference IN-TECH-ED ’05, Budapest, Hungary, 7-9 September, p. 366. ˇ Gersak, J. and Grujic´, D. (2000), “Influence of thermophysiological characteristics of clothes on the human comfort during different burden exposure and climate condition” (“Vpliv oblacˇila na toplotno fiziolosˇko udobje cˇloveka pri razlicˇnih obremenitvah in klimatskih pogojih”), Tekstilec, Vol. 46 Nos 7/8, pp. 183-90. Holmer, I. and Elna¨s, S. (1981), “Physiological evaluation of the resistance to evaporative heat transfer by clothing”, Ergonomics, Vol. 24 No. 1, ISSN: 0014-0139, pp. 63-74. Huizenga, C., Hui, Z. and Arens, E. (2001), “A model of human physiology and thermal comfort for assessing complex thermal environments”, Building and Environment, Vol. 36 No. 6, pp. 691-9. Huizenga, C., Hui, Z., Duan, T. and Arens, E. (1999), “An improved multinode model of human physiology and thermal comfort”, in Masaya, O. (Ed.) IBPSA Conference Proceedings, Kyoto, 13-15 September. Mecheels, J. (1991), Ko¨rper – Klima – Kleidung: Grundzu¨ge der Bekleidungsphysiologie, Schiele & Scho¨n, Berlin, ISBN 3 7949-0531. Mecheels, J. (1998), Ko˝rper – Klima – Kleidung: Wie funktioniert unsere Kleidung?, Schiele & Scho¨n GmbH, Berlin, ISBN 3 7949-619-5. Negra˜o, C.O.R., Calvalho Fo, C.O. and Melo, C. (1999), “Numerical analysis of human thermal comfort inside occupied spaces”, in Masaya, O. (Ed.) IBPSA Conference Proceedings, Kyoto, 13-15 September. oSIST prEN ISO 9920: 2005 (2005), Ergonomics of the Thermal Environment – Estimation of the Thermal Insulation and Evaporative Resistance of a Clothing Ensemble, Annex H Guidance on the Determination of the Covered Body Surface Area, International Standards Organization, Geneva, pp. 89-90. Plazl, K. (2004), “Raziskava vpliva vrste lezˇisˇcˇa na fiziolosˇko udobje cˇloveka (Research of the influence of bed lines on physiological comfort of human being)”, Master’s thesis, Faculty of Mechanical Engineering, University of Maribor, Maribor. SIST EN ISO 7730 (2006), “Ergonomics of the thermal environment – analytical determination and interpretation of the thermal comfort using calculation of the PMV and PPD indices and local thermal comfort criteria”, November. Song, G. et al. (2004), “Modelling the thermal protective performance of heat resistant garments in flash fire exposures”, Textile Res. J., Vol. 74 No. 12, pp. 1033-40, ISSN: 0040-5175. Stolwijk, J.A.J. (1971), A Mathematical Model of Physiological Temperature Regulation in Man, National Aeronautics and Space Administration (NASA CR-1855), Washington, DC. Stolwijk, J.A.J. and Hardy, J.D. (1966), “Temperature regulation in man – a theoretical study”, Pflugers Archive Ges. Physiology, No. 291, pp. 129-62. Tanabe, S., Stuzuki, T., Kimura, K. and Horikawa, S. (1995), “Numerical simulation model of thermal regulation of man with 16 body parts of evaluating thermal environment (Part I Heat transfer at skin surface and comparison with SETT and Stolwijk model)”, paper presented at Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan. Corresponding author Jelka Gersˇak can be contacted at: [email protected] To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Development of a mathematical model 241

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,3/4

A smart sensor for compression measurement in automotive textiles

242

Emilie Drean, Laurence Schacher and Dominique Adolphe Laboratoire de Physique et Me´canique Textiles, Mulhouse, France, and

Franc¸ois Bauer PIEZOTECH SA, Hesingue, France Abstract Purpose – For size reasons, adapted sensors, able to measure intrinsic mechanical properties of fabrics, have not been developed yet. The study aims at developing a sensor that can be inserted within a specific textile, a “complex” fabric used as seat-cover fabrics, and consisting of an assembly of three layers. Design/methodology/approach – Piezoelectric polymer sensors containing polyvinylidene fluoride (PVDF) were chosen. A total of 20 “complex” were studied. A characterisation in compression was achieved, using the Kawabata Evaluation System. The best-adapted measurement method using a PVDF sensor has been required. The method consists in analyzing the response under compressive stress of a PVDF disc using the resonant frequency of the material. A constraint series is applied to the fabric in the sensor area; the maximal phase at the sensor’s resonant frequency is taken up for each one. Findings – Phase variation is linear and differs according to the studied “complex”. A correlation study between Kawabata compression parameters and slopes did not show any relationship between slope values and compression properties when the surface fabrics of “complex” are compared, but a classification in “families” is possible when different foams are considered. Research limitations/implications – Further studies should demonstrate whether these “smart” textiles could find applications in the automotive field, to measure accurately the mass of a passenger. The influence of the external parameters (vibrations, temperature variations) has to be checked, knowing that the sensor is not depending on moisture. To complete the study, the sensor has to be tested in a real situation, i.e. inserted in a car-seat, in contact with a human body. Originality/value – This study promises development of a sensor that can be inserted into a specific textile, a “complex” fabric used as a seat-cover, consisting of an assembly of three layers. Keywords Textiles, Automotive industry, Piezo electricity, Compressive strength Paper type Research paper

International Journal of Clothing Science and Technology Vol. 19 No. 3/4, 2007 pp. 242-252 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710741713

Introduction Induced stresses in textiles play a major role in their mechanical properties. Optimization of these stresses during fabrication can improve mechanical properties of finished product and can also have an impact on durability. On the other hand, induced stresses are responsible for worn fabric behaviour and a best knowledge could improve comfort. But for size reasons, adapted sensors, able to measure intrinsic mechanical properties of fabrics, have not been developed yet. The key of the problem is the design of functional textiles generally called “smart textiles”. Polyvinylidene fluorure and its copolymer exhibit tailorable ferroelectric, piezoelectric and structural properties, and are useful in a variety of transducer and sensor applications. Submicronic shaped

sensors made from these materials and directly integrated in a fabric, could be used for pressure measurement for example. Since, 1975, ISL studied PVDF polymer films. These films are used as active elements in stress gauges to provide nanosecond, time-resolved stress measurements thanks to a patented polarisation process (Bauer, 1986), which confers to PVDF reproducible piezoelectric and pyroelectric properties. PVDF is a semicrystalline polymer in which each monomer unit – (CH2-CF2) – has two dipole moments, one due to CF2 and the other to CH2. PVDF exhibits a variety of molecular conformations and crystal structures depending on the method of preparation. The most common and most studied of these are the a phase which has a trans/gauche conformation and the b phase in which the conformation is the all-trans planar zigzag with the dipole moments perpendicular to the chain axis. The b phase is the ferroelectric and the most useful phase. Among ferroelectric polymers, b-PVDF has the unusual property that the forces responsible for the ordering of dipoles in the ferroelectric phase are sufficiently strong that the polymer melts before it undergoes a ferroelectric to paraelectric phase transition. Copolymers of VDF and TrFE exhibit ferroelectric transitions below the melting point. Because, induced stresses in textiles are very low, a high-sensitivity sensor is necessary. Particular properties of PVDF make this material really interesting for “smart textile” applications.

Sensor for compression measurement 243

Sensor integration The problem that has to be solved to design an instrumented fabric is the implantation of the sensor in the embedded fabric. To facilitate a first development, the study has been focused on a specific type of fabric called “complex” fabric. Such fabrics are composed of three layers bonded by different processes. These “complex” fabrics are used in automotive field for seat upholstery. The “complex” fabric is the cap of the car-seat, that is to say the covering part. It is composed of three layers, which are a surface polyester fabric (weaving, knitting, etc.), a thin polyurethane foam (0.5-5 mm), and a polyamide knitting which is used to improve cap slip on the cushioning, and also reinforce mechanical properties (Figure 1). These three layers are bonded thanks to a laminating process (flame or adhesive). Taking into account this particular structure, a sensor can easily be inserted between two layers of the complex, in our case between surface fabric and foam. Method used for sensor implantation is the following. First, the surface fabric and the foam are separated by immersion in a solvent (acetone). This solvent has no influence on fabric or foam properties. A flexible adhesive of small thickness is then applied on the foam to bond the sensor and the surface fabric on it.

fabric

PU foam

knitting

Figure 1.

IJCST 19,3/4

244

Fabrics characterization “Complex” fabrics, as used in seat upholstery, are mostly stressed in compression. PVDF sensors behaviour when inserted in textiles is not known, so the first step of the study consists in mechanical characterization of the “complex” under compression stress, to obtain comparison data. Characterization in compression was achieved with the compression device of the Kawabata Evaluation System (Kawabata, 1980). The specimen is compressed by two circular plates (2 cm2) until pressure reaches 5 kPa. This pressure is similar to the maximal pressure that can be applied by a human body on a car-seat. Measurement speed depends on “complex” thickness: 0.02 mm/s for “complex” thickness until 5 mm and 0.2 mm/s between 5 and 50 mm. During measurement, the compression and recovery processes are recorded. All measurements were performed in textile standard conditions (208C and 65 per cent HR). A total of 20 different “complex” fabric combinations have been studied. These “complex” fabrics are only composed of two layers, i.e. the surface fabric and the foam, as previous studies on such specific fabrics have shown that the knitting has no significant influence on compression properties of the whole “complex” (Schacher and Adolphe, 1999). The 20 “complex” studied are composed of seven different surface fabric, one fabric (satin weave, No. 1), two Jacquard knitting (Nos 2 and 3), and four Raschel velvets (Nos 4, 5, 6, and 7). Eight foams were selected, of different compression behaviour. These elements are representative of classical industrial upholstering. Every selected material has first been tested separately (both fabrics and foams), and bonded. A comparison of curves obtained has shown that in every combination, “complex” fabric behaviour follows foam behaviour. Surface fabric behaviour is not preponderant. Following figures show two compression curves obtained for “complex” fabrics with foams of opposed compression behaviour. The curve, shown in Figure 2, is representative of polyurethane foams behaviour under compression load. The upper curve can be divided in two parts, a linear deformation and a constraints plateau. This foam can be considered as “compressive” and by comparison, the foam 1 (compression curve shown in Figure 3) can be considered as “non-compressive”. 5

pressure (kPa) "complex" fabric velvet 3 + foam 5

4

3

2

1 thickness (mm)

Figure 2.

8

7

6

5

4

0

Sensor for compression measurement

5

pressure (kPa)

4

"complex" fabric velvet 3 + foam 1

245

3

2

1 thickness (mm) 8

7

6

Figure 3.

0

Table I summarizes Kawabata parameters measured for the 20 “complex” combinations studied. Each combination has been tested ten times. Results are presented with following abbreviations: . “sf” for surface fabric, the different fabrics are numbered as previously explained; . “F” for foam, the different foams are numbered from 1 to 8; and

“Complex” fabric combination Satin weave (sf1) Knitting 1 (sf2) Knitting 2 (sf3) Velvet 1 (sf4) Velvet 2 (sf5) Velvet 3 (sf6)

Velvet 4 (sf7)

Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam Foam

WC 5 1 5 1 5 1 5 1 5 1 1 2 3 4 5 6 7 8 5 1

10.41 0.87 9.03 0.82 8.85 0.73 9.21 1.43 9.43 1.31 0.87 1.25 0.78 4.29 9.20 1.58 0.73 7.95 9.00 1.21

Kawabata compression parameters EMC RC To 45.24 6.99 51.95 6.86 51.65 5.79 43.00 6.42 48.73 10.28 6.39 7.70 5.31 30.88 51.65 8.77 4.42 33.75 48.88 8.81

56.28 45.89 58.34 49.51 57.53 53.28 54.78 50.19 53.56 40.15 45.75 53.71 48.47 42.56 50.60 55.96 34.77 49.06 53.09 47.81

6.78 6.06 6.40 5.77 6.15 5.65 7.90 7.33 7.34 6.68 6.70 6.41 6.40 4.52 7.27 6.82 7.34 6.31 7.03 6.39

Tm 3.72 5.64 3.07 5.38 2.97 5.30 4.50 6.86 3.76 5.99 6.27 5.90 6.06 3.10 2.97 6.23 7.01 4.18 3.59 5.83

Table I. Kawabata compression parameters measured

IJCST 19,3/4

246

.

WC for compression energy (Nm/m2), EMC for compressibility ratio (per cent), RC for resiliency (per cent), To for thickness at 0.5 kPa (mm) and Tm for thickness at 5 kPa (mm).

Compression energy and compressibility ratio seem to be preponderant factors as their values are ten times higher for “complex” fabrics with foam 5 (the most compressive) than for “complex” fabrics with foam 1. Resiliency is similar in the two cases, and if we look at the thickness difference between To and Tm, Tm is two times lower for foam 5, and approximately the same for foam 1. Considering compressibility ratio, “complex” fabrics with velvet 3 can be divided in three groups, the less compressive, F1sf6, F2sf6, F3sf6, F6sf6 and F7sf6, the most compressive, F5sf6, and in the middle F4sf6 and F8sf6. Measurement: phase variation at sensor’s resonant frequency PVDF sensors are usually used for pressure measurement, but are sensitive to the electrical environment. In case of textiles, the sensor should be able to detect very low stresses. Metrology method generally used with PVDF sensors is to measure charge variations of the sensor, generated under stress. Preliminary tests have shown that this method is not well adapted to measure induced stresses in textiles. Use of PVDF in resonant frequency mode could be interesting. It has been demonstrated that resonators show a linear variation of its resonant frequency according to load applied (Ngamalou et al., 1995). Moreover, phase decreases linearly versus resonant frequency. The sensor used is a P (VDF-TrFE) transducer, of low crystallinity and high yieldingness. It is circular (20 mm in diameter) to avoid unwanted resonance, 500 mm thick and gold sputtered on both sides; gold deposit acts as electrods. Connexion is assumed by two thin wires soldered on each side of the sensor. The sensor is connected with an impedance/gain phase analyzer, which measures frequency, phase and impedance in a frequency range of 100 Hz-40 MHz. Raw measurements are adjusted thanks to a short circuit and an open circuit measurement, done in the same frequency range. Vertical loads are applied on the sensor area confined in the fabric, from 0 to 150 cN. A composite disk of 5 cm2 is placed on the “complex” pressure applied is thus included between 0 and 3 kPa (Figure 4). For each load applied the maximal phase at sensor resonant frequency (2.2 MHz for the sensor) is taken in, and reported on a graph. Measurement is repeated 10 times for each “complex” combination studied, and a mean slope is calculated. Results and discussion Measurement method detailed below has been applied for each “complex” fabric combination. Figure 5 shows the slopes obtained for three different one. Several observations can be done. First, slopes of all curves are clearly different. This difference is the most marked when the foams constituting the “complex” fabric are differing. As an illustration, the foam 5 is a soft foam, and the foam 1 is rigid. Results obtained seem to be coherent with these observations. The slope is lower for the softer foam which should more damp vibrations. Moreover, the curve obtained for the “complex” fabric F5-sf3 is slightly curved, whereas linearity is excellent in the other cases. This could be due to foam 5 behaviour, which is more compressive than foam 1.

Sensor for compression measurement

counterweight device

247 impedance/ gain-phase analyzer

"complex"

connection with the sensor

polymer sensor

0

0,5

1

−66,1

1,5

2

2,5

Figure 4.

3

pressure (kPa)

−66,3

phase (°)

−66,5 −66,7 "complex" − F5 + sf3 −66,9

"complex" − F1 + sf3 "complex" − F1 + sf6

−67,1 −67,3 −67,5

y = −0,4365x − 66 R2 = 0,9993

y = −0,1993x − 66 R2 = 0,9917

y = −0,4675x − 66 R2 = 0,9996

These observations are similar for every “complex” fabric combinations studied but not presented here. Values measured for these combinations are summarized in Table II. Results previously obtained show that measurement method using resonant PVDF is sensitive enough to detect different surface fabrics in a “complex” as every slope is different. When different foams are used, this difference is more marked. But, taking into account measurement errors, a part of slope values could be considered as equivalent. A statistical analysis has been performed to see if differences between slopes values are really significant. Statistical analysis consists first in an F-test, to test difference between standard deviations. Then, if F-test is accepted, a t-test can be performed.

Figure 5.

IJCST 19,3/4

248

Table II. Slopes mean values calculated for the “complex” fabrics combinations studied

“Complex”

Slope mean value

F1sf1 F1sf2 F1sf3 F1sf4 F1sf5 F1sf6 F1sf7 F5sf1 F5sf2 F5sf3 F5sf4 F5sf5 F1sf7 F2sf3 F3sf3 F4sf3 F5sf3 F6sf3 F7sf3 F8sf3

20.4041 20.4846 20.4365 20.4486 20.4246 20.4675 20.3896 20.2085 20.2398 20.1992 20.1999 20.2072 20.2254 20.7110 20.6191 20.4532 20.2058 20.6796 20.5740 20.3762

The sample size is ten for every slope, and reliability index of 5 per cent. Consequently, range bounds for F-test are (0.248; 4.03), and for t-test (2 2.262; 2.262). Results of F-test and t-test are, respectively, depicted in Tables III and IV. Values included in range bounds are marked in grey. The “complex” fabrics with various surface fabrics are first considered. In every cases, F-test is accepted, i.e. difference between standard deviations is non-significant. Concerning t-test, difference between slopes mean values is significant for “complex” fabrics with foam 1, but for “complex” fabrics with foam 5, difference is non-significant in five cases. This could be explained by the foam behaviour. Foam 5 is more compressive; “complex” fabric behaviour depends, therefore, more on foam than on surface fabric. Moreover, sensor phase variation is lower and measurement errors are more important. Concerning “complex” fabrics with various foams, differences between slopes mean values are significant in every cases, as expected.

Table III. Results of F-test

F1 þ sf 1-7 1-2 0.377 3-4 1.358 F5 þ sf 1-7 1-2 0.673 3-4 0.672 sf6 þ F 2-8 2-3 1.269 4-5 1.196

1-3 0.323 3-5 2.482 1-3 0.607 3-5 0.575 2-4 2.464 4-6 0.100

1-4 0.438 3-6 0.902 1-4 0.408 3-6 2.135 2-5 2.946 4-7 0.348

1-5 0.800 3-7 1.133 1-5 0.349 3-7 1.681 2-6 0.247 4-8 1.437

1-6 0.291 4-5 1.828 1-6 1.297 4-5 0.856 2-7 0.857 5-6 0.084

1-7 0.365 4-6 0.664 1-7 1.021 4-6 3.176 2-8 3.541 5-7 0.291

2-3 0.856 4-7 0.835 2-3 0.903 4-7 2.502 3-4 1.942 5-8 1.202

2-4 1.162 5-6 0.363 2-4 0.607 5-6 3.711 3-5 2.321 6-7 3.476

2-5 2.124 5-7 0.457 2-5 0.519 5-7 2.923 3-6 0.194 6-8 14.361

2-6 0.772 6-7 1.257 2-6 1.928 6-7 0.788 3-7 0.675 7-8 4.132

2-7 0.970 2-7 1.518 3-8 2.791

sf6 þ F 2-8

F5 þ sf 1-7

F1 þ sf 1-7

1-2 25.72 3-4 3.19 1-2 9.99 3-4 0.17 2-3 211.56 4-5 248.17

1-3 9.76 3-5 23.47 1-3 22.86 3-5 1.88 2-4 236.57 4-6

1-4 14.99 3-6 7.42 1-4 2 2.32 3-6 5.05 2-5 2 73.40 4-7 16.20 4-8 2 15.59

1-5 8.35 3-7 2 11.85 1-5 2 0.33 3-7 8.11 2-6

1-6 18.38 4-5 27.79 1-6 2.41 4-5 1.58 2-7 215.65 5-6

1-7 24.59 4-6 4.83 1-7 6.04 4-6 4.19 2-8 249.71 5-7 50.44

2-3 212.25 4-7 216.09 2-3 211.52 4-7 6.91 3-4 225.53 5-8 36.34

2-4 29.89 5-6 12.10 2-4 210.11 5-6 2.03 3-5 265.44 6-7 27.77 6-8

2-5 2 18.54 5-7 2 10.72 2-5 2 7.87 5-7 4.67 3-6

2-6 2 4.23 6-7 2 19.15 2-6 2 8.35 6-7 4.00 3-7 2 5.42 7-8

3-8 2 39.47

2-7 2 4.62

2-7 2 25.00

Sensor for compression measurement 249

Table IV. Results of t-test

IJCST 19,3/4

250

This statistical analysis shows that compression behaviour of “complex” fabrics seems to be correlated with phase variation of the sensor, as we observed a lower measurement precision in case of softer foam. Correlation between slope values and Kawabata compression parameters has therefore been studied. Examples of correlation study with compressibility ratio EMC are shown on Figures 6 and 7. Figure 6 clearly shows that there is no correlation between slopes mean values and compressibility ratio for the “complex” fabrics composed of foam 1 and various surface fabrics. On the contrary, for “complex” fabrics composed of surface fabric 6 and various foams (Figure 5(b)), points can be divided in three groups, corresponding to the groups defined thanks to Kawabata measurement. These groups were the less compressive, F2sf6, F3sf6, F6sf6 and F7sf6, the most compressive, F5sf6, and in the middle F4sf6 and F8sf6. Similar results are obtained with the compression parameters WC and RC, that is to say no correlation when the surface fabric varies, and a correlation when foam varies. Concerning parameters To and Tm, no correlation is observed whatever the “complex” fabric studied. Previous observations seem to attest measurement method. The sensor used is linked with the foam but not with the surface fabric. The sensor should therefore “move” with the foam and not with surface fabric. Conclusion and perspectives The aim of the study was to design a sensor able to measure induced stresses in textiles. We chose to use PVDF sensors because of their insertion capability in fabrics. The general metrology method has been adapted to avoid noise problems. So, measurement has been oriented on study of resonant frequency variations of the sensor. Sensor behaviour is linear, either with compressive and non-compressive foam. Statistical study shows that the difference between slopes for the different “complex” fabrics studied is significant in most cases. There is no correlation between Kawabata EMC % −0,36 6

5

8

7

9

10

11

−0,38

−0,42 −0,44

F1 sf7

Slope °/Pa

−0,40

F1 sf1 F1 sf5

F1 sf3 F1 sf4

−0,46 F1 sf6 −0,48

Figure 6.

−0,50

F1 sf2

EMC % −0,15 0

10

20

30

40

−0,25

−0,45 −0,55

60

F5 sf6

251

Slope °/Pa

−0,35

50

Sensor for compression measurement

F8 sf6 F4 sf6

F7 sf6 F3 sf6

−0,65

F6 sf6 F2 sf6 −0,75

compression parameters and slopes when the surface fabric varies. When the foam constituting the “complex” fabric varies, slopes can be divided in three groups corresponding to the general compression behaviour of the foam. A future application of this instrumented fabric could be in automotive field to contribute to “smart” airbags development, able to provide appropriate inflation depending on passenger morphology and positioning. In fact, airbags are really efficient but present a number of hazards. These hazards are firstly linked with the own system: burns due to high-temperature gas propulsion, auditory lesions caused by explosion noise, explosive toxicity. All other hazards are linked with a “bad use” of the airbag, i.e. car occupant is “out of position”. What are called “out of position” are all the cases when car-seat is abnormally occupied. For example, when car-seat is empty (airbag inflation is not necessary), when passenger is not far enough from airbag (he can face hurt), or when an infant safety seat is placed on car-seat (absolutely no inflation). That’s why information about car occupant is necessary. Such information could be obtained by integrating sensors in car-seat fabrics. Sensor function will therefore be to detect passenger occupancy and to weigh him. Depending on this information, airbags will be disconnected, partially or totally inflated. Taking into account results obtained by the metrology method developed, study of phase should work to weigh car passenger. To complete the study, the sensor has to be tested in real case, i.e. inserted in a car-seat, a human body in contact with. Influence of external parameters like temperature or vibrations has to be checked, knowing that the sensor is not depending on moisture. References Bauer, F. (1986), “Method and device for polarizing ferroelectric materials”, US Patent 4, Vol. 611, p. 260. Kawabata, S. (1980), “The standardization and analysis of hand evaluation”, Hand Evaluation and Standardization Committee, 2nd ed., Textile Machinery Society of Japan, Osaka.

Figure 7.

IJCST 19,3/4

Ngamalou, L., Benech, P. and Chamberod, E. (1995), “Pressure measurement with resonant PVDF or copolymer”, Ferroelectrics, Vol. 171 Nos 1/4, pp. 217-24. Schacher, L. and Adolphe, D.C. (1999), “Etude de la compression de mate´riaux multicouches utilise´s en garnissage automobile”, Proceedings of 14e`me Congre`s Franc¸ais de Me´canique, CD-ROM, N. 37, Toulouse, France, September.

252

Further reading Drean, E., Schacher, L., Bauer, F. and Adolphe, D.C. (2005), “A smart sensor for induced stresses measurement in automotive textiles”, Proceedings of The Fall 2005 Meeting of the Fiber Society, Newark, New Jersey, USA, October, pp. 87-9. Corresponding author Emilie Drean can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Artificial neural network-based prediction of aesthetic and functional properties of worsted suiting fabrics

Artificial neural network-based prediction

B.K. Behera and Rajesh Mishra

Received March 2006 Accepted December 2006

Department of Textile Technology, Indian Institute of Technology, New Delhi, India

259

Abstract Purpose – The purpose of this paper is to investigate an alternative approach that can predict non-linear relations. Design/methodology/approach – An engineered approach to fabric development is described in which a radial basis function network is trained with worsted fabric constructional parameters to predict functional and aesthetic properties of fabrics. An objective method of fabric appearance evaluation with the help of digital image processing is introduced. The prediction of fabric properties by the network with changing basic fibre characteristics and fabric constructional parameters is found to have good correlation with the experimental values of fabric functional and aesthetic properties. Findings – The radial basis function network can successfully predict the fabric functional and aesthetic properties from basic fibre characteristics and fabric constructional parameters with considerable accuracy. The network prediction is in good correlation with the actual experimental data. There is some error in predicting the fabric properties from the constructional parameters. The variation in the actual values and predicted values is because of small sample size. Moreover, the properties of worsted fabrics are greatly influenced by the finishing parameters which are not taken into consideration in the training of the network. Prediction performance can be further improved by including these parameters as input, during the training phase. In few cases, the network has predicted contradictory trends, which are found difficult to be explained. Originality/value – The paper describes a radial basis function neural network model that can be used for the prediction of the fabric appearance values and comfort properties using fabric constructional parameters and some primary fibre mechanical properties as input parameters of the network. Keywords Neural nets, Image processing, Fabric testing Paper type Research paper

1. Introduction In today’s competitive market situation it is very essential to reduce lead time for any type of product development activity. The manufacturers not only have to deliver fabrics with specified quality but also within the shortest possible time as well. The introduction of fabric objective measurement technology has enabled the garment manufacturers to specify the required mechanical properties of the woven fabrics. However, the problem of the weaver, who must now produce fabrics to the desired specifications, has not been considered in any detail. Empirical methods can result in the production of excessively expensive solutions, since it is impossible to be sure that one has produced a unique solution. Admittedly, there is no reliable predicting tool to

International Journal of Clothing Science and Technology Vol. 19 No. 5, 2007 pp. 259-276 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710819483

IJCST 19,5

260

be really useful to the weaver to design and manufacture apparel fabric to meet the fabric property requirements for a given end use (Grosberg and Leaf, 1988). With continuous improvement in standard of living, aesthetic characteristics of clothing have become equally important as functional properties in determining quality as well as serviceability of apparel fabrics. Fabric handle as a measure of mechanical comfort has been evaluated by Kawabata evaluation system (KES) instrument for more than three decades (Kawabata, 1980; Kawabata and Niwa, 1991). However, fabric appearance in general is evaluated from traditional appearance attributes like wrinkling, creasing, texture, pilling, drape and colours. Most of these attributes are measured using methods which lack reproducibility. Therefore, fabric manufacturers need a reliable and objective method for measuring aesthetic appearance of apparel fabrics. For engineering design of textile structures, many models are derived from first principles which normally yield unacceptable errors due to assumptions and therefore become unsuitable as prediction tools under real world dynamic conditions. In this context, artificial neural networks (ANN) offer an efficient tool to model the complexities of fabric structure-property relationships. In the past, weaving technologists used/preferred one or the other empirical relationships for calculating sett/count relationships in fabrics such as – Ashenhurst’s “ends plus intersections theory” Armitage and Brierley’s maximum sett theory, or mathematical theories based on Peirce’s geometric model, graphical solutions, to optimize fabric design (Ashenhurst, 1885; Armitage, 1907; Brierley, 1953; Peirce, 1937; Love, 1954). The advantage of applying ANN is that, it can model the input-output relationships at any level of interactions without any assumptions or constraints on parameters. It has the ability to predict the dynamics of real fabrics (Behera and Muttagi, 2002; Muttagi, 2002). This research work describes ANN approach to engineer structural design of woven wool and wool blended suiting fabrics to be useful for the weavers. The radial basis function network which is proved to be one of the most efficient tools, is used to predict the aesthetic and functional properties of suiting fabrics from the basic raw material characteristics and fabric constructional parameters (Behera and Muttagi, 2004a, b). 2. Methodology In this study, 58 worsted suiting fabric samples were tested for their constructional parameters as well as aesthetic and functional properties. Aesthetic properties include total appearance value, fabric appearance index (FAI) and drape coefficient. The functional properties are restricted to handle and comfort properties like total hand value, formability, air permeability, relative water vapour permeability and thermal insulation. Functional properties are taken from KES and fabric assurance by simple testing (FAST) equipment where as for appearance an image processing-based system was developed for evaluation of FAI by objective measurement of appearance attributes. A radial basis function neural network model was used for prediction of the fabric appearance values and comfort properties using fabric constructional parameters and some primary fibre mechanical properties as input parameters of the network. 2.1 Materials The samples tested were all suiting fabrics with areal densities ranging between 130 and 300 GSM. A variety of blends were taken ranging from 100 per cent wool to blends of wool with other natural fibres like silk, tassar, cotton, linen, etc. Some commercial

worsted suiting fabrics of wool: PET, wool: nylon and wool: acrylic blends were also selected for analysis. 2.2 Methods 2.2.1 Evaluation of functional properties. 2.2.1.1 Total hand value. The total hand value is determined using KES instrument for winter as well as summer applications separately using standard procedures prescribed by the inventor (Kawabata, 1983). The Kawabata equation for calculating total hand value is given below: " # 16 X C 1i ðX i 2 m1i Þ C 2i ðX 2i 2 m2i Þ þ THV ¼ C 0 þ s1i s2i i¼1 where C0, C1i and C2i are coefficients/constants for the ith parameter, Xi is the mechanical property of the ith variable term, m1i and s1i are the population mean and standard deviation, m2i and s2i are the square mean and standard deviation. 2.2.1.2 Air permeability. Air permeability of all the fabric samples was measured on Textest FX 3300 tester. The instrument measures the volume of air passing through the fabric per unit area per unit time. The parameters of this test are: air pressure – 100 Pa, diameter of sample under test – 2.54 cm, test area – 5.08 cm2. 2.2.1.3 Thermal insulation. The thermal insulation test was carried out on Kawabata evaluation system (KES-FB7) to measure the resistance to heat flow or thermal insulation value of the fabric. The testing parameters are: sample dimensions – 20 £ 20 cm, method involved – dry contact, air velocity – 30 cm/s. 2.2.1.4 Water vapour permeability. The relative water vapour permeability was evaluated on the Alambeta Permetest equipment. Relative water vapour permeability : P wv ¼ ¼

q1 q0

Heat lost when fabric is placed on the measuring head Heat lost from bare measuring head

Therefore: P wv ¼

u1 u0

where u1 is the output voltage when fabric is placed on the measuring head; u0 is the output voltage from bare measuring head, sample dimensions – 10 £ 10 cm, minimum sample thickness required – 0.5 mm 2.2.1.5 Formability. Fabric formability is the product of bending rigidity and low stress extensibility as introduced by Lindberg and is measured by FAST system (Lindberg et al., 1960). Bending rigidity is measured by FAST-2 and low stress extensibility by FAST-3. The system calculates the formability and gives results in warp and weft directions separately. 2.2.2 Evaluation of aesthetic properties. 2.2.2.1 Total appearance value. The total appearance value was calculated from the fabric low stress mechanical properties tested on the KES. Kawabata equation for

Artificial neural network-based prediction 261

IJCST 19,5

262

calculating total appearance value is given below (Kawabata, 1983): " # 8 X C 1i ðP i 2 m1i Þ C 2i ðP 2i 2 m2i Þ TAV ¼ C 0 þ þ s1i s2i i¼1 where C0, C1i and C2i are coefficients/constants for the ith parameter, Pi is the mechanical property of the ith variable term, m1i and s1i are the population mean and standard deviation, m2i and s2i are the square mean and standard deviation. 2.2.2.2 Fabric appearance index. The properties which are commonly used to access the aesthetic appearance of an apparel fabric are drape (mechanical properties), texture (constructional parameters), wrinkle (irregular surface deformations), and pilling (surface abrasion). Each of these four parameters is measured by using digital image processing system and a standard procedure is developed to evaluate fabric appearance by combining above parameters into a weighted average approach. This integrated appearance index is termed as FAI. In order to determine weightage of each attribute, an expert panel is constituted and a survey was conducted to decide the contribution of each element to fabric appearance: FAI ¼

n X

Ai W i

i¼1

where n – total no. of properties, Ai – grade of the ith property obtained by digital image processing, Wi – weightage of the ith property 2.2.2.3 Drape coefficient. The drape coefficient for the fabric samples is evaluated by using the digital image processing-based drapemeter. Drape coefficient ¼

Ad 2 A1 £ 100 A2 2 A1

where Ad is the area of the draped fabric image, A1 is area of the fabric supporting disc and A2 is the area of the undraped fabric sample. 2.2.3 Evaluation of fibre properties. 2.2.3.1 Evaluation of tensile properties. The fibre tensile properties were evaluated on the Instron tensile tester. Paper windows were prepared with a dimension of 2 cm length and 1 cm width. Single fibres were tested for their tenacity, extensibility and initial modulus: Gauge length – 2 cm, jaw speed – 1 cm/min, time to break – 20 ^ 2 s. 2.2.3.2 Evaluation of bending rigidity. Kawabata evaluation system (KES-FB2), bending tester was used to measure the bending rigidity of fibres. Paper windows of 10 cm length and 1 cm width were prepared for this purpose: no of fibres per window – 10, maximum curvature – 2.5 cm2 1. 2.2.3.3 Measurement of fibre diameter. Fibre diameter was measured using projectina microscope. 2.2.3.4 Measurement of fibre denier. Denier value for all types of fibre samples was calculated by taking weight of a known length of fibres using a digital balance. 2.3 Neural network model The radial basis function neural network model was used for training and prediction of the parameters. In the first stage, the networks were trained to predict the fabric

properties from fibre, yarn and fabric constructional parameters as inputs. In the second stage, the neural networks were tested to prescribe desired worsted fabric property from the input parameters. The strategy used in RBF networks consists of approximating an unknown function with a linear combination of non-linear functions, called basis functions. The basis functions are radial functions, i.e. they have radial symmetry with respect to a centre. The schematic of the RBF network with n inputs and a scalar output is shown in the Figure 1. The input parameters for the neural network are given in Table I and the output parameters in Table II. Before feeding to the network, the input-output data set was scaled down to be within (0-1), by dividing each value of the data by the maximum value of the overall data in order to decrease the error and learning time. The output was descaled by multiplying with the maximum value again. . Thus, the number of input nodes is 18 and number of output nodes is 10. . Sum square error goal was set to 1.0. . Spread constant was 2.2. . Number of neurons in the hidden layer, i.e. epochs was 21 and training time 3.85 s.

Artificial neural network-based prediction 263

MATLAB 6.5 of Mathworks Inc. was used for the purpose. λ1

x1

λ0

x2

λ2

∑

xn

λn

linear combiner

fr (x)

Figure 1. Schematic of RBF network

nonlinear transformation

Input parameters

Maximum

Minimum

1. Wool content (per cent) 2. Percentage of other component fibre 3. Fibre denier 4. Fibre diameter (microns) 5. Fibre bending rigidity (cNcm2/tex2) 6. Fibre initial modulus (cN/denier) 7. Fibre extensibility (per cent) 8. Fibre tenacity (cN/denier) 9. Warp yarn resultant count (Nm) 10. Single or 2-ply 11. Weft yarn resultant count (Nm) 12. Single or 2-ply 13. Ends per cm 14. Picks per cm 15. Warp crimp (per cent) 16. Weft crimp (per cent) 17. Fabric areal density (g/m2) 18. Fabric thickness (mm)

100 80 3.5 22.5 0.00105 201.39 29.36 6.12 50 1 50 1 34 27 25.5 25.5 311 0.6

20 0 1.5 12.5 0.00020 55.23 2.48 1.61 11 0 11 0 15 13 2.9 1.3 120 0.25

Table I. Input parameters

IJCST 19,5

Output parameters

264

1. Total appearance value (TAV) 2. Fabric appearance index (FAI) 3. Total hand value (Winter) – THV(W) 4. Total hand value (Summer) – THV(S) 5. Warp way formability (mm2) 6. Weft way formability (mm2) 7. Drape coefficient (per cent) 8. Air permeability (cm3/cm2/s) 9. Relative water vapour permeability (per cent) 10. Thermal insulation (per cent)

Table II. Output parameters

Maximum

Minimum

3.99 4.53 4.34 3.9 0.46 0.81 53.7 46.85 38.33 28.57

2.03 2.97 2.46 0.55 0.12 0.07 22.27 8.23 12.85 4.52

3. Results and discussion The radial basis function network was trained with the fabric data set as input and output pairs and was tested to predict the fabric properties from the input parameters. 3.1 Prediction performance The prediction performance of the neural network was evaluated by testing with ten sample fabrics data which were not used during the training step. The error of prediction was calculated for each parameter with respect to the actual experimental data. Grand mean error per cent was also calculated so as to evaluate the prediction performance of the network as a whole. At the same time, the correlation coefficient for each parameter was found for the predicted values and the actual values. The results are given in Table III. 3.2 Trend evaluation The network was evaluated to see how it predicts the fabric properties if changes in one input parameter are made keeping the other parameters fixed at the mean value. The input parameter was varied in ten equal steps and fed to the network. Output curves were plotted against the actual values from the fabric data set (Table IV). 3.2.1 Prediction of worsted fabric properties from fibre parameters as input. 3.2.1.1 Effect of wool content on worsted fabric properties. Figure 2 shows the effect of increasing wool per cent in the blend on various fabric functional and aesthetic properties. Figure 2(a) and (b) show an increase in appearance value with increasing Fabric parameters

Table III. Prediction performance of the neural network

1. Total appearance value 2. Fabric appearance index 3. Total hand value (Winter) 4. Total hand value (Summer) 5. Formability1 6. Formability2 7. Drape coefficient 8. Air permeability 9. Relative water vapour permeability 10. Thermal insulation Mean

Error (per cent)

R2

21.36 16.5 14.2 16.63 12.3 14.12 10.24 14.13 16.88 6.25 14.26

0.76 0.88 0.86 0.83 0.89 0.75 0.91 0.79 0.76 0.96 0.84

5

35 2

0

50

4 3 2

100

0

50

(a) TAV

3 2

40 30 60

80

100

50

(e) Formability1

100

265

40

60

80

100

(f) Formability2

50

50

40 30 20

80

0.5 0 20

100

RWVP %

50

0

AP (cm3 /cm2 /sec)

(d) THV(S) DC %

0.4

100

60

1

0.3 50

40

(c) THV(W) F2mm2

F1 mm2

THV(S)

4

40

3 2 20

100

0.5

20 20

4

(b) FAI

5

0

Artificial neural network-based prediction

5

THV(W)

FAI

TAV

5

0

(g) Drape Coefficient

50

100

(h) Air Premeability

40 30 20

0

50

100

(i) Relative Water Vapour Premeability

TI %

30 20 10 0 20

40

60

80

100

(j) Thermal Insulation Notes: X-axis-Wool %, Y-axis- Fabric Properties, (++++ Actual Values, ****Predicted Values)

Figure 2. Effect of wool content on worsted fabric properties predicted by the neural network

wool proportion in the fibre mix. This is due to the fact that wool is a longer staple fibre and more uniform in its cross-section, as compared to other natural fibres taken in the blend. Thus, a higher percentage of wool generates more uniformity in the yarn and the fabric leading to better appearance. Figure 2(c) shows increase in total hand value for winter application with increasing wool ratio due to smoothness and fullness resulting from higher bulk. However, an adverse effect is seen with respect to summer Name of input fabric parameter to the neural network 1. Wool content (per cent) 2. Fibre denier 3. Fibre bending rigidity (cN.cm2/tex2) 4. Fibre initial modulus (cN/denier) 5. Fibre extensibility (per cent) 6. Fibre tenacity (cN/denier) 7. Warp or weft resultant count (Nm) 8. Thread density (ends/cm and picks/cm) 9. Warp and weft crimp percentage 10. Fabric areal density (g/m2)

Fixed value 60 2.5 0.00060 130 15 4 30 20 13.5 200

Table IV. Fixed values of fabric input parameters for trend evaluation

266

application as because bulk and smoothness go against the requirement of a crisp feel desired. This is shown in Figure 2(d). Figure 2(e) and (f) shows increasing trend in formability which is product of bending rigidity and low stress extensibility. Though the bending rigidity is low for wool, but it has a very high extensibility thus contributing towards increase in formability. Drape coefficient in wool rich fabrics is lower due to higher crimp and low bending rigidity as shown in Figure 2(g). An increased wool percentage increases air permeability due to more openness of the fibre assembly resulting from highly crimped wool fibres shown in Figure 2(h). Water vapour permeability is a combination of adsorption, diffusion and evaporation through the micro pores in the yarns and macro thread spacings in the fabric which depends on the openness of structure thus being facilitated by higher crimp with increase in wool per cent as is shown in Figure 2(i). Figure 2(j) shows increasing thermal insulation which again is a function of crimp generating more interfibre air gaps responsible for the insulation behaviour. Moreover, the micropores in the scaly wool fibre structure have a great to contribute to this effect. 3.2.1.2 Effect of fibre denier on worsted fabric properties. The effect of increasing denier/resultant linear density of the fibre blend on fabric properties is shown in Figure 3(a) and (b) shows a decrease in appearance value due to the fact that coarse fibres in the yarn lead to less number of fibres in the cross-section and thereby more

2.5 0 1.5

5

5

FAI

TAV

5

2.5

THV(W)

IJCST 19,5

4 3 2 1.5

3.5

2.5

(a) TAV

2.5

0.4

0.3 1.5

3.5

(d) THV(S)

0 1.5

3.5

30

50

50

40

40

3.5

30

30 20 1.5

3.5

2.5

(f) Formability2

%

cm3 /cm2 /sec

DC %

40

2.5

2.5

0.5

(e) Formability1

50

3.5

1

F2 mm2

F1 mm2

THV(S)

3

2.5

(c) THV(W)

0.5

4

20 1.5

0 1.5

3.5

(b) FAI

5

2 1.5

2.5

(g) Drape Coefficient

2

2.5

3

(h) Air Premeability

3.5

20 1.5

2.5

3.5

(i) Relative Water Vapor Premeability

%

30

Figure 3. Effect of fibre denier on worsted fabric properties predicted by the neural network

15

0 1.5

2.5

3.5

(j) Thermal Insulation Notes: X-axis-Fibre Denier, Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

susceptible to diametric variations. Coarser natural fibres are normally shorter in length and generate more pills on abrasion. Coarser fibres generate less smoothness and softness leading to lower winter value of handle shown in Figure 3(c). The hand value for summer application is increased with coarser fibres because of crispness and higher bending rigidity leading to anti drape properties as is shown in Figure 3(d). Figure 3(e) and (f) shows a decrease in formability in spite of higher bending rigidity due to low extensibility of coarser fibre spun yarns. A yarn spun from coarser fibres has less number of fibres in the cross-section and thereby a lower degree of relative movement between adjacent fibres. As shown in Figure 3(g), the drape coefficient increases due to higher bending rigidity of coarser fibres. Figure 3(h) and (i) shows an increasing trend in air and water vapour permeability, respectively, which is attributed to lower cohesion between coarser fibres in the yarn cross-section resulting in a more open structure. Thermal insulation property however is diminished due to shortage of micro air pores as shown in Figure 3(j). 3.2.1.3 Effect of fibre bending rigidity on worsted fabric properties. Figure 4 shows the prediction trend for the fabric properties resulting from an increasing bending rigidity of fibres. As is shown in Figure 4(a) and (b), the fabric appearance improves with increase in fibre bending rigidity. This is because of higher drape coefficient value and higher resistance to creasing/wrinkling. Figure 4(c) shows a decrease in hand 5

2.5 0

4 3

0.0005

0.001

0

0

4 3 0.0005

0.4

0.001

0.001

0

0.0005

0.001

(f) Formability2 50

50

40

40

%

cm3/cm2 /sec

30 0.0005

0.0005

(e) Formability1

40

20 0

0.5 0

0

(d) THV(S) 50

0.001

1

0.3

2

0.0005

(c) THV(W) Fl mm2

Fl mm2

THV(S)

0.001

0.5

5

DC %

0.0005

(b) FAI

(a) TAV

0

2.5 0

2 0

267

5

THV(W)

FAI

TAV

5

Artificial neural network-based prediction

30

30

20

20 0

0.001

(g) Drape Coefficient

0.0005

0.001

(h) Air Permeability

0

0.0005

0.001

(i) Relative Water Vapor Permeability

%

30 15 0

0

0.0005

0.001

(j) Thermal Insulation Notes: X-axis-Fibre Bending Rigidity (cN.cm2/tex2), Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

Figure 4. Effect of fibre bending rigidity on worsted fabric properties predicted by the neural network

268

value for winter application which is because of lower smoothness and softness. However, an improvement in summer hand value is resulted from crispness and anti-drape stiffness due to higher bending rigidity shown in Figure 4(d). Increased bending rigidity always increases formability and the same is shown in Figure 4 (e) and (f) for warp and weft directions, respectively. A higher bending rigidity results in maintaining the anti-drape flatness of fabric and less hanging along gravity thus resulting in an increase in drape coefficient value shown in Figure 4(g). Air and water vapour permeability are favoured by increasing fibre bending rigidity because of lower cohesion between fibres and a loose yarn and fabric structure shown in Figure 4(h) and (i). Fibres with higher bending rigidity will crimp less and thus a yarn spun from these is susceptible to less crimping resulting in an excessively open structure and absence of micro air particles for thermal insulation shown in Figure 4(j). 3.2.1.4 Effect of fibre initial modulus on worsted fabric properties. Fibre initial modulus is yet another important parameter deciding the low stress mechanical properties responsible for the functionality and aesthetics of apparel fabrics. It is shown in Figure 5(a) and (b) that there is an increase in fabric appearance value resulting from increase in fibre initial modulus. This is attributed to the lower pilling, higher drape coefficient and lower distortion in the fabric structure resulting from higher initial modulus. However, as far as winter suitings are concerned, the hand value is decreased. This is due to loss of smoothness, softness and fullness required for 5

FAI

TAV

5 2.5 0 50

5

THV(W)

IJCST 19,5

4 3

250

50

(a) TAV

Fl mm2

Fl mm2

THV(S)

3

0.4

250

50

150

50

250

30

250

50

50

40

40

30

250

30

20 50

(g) Drape Coefficient

20 150

250

50

150

250

(h) Air Permeability (i) Relative Water Vapor Permeability

%

30

Figure 5. Effect of fibre initial modulus on worsted fabric properties predicted by the neural network

150

(f) Formability2

%

cm3/cm2 /sec

40

150

0.5

(e) Formability1

50

250

0

0.3 150

150

(c) THV(W) 1

(d) THV(S) DC %

250

0.5

4

20 50

150

(b) FAI

5

2 50

0 50

2 150

2.5

15 0 50

150

250

(j) Thermal Insulation Notes: X-axis-Fibre Initial Modulus (cN/Denier), Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

winter applications. The trend is shown by Figure 5 (c). In summer applications as shown in Figure 5(d), the handle is improved by increased anti-drape property and crispness with increasing initial modulus. Figure 5 (e) and (f) show an increase in formability which is due to higher bending rigidity associated with a higher initial modulus. It is well established that bending rigidity is related to modulus of elasticity as:

Artificial neural network-based prediction 269

G¼E £I

5

4

4

3 2

5

THV(W)

5

FAI

TAV

where G – bending rigidity, E – Young’s modulus, I – inertia and for a circular cross-section it is equal to: r 4 =4 Drape coefficient as shown in Figure 5(g) also has an increasing trend due to stiffness of fabric resulted from high initial modulus of fibres. Transmission properties are favoured by higher initial modulus as a result of more open structure in yarn and fabric due to lower interfibre cohesiveness. Air and relative water vapour permeability are improved as shown in Figure 5 (h) and (i). As shown in Figure 5(j) the thermal insulation is hampered due to a high modulus fibre which is less prone to deformation and crimping and does not allow air pores inside yarn structure. 3.2.1.5 Effect of fibre extensibility on worsted fabric properties. The effect of fibre extensibility on fabric properties is shown in Figure 6. In Figure 6(a) and (b) it is shown

3

15

30

3 2

2 0

4

0

(a) TAV

15

30

0

15

30

(c) THV(W)

(b) FAI 1

0.5 4 3 2

0.4

F2 mm2

Fl mm2

THV(S)

5

0.3 0.2

0

15

30

30

0

40 30 20

50

50

40

40

30

30

30 20

20

30

15

(f) Formability2

%

cm3/cm2 /sec

DC %

15

(e) Formability1

50

15

0.25 0

0

(d) THV(S)

0

0.5

0

(g) Drape Coefficient

15

30

(h) Air Permeability

0

15

30

(i) Relative Water Vapor Permeability

30

%

20 10 0 0

15

30

(j) Thermal Insulation Notes: X-axis-Fibre Extensibility (%), Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

Figure 6. Effect of fibre extensibility on worsted fabric properties predicted by the neural network

IJCST 19,5

270

that fabric appearance is favoured by a higher extensibility of fibres. This is due to the fact that a low stress elongation is desirable for all appearance attributes like low pilling, good draping, and anticrease property, etc. Total hand value for winter application is also increased as shown by the prediction curve in Figure 6(c). This is because smoothness, fullness and softness are resulted from higher extensibility at low load. Higher extensibility hampers crispness and anti-drape stiffness thus reducing the total hand value for summer suits. This fact is supported by Figure 6(d). Warp way formability increases with increasing extensibility as shown in Figure 6(e). Formability is the product of fabric bending rigidity and low stress longitudinal compressibility. Within the elastic limit, longitudinal compressibility is assumed to be equal to low stress extensibility. Thus, an increase in any of these two component factors would result in higher formability. However, at the fibre level, the inherent structural parameters like crystallinity and orientation have a major role in deciding both of these mechanical properties. In the weft direction, the formability initially decreases and then increases with increasing extensibility under low load condition as in (f) of Figure 6. This may be attributed to the fact that in the initial conditions a higher extensibility of fibres in the yarn results in lower bending rigidity of the yarn and thereby reducing the formability of fabric. However, for still higher extensibility, the value itself becomes substantially higher to contribute towards higher formability. And all this is more relevant in case of a weft yarn that is comparatively low twisted and slight variation in fibre properties may vary the yarn properties. Drape coefficient values show a decreasing trend with increasing fibre extensibility which helps in easy draping of fabric under gravity. This is shown by Figure 6(g). Air and relative water vapour permeability as shown in Figure 6(h) and (i), respectively, show a decreasing trend. This is due to the fact that a higher fibre extensibility results in a compact yarn and compact fabric thereby. Thus, the transmission of air and water vapour are restricted. The thermal insulation is also decreased initially because of the same reason but with further increase in fibre extensibility, the recovery also increases, causing more crimp in fibre and thus creating micro openings to entrap air pockets inside the fibre assembly. This improves thermal resistance as is shown in Figure 6 (j). 3.2.1.6 Effect of fibre tenacity on worsted fabric properties. The influence of fibre tenacity on worsted fabric functionality and aesthetics is shown in Figure 7. As shown in Figure 7(a) and (b), the appearance of fabric is improved with increasing fibre tenacity. This is because a high tenacity fibre maintains the dimensional stability of the structure thereby improving the low stress mechanical properties as well as pilling resistance and crease/wrinkle resistance, etc. Drape coefficient will be higher for these kind of fabrics and also there will be uniformity in the thread density distribution improving the texture. All these factors are responsible for an improved appearance. In Figure 7(c) it is shown that the THV(W) is decreasing with increase in fibre tenacity. This is mainly because of loss in smoothness and softness of fabric developed from higher tenacity fibres. However, THV(S) shows an increasing trend due to a crisp feel developed in the fabric as well as a higher anti-drape property associated with the fabric as is shown in Figure 7(d). Fabric formability increases with increase in fibre tenacity as because a higher tenacity is associated with higher bending rigidity as is shown in Figure 7(e) and (f). Drape coefficient is a function of the spread of fabric against gravitational force which

0 0

4

THV(W)

2.5

4 3 2 0

8

(a) TAV

0

3 4

0.4 0.3

8

30 0

4

8

271

0 0

4

8

0

4

8

(f) Formability2

50

50

40

40

%

cm3/cm2 /sec

40

8

0.5

(e) Formability1

50

4

(c) THV(W) 1

(d) THV(S)

DC %

8

F2 mm2

Fl mm2

THV(S)

4

20

0 4

0.5

0

2.5

(b) FAI

5

2

Artificial neural network-based prediction

5

5

FAI

TAV

5

30

30 20

0

(g) Drape Coefficient

4

8

(h) Air Permeability

20

0

4

8

(i) Relative Water Vapor Permeability

%

30 15 0 0

4

8

(j) Thermal Insulation Notes: X-axis-Fibre Tenacity (cN/Denier), Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

largely depends on the bending rigidity. Thus, fabrics produced from high tenacity fibres show a higher drape coefficient and the same is shown by Figure 7(g). Figure 7(h) and (i) show increasing trend in air permeability and relative water vapour permeability, respectively, with an increase in fibre tenacity. This is because of the openness of yarn structure resulting from less crimp in higher tenacity fibres. However, thermal insulation property depends on air spaces which is lacking in the absence of crimp as shown in Figure 7(j). 3.2.2 Prediction of worsted fabric properties from yarn and fabric constructional parameters as input. 3.2.2.1 Effect of yarn linear density on worsted fabric properties. Figure 8 shows the prediction of worsted fabric properties by changing the yarn linear density. In Figure 8(a) and (b) it is shown that the fabric appearance deteriorates as the yarn becomes finer. This is due to more unevenness in the finer yarns as the number of fibres in the yarn cross-section goes down. Total hand value for winter applications also falls down because of decreasing smoothness, fullness and softness as shown in Figure 8(c). Summer hand value shows an increasing trend because of crisp feeling of light weight fabrics woven from finer yarns Figure 8(d). As because finer yarns have less number of fibres in the cross-section, there is less interfibre friction and a lower bending rigidity is observed, which is responsible for

Figure 7. Effect of fibre tenacity on worsted fabric properties predicted by the neural network

0

25

4 3 2

50

THV(W)

FAI

3.2 2

5

5

5

TAV

IJCST 19,5

0

25

(a) TAV

0

25

0.4 0.3

50

0

20 10

20

30

40

25

0 10

50

50

50

20

30

40

50

50

40 30 20

50

(f) Formability2 FWP %

30

AP(cm3/cm2 /sec)

40

40

0.5

(e) Formability1

50

30

(c) THV(W) F2 mm2

Fl mm2

3

20

1

(d) THV(S)

DC %

2 10

50

0.5

4

2

3

(b) FAI

5

THV(S)

272

4

40 30 20

0

(g) Drape Coefficient

25

50

(h) Air Permeability

0

25

50

(i) Relative Water Vapor Permeability

Figure 8. Effect of yarn linear density on worsted fabric properties predicted by the neural network

TI %

30 20 10 0

10

20

30

40

50

(j) Thermal Insulation Notes: X-axis-Yarn Linear Density (Nm), Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

decreasing trend in fabric formability as is shown in Figure 8(e) and (f). Drape coefficient in Figure 8(g) also shows a decreasing trend because of the same reason. Fabrics woven from finer yarns are more open in their structure and therefore favour the air and water vapour permeability. The increasing permeabilities are shown in Figure 8(h) and (i), respectively. With respect to thermal insulation, the finer yarn woven fabrics possess less air pockets due to less number of fibres in yarn cross-section and thus there is a decrement in heat resistance Figure 8(j). 3.2.2.2 Effect of thread density on worsted fabric properties. Thread density is yet another major factor influencing the fabric properties at large. Figure 9 shows effect of increasing thread densities both in warp and weft directions on fabric appearance, handle and comfort. In Figure 9(a) and (b) it is shown that the fabric appearance value increases with increase in thread density. This is because of uniformity of texture, higher drape coefficient, pilling resistance and better low stress mechanical properties which ultimately improves the appearance. Figure 9(c) shows that the total hand value for winter application is also increased with increasing thread density. This can be attributed to improvement in bulk, softness, fullness and smoothness in the fabric with higher thread density. However, summer handle is deteriorated with higher thread density due to loss of crispness which is clearly shown in Figure 9(d). Figure 9(e) and (f) show an increase in formability with higher thread densities. This is because of increase in bending rigidity with higher warp and weft densities. The same is responsible for increase in drape coefficient which is shown in Figure 9(g). Transmission properties are adversely affected by increasing thread densities. This is because the inter yarn spacing is reduced thereby reducing the permeability of air and

5

3.5 2 0

20

4 3 2

40

0

20

(a) TAV

20

0.4

0.3

40

0

10

20

30

20

40

40

0

10

20

30

40

(f) Formability2 50

50 40 30 20

30

273 0

40

FWP %

30

AP(cm3/cm2 /sec)

DC %

40

20

0.5

(e) Formability1

50

10

1

(d) THV(S)

0

0

(c) THV(W) F2 mm2

Fl mm2

THV(S)

3

20

3 2

40

0.5

4

0

4

(b) FAI

5

2

Artificial neural network-based prediction

5 THV(W)

FAI

TAV

5

0

20

40

(h) Air Permeability

(g) Drape Coefficient

40 30 20

0

20

40

(i) Relative Water Vapor Permeability

TI%

30 20 10 0

0

10

20

30

40

(j) Thermal Insulation Notes: X-axis-Thread Density (per cm), Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

water vapour as is shown by Figure 9(h) and (i). Thermal insulation property in improved with increasing thread densities because of increase in fabric bulk, increasing number of micro air pockets responsible for thermal resistance. It is shown in (j) of Figure 9. 3.2.2.3 Effect of yarn crimp on worsted fabric properties. The effect of yarn crimp on properties of the fabrics as shown by the neural network is plotted in Figure 10 against the actual values. In Figure 10(a) and (b) the effect of increase in yarn crimp on fabric appearance is shown. Figure 10(b) shows an increasing trend in the fabric appearance. Higher crimp generally is associated with higher thread density and uniformity of texture. Thus, better appearance is predicted in terms of FAI. But, as shown in Figure 10(a), TAV which is purely calculated from low stress mechanical properties shows initial decrease and increase with further increase in crimp per cent. This is because, initially when the crimp increases, the bending rigidity of fabric decreases along with other low stress mechanical properties and thus decreasing drapability and overall appearance. However, further increase of crimp, makes the fabric firm and compact to improve appearance. The total hand value for winter application also increases with increase in yarn crimp due to increase in fabric bulk resulting in improved softness and fullness as is shown in (c) of Figure 10. In Figure 10(d) it is shown that the total hand value for summer application is diminished and this is due to loss of crispness and stiffness when the yarns are highly crimped in the fabric. As is shown in Figure 10(e), warp way formability initially increases with increase in crimp of yarn because of increasing firmness and resulting fabric bending rigidity but further increase of crimp lowers the bending rigidity due to excessive deformation of yarn and thus the formability decreases. Weft way formability shows initial decreasing trend due to the fact that the weft yarn is normally low twisted and

Figure 9. Effect of thread density on worsted fabric properties predicted by the neural network

5

3.5 2 0

20

4 3 2

40

5

THV(W)

FAI

5

TAV

IJCST 19,5

0

(a) TAV

Fl mm2

THV(S)

3

0

30 0

20

0

40

0

40

50

40

50

40 30 20

20

(f) Formability2 FWVP %

DC %

40

20

20

0.5

(e) Formability1 AP(cm3/cm2 /sec)

(d) THV(S) 50

40

1

0.4

0.3

40

20

(c) THV(W)

0.5

4

20

0

(b) FAI

5

2 0

3 2

40

F2 mm2

274

20

4

0

(g) Drape Coefficient

20

(h) Air Permeability

40

40 30 20

0

20

40

(i) Relative Water Vapor Permeability

Figure 10. Effect of yarn crimp per cent on worsted fabric properties predicted by the neural network

TI%

30 20 10 0 0

20

40

(j) Thermal Insulation Notes: X-axis-Yarn Crimp %, Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

increased deformation of yarn lowers the bending rigidity. However, further increase in the yarn crimp generates firmness in the fabric and increases bending rigidity as well as formability thereby. This fact is shown with Figure 10(f). Drape coefficient in Figure 10(g) shows an increasing trend due to increasing bending rigidity with a more compact and firm fabric structure resulting from higher yarn crimp. Air permeability initially increases slightly due to increase in openness of structure when the yarns try to deform slightly higher but with further increase in yarn crimp a firmer structure shows which obstructs air passage through the fabric Figure 10(h). Water vapour permeability is a more complex phenomenon as compared to air permeability as it involves adsorption, diffusion and evaporation. Increasing crimp always decreases openness needed for water vapour transmission. Sometimes air may pass through but water vapour cannot as is shown in Figure 10(i). Thermal insulation behaviour can be explained in the same light of air permeability as because the former depends on the presence of air pores in the structure. Initial increase of crimp slightly increases fabric openness for air permeability but the openings are too big for thermal resistance. Further, crimping the yarn, increases fabric bulk and more air pores are generated to facilitate thermal insulation property Figure 10(j). 3.2.2.4 Effect of fabric areal density on worsted fabric properties. Fabric areal density is probably the most important parameter describing almost all the properties. This, of course, is the function of many other constructional parameters like fibre type, fibre crimp, yarn linear density, thread density and yarn crimp, etc. Figure 11 shows the influence of

5

2.5 2 120

220

4 3 2 120

320

(a) TAV

220

0.4

0 120

320

30 220

320

220

320

(f) Formability2 50

40

40

%

cm3/cm2 /sec

DC %

220

50

40

275

0.5

(e) Formability1

(d) THV(S) 50

320

1

0.3 120

320

220

(c) THV(W)

F2 mm2

Fl mm2

THV(S)

3

20 120

0 120

320

0.5

4

2 120

220

2.5

(b) FAI

5

Artificial neural network-based prediction

5

THV(W)

FAI

TAV

5

30

30 20 120

(g) Drape Coefficient

220

320

(h) Air Permeability

20 120

220

320

(i) Relative Water Vapor Permeability

30

%

20 10 0 120

220

320

(j) Thermal Insulation Notes: X-axis-Fabric Areal Density (g/m2), Y-axis- Fabric Properties (++++ Actual Values, ****Predicted Values)

change in GSM on the functional and aesthetic attributes. Figure 11(a) and (b) shows the increase in fabric appearance value with increasing areal density. This is because of more uniform mass distribution in a heavier fabric and higher low stress mechanical properties along with higher pilling resistance, higher drape coefficient and wrinkle resistance. With respect to handle, as is normally anticipated a higher areal density favours winter applications shown in Figure 11(c) and deteriorates summer hand value shown in Figure 11(d) because of better fullness and smoothness but inferior crispness. Formability Figure 11(e) and (f) as well as drape coefficient Figure 11(g) show increasing trend due to increasing bending rigidity of fabric with higher GSM. Both air permeability and relative water vapour permeability show decreasing trend for higher areal densities because of increasing compactness shown in Figure 11(h) and (i). Thermal insulation property as shown in Figure 11(j) is however favoured because of more air spaces in a fabric of higher areal density and related higher bulk. 4. Conclusion It can be concluded that the radial basis function network can successfully predict the fabric functional and aesthetic properties from basic fibre characteristics and fabric constructional parameters with considerable accuracy. The network prediction is in good correlation with the actual experimental data. There is some error in predicting the fabric properties from the constructional parameters. The variation in the actual values and predicted values is because of small sample size. Moreover, the properties of worsted fabrics are greatly influenced by the finishing parameters which are not taken into

Figure 11. Effect of fabric areal density on worsted fabric properties predicted by the neural network

IJCST 19,5

276

consideration in the training of the network. Prediction performance can be further improved by including these parameters as input, during the training phase. In few cases, the network has predicted contradictory trends, which are found difficult to be explained. References Armitage, E. (1907), “Structure Property Relationships for Fabrics”, Huddersfield Textile Society Journal, Vol. 8, p. 297. Ashenhurst, T.R. (1885), Weaving and Designing Textile Fabrics, Broadbend and Co., London. Behera, B.K. and Muttagi, S.B. (2002), “Engineering design of woven fabrics-a recent approach”, Indian Journal of Fibre and Textile Research, Vol. 27 No. 3, pp. 315-22. Behera, B.K. and Muttagi, S.B. (2004a), “Performance of error back propagation vis-a`-vis radial basis function neural network, part-i: prediction of properties for design engineering of woven suiting fabrics”, Journal of Textile Institute, Vol. 95 Nos 1/6, pp. 283-300. Behera, B.K. and Muttagi, S.B. (2004b), “Performance of error back propagation vis-a`-vis radial basis function neural network, part-ii: reverse engineering of woven fabrics”, Journal of Textile Institute, Vol. 95 Nos 1/6, pp. 301-17. Brierley, S. (1953), “Cloth settings reconsidered”, Textile Manufacture, Vol. 78 No. 7, pp. 349-51, No. 8, pp. 431-3, No. 9, pp. 449-53, No. 10, pp. 533-7, No. 12, pp. 653-5, Vol. 79 No. 6, pp. 293-4. Grosberg, P. and Leaf, G.A.V. (1988), “Modern computational methods in the design of wool textile structures”, in Carnaby, G.A., Wood, E.J. and Story, L.F. (Eds), Application of Mathematics and Physics in the Wool Industry, WRONZ and The Textile Institute (New Zealand Section), Christchurch, pp. 83-91. Kawabata, S. (1980), Standardization and Analysis of Hand Evaluation, Textile Machinery Society of Japan, Osaka. Kawabata, S. (1983), “Recent developments in the objective measurement of apparel fabric handle, quality and physical properties”, in Postle, R., Kawabata, S. and Niwa, M. (Eds), Objective Evaluation of Apparel Fabrics, Textile Machinery Society of Japan, Osaka, pp. 15-28. Kawabata, S. and Niwa, M. (1991), “Objective measurement of fabric mechanical property and quality, its application to textile & clothing manufacture”, International Journal of Clothing Science & Technology, Vol. 1, pp. 7-18. Lindberg, J., Waesterberg, L. and Svenson, R. (1960), “Wool fabrics as garment construction materials”, Journal of Textile Institute, Vol. 51, pp. T1475-93. Love, L. (1954), “Graphical relationships in cloth geometry for plain, twill and satin weaves”, Textile Research Journal, Vol. 24, pp. 1073-83. Muttagi, S.B. (2002), Artificial Neural Network Embedded Expert System for Woven Fabrics, PhD thesis, IIT, New Delhi. Peirce, F.T. (1937), “The geometry of cloth structure”, Journal. of Textile Institute, Vol. 28, pp. T45-T112. Corresponding author B.K. Behera can be contacted at [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Characterization of commercial clothes dryers based on energy-efficiency analysis W.M. To, Y.W. Yu, T.M. Lai and S.P. Li Macao Polytechnic Institute, Macao SAR, People’s Republic of China

Characterization of commercial clothes dryers 277 Received March 2007 Accepted June 2007

Abstract Purpose – The purpose of this paper is to investigate the effectiveness of commercial clothes dryers based on energy-efficiency and environmental impact analyses. Design/methodology/approach – The study consisted of reviewing and examining the energy consumption of clothes dryers. It adopted the US Department of Energy’s uniform test method for measuring the energy consumption of electric and gas-fired commercial clothes dryers. Time efficiency, the cost of drying clothes and the environmental impact caused by the operation of commercial clothes dryers were studied. Findings – Electric commercial clothes dryers were found to be the most efficiency one in terms of energy per kilogram loadings. However, gas-fired commercial clothes dryers were found to produce less environmental impact in global terms. Liquefied petroleum gas-fired commercial clothes dryer was the most cost effective and time efficient. Research limitations/implications – All tests were conducted using three types of testing loads, namely towels, jeans and thermal clothing. They exhibited homogeneity in the tests but did not reflect the real-life situations. Practical implications – The implications of this study include assessing the efficiency of commercial clothes dryers based on standard testing loads for product comparison and increasing the awareness of environmental performance in the laundry industry as well as the clothing industry. Originality/value – Energy-efficiency and environmental performance become critical factors in determining competitiveness in all industries. This paper presents a detailed investigation on these aspects and provides insights for the laundry industry and appliance manufacturers. Keywords Driers, Laundry equipment, Energy consumption, Environmental testing Paper type Research paper

1. Introduction As non-renewable energy sources become less abundant and oil price rises continuously, it becomes more important to reduce energy consumption in all domestic and commercial sectors. In the USA, the National Energy Information Centre (US Energy Information Administration, 2005) estimated that kitchen and laundry appliances accounted for about one-third of household electricity consumption. A countrywide survey (US Department of Energy, 2001) indicated that 61 millions, i.e. 57 per cent, of US households had clothes dryers in 2001. The vast majority of those clothes dryers were electric and they used up 66 billion kWh that contributed to 5.8 per cent of the overall household electricity consumption in that year. Nowadays, clothes dryers are found in about 70 per cent of homes in the US. In addition, there are over 30,000 commercial laundry facilities in the US, located in hotels, resorts, health

International Journal of Clothing Science and Technology Vol. 19 No. 5, 2007 pp. 277-290 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710819492

IJCST 19,5

278

service facilities, prisons, and stand-alone companies and consume a significant amount of energy resources. In Australia, the Australian Bureau of Statistics (ABS) has collected ownership data for clothes dryers since 1994. In 2005, it reported (ABS, 2005) that 55 per cent of households across Australia owned clothes dryers, up from 52 per cent in 1994. However, clothes dryers’ usage varied quite substantially with weather conditions having a significant impact on patterns of use. In the 2005 ABS report 21 per cent of the respondents indicated that they used clothes dryers at least once a week. 8 per cent of respondents indicated that they used clothes dryers at least once a month; one third indicated that usage depended upon weather conditions while one third indicated they rarely used clothes dryers. During the same period of time, the number of commercial clothes dryers increased dramatically as new casinos, hotels, resorts, university hostels and health service facilities were flourished all over Australia. An energy audit report provided by the Energy Management Association of New Zealand (EMA, 2002) indicated space heating was the largest end-use in a typical inn in New Zealand, accounting for about one-third of the total annual energy consumption. The second largest end-use was kitchen and the third was laundry. The energy demand of clothes dryers in laundry was estimated to be about 118,000 kWh/year of electricity, representing a 10 per cent of the total annual energy. The audit report also indicated that the night time electricity use was dominated by commercial clothes dryers in laundry. In Asian cities, people live in a congested environment and would not have sufficient spaces to place domestic clothes dryers. In Japan, the ownership of clothes dryers is about 20 per cent. The ownership of clothes dryers in other Asian cities is much lower. As a result, many people would send their clothes to commercial laundries for washing and drying, especially in more affluent cities such as Macao, Hong Kong, Singapore and cities in Japan. The situation is even worse in the summer time because of high humidity. In Macao, there are hundred of laundry facilities. Most of them are located in hotels and health service facilities. In Hong Kong, there were around 900 commercial laundries and thousands of commercial clothes dryers according to the statistical data published by the Hong Kong Census and Statistics Department (Hong Kong Productivity Council, 1998). Most commercial laundries used gas-fired clothes dryers because of the relatively low fuel costs and unpopularity of electric clothes dryers in the 90s. However, gas-fired clothes dryers require a vented system that may produce local environmental impact in its vicinity. In the past few years, many commercial laundries started switching to electric clothes dryers because electric clothes-dryers are much easier to install and claim to have high energy-efficiency, such as 3 per cent to 20 per cent energy saving vs gas dryers (China Light and Power, 2004). The objective of this study was to measure, analyse and report on the performance of electric and gas-fired clothes dryers in commercial laundries based on energy-efficiency and environmental performance. 2. Description of commercial clothes dryers and testing sites Clothes dryers all operate basically the same way. Once a clothes dryer is activated, a fan distributes the electrically/gas-fired heated air through a drum that rotates the clothes placed inside it. The heated air then infiltrates or is in contact with the clothes

to evaporate water. The drying process, which involves turbulent flow, forced convection heat transfer, and mass transfer with a phase change, is then taken place. The moist air leaves the dryer through a pipe at the rear/side of the unit and is vented to outdoor. Some electric clothes dryers have equipped with a condenser. The function of the condenser is to extract moisture from the air leaving the drum. The extracted water/condensate is held to be emptied manually, or pumped away to a drain. The dehumidified air is then reused in the dryer. However, most commercial clothes dryers would not equip with such a condenser and use vented systems. In order to provide, in situ evaluation of clothes dryers, four series of tests were conducted in two commercial laundries: one located at street level and another situated in an industrial building. The first laundry provides washing, drying, ironing, and storage services to customers who are mostly residents living in the vicinity. It is equipped with two washing machines, one spin-drying machine and two electric clothes dryers. The second laundry mainly provides services to restaurants, hotels, garment factories, and chain laundries that collect clothes from outlets all over Hong Kong. It is equipped with six washing machines, one spin-drying machine, eight gas-fired clothes dryers using town gas (consisting mainly of methane and hydrogen derived from naphtha) and two gas-fired clothes dryers using liquefied petroleum gas (LPG). The characteristics of those clothes dryers evaluated in this study are shown in Table I.

3. Testing procedures The testing procedure adopted in this study was similar to the US Department of Energy (US Department of Energy, 1981; Kao, 1998) uniform test method for measuring the energy consumption of clothes dryers. The test samples were three types of clothes newly purchased. They were: (1) about 21 pieces of large double-hemmed, cotton bath towels (70 £ 140 cm) having a total weight of around 7 kg; (2) about 12 pieces of jeans (size: 32-36) having a total weight of around 7 kg; and (3) about 30 pairs of man/woman’s thermal clothing (various sizes) that consist of blended fabric of 50 per cent cotton and 50 per cent polyester having a total weight of around 6 kg. Before each series of tests, the bone-dry weight of each test load was recorded. The load was dampened by agitating in water using a washing machine with a fixed amount (half-cup ¼ 100 ml) of commercial softener. The excessive water content was then extracted from the wet test load by spinning the load until the moisture content of the load is between 66.7 and 75 per cent of the bone-dry weight of the test load. The wet test load was transferred to the commercial clothes dryer in a random manner. The dryer’s timer was set to 30 min but the clothes dryer was operated for 13-27 min until the moisture content of the test load was close to the bone-dry weight of the test load. Each test was repeated three times to verify the repeatability and accuracy of the measured data. For electric clothes dryers, precision instruments were used to record electrical and physical parameters during the tests. They included:

Characterization of commercial clothes dryers 279

Electric clothes dryers

Electric clothes dryers

Gas-fired clothes dryers Gas-fired clothes dryers

1

2

3 13.6 kg

13.6 kg

13.5 kg

13.5 kg

Capacity (kg)

309

312

286

286

Volume (l)

Unidirectional

Bi-directional (revising drum) Unidirectional

Unidirectional

Rotation

Drum Main energy source

Time-controlled Liquefied petroleum gas

Time-controlled Electricity 3 phase (3 £ 32A) Time-controlled Electricity 3 phase (3 £ 32A) Time-controlled Town gas

Operational mode

280

4

Type

Table I. Characteristics of the commercial clothes dryers tested

Dryer no.

USA

USA

Sweden

Sweden

Country of origin

IJCST 19,5

.

.

.

.

a standard three-phase kilowatt-hour meter having a resolution of 0.001 kWh (clamped at power cable for electrical power consumption measurement); a MARK II thermo-hygrometer TRH-5S having a resolution of 0.18C and 0.1 per cent RH (placed in laundry for humidity measurements); a Lutron TM-903 dual channel thermometer TM-905 having a resolution of 0.18C (placed in laundry for monitoring ambient conditions); and a precision electronic balance with a weighing indicator AD-4326A having a resolution of 0.01 kg (used for weight measurements).

In addition to, the above instruments, Ricoh Elemex MPD-23A gas-meter with a resolution of 0.001 m3 was deployed to measure the town gas or LPG consumed by gas-fired clothes dryers. 4. Cost effectiveness and environmental performance The Macao Statistics and Census Services (2006) indicated that as at the third quarter of 2006 the average list prices of fuels were US$0.163 per kWh for electricity and US$1.266 per kg for bottled LPG. As 1kg of LPG has a calorific value of 45 MJ (or 12.5 kWh), one can determine that the fuel price for bottled LPG is US$0.101 per kWh. In Hong Kong, the fuel prices are similar. The average prices of fuels are 0.166 per kWh for electricity, 0.101 per kWh for bottled LPG, and 0.117 per kWh for town gas (all in US$). In characterizing the environmental performance of various commercial clothes dryers, it is necessary to understand the environmental impact caused by the generation of electricity and the combustion of gaseous fuels. In this study, the widely used Global Warming Potential (GWP) is employed. The GWP was first proposed by Lashof and Ahuja (1990) and later refined by Houghton et al. (1996, 2001) to compare the ability of each greenhouse gas to trap heat, i.e. infrared radiation, in the atmosphere relative to another gas. The definition of a GWP for a particular greenhouse gas is the ratio of heat trapped by one unit mass (mostly in 1 kg) of the greenhouse gas to that of one unit mass of carbon dioxide (1 kg of CO2) over a specified time period, say 100 years. As the major pollutants discharged from burning of fossil fuels including coal, natural gas, LPG and synthetic gases such as town gas are CO2, water vapour (H2O), nitrogen oxides (NOx), and sulphur dioxide (SO2), we normalize all these pollutants to CO2 equivalent which is the measurement unit of GWP as shown in Table II. In Macao, Companhia de Electricidade de Macau (CEM, 2005) showed that it generates 0.633 kg of CO2, 0.011 kg of NOx, and 0.008 kg of SO2 per kWh electricity. In Hong Kong, China Light and Power (2004) generates 0.871 kg of CO2, 0.002 kg of NOx, and 0.004 kg of SO2 per kWh electricity (Liao, 2006). Hong Kong Towngas, the company producing synthetic gas from naphtha, states that town gas is comprised of hydrogen (49 per cent), methane (28.5), CO2 (19.5 per cent), and carbon monoxide (3 per cent). By performing a stoichiometric combustion analysis, one can determine that in complete combustion 1 kg of town gas will produce 1.026 kg of CO2 and 5.05 kg of H2O. As the calorific value of town gas is 27.56 MJ/kg, one can determine that generation of 1 kWh-energy from burning of town gas will produce 0.134 kg of CO2 and 0.660 kg of H2O. LPG is comprised mainly of n-butane, propane, and i-butane. According to the recent study completed by Tsai et al. (2006), 98 per cent of the LPG sold in this region was saturated hydrocarbons, with n-butane (46.4 per cent), propane (26.0 per cent), and

Characterization of commercial clothes dryers 281

IJCST 19,5

i-butane (22.4 per cent) as the main constituents. After combustion, saturated hydrocarbons were converted to CO2 and H2O. By performing a stoichiometric combustion analysis, one can determine that burning of 1 kg of LPG produces 2.96 kg of CO2 and 1.55 kg of H2O. In terms of pollutants per unit of energy, the combustion of LPG produces 0.237 kg of CO2 and 0.125 kg of H2O per 1 kWh-energy. The environmental impacts caused by various fuels are summarized in Table II.

282 5. Test results and analyses The ambient conditions were steady during the tests conducted in the two commercial laundries. Although the room temperatures, 258C-278C, and values of relative humidity, 66 per cent-74 per cent, exceeded the recommended values of the DOE test procedure, they should have a minimal effect on the analysed results because the study focused on comparison between energy-efficiency and environmental performance of clothes dryers. Moreover, it should be noted that the DOE test procedure was established in a test room while our measurements were conducted in situ to evaluate actual performance of clothes dryers. Table III shows the measured and the calculated parameters of the performance tests undertaken for unidirectional electric commercial clothes dryers, Dryer No.1. The average room temperature and humidity for this series of tests were 27.38C and 65.5 per cent RH, respectively. The average energy required to dry one kilogram of moisture, i.e. energy effectiveness, and the average standardized cycle time were 1.179 kWh/kg and 25.71 min. for towels, 1.439 kWh/kg and 31.19 min. for jeans, and 1.285 kWh/kg and 24.21 min. for thermal clothing, respectively. The average cost of drying 1 kg of moisture and its associated environmental impact were US$ 0.196 (per kg moisture) and 1.027 kg CO2 equivalents (per kg moisture) for towels, US$ 0.239 (per kg moisture) and 1.253 kg CO2 equivalents (per kg moisture) for jeans, and US$ 0.213 (per kg moisture) and 1.120 kg CO2 equivalents (per kg moisture) for thermal clothing, respectively. The second, third, and four series of tests were carried on bi-directional electric clothes dryers, town gas-fired clothes dryers, and LPG-fired clothes dryers, respectively. As mentioned in Section 3, each setting was repeated three times to ensure the repeatability of the measured values. The calculated energy-effectiveness and the standardized cycle time of all tests were summarized in Table IV. The cost required to dry 1 kg of moisture and the environmental impact in terms of GWP per kg moisture removal were presented in Table V. Greenhouse gases (kg per kWh-energy) CO2 H2O NOx SO2 GWP(b) (kg CO2 equivalents per kWh-energy) 0.002 0.004 Electricity (Hong Kong) 0.871 a Town gas 0.134 0.660 – – Liquefied petroleum gas 0.237 0.125 – – Table II. Global warming potentials of different fuels

0.871 0.134 0.237

Notes: aData are not provided in Liao (2006); balthough H2O, NOx and SO2 have direct and/or indirect positive radiative forcing effect, they are neither long-lived nor well mixed in the atmosphere and normally ignored in the determination of global warming potentials (US Environmental Protection Agency, 2002)

27.4 65.6 7.00 12.22 7.38 5.6 – 26 74.6 5.4 4.84 0.864 1.157 0.186 5.35 1.310 24.82 0.192 1.008

1 27.3 65.0 7.00 12.25 7.46 5.7 – 27 75.0 6.6 4.79 0.840 1.190 0.177 5.50 1.273 26.04 0.198 1.036

Towels 3

27.1 66.0 7.00 12.09 7.34 5.65 – 27 72.7 4.9 4.75 0.841 1.189 0.176 5.50 1.274 26.26 0.197 1.036

2

0.848 1.179 0.180 5.45 1.286 25.71 0.196 1.027

27.3 65.5 7.00

Average 27.3 65.5 6.85 11.18 7.30 5.5 – 26 63.2 6.6 3.88 0.705 1.418 0.149 6.41 1.069 30.30 0.235 1.235

1 27.2 65.0 6.85 11.62 6.80 5.7 – 27 69.6 9.3 4.13 0.725 1.380 0.153 6.24 1.098 29.56 0.229 1.202

Jeans 3

27.3 66.0 6.85 10.54 6.92 5.5 – 27 53.9 1.0 3.62 0.658 1.519 0.134 6.87 0.997 33.72 0.252 1.323

2

0.696 1.439 0.145 6.51 1.055 31.19 0.239 1.253

27.3 65.5 6.85

Average 27.4 65.0 6.08 10.33 6.14 5.7 – 27 69.9 0 4.25 0.746 1.341 0.157 5.38 1.130 25.49 0.223 1.168

1 27.3 65.9 6.08 10.17 6.82 4.1 – 19 67.3 12.1 3.35 0.817 1.224 0.176 4.91 1.238 22.76 0.203 1.066

24 68.4 3.5 3.95 0.775 1.291 0.165 5.18 1.173 24.38 0.214 1.125

27.2 65.5 6.08 10.24 6.29 5.1

0.779 1.285 0.166 5.16 1.180 24.21 0.213 1.120

27.3 65.5 6.08

Thermal clothing 2 3 Average

Notes: aNormalized by a reduction of 66 per cent moisture content in accordance with the formula given in Federal Register 46:27326 (US Department of Energy, 1981); bnormalized by a reduction of 66 per cent moisture content for a hypothetical cycle

Room temperature (8C) Room humidity (per cent) Measured bone-dry weight (kg) Measured wet weight (kg) Measured dry weight (kg) Measured electricity (kWh) Measured gas-consumed (m3) Duration (min) Calculated Moisture content, wet (per cent) Moisture content (dry %) Moisture removed (kg) Moisture removal rate (kg/kWh) Energy effectiveness (kWh/kg) Evaporation rate in (kg/min) Per-cycle energya (kWh) Energy factor (kg/kWh) Standardized cycle timeb (min) Cost effectiveness (US$/kg) Environmental impact (GWP/kg)

Measured

Test No.

Characterization of commercial clothes dryers 283

Table III. The measured and calculated parameters for the unidirectional electric clothes dryers (Dryer No.1)

Table IV. The calculated energy-effectiveness and the standardized cycle time of all tests

LPG-fired dryer

Town gas-fired dryer

Bi-directional electric dryer

Electric dryer

284

1.157 1.189 1.190 1.236 1.244 1.264 2.140 2.210 2.190 1.620 1.520 1.530

25.83 26.28 26.05 26.61 27.37 27.08 26.52 28.36 26.74 22.21 20.67 21.94

1.418 1.519 1.380 1.446 1.441 1.425 2.117 2.288 2.500 1.719 1.764 1.524

31.26 33.50 29.31 30.25 30.41 29.85 25.61 30.41 31.10 24.62 24.19 24.00

1.360 1.224 1.291 1.333 1.348 1.356 2.180 2.328 2.342 1.571 1.567 1.469

26.13 23.01 24.06 24.73 25.15 24.81 23.76 24.86 26.74 15.80 17.40 15.82

Towels Jeans Thermal clothing Energy effectiveness Standardized cycle Energy effectiveness Standardized cycle Energy effectiveness Standardized cycle (kWh/kg) time (min) (kWh/kg) time (min) (kWh/kg) time (min)

IJCST 19,5

LPG-fired dryer

Town gas-fired dryer

Bi-directional electric dryer

Electric dryer

0.192 0.197 0.198 0.205 0.207 0.210 0.253 0.261 0.259 0.165 0.155 0.157

1.008 1.036 1.036 1.076 1.084 1.101 0.328 0.340 0.334 0.404 0.380 0.385

Towels Environmental impact, (kg CO2 Cost of drying clothes (US$ per kg equivalents per kg moisture) moisture) 0.235 0.252 0.229 0.240 0.239 0.237 0.250 0.271 0.296 0.176 0.181 0.157

Cost of drying clothes (US$ per kg moisture) 1.235 1.323 1.202 1.260 1.255 1.241 0.324 0.354 0.383 0.433 0.444 0.388

Environmental impact, (kg CO2 equivalents per kg moisture)

Jeans

0.223 0.203 0.214 0.221 0.224 0.225 0.262 0.275 0.277 0.161 0.161 0.151

1.168 1.066 1.125 1.161 1.174 1.181 0.340 0.356 0.364 0.393 0.393 0.369

Thermal clothing Environmental impact, (kg CO2 Cost of drying clothes (US$ per kg equivalents per kg moisture) moisture)

Characterization of commercial clothes dryers 285

Table V. The cost of drying clothes and the environmental impact in terms of kg moisture removal

IJCST 19,5

286

Figure 1 shows the energy-effectiveness of different dryers. One can observe that electric clothes dryers are the most effective dryers in terms of the energy required to dry one kilogram of moisture (as they have the lowest values). They were followed by LPG-fired clothes dryer and town gas-fired clothes dryer. A one-way ANOVA followed by Tukey’s multiple comparison post-tests were performed with the four clothes dryer types and values of the energy effectiveness as the dependent variable using SPSS. This one-way ANOVA demonstrated that there was significant difference (F ¼ 153.19, df ¼ 3.32, p , 0.001) between various types of clothes dryers. Multiple comparisons using Tukey’s Wholly Significantly Different (WSD) test revealed significant difference between the mean energy-effectiveness values of town gas-fired clothes dryer, LPG-fired clothes dryer and electric clothes dryers. However, the mean difference between the energy-effectiveness values of unidirectional and bi-directional electric dryers was not statistically significant at the 0.05 level. Figures 2-4 shows the standardized cycle time, the cost of drying clothes in terms of US$ per kg moisture removal, and the environmental impact using GWP as a measurement unit of commercial clothes dryers. Again, one-way ANOVA tests demonstrated that there were significant differences between various types of clothes dryers based on these three parameters (F ¼ 11.87, df ¼ 3.32, p , 0.001 for the standardized cycle time; F ¼ 72.77, df ¼ 3.32, p , 0.001 for cost; F ¼ 414.70, df ¼ 3.32, p , 0.001 for environmental impact). Multiple comparisons using Tukey’s WSD test revealed significant difference between the standardized cycle time of LPG-fired clothes dryer and that of other clothes dryers. In terms of the cost of drying clothes, Town gas-fired clothes dryer has the highest cost per kg moisture removal, followed by electric clothes dryers and LPG-fired clothes dryer. Multiple comparisons using Tukey’s WSD test also revealed significant difference between the mean GWP generated by the operation of electric clothes dryers and that generated by the 2.6

Energy Effectiveness (kWh/kg)

2.4

Figure 1. The calculated energy effectiveness of commercial clothes dryers

2.2

Towels Jeans Thermal clothing

2.0 1.8 1.6 1.4 1.2 1.0

Electric Dryer

Bi-directional Electric Dryer

Town Gas-Fired Dryer

LPG-Fired Dryer

Characterization of commercial clothes dryers

35.0

Standardized Cycle Time (min.)

30.0

25.0

287

20.0

15.0 Towels 10.0 Jeans 5.0

0.0

Thermal clothing

Electric Dryer

Bi-directional Electric Dryer

Town GasFired Dryer

LPG-Fired Dryer

Figure 2. The standardized cycle time of commercial clothes dryers

0.30

Cost (US$ per kg moisture removal)

0.25

0.20

0.15

0.10

Towels Jeans

0.05 Thermal clothing 0.00

Electric Dryer

Bi-directional Electric Dryer

Town GasFired Dryer

LPG-Fired Dryer

operation of gas-fired dryers. Table VI shows the homogenous subsets of those parameters. Although the four parameters, namely energy-effectiveness, the standardized cycle time, the cost of drying clothes and the environmental impact, of commercial

Figure 3. The cost of drying clothes in terms US dollars per kg moisture removal

IJCST 19,5

Figure 4. The environmental impact of drying clothes

GWP (kg CO2 equ. per kg moisture removal)

288

1.4

1.2

1.0

0.8

0.6 Towels 0.4 Jeans 0.2

Thermal clothing

0.0 Electric Dryer

Electric dryer mean (SD)

Table VI. Homogenous subsets for the analysed data

Energy effectiveness in kWh Cost per kg moisture removal in US$ Standardized cycle time in minutes Environmental impact in GWP

Bi-directional Electric Dryer

Town Gas-Fired Dryer

LPG-Fired Dryer

Dryer type Bi-directional electric Town gas-fired dryer mean (SD) dryer mean (SD)

LPG-fired dryer mean (SD)

1.30 (0.12)

1.34 (0.08)

2.26 (0.12)

1.59 (0.10)

0.22 (0.02)

0.22 (0.01)

0.27 (0.01)

0.16 (0.01)

27.04 (3.46)

27.49 (2.34)

27.32 (2.41)

20.46 (3.49)

1.13 (0.11)

1.17 (0.07)

0.35 (0.02)

0.40 (0.02)

clothes dryers have different measurement units, they have a common directional characteristic – the smaller the better. By normalizing all the mean calculated values to the ones of unidirectional commercial clothes dryer, Dryer No.1, we produce a radar diagram as shown in Figure 5. This figure shows LPG-fired clothes dryer having the smallest area based on this four attributes. Both gas-fired clothes dryers produce much less environmental impact in comparison with electric clothes dryers in terms of GWP. However, those gas-fired clothes dryers produce unfavourable by-products, such as CO2 and H2O, affecting people living in the vicinity. Besides, electric clothes dryers can produce less environmental impact if local power companies switch from fossil fuels such as coal, natural gas and oil to renewable resources/non-fossil fuels as their primary energy sources.

Electric Dryer

Bi-directional Electric Dryer

Town Gas-fired Dryer

LPG-fired Dryer

Normalized energy effectiveness 2 1.6 1.2 0.8 0.4 0 Normalized cost of drying clothes Normalized standardized cycle time

Normalized environmental impact

6. Conclusion Traditionally people consider either the energy effectiveness or the fuel cost as a selection criterion for choosing equipment. In the clothing and laundry industries, clothes dryers play a significant role in a company’s energy consumption and costs. This work shows that electric commercial clothes dryers have the highest energy effectiveness in terms of kWh per kg of moisture removal. However, electricity is a secondary energy source and local power companies use traditional fossil fuels such as coal, natural gas and oil as their primary energy sources. Inevitably, the consumption of electricity would generate the highest global warming potential in comparison with the direct burning of town gas and LPG in commercial clothes dryers. When the fuel cost and the standardized cycle time are considered, the results indicate that LPG-fired commercial clothes dryer has its competitive edges because of its lowest fuel cost and shortest standardized cycle time, i.e. more batches per time unit. References ABS (2005), Environmental Issues: People’s Views and Practices, ABS Publication No.4602.0, Australian Bureau of Statistics, Canberra, March. CEM (2005), 2004 Safety, Health, Environmental and Quality – Statistics, Companhai de Electricidada de Macau, Macau, p. 2. China Light and Power (2004), Social and Environmental Report, CLP Power, Kowloon. EMA (2002), Bloggsville Motor Inn/Restaurant Energy Audit for 2002, Energy Management Association of New Zealand, Wellington, available at: www.ema.org.nz/ Hong Kong Productivity Council (1998), Modification, Compliance, Checking and Surveys of Existing Dry-Cleaning Machines – Final Report, Environmental Management Division, Hong Kong Productivity Council, Kowloon. Houghton, J.T., Meiro Filho, L., Callander, B., Harris, N., Kattenberg, A. and Maskell, K. (1996), Climate Change 1995. The Science of Climate Change, Cambridge University Press, Cambridge.

Characterization of commercial clothes dryers 289

Figure 5. Attributes of commercial clothes dryers

IJCST 19,5

290

Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K. and Johnson, C.A. (2001), Climate Change 2001: The Scientific Basis, Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge. Kao, J.Y. (1998), “Energy test results of a conventional clothes dryer and a condenser clothes dryer”, Proceedings of the 49th International Appliance Technical Conference, Columbus, Ohio, USA, pp. 11-21. Lashof, D.A. and Ahuja, D.R. (1990), “The relative contributions of greenhouse gas emissions to global warming”, Nature, Vol. 344, pp. 529-31. Liao, S. (2006), “Air pollutants emitted from CLP power’s coal-fired generating units”, press release by the Secretary for the Environment, Transport and Works, the Government of Hong Kong Special Administrative Region, 29 March. Macao Statistics and Census Services (2006), Balance of Energy, 3rd Quarter, Macao Statistics and Census Services, December, p. 15. Tsai, W.Y., Chan, L.Y., Blake, D.R. and Chu, K.W. (2006), “Vehicular fuel composition and atmospheric emissions in South China: Hong Kong, Macau, Guangzhou, and Zhuhai”, Atmospheric Chemistry and Physics Discussions, Vol. 6, pp. 3687-707. US Department of Energy (1981), “Uniform test method for measuring the energy consumption of clothes dryers”, Federal Register, Vol. 46, p. 27326. US Department of Energy (2001), End-Use Consumption of Electricity 2001, US Department of Energy, Washington, DC, available at: www.eia.doe.gov/emeu/recs/recs2001/enduse2001/ enduse2001.html US Energy Information Administration (2005), U.S. Household Electricity Report, US Energy Information Administration, Washington, DC, available at: www.eia.doe.gov/emeu/reps/ enduse/er01_us.html US Environmental Protection Agency (2002), Greenhouse Gases and Global Warming Potential Values-Excerpt from the Inventory of US Greenhouse Emissions and Sinks: 1900-2000, EPA 430-R-02-003, US Greenhouse Gas Inventory Program, Office of Atmospheric Programs, Washington, DC, April, pp. 5-9. Corresponding author W.M. To can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Application of artificial neural networks to the prediction of sewing performance of fabrics Patrick C.L. Hui Institute of Textiles and Clothing, The Hong Kong Polytechnic University, Kowloon, Hong Kong

Application of artificial neural networks 291 Received 12 January 2006 Accepted 5 February 2007

Keith C.C. Chan Department of Computing, The Hong Kong Polytechnic University, Kowloon, Hong Kong, and

K.W. Yeung and Frency S.F. Ng Institute of Textiles and Clothing, The Hong Kong Polytechnic University, Kowloon, Hong Kong Abstract Purpose – This paper aims to investigate the use of artificial neural networks (ANN) to predict the sewing performance of fabrics. The purpose of this study is to verify the ANN techniques that could be emulated as human decision in the prediction of sewing performance of fabrics. Design/methodology/approach – In order to verify the ANN techniques that could be emulated as human decision in the prediction of sewing performance of fabrics, 109 data sets of fabrics were tested by using fabric assurance by simple testing system and the sewing performance of each fabric’s specimen was assessed by the domain experts. Of these 109 input-output data pairs, 94 were used to train the proposed backpropagation (BP) neural network for the prediction of the unknown sewing performance of a given fabric, and 15 were used to test the proposed BP neural network. Findings – After 10,000 iterations of training of BP neural network, the neural network converged to the minimum error level. The experimental results reveal the great potential of the proposed approach in predicting the sewing performance of fabrics for apparel production. Originality/value – Generally, the fabric’s performance in the manufacturing process is judged subjectively by the operators and/or their supervisors. Current methodologies of acquiring fabric property information and predicting fabric sewing performance are still incapable of providing a means for efficient planning and control for the sewing operation. Further, development of techniques to predict the sewing performance of fabric is essential for the current apparel production environment. In this paper, the use of ANN to predict the sewing performance of fabrics in garment manufacturing is investigated. Keywords Neural nets, Fabric testing Paper type Research paper

In apparel production, the sewing process is one of the critical processes in the determination of productivity and the quality of the finished garment. Consistency of sewing quality is essential if the apparel manufacturer is to satisfy the customer. Failings or variations in sewing quality may be caused by one or both of the following factors: mechanical factor (the sewing machine) and human factor (the operator).

International Journal of Clothing Science and Technology Vol. 19 No. 5, 2007 pp. 291-318 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710819500

IJCST 19,5

292

At the center of the sewing operation is the sewing machine, and an experienced operator is required to set up the sewing machine to properly sew each fabric type. Good sewing performance depends on the setting of sewing parameters and the skill of operators when handling the sewing materials. Fabric properties and the behaviour of the fabric during sewing operations determines the likelihood of sewing problems, and also indicates the degree of difficulty in processing and controlling the sewing operations. If the fabric type or properties change, the supervisor should detect this change and make decisions regarding the sewing skills or alter the sewing parameters to optimise seam quality. The prediction of fabric performance during sewing process (such as fabric stretch, distortion, pucker, needle damage and overfeeding) provide the basis for maintaining consistency of sewing quality among different types of fabrics, as well as maintaining the quality standard throughout the sewing up process. Generally, the fabric’s performance in the manufacturing process is judged subjectively by the operators and/or their supervisors. As the manufacturing industry moves towards smaller lots with greater product variability, the sewing operation becomes increasingly inefficient due to frequent trial and error alterations of sewing machine parameters to match fabric properties. In addition, subjective judgement of fabric quality and performance is unreliable and unrepeatable. The problem is exacerbated by the ever-increasing diversity of fabrics and clothing style, the need for quick response and automation, and the retirement and non-replacement of experienced personnel. In this paper, we investigate the use of artificial neural networks (ANN) to predict the sewing performance of fabrics in garment manufacturing. The purpose of this study is to verify the ANN techniques that could be emulated as human decision in the prediction of sewing performance of fabrics. Physical and mechanical properties of fabrics and its sewing performance The physical and mechanical properties of the fabrics used determine the method of production and the quality of the finished garment. The physical properties of fabrics are the static physical dimensions of a fabric, inclusive of fabric structure, fibre content, weight, thickness, yarn size, colour and finishes, etc. Physical properties often control the nature and extent of the fabric characteristics. An understanding of the properties and characteristics of the fabrics being sewn is a prerequisite to controlling apparel processes. For example, fabrics with dense structures usually create more resistance and more friction between the fabric and needle in the sewing operation, which may lead to needle damage on the fabrics being sewn. Puckering occurs easily in light-weight fabrics with dense structures. The mechanical properties of apparel fabrics are important from the point of view of stresses applied to fabrics in the making up process. They are of critical importance in the forming and sewing of flat pieces of fabrics into three-dimensional shaped garments. Some typical examples are provided below. Extensibility is a measure of the fabric’s ability to be stretched during handling and assembling which is highly related to the sewing operation. Both excessive and insufficient extensibility will cause problems in the sewing process (CSIRO Division of Wool Technology, 1989). If the extensibility of fabric is too high, it could be easily stretched during the sewing of unsupported seams. Therefore, special attention and

control have to be exercised by the operative. Generally, the greater the extensibility, the more difficulty will be experienced in sewing. Bending rigidity determines the fabric’s resistance to bending. Rigid fabrics have greater resistance when bent by an external force such as that encountered during fabric manipulation in sewing. Fabrics with very high bending rigidity values may lead to sewing and handling problems as they are too stiff to be manipulated and controlled. Fabric formability, which is defined on the basis of fabric extensibility and rigidity, can be used to predict the limit of overfeed before the fabric becomes buckled. Puckering or sleeve-setting problems occur easily in the sewing process if the fabric being sewn has a very low formability value. The lower the formability the more likelihood of seam pucker (Ly and De Beos, 1990; Kawabata et al., 1990), because a fabric is unable to accommodate the degree of compression applied by the sewing thread. Fabric shear is another major mechanical property that determines the degree to which the fabric may be stretched and sheared when it is conformed to the intended garment shape. If the shear rigidity is too low, then the fabric is easily distorted and will skew or bow during handling and sewing. Fabrics with very high-shear rigidity tend to be difficult to shape and mould during sewing operations. The sewing performance of a particular fabric is mainly judged subjectively by the operatives and/or their supervisors. According to opinions of the sewing experts consulted on this topic, sewing quality is directly associated with the likelihood of sewing problem occurrence caused by the fabric being sewn. Therefore, the sewing performance of a fabric is classified according to four main factors: (1) pucker; (2) needle damage; (3) fabric distortion; and (4) overfeeding of the fabric during the sewing operation. Based on the practical experience of the operators and/or their supervisors in the sewing room, fabric performance and fabric behaviour during the sewing operation can be divided into four levels according to its value and the degree of control required in the particular sewing operations: (1) fabrics which could be processed without difficulty and for which no special control was required; (2) fabrics which could be processed with a little difficulty and some control was required; (3) fabrics which were considered to have poor performance in processing and needed to be handled with care; and (4) fabrics which caused serious problems in processing and for which manual assistance was extremely important. This classification scheme resulted from subjective judgement on the part of the operators and/or their supervisors and was customarily used as a reference to determine fabric sewing performance during production.

Application of artificial neural networks 293

IJCST 19,5

294

Review of methods used to solve the problem There are many means by which information on fabric properties can be obtained. Some of the most popular methods include the HATRA sewability tester, the L&M sewability tester, the Kawabata testing system (KES-F), and the fabric assurance by simple testing (FAST) system. The HATRA sewability tester was designed to measure the temperature of the sewing needle while the fabric is being sewn using an infrared sensor. An F-index is computed to represent the characteristics of frictional properties of the fabric (Roberts, 1981). The L&M sewability tester is a contact device that measures fabric sewability by using peak needle penetration force. The peak force is recorded for 100 needle penetrations at a speed of 100 stitches per minute. The fabric damage that may occur during apparel assembly is assessed by measuring the number of times the peak force exceeds a certain threshold. Both KES-F and FAST are fabric objective measurement systems which have been adopted for the prediction of fabric performance in apparel manufacturing. KES-F, a method of fabric objective hand evaluation, consists of four testers measuring respective basic fabric mechanical properties, i.e. tensile and shear properties; bending properties; lateral compression, and surface properties. A series of correlation equations has been derived so as to provide an objective specification of fabric handle and they are related to sewing performance of some fabrics (Radhakrishnaiah and Jayaraman, 1990). The FAST system uses a dimensional stability test and three instruments: a compression meter, a bending meter, and an extension meter, for assessing aspects of the appearance, handle and performance properties of the fabrics (Australian Wool Corporation, 1992). All of the above-mentioned methods are off-line approaches to the acquisition of data about fabrics. The HATRA sewability tester’s main function is to act as a rapid screening test, and the L&M tester performs at only 100 stitches per minute which is far below the industrial speed of 4,500 stitches or more per minute. Both KES-F and FAST system are limited to the application of lightweight suiting or shirting fabrics. As the KES-F system is both sensitive and complex, it is normally used for research or academic purposes, rather than industrial applications. Little and Matthews (1989) developed a sewing dynamometer which is used to measure the fabric/machine interaction for the objective prediction of fabric sewing performance. All the forces during the fabric/machine interaction are captured and analysed in real time with a high-speed analog-to-digital converter equipped microprocessor. However, the data collection system is too slow for continuous online use. A correlation between fabric properties and a subjective assessment of seam pucker and sewing damage of fabric being sewn was recently discovered by Stylios and Sotomi (1994) and Stylios and Moore (1993). The fabric information is collected off-line, and the correlation is learnt by neural networks, albeit in a limited way. Barrett et al. (1996) developed an online fabric classification technique for the adoption of sewing parameters to improve stitch formation and seam quality during the assembly process. Both the fabric type and the number of plies being sewn were identified for measuring the real-time needle penetration force using the wavelet transform method. A wavelet-based neural network was used to learn the mapping of the significant wavelet coefficients of measured forces to the fabric/ply combination.

However, the speed of data acquisition used in this system is relatively slow, and if dedicated digital signal processing hardware is used to improve the speed of data acquisition, the cost of this hardware is relatively high. Recently, an “online adaptive control system” was developed by Rocha et al. (1996) to control the sewing operation. This system includes a sewing machine with sensors to understand the dynamics of sewing at the various interfaces – sewing thread/fabric, sewing thread/sewing machine, fabric/sewing machine. It was, however too costly because such a system requires dedicated hardware, software and special machinery. In addition, it is impractical in a real production environment. Meanwhile, Ramesh et al. (1995) used an ANN to predict yarn tensile properties based on yarn processing and material variables. Pynckels et al. (1995) used an ANN to determine the spinning performance of fibers based on fiber properties. Cheng and Adams (1995) predicted yarn strength according to fiber properties with ANN. All attempts have recently been made to apply ANN techniques to textile production but not apparel manufacturing. Although Gong and Chen (1999) have attempted to apply an ANN for predicting the performance of fabrics in garment manufacturing, his work focused on the effects of different ANN architectures on their training speed and prediction accuracy. As the number of fabrics used to train the ANNs in his study was small, their ability to generalize and provide very accurate predictions was limited. Moreover, they failed to verify the capability of ANN techniques on emulation of human decision. Current methodologies of acquiring fabric property information and predicting fabric sewing performance are still incapable of providing a means for efficient planning and control for the sewing operation. Further, development of techniques to predict the sewing performance of fabric is essential for the current apparel production environment.

Selection of an artificial neural network The selection of an ANN was based on the consideration of the problem structure according to the following: . The prediction of the sewing performance of a fabric is a state-type problem that determines the performance and behaviour of the fabric in the sewing process. . The knowledge acquisition is complex and time consuming because it includes many attributes for one set of data including both numerical information and human knowledge. A large amount of data is required to make the prediction reliable, but it is impossible to acquire a full range of knowledge for the prediction. . There are many possible solutions for predicting the sewing performance of a fabric. Thus, the solution space is unpredictable and it is difficult to reduce. . Since, the prediction of sewing performance of the fabric mainly relies on its physical and mechanical properties, the knowledge is represented in the numerical form directly. In this case, the result is generally induced from training data set without any noisy (or unwanted) data.

Application of artificial neural networks 295

IJCST 19,5

296

As a result, an ANN was selected to handle the problem because this learning method provides a robust approach to approximating real-valued, discrete- and vector-valued target functions, and it has the power to distinguish complex patterns (Pao, 1989). Experimental Of all the ANN models, the BP network developed by Rumelhart (Tou and Gonzalez, 1974) is the most commonly used learning technique. The BP neural network was adopted for the purpose of the experiment because it is appropriate for problems where the input/output relationships are nonlinear, and/or involve high-order correlation among the input variables. This learning algorithm uses a gradient – descent heuristic, which enables a network to improve its performance over time. In addition, it is capable of handling noisy data, and it can provide significant speed advantages in applications over the other neural network learning methods. Figure 1 shows the structure of a BP neural network. There are two kinds of elements in the network – the neuron node and the connection weight. A neuron node is the basic processing unit that has an activation function. Neuron nodes are arranged in a layered structure, and the neuron nodes in consecutive layers are fully connected by connection weights. The first layer is called the input layer, and the second and third are called hidden and output layers, respectively. A connection weight has a weighting value for the node. Connection weights are important because their values determine the behaviour of the BP network. The BP model comprises of a two-phase computation process: forward and backward. In the “forward” phase, the BP network receives the input data pattern and directly passes it onto the hidden layer. Each element of the hidden layer calculates an activation value by summing up the weighted inputs of each element of the input layer, and then by using an activation function such as sigmoid function, an activation value is transformed into the threshold output in each element of the hidden layer. Each element of the output layer is used to calculate an activation value by summing up the threshold outputs of each element of the hidden layer. An activation function is then used to calculate the network output. In the “backward” phase, the actual network output is compared with the target value by using the BP formulae. If a difference input layer

hidden layer H1

output layer

I1 I2

O1

I3

O1

I4 Ok

Figure 1. Structure of a BP neural network

Ii Wij

Wjk Hj

arises between the output of the network and the desired output, the gradient-descent algorithm is used to adjust the connected weights. Otherwise, there is no learning proceeded with. In summary, the operating process of the BP is given in the steps as follows: (1) Initialise the weights between layers to a small random number (e.g. between 2 0.05 and 0.05). (2) Select the learning parameters for a neural network (e.g. learning rate, learning count, error criterion, etc.). (3) Repeat the following procedure until learning counts or the error criterion has arrived: . calculate the output of each hidden node; . calculate the output of each output node; . calculate the error between target output and network output; . calculate the modified gradient for output node; . calculate the modified gradient for hidden node; . modify the weight between output layer and hidden layer; and . modify the weight between hidden layer and input layer. In this experiment, the input units were those fabric physical and mechanical properties that covered the most important fabric constructional and behavioural parameters, and such parameters are important to evaluate the sewing performance of a fabric. A total of 21 parameters such as the fabric content, weight, thickness, formability and extensibility were used for the input units (details are listed in Table I). The output units were the four control levels of sewing performance (Level 1-4, as described in the previous section) of each of the four main criteria: puckering, needle damage, distortion, and overfeeding. Since, the fabric content, weave structure, and the control levels of sewing performance are discrete variables, they were represented in the binary form. For example, if the composition of fabric is pure cotton, the value of the first input unit (X1) is 1; otherwise, the value of the first input unit (X1) is 0. For continuous input variables, the data was normalised within (0, 1) to avoid any “over-float” problem in computation. Output data were also normalised within (0, 1) to correspond with the output range of the binary sigmoid activation function (namely, f(x) ¼ 1/(1 þ exp2 x), which was used in the hidden and output layers. Determination of the number of hidden units and learning rate is important for the BP neural network model. Too few hidden units and too high a learning rate will result in poor learning performance. Too many hidden units and too low a learning rate will take unacceptable training time and the many derived weights cannot be reliably estimated from the available training data. As no proven algorithm can determine the optimum number of hidden units and learning rate, hence, the number of hidden units is set same as the number of input units, and the learning rate is 0.1, that is a general implementation. Data for a total of 109 fabrics were collected in order to train the model. The fabrics selected in this study ranged widely in content, weight, structure, and mechanical properties. The physical properties of each fabric were analysed in the testing laboratory of The Hong Kong Polytechnic University. All the fabrics were tested for their mechanical properties using the FAST system. The assessment of the control

Application of artificial neural networks 297

IJCST 19,5

298

Table I. Particulars of input and output units

Unit name Input unit X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16 X17 X18 X19 X20 X21 Output unit C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16

Particulars Composition of fabric – pure cotton (1:Yes;0:No) Composition of fabric – pure silk (1:Yes;0:No) Composition of fabric – pure linen (1:Yes;0:No) Composition of fabric – pure polyester (1:Yes;0:No) Composition of fabric – 55% polyester/45% cotton (1:Yes;0:No) Plain weave structure (1:Yes;0:No) Twill weave structure (1:Yes;0:No) Satin weave structure (1:Yes;0:No) Threads density in warpwise direction per cm Threads density in weftwise direction per cm Warp yarn count in Tex system Weft yarn count in Tex system Fabric weight in g/m2 Fabric thickness in mm Formability (warp) in %mm2 Formability (weft) in %mm2 Extensibility (warp) in percent Extensibility (weft) in percent Bending rigidity (warp) in mN.m Bending rigidity (weft)in mN.m Shear rigidity in N/m Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control Control

level level level level level level level level level level level level level level level level

1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

in seam puckering in seam puckering in seam puckering in seam puckering in needle damage in needle damage in needle damage in needle damage in distortion in distortion in distortion in distortion in overfeed in overfeed in overfeed in overfeed

level of the sewing performance of all the fabrics was judged subjectively by an assessment panel. The assessment panel comprises of eight experts. They all have over four years experience in apparel manufacturing. It includes a merchandising director, a merchandising manager, a factory manager and five supervisors of sewing room. Of these 109 input-output data pairs, 94 were used as training sets and 15 were used as testing sets. The particulars of all 109 fabrics are shown in Appendix 1. To describe the prediction process of the BP network designed in the experiment, neuron nodes in input, hidden, and output layers are denoted as Ii, Hj, and Ok, respectively. The connection weights between Ii and Hj are denoted as Wij, and the

connection weights between Hj and Ok are Wjk. Assume that the sewing performance of an unknown fabric, with 21 experimentally determined fabric properties pi ; 1 # i # 21; is going to be predicted according to four main criteria and at four different levels – ck ; 1 # k # 16: A neural network is then designed by setting the number of neuron nodes in the input layer at 21 and the number of neuron nodes in the output layer at 16. This neural network is shown in Figure 2. Each input node corresponds to a physical and mechanical properties of the fabric and each output node to a level of sewing performance of the fabric according to four main criteria (puckering, needle damage, distortion, and overfeeding). Suppose the connection weights of neural networks Wij and Wjk are obtained by the following training process. The prediction of performance level in which fabric being sewn can then be obtained by first calculating the values of hidden nodes Hj and output nodes Ok with the following formulae: ð1Þ I i ¼ pi H j ¼ f ðSi W ij I i 2 uj Þ

ð2Þ

Ok ¼ f ðSj W jk H j 2 uk Þ

ð3Þ

Application of artificial neural networks 299

where f ðxÞ ¼ 1=1 þ e 2x Þ is a sigmoid function as shown in Figure 3 and u is a bias to shift the sigmoid function along the horizontal axis. The value of Ok lies between 0 and 1, which represents the probability that the unknown fabric belongs to the level of sewing performance in the four main aspects ck. The decision rule for the prediction of p1

1

1

p2

2

2

p20

20

20

p21

21

21

1

c1

2

c2

15

c15

16

c16

Figure 2. Proposed BP neural network

f (x) 1.0

0.5

x q

Figure 3. Sigmoid function

IJCST 19,5

300

sewing performance in a particular aspect is that the highest likelihood among the four different levels acts as the predicted performance level. In the above equations, pi are the given input variables, Ok are the calculated outputs, and the parameters W ij ; W jk ; uj ; uk are available based on a training process. Hence, training is the process of adjusting these parameter values W ij ; W jk ; uj ; uk so that the network behaviour matches some desired behaviour. The BP formulae as below were used to adjust the connection weight and bias if there is a difference between the network output generated by equations (1)-(3). ð4Þ DW jk ¼ hdOk H j ; DuOk ¼ 2hdOk ; where

dOk ¼ ðO 0k 2 Ok ÞOk ð1 2 Ok Þ: W 0jk ¼ W jk þ DW jk DW jk ¼ hdHj I j ;

where

DuHj ¼ 2hdHi ;

ð5Þ ð6Þ

dHj ¼ ðSk dOk W jk ÞH j ð1 2 H j Þ: W 0ij ¼ W ij þ DW ij :

ð7Þ

Equation (4) adjusts DW ij of the weights Wjk between Hj and Ok by the errors between O0k and Ok. The new weight value W 0jk is equal to the original weight value plus the adjustment. Moreover, the errors between O0k and Ok will be propagated so as to adjust the weights Wij between Ii and Hj based on equations (6) and (7). (See reference CSIRO Division of Wool Technology (1989) for details of the derivations of equations (4)-(7)). Results and discussion For the measuring of the physical and mechanical properties of fabrics, five specimens were tested for each of the 109 fabrics and the mean values of each parameter were obtained. The result of the control level of the sewing performance for each of the 109 fabrics is the average of the individual ranking among the members of the assessment panel. Details of the results are shown in Appendix 2. A total of 94 sets of input-output data pairs were used to train the proposed BP neural network for the prediction of the unknown sewing performance of a given fabric. After 10,000 iterations of training of the neural network, it was found that the BP neural network converged to the minimum error level. The connection weights of Ii, Hj, and Ok were derived and are shown in Appendix 3. Given a set of connection weights learnt by a proposed BP algorithm, if there is a fabric of unknown sewing performance when the physical and mechanical properties of such fabric are available, the above neural network would predict the sewing performance of a fabric systematically. Validation For the evaluation of the accuracy and effectiveness of the proposed BP neural network model, a validation panel was formed to carry out a sewing test on 15 tested fabrics. The validation panel comprised of another eight garment experts who were working at either supervisory or operative levels in the sewing room. The sewing test of the 15 tested fabrics

was carried out using a single needle lockstitch machine (B.S. 301). The superimposed seam (B.S. 1.01.01) was selected for the experiment (Appendix 4 for the construction of stitch B.S.301 and seam B.S.1.01.01). Three sizes of sewing needle (Singer No. 9, 12, and 16) and three sizes of sewing thread (100 per cent Polyester in Ticket No. pp 180, 140 and 100) were used for the sewing test. The members of the validation panel were asked to determine the appropriate needle size and thread size for each of the 15 tested fabrics. They were also asked to adjust the stitch tension, the stitch density, and the pressure of the pressure foot for individual tested fabrics according to their experience. Each member of the validation panel was asked to rank the sewing performance of each tested fabric according to the amount of effort they spent in setting the optimal sewing conditions, as well as the degree of attention they needed to give to the fabric during the sewing process. The actual sewing performance of the fabrics determined by the validation panel was used as the basis for evaluation. With respect to the actual sewing performance of fabrics, the accuracy of prediction using the BP neural network model and human judgement could be determined, respectively. The same 15 sets of fabric data which comprised of the value of each input unit but of unknown sewing performance was fed into the BP neural network for testing. The value of each output unit generated by the trained neural network is listed in Appendix 5. For each selected factor (puckering, needle damage, distortion, and overfeeding), the control level of sewing performance of 15 tested fabrics could be interpreted according to these output values. The prediction of the sewing performance of tested fabrics was tested six times each with different samples of data for training and testing, and the results of these six runs are shown in Appendix 6. Table II shows the results of comparison between the prediction made by BP neural network, the prediction made by domain experts and the actual sewing performance of fabrics determined by the validation panel. The prediction made by the BP neural network and that judged subjectively by the domain experts of the assessment panel both had 12 out of 15 testing data matched exactly to the actual sewing performance determined by the validation panel. When the four main factors (puckering, needle damage, distortion, and overfeeding) of sewing performance were examined individually, it was observed that the prediction made by the BP neural network achieved higher accuracy for factors “puckering” and “needle damage”. Conversely, the prediction made by domain experts was relatively better for factors “distortion” and “overfeeding”. Conclusions The properties of the fabric used in apparel production determine the method of manufacturing and the quality of the finished garment. Maintaining the sewing standards of individual fabrics through prediction of fabric performance is vital to the achievement of consistency of sewing quality. The operators or supervisors of the sewing room must be able to determine changes in fabric properties. Appropriate adjustment in sewing parameters and/or sewing skill is essential in optimising sewing quality. The experiment focused on the introduction of a BP neural network model for the prediction of fabric sewing performance based on the physical and mechanical properties of fabrics. The BP neural networks were selected from amongst the most effective learning methods currently known because of advantages in handling pattern recognition and problem prediction. A three-layered BP neural network that consists of 21 input units, 21 hidden units, and 16 output units was developed. All input units are

Application of artificial neural networks 301

IJCST 19,5 Testing data set

302

Table II. Comparison of the results of fabric sewing performance predicted by BP neural network, predicted by expert’s assessment, and the actual sewing performance

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Accuracy of prediction with respect to the actual sewing performance Puckering Needle damage Distortion Overfeed Overall

Average predicted result of a BP neural network P 1 1 4 3 1 1 4 3 4 4 3 4 2 3 2 BP neural network’s prediction (percentage) 93 100 93 87 93

Assessment result of the Actual sewing domain experts performance N D O P N D O P N D O 1 1 4 4 1 1 4 1 3 3 3 3 2 3 2

1 1 4 3 1 3 2 1 4 3 4 4 2 2 4

1 1 4 4 1 3 4 1 4 3 3 4 2 3 2

1 1 4 3 1 1 4 1 3 4 3 4 2 3 2

1 1 4 4 1 1 3 1 2 3 3 3 2 3 2

1 1 4 3 1 3 2 1 3 3 4 4 2 2 4

1 1 4 4 1 3 3 1 3 3 3 4 2 3 2

1 1 4 3 1 1 4 2 4 4 3 4 2 3 2

1 1 4 4 1 1 4 1 3 3 3 3 2 3 2

1 1 4 3 1 3 2 1 3 3 4 4 2 2 4

1 1 4 4 1 3 3 1 3 3 3 4 2 3 2

Expert’s judgement (percentage) 87 87 100 100 93

Notes: P – control level in seam puckering; N – control level in needle damage; D – control level in distortion; O – control level in overfeed; 1 – fabrics which could be processed without difficulty and for which no special control was required; 2 – fabrics which could be processed with a little difficulty and some control was required; 3 – fabrics which were considered to have poor performance in processing and needed to be handled with care; and 4 – fabrics which caused serious problems in processing and for which manual assistance was extremely important)

the physical and mechanical properties of fabrics, i.e. the critical parameters of a fabric’s constructional and behavioural pattern. The output units of the model are the control levels of sewing performance in the areas of puckering, needle damages, distortion, and overfeeding. There were 94 full data sets of fabric entries in the training group and 15 sets of fabric data of unknown sewing performance entries in the test groups. The testing data set was used to evaluate the effectiveness of the trained BP neural network model. The evaluation of the model showed that the overall prediction accuracy of the developed BP model was at 93 per cent, that is same as the accuracy of prediction made by human assessment. The predicted values of most fabrics were found to be in good agreement with the results of sewing tests carried out by domain experts. When the prediction made by the BP neural network was compared to the human assessment, it was noted that the BP neural network was better in the areas of “puckering” and “needle damage” but the human judgements showed higher accuracy in the areas of “distortion”

and “overfeeding”. The experimental results reveal the great potential of the proposed approach in predicting the sewing performance of fabrics for apparel production. References Australian Wool Corporation (1992), FAST Instruction Manual, CSIRO Australia, Australian Wool Corporation, Sydney. Barrett, G.R., Clapp, T.G. and Titus, K.J. (1996), “An on-line fabric classification technique using wavelet-based neural network approach”, Textile Research Journal, Vol. 66 No. 8, pp. 521-8. CSIRO Division of Wool Technology (1989), The FAST System for the Objective Measurement of Fabric Properties – Operation, Interpretation and Applications, CSIRO, Sydney. Cheng, L. and Adams, D.L. (1995), “Yarn strength prediction using neural networks”, Textile Research Journal, Vol. 65, pp. 495-500. Gong, R.H. and Chen, Y. (1999), “Predicting the performance of fabrics in garment manufacturing with artificial neural networks”, Textile Research Journal, Vol. 69, pp. 477-82. Kawabata, S., Masako, N., Ito, K. and Nitta, M. (1990), “Application of objective measurement to clothing manufacture”, International Journal of Clothing Science & Technology, Vol. 2 Nos 3/4, pp. 18-31. Little, T.J. and Matthews, B.A. (1989), North Carolina State University, US Patent 4,869,187, September 26. Ly, N.G. and De Beos, A.G. (1990), “Application of the FAST system to the manufacture of fabrics and garments”, Wool Research Organization of New Zealand, Vol. 5, pp. 370-409. Pao, Y.H. (1989), Adaptive Pattern Recognition and Neural Network, Addison-Wesley, Reading, MA. Pynckels, F., Kiekens, P., Sette, S., Van Langenhore, L. and Impe, K. (1995), “Use of neural nets for determining the spinnability of fibers”, Journal of the Textile Institute, Vol. 86, pp. 425-37. Radhakrishnaiah, P. and Jayaraman, S. (1990), “Relationship between KES properties and sewing performance of difficult-to-sew woven fabrics”, Knitting Int., Vol. 97 No. 12, p. 99. Ramesh, M.C., Rajamanickam, R. and Jayarama, S. (1995), “The prediction of yarn tensile properties by using artificial neural networks”, Journal of the Textile Institute, Vol. 86, pp. 459-69. Roberts, Z. (1981), “Finding solutions to needle problems”, Bobbin, Vol. 22, pp. 180-7. Rocha, A.M., Lima, M.F., Ferreira, F.N. and Araujo, M.D. (1996), “Developments in automatic control of sewing parameters”, Textile Research Journal, Vol. 66 No. 4, pp. 251-6. Stylios, G. and Moore, R. (1993), “Seam Pucker prediction using neural computing”, Proceedings of 22nd Textile Research Symposium at Mt Fuji, pp. 233-8. Stylios, G. and Sotomi, O.J. (1994), “A neural network approach for the optimisation of the sewing process of wool and wool mixture fabrics”, Proceedings of A Collection of Theses for the First China International Wool Textile Conference, April 1, Vol. 2. Tou, J.T. and Gonzalez, R.C. (1974), Pattern Recognition Principles, Addison-Wesley, Reading, MA. Corresponding author Patrick C.L. Hui can be contacted at: [email protected] To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Application of artificial neural networks 303

IJCST 19,5

Appendix 1

Fabric sett Tex Weight Thickness (Warp) (Weft) Weave (Warp) (Weft) (g/m2) (mm)

Fabric no. Material

304

Table AI. Particulars of the 109 fabrics for this experiment

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

100% Polyester 100% Cotton 100% Silk 100% Silk 100% Silk 100% Silk 100% Cotton 100% Cotton 100% Nylon 100% Nylon 100% Cotton 45% Cotton 55% 100% Polyester 100% Polyester 100% Nylon 100% Cotton 100% Polyester 100% Silk 100% Silk 45% Cotton 55% 45% Cotton 55% 45% Cotton 55% 100% Cotton 100% Cotton 100% Silk 100% Linen 100% Linen 100% Linen 100% Linen 100% Linen 100% Linen 45% Cotton 55% 100% Linen 100% Linen 100% Linen 100% Cotton

Polyester

Polyester Polyester Polyester

Polyester

53 35 53 84 39 38 31 28 41 48 32 75 31 62 50 25 61 33 37 30 22 39 30 31 44 31 21 17 20 20 24 52 24 21 21 26

33 31 51 65 37 28 29 28 33 35 21 36 35 33 49 39 34 27 28 24 20 23 28 30 27 27 21 17 21 23 22 35 23 20 20 20

Satin Plain Satin Satin Plain Plain Plain Plain Satin Satin Plain Plain Plain Plain Satin Plain Plain Plain Plain Twill Plain Twill Plain Plain Plain Plain Plain Plain Plain Plain Plain Plain Plain Plain Plain Plain

7 20 6 4 3 9 16 13 7 8 16 6 8 10 8 18 10 12 14 40 32 9 16 20 12 12 42 66 35 28 28 47 28 18 28 30

19 30 10 4 3 12 20 13 7 16 20 11 4 10 16 22 10 14 20 42 36 20 18 20 22 14 42 66 28 28 40 20 28 37 45 37

110 187 89 63 24 73 118 75 54 100 106 86 38 99 122 149 97 81 111 240 158 87 109 137 120 78 188 236 130 129 150 119 142 116 156 175

0.21 0.51 0.16 0.19 0.07 0.11 0.26 0.41 0.05 0.13 0.25 0.2 0.1 0.24 0.16 0.45 0.25 0.16 0.19 0.47 0.31 0.16 0.18 0.36 0.2 0.18 0.31 0.38 0.24 0.24 0.33 0.19 0.33 0.22 0.36 0.343 (continued)

Fabric no. Material 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

100% Cotton 100% Cotton 100% Cotton 45%Polyester 55%Cotton 45%Polyester 55%Cotton 45%Polyester 55%Cotton 100% Cotton 100% Cotton 45%Cotton 55%Polyester 100% Cotton 100% Cotton 100% Cotton 100% Cotton 45%Cotton 55%Polyester 45%Cotton 55%Polyester 100% Cotton 100% Cotton 100% Cotton 100% Cotton 45%Cotton 55%Polyester 45%Cotton 55%Polyester 100% Cotton 100% Cotton 45%Cotton 55%Polyester 45%Cotton 55%Polyester 100% Polyester 100% Polyester 100% Polyester 100% Polyester 100% Linen 100% Linen 100% Linen 100% Linen 100% Silk 100% Silk 100% Silk 100% Silk

Fabric sett Tex Weight Thickness (Warp) (Weft) Weave (Warp) (Weft) (g/m2) (mm) 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 56 56 56 56 56 56 22 21 22 12 47 50 55 82

20 20 20 20 20 20 20 20 32 32 32 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 20 21 22 22 50 43 53 67

Plain Twill Twill Twill Plain Satin Satin Satin Satin Satin Satin Plain Plain Plain Satin Satin Satin Twill Twill Twill Twill Twill Twill Plain Plain Plain Plain Twill Twill Plain Plain Plain Plain Satin Satin Satin Satin

30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 17 17 17 17 17 17 18 45 45 35 10 8 6 6

59 59 37 18 18 18 37 59 18 18 15 15 18 18 18 18 15 15 18 18 18 18 15 15 18 18 11 11 18 39 45 45 28 8 10 8 4

227 220 167 126 128 125 161 209 147 145 132 136 148 143 140 143 134 137 145 145 145 149 138 116 125 123 105 131 121 146 174 193 152 91 88 79 80

0.429 0.543 0.424 0.324 0.243 0.391 0.591 0.737 0.63 0.54 0.5 0.41 0.43 0.41 0.6 0.54 0.48 0.42 0.43 0.38 0.51 0.4 0.46 0.22 0.21 0.2 0.18 0.22 0.25 0.34 0.35 0.35 0.3 0.19 0.18 0.17 0.21 (continued)

Application of artificial neural networks 305

Table AI.

IJCST 19,5

306

Table AI.

Fabric sett Tex Weight Thickness (Warp) (Weft) Weave (Warp) (Weft) (g/m2) (mm)

Fabric no. Material 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 2109

100% Silk 100% Silk 100% Silk 100% Cotton 100% Polyester 100% Polyester 100% Polyester 100% Polyester 100% Polyester 100% Polyester 100% Polyester 100% Polyester 100% Polyester 45% Cotton 55% 45% Cotton 55% 45% Cotton 55% 45% Cotton 55% 45% Cotton 55% 45% Cotton 55% 100% Linen 100% Linen 100% Linen 100% Linen 100% Silk 100% Silk 100% Cotton 100% Cotton 100% Cotton 100% Cotton 100% Silk 100% Silk 100% Polyester 100% Polyester 100% Polyester 45% Cotton 55% 45% Cotton 55%

Polyester Polyester Polyester Polyester Polyester Polyester

Polyester Polyester

78 43 37 35 46 43 50 45 46 50 58 58 35 33 31 27 28 21 25 29 27 24 12 78 42 32 28 50 31 42 35 31 61 56 41 26

57 39 34 28 37 33 41 34 38 43 31 39 34 25 28 24 24 18 24 27 24 20 18 60 34 27 26 41 31 23 26 33 35 30 21 30

Satin Plain Plain Plain Satin Satin Satin Satin Satin Satin Plain Plain Plain Twill Twill Twill Plain Plain Plain Plain Plain Plain Plain Satin Plain Plain Plain Plain Plain Plain Plain Plain Plain Twill Twill Satin

4 4 4 12 7 6 10 10 10 8 10 10 10 36 40 34 30 34 36 14 14 45 35 4 3 22 12 10 15 12 14 8 8 15 8 30

6 4 6 12 7 7 14 16 18 18 12 8 10 38 40 36 34 38 40 14 16 45 28 6 4 22 12 10 15 18 18 6 10 15 18 15

68 34 38 80 61 51 113 104 119 122 101 100 97 241 261 197 174 151 205 85 84 208 97 71 28 136 68 93 102 96 101 48 90 138 74 129

0.20 0.09 0.10 0.14 0.06 0.05 0.14 0.15 0.17 0.17 0.26 0.23 0.25 0.42 0.44 0.40 0.28 0.37 0.41 0.20 0.22 0.36 0.29 0.19 0.08 0.31 0.2 0.22 0.34 0.18 0.17 0.12 0.23 0.23 0.15 0.61

X1

Training Data Set 1 0 2 1 3 0 4 0 5 0 6 0 7 1 8 1 9 0 10 0 11 1 12 0 13 0 14 0 15 0 16 1 17 0 18 0 19 0 20 0 21 0 22 0 23 1 24 1 25 0 26 0 27 0 28 0 29 0 30 0 31 0

Fabric no.

0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0 0 0 0

X2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1

X3 1 0 0 0 0 0 0 0 1 1 0 0 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X4 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0

X5 0 1 0 0 1 1 1 1 0 0 1 1 1 1 0 1 1 1 1 0 1 0 1 1 1 1 1 1 1 1 1

X6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0

X7 1 0 1 1 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X8 53 35 53 84 39 38 31 28 41 48 32 75 31 62 50 25 61 33 37 30 22 39 30 31 44 31 21 17 20 20 24

X9 33 31 51 65 37 28 29 28 33 35 21 36 35 33 49 39 34 27 28 24 20 23 28 30 27 27 21 17 21 23 22

X10 7 20 6 4 3 9 16 13 7 8 16 6 8 10 8 18 10 12 14 40 32 9 16 20 12 12 42 66 35 28 28

X11 19 30 10 4 3 12 20 13 7 16 20 11 4 10 16 22 10 14 20 42 36 20 18 20 22 14 42 66 28 28 40

X12 110 187 89 63 24 73 118 75 54 100 106 86 38 99 122 149 97 81 111 240 158 87 109 137 120 78 188 236 130 129 150

X13 0.21 0.51 0.16 0.19 0.07 0.11 0.26 0.21 0.05 0.13 0.25 0.2 0.1 0.24 0.16 0.32 0.25 0.16 0.19 0.47 0.31 0.16 0.18 0.3 0.2 0.18 0.31 0.38 0.24 0.24 0.33

X14 0.21 2.4 0.59 0.39 0.09 0.17 0.19 0.1 0.05 0.08 0.24 0.29 0.13 0.1 0.15 0.18 0.08 0.34 0.18 0.34 0.23 0.1 0.14 0.28 0.24 0.13 0.2 0.45 0.18 0.33 0.24

X15 0.12 0.55 0.21 0.28 0.09 0.06 0.26 0.26 0.07 0.11 0.18 0.18 0.05 0.14 0.19 0.32 0.18 0.12 0.27 0.72 0.34 0.11 0.09 0.23 0.1 0.41 0.68 0.91 0.79 0.71 0.67

X16 1 3.3 2.9 3.2 0.4 5.1 1.5 2.4 1.2 0.5 2.6 2 0.4 1.9 0.5 2.2 2.5 4.4 2.2 1.4 1.7 1.3 2.3 2.9 3.2 1 0.8 1 1.4 0.9 0.8

X17 0.8 2 3.1 5.6 1.5 1.2 2.7 6.6 2.2 1.1 4.7 2.8 1.6 4.7 2 7 6.5 3 3.1 3.8 2.7 2 3.1 3.1 3.2 4.9 5.2 7.8 6.3 5.1 5

X18 8.7 50.3 7.5 5.7 12.9 2.1 17.2 3 4.1 11.8 5.9 8.7 18.9 2.2 16.2 6 1.6 3.6 5.3 15.1 9.2 6.2 6.7 9.5 4.7 9.9 30 33 13.4 24.1 11.8

X19

X21

20.3 35 22.1 61 3.9 23 2.3 14 4.6 20 5 20 8 62 2.5 15 1.8 80 9.3 75 2.4 21 2.9 28 2.3 35 1.6 19 5.5 56 3.7 36 1.4 14 2.4 24 7 63 10.9 43 7.1 28 4.1 20 3.3 58 6.9 80 2 21 5.8 18 12.6 47 13.4 50 7.3 30 8.7 28 9.8 26 (continued)

X20

Appendix 2

Application of artificial neural networks 307

Table AII. Details of training data set and test data set

0 0 0 0 1 1 1 1 0 0 0 1 1 0 1 1 1 1 0 0 1 1 1 1 0 0 1 1 0 0 0

32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62

Table AII.

X1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X2 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1

X4 1 0 0 0 0 0 0 0 1 1 1 0 0 1 0 0 0 0 1 1 0 0 0 0 1 1 0 0 1 1 0

X5 1 1 1 1 1 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 1 1 1

X6 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0

X7 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0

X8 52 24 21 21 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 56 56 56

X9 35 23 20 20 20 20 20 20 20 20 20 20 20 32 32 32 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30

X10 47 28 18 28 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 17 17 17

X11 20 28 37 45 37 59 59 37 18 18 18 37 59 18 18 15 15 18 18 18 18 15 15 18 18 18 18 15 15 18 18

X12 119 142 116 156 175 227 220 167 126 128 125 161 209 147 145 132 136 148 143 140 143 134 137 145 145 145 149 138 116 125 123

X13 0.19 0.33 0.22 0.36 0.34 0.43 0.54 0.42 0.32 0.24 0.39 0.59 0.74 0.63 0.54 0.5 0.41 0.43 0.41 0.6 0.54 0.48 0.42 0.43 0.38 0.51 0.4 0.46 0.22 0.21 0.2

X14 0.24 0.3 0.19 0.28 1.09 0.9 1.73 1.32 0.5 0.5 0.49 0.57 0.71 0.54 0.43 0.38 0.39 0.52 0.42 0.53 0.53 0.49 0.5 0.55 0.54 0.49 0.59 0.54 0.27 0.28 0.22

X15 0.12 0.76 0.56 0.78 0.41 0.97 0.6 0.33 0.58 0.24 0.53 0.8 1.37 0.47 0.36 0.22 0.26 0.23 0.22 0.42 0.33 0.32 0.19 0.17 0.13 0.13 0.16 0.17 0.31 0.15 0.11

X16 2.7 2.5 1.5 1.4 1.4 1.4 1.3 1.3 1.2 1.2 1.2 1.3 1.4 1.3 1.2 1.3 1.5 1.4 1.4 1.4 1.3 1.2 1.3 1.3 1.4 1.2 1.4 1.2 1 0.9 1

X17 2.2 7.3 4.7 5.1 3.4 4.1 4.7 4.4 3.8 3.2 3.3 3.6 4.1 2.3 2 1.2 2.2 2.3 2.3 2.6 2.5 2.3 2.4 2.5 2.5 2.9 3 2.6 2.2 2.3 2.3

X18 5.9 11.1 9.4 13.7 80 132 127 97.2 73.3 74.5 72.7 83.3 104 79.1 63.6 56.5 57.2 77 62.2 77.7 78.2 72.7 73.7 81.1 79.3 72.3 86.7 79.1 39.4 41.7 32.5

X19

308

Fabric no.

X21 8.7 123 10.1 38 10.3 30 11.4 37 30.2 95 143 137 88.6 62 24.1 38 17.1 32 17.5 62 13.1 14 16.8 16 40.3 19 11.5 25 10.5 23 6.4 21 6.3 34 11.2 41 10.8 44 10.2 23 9.6 21 5.9 20 5.5 34 8.3 44 6.4 43 6.5 40 7.8 41 6.1 36 7.6 88 11.3 103 8.3 154 (continued)

X20

IJCST 19,5

63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94

Fabric no.

X2

0 0 0 0 0 0 0 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1

X3 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0

X4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0

X5 1 0 0 1 1 1 1 0 0 0 0 0 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 1

X6 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0

X7 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0

X8 56 56 56 22 21 22 12 47 50 55 82 78 43 37 35 46 43 50 45 46 50 58 58 35 33 31 27 28 21 25 29 27

X9 30 30 30 20 21 22 22 50 43 53 67 57 39 34 28 37 33 41 34 38 43 31 39 34 25 28 24 24 18 24 27 24

X10 17 17 17 18 45 45 35 10 8 6 6 4 4 4 12 7 6 10 10 10 8 10 10 10 36 40 34 30 34 36 14 14

X11 11 11 18 39 45 45 28 8 10 8 4 6 4 6 12 7 7 14 16 18 18 12 8 10 38 40 36 34 38 40 14 16

X12 105 131 121 146 174 193 152 91 88 79 80 68 34 38 80 61 51 113 104 119 122 101 100 97 241 261 197 174 151 205 85 84

X13 0.18 0.22 0.25 0.34 0.35 0.35 0.3 0.19 0.18 0.17 0.21 0.2 0.09 0.1 0.14 0.06 0.05 0.14 0.15 0.17 0.17 0.26 0.23 0.25 0.42 0.44 0.4 0.28 0.37 0.41 0.2 0.22

X14 0.18 0.31 0.28 0.17 0.18 0.2 0.18 0.4 0.45 0.38 0.52 0.34 0.14 0.14 0.28 0.07 0.05 0.1 0.08 0.11 0.11 0.07 0.05 0.36 0.47 0.45 0.28 0.3 0.2 0.44 0.15 0.2

X15 0.07 0.13 0.13 0.72 0.77 0.68 0.69 0.23 0.23 0.29 0.31 0.26 0.16 0.25 0.07 0.1 0.08 0.15 0.13 0.16 0.18 0.16 0.15 0.52 0.63 0.81 0.42 0.39 0.35 0.56 0.44 0.4

X16 0.9 1.1 1 1.7 1.1 1 1.2 3.1 3.3 2.8 3 3.6 0.5 0.8 5.5 1 1.5 0.5 0.6 0.6 0.8 2.1 2 0.5 1.6 1.5 2 1.5 1.9 1.6 1.2 1

X17 2.4 2.5 2.6 4.9 5.4 5.2 6.2 3.4 3.6 3 4.8 6.1 1.3 1.8 2 1.9 2.3 1 1.4 2.1 2.4 5 4.5 1.5 4 3.8 4.2 2.4 2.8 1.3 4.6 4.8

X18 3.5 88 6.3 41 9.5 72 15.2 28 12.7 53 14.3 56 10.2 25 3.7 23 3.4 22 3.9 26 3 15 2.5 13 7.7 22 9.1 21 5.5 21 2.5 95 1.7 72 10.9 88 9.5 82 4.8 62 5.3 54 1.7 20 1.6 19 25.7 38 9.3 47 11.9 44 6.8 35 8.1 29 7.3 29 11.7 31 6.4 18 6.5 17 (continued)

26.4 22.6 20.9 8.5 26.8 29.6 13.4 6.6 7.3 7 7.7 5.6 21.3 20.8 3.7 5.1 3.4 14.9 12.1 15.9 16.1 1.6 1.6 52.2 13.8 16.7 8.2 11 9.7 16.1 11.2 10

X21

X20

X19

Application of artificial neural networks 309

Table AII.

C1 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 1 1 0 0 1 0 1 1 1 1 1 1 1

Fabric no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Table AII.

X1

C2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0

X2 C3 0 0 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0

X3 C4 1 0 1 1 0 1 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X4 C5 0 1 0 0 0 0 1 1 0 0 1 0 0 0 0 1 0 0 0 1 1 0 1 1 0 1 1 1 1 1 1 1

X5 C6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0

X6 C7 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0

X7 C8 1 0 1 1 1 1 0 0 1 1 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X8 C9 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 1 1 0 1 1 0 0 1 1 1 1 1 1

X9 C10 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 0 0 0 0 0 0

X10 C11 1 0 0 0 1 0 0 0 0 1 1 1 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0

X11 C12 0 0 1 1 0 1 0 0 1 0 0 0 1 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X12 C13 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 1 1 0 1 1 0 1 1 1 1 1 1 1

X13 C14 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0

X14 C15 1 0 0 0 0 0 0 0 0 1 0 0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0

X15 C16 0 0 1 1 1 1 0 0 1 0 0 0 0 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X16

X17

X18

X19

310

Fabric no.

X21

(continued)

X20

IJCST 19,5

33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

Fabric no.

X2

0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 1

X1

1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 1 1 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0

X3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0

X4 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 0 1 1 0 0 0 0 0 0

X5 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1 0 0 1 0 0 1 0 0 0 0 1

X6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0

X7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0

X8 1 1 1 1 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 1 1 1 1 1 1 1

X9 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 1 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 0

X10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X11 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0

X12 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 1 1 1 1 0 0 1

X13 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1 0 0 0 0 0 0 0 0 1 1 0

X14 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X15 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X16

X17

X18

X19

X21

(continued)

X20

Application of artificial neural networks 311

Table AII.

0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1

65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94

Table AII.

X1

1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X2 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0

X3 0 0 0 0 0 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0

X4 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1

X5 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X6 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0

X7 0 0 0 0 0 1 1 1 1 1 1 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0

X8 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0

X9 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1

X10 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0

X11 0 0 0 0 0 1 1 1 1 1 0 0 1 1 1 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0

X12 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1

X13 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

X14 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 1 0 0 0 0 0 0 0 0

X15 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0

X16

X17

X18

X19

312

Fabric no.

X21

(continued)

X20

IJCST 19,5

X3 X3 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 C3 0 0 0 1 0 0 0 0 0 1 1 0 0 1 0

Testing data set X1 X2 95 0 0 96 0 0 97 0 1 98 0 1 99 1 0 100 1 0 101 1 0 102 1 0 103 0 1 104 0 1 105 0 0 106 0 0 107 0 0 108 0 0 109 0 0 C1 C2 95 1 0 96 0 1 97 0 0 98 0 0 99 1 0 100 1 0 101 1 0 102 1 0 103 0 1 104 0 0 105 0 0 106 0 0 107 0 1 108 0 0 109 0 1

X1

X2

Fabric no. X4 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 C4 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0

X4 X5 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 C5 1 1 0 0 1 1 1 1 0 0 0 0 0 0 0

X5 X6 1 1 0 1 1 1 1 1 1 1 1 1 0 0 0 C6 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0

X6 X7 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 C7 0 0 0 0 0 0 0 0 0 1 1 1 0 1 1

X7 X8 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 C8 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0

X8 X9 24 12 78 42 32 28 50 31 42 35 31 61 56 41 26 C9 1 0 0 0 1 0 1 1 0 0 0 0 1 0 0

X9 X10 20 18 60 34 27 26 41 31 23 26 33 35 30 21 30 C10 0 1 0 0 0 1 0 0 1 0 0 0 0 1 0

X10 X11 45 35 4 3 22 12 10 15 12 14 8 8 15 8 30 C11 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0

X11 X12 45 28 6 4 22 12 10 15 18 18 6 10 15 18 15 C12 0 0 1 0 0 0 0 0 0 0 1 1 0 0 1

X12 X13 208 97 71 28 136 68 93 102 96 101 48 90 138 74 129 C13 1 1 0 0 1 0 1 1 0 0 0 0 1 0 0

X13 X14 0.36 0.29 0.19 0.08 0.31 0.2 0.22 0.34 0.18 0.17 0.12 0.23 0.23 0.15 0.61 C14 0 0 0 0 0 1 0 0 1 0 0 0 0 0 1

X14 X15 0.2 0.12 0.33 0.11 0.27 0.12 0.14 0.16 0.17 0.16 0.16 0.11 0.29 0.11 0.46 C15 0 0 0 0 0 0 0 0 0 1 1 0 0 1 0

X15 X16 0.7 0.49 0.27 0.12 0.29 0.3 0.14 0.19 0.07 0.28 0.08 0.15 0.18 0.11 0.36 C16 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0

X16 X17 1.1 1.4 3.1 0.4 2.5 2.7 3 3 3.3 2.3 0.5 2.2 1.2 1.4 1.4

X17 X18 5.1 6.5 5.3 1.3 3 6.1 3.5 2.7 3.8 3.5 1.8 4.4 2.5 2.3 2.8

X18

X20 X20 15.2 6.5 2.8 5.9 7 2.7 3.5 4.8 1.2 5.8 3 1.5 9 3.1 8.7

X19 X19 31.9 8.6 6 15.6 10 3 4.7 4.7 3.6 4.7 23.8 2.3 21.1 5.5 68.3

X21 59 23 15 21 88 14 44 88 20 56 38 19 45 19 23

X21

Application of artificial neural networks 313

Table AII.

IJCST 19,5

314

Table AIII. The result of the connection weight between input nodes, hidden nodes and output nodes after 10,000 iterations of training

Appendix 3

Application of artificial neural networks

Appendix 4

301 1

315 a

This stitch type is formed with two threads : one needle thread (1), and one bobbin thread (a), A loop of thread 1 is passed through the material from the needle side and is interlaced with thread a on the other side. Thread 1 is pulled back so that the interlacing comes midway between the surfaces of the material being sewn. This stich type is sometimes produced from a single thread, in which case the first stitch differs from subsequent stitches. A minimum of two stitches describes this stitch type.

Material configuration Configuration des matériaux

1.01

Location of needle penetration or passage Emplacement des points de pénétration de I'aiguille

Numerical designation Désignation numérique

1.01.01

Figure A1. Constructions of Stitch B.S. 301 and Seam B.S. 1.01.01

Table AV. The value of each output unit generated by the trained neural network

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

0.999986 1 0 0 0.995741 0.956161 0 0.674428 0 0.000047 0.000002 0 0.003485 0.000138 0.070658

C1

0.000047 0.000001 0.000072 0.000015 0.000002 0.000012 0 0.000003 0.802836 0.000036 0.003828 0.000616 0.979482 0.000112 0.99844

C2 0.000001 0.000001 0 0.999996 0.000001 0 0.000005 0 0 0.047984 0.996863 0 0.000036 0.999119 0

C3 0.000001 0.000005 0.999992 0.000239 0.004295 0.338285 0.999713 0.353463 0.999014 0.034453 0.007661 0.999998 0.000502 0.000007 0.000041

C4 1 1 0 0 0.999998 0.999996 0.259495 0.999992 0.000032 0.517866 0 0.000121 0.001555 0.007194 0

C5 0 0 0.000099 0.000031 0 0 0 0 0.025868 0 0.018044 0 0.16958 0.000016 1

C6 0.000012 0.010077 0.000019 0.013169 0 0.007233 0 0 0.035619 0.999676 0.995549 0.996891 0 0.998604 0

C7 0 0 0.999762 0.983204 0.000006 0 0.999994 0.00034 0.002263 0 0.011258 0.000428 0.001336 0 0

C8 0.999341 0.982611 0 0 1 0 0.00005 1 0 0.016437 0.000008 0.000021 0.977142 0 0.000228

C9 0.000006 0.004438 0 0.002554 0 0.109121 0 0 0 0 0.000055 0 0.000002 0.999975 0.000002

C10 0 0 0 1 0.002204 0 0.019804 0.035977 0 0.271961 0.037911 0 0.020342 0.010096 0

C11 0 0.000001 1 0 0 0.178667 0.000032 0 0.999922 0 0.838185 0.999997 0 0 0.855694

C12

1 0.999811 0 0.000029 0.979532 0.000083 0.000002 0.058863 0 0.000005 0.000383 0.000001 0.440851 0.003243 0.018882

C13

0.000001 0 0.00068 0.000019 0.000057 0 0.000006 0.000069 0.280005 0.000045 0.004324 0.000004 0.02384 0 0.999846

C14

0.000021 0.000805 0.000002 0.010943 0.000001 0.002949 0.000001 0.000039 0.002305 0.997255 0.994034 0.523854 0.023154 0.999676 0.000055

C15

316

Testing

0 0 0.999819 0.968798 0 0.000031 0.981099 0 0.36678 0.000006 0.000048 0.939057 0 0 0

C16

IJCST 19,5 Appendix 5

Appendix 6 Test 1st run 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 2nd run 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 3rd run 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 4th run 46 47 48 49 50 51 52 53 54 55 56 57 58

Puckering

Needle

Distort

Handling

1 2 4 3 1 1 1 1 2 3 3 4 2 3 2

1 1 4 4 1 1 1 1 2 3 3 3 2 3 2

1 2 4 3 1 2 1 1 2 3 4 4 2 2 4

1 1 4 4 1 2 1 1 2 3 3 4 2 3 2

1 1 4 3 1 1 4 4 4 4 3 4 2 3 2

1 1 4 4 1 1 4 1 3 1 3 3 2 3 2

1 1 4 3 1 2 1 1 4 1 4 4 2 2 4

1 1 4 4 1 4 4 1 4 3 3 4 1 3 2

1 1 4 3 1 1 4 4 4 3 3 4 2 3 2

1 1 4 4 1 1 4 1 3 3 3 3 2 3 2

1 1 4 3 1 4 4 1 4 4 4 4 2 2 4

1 1 4 4 1 3 4 1 4 3 3 3 2 3 2

1 1 4 3 1 1 4 4 4 4 3 4 2

1 1 4 4 1 1 4 1 3 3 3 3 2

1 1 4 3 1 2 1 1 4 1 4 4 1

1 1 4 4 1 4 4 3 4 3 3 3 2 (continued)

Application of artificial neural networks 317

Table AVI. The experimental results of the prediction of fabric performance by using the BP neural network

IJCST 19,5

318

Table AVI.

Test 59 60 5th run 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 6th run 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 Average 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 SD 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120

Puckering 3 2

Needle 3 2

Distort 2 4

Handling 3 2

1 1 4 3 1 1 4 4 4 4 3 4 2 3 2

1 1 4 4 1 1 4 1 2 3 3 3 2 3 2

1 1 4 3 1 2 4 1 4 4 4 4 2 2 4

1 1 4 4 1 4 4 1 4 3 3 3 2 3 2

1 1 4 3 1 1 4 1 4 3 3 4 2 3 2

1 1 4 4 1 1 4 1 3 3 3 3 2 3 2

1 1 4 3 1 4 3 1 4 3 4 4 1 2 4

1 1 4 4 1 3 4 1 4 3 3 4 1 3 2

1 1 4 3 1 1 4 3 4 4 3 4 2 3 2

1 1 4 4 1 1 4 1 3 3 3 3 2 3 2

1 1 4 3 1 3 2 1 4 3 4 4 2 2 4

1 1 4 4 1 3 4 1 4 3 3 4 2 3 2

0 0.4 0 0 0 0 1.2 1.5 0.8 0.5 0 0 0 0 0

0 0 0 0 0 0 1.2 0 0.5 0.8 0 0 0 0 0

0 0.4 0 0 0 1 1.5 0 0.8 1.4 0 0 0.5 0 0

0 0 0 0 0 0.8 1.2 0.8 0.8 0 0 0.5 0.5 0 0

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Anthropometric measurement of premature infants

Anthropometric measurement

Yi-lin Kwok Institute of Textiles and Clothing, The Hong Kong Polytechnic University, Hong Kong, China

Kar-yin Wong Department of Paediatrics, The Hong Kong Queen Mary Hospital, Hong Kong, China

319 Received April 2006 Revised March 2007 Accepted March 2007

Bo-an Ying Clothing and Art Design College, Xi’an Polytechnic University, Xi’an, Shaanxi, People’s Republic of China

Kit-lun Yick and Li Yi Institute of Textiles and Clothing, The Hong Kong Polytechnic University, Hong Kong, China, and

Yeung Chap-yung Department of Paediatrics and Adolescent Medicine, The University of Hong Kong, Hong Kong, China Abstract Purpose – The purpose of this paper is to present anthropometric measurements on 42 premature infants nursed in the neonatal intensive care unit of Queen Mary Hospital, Hong Kong. Design/methodology/approach – Birth information, including maturity, age, gender, birth weight and present weight, were recorded. About 13 body size measurements, including stature, hand girth, armscye girth, chest girth, arm length, max girth, abdomen girth, hand length, thigh girth, shoulder width, head to nape length, inside leg to heel length and foot length, were measured for each infant. Using these data, the body size distribution, the correlation between each body size measurement, and linear regressions of present weight and stature with other body size measurement were analyzed. Findings – It was found that present weight and stature of premature infants were the most desirable and significant size parameters for the development of a measurement chart for premature infants. Originality/value – The paper provides anthropometric measurement details of premature infants. Keywords Infants, Measurement, Hong Kong Paper type Research paper

1. Introduction A premature infant is defined as a subject born at or before 37 weeks of gestation. That is approximately one month or more before the estimated date of confinement. The research is supported by a grant from the Research Grant Council of the Hong Kong Special Administrative Region, China (No.: POLYU 5280/03E).

International Journal of Clothing Science and Technology Vol. 19 No. 5, 2007 pp. 319-333 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710819519

IJCST 19,5

320

The birth weights of these subjects are normally below 2,500 g, which is the smallest of the normal weights of full-term infants of 40 weeks gestation age (Whitelaw and Cooke, 1988; Redshaw et al., 1985; Lam and Tam, 1991). Normally, premature infants need the medical care in a neonatal intensive care unit (NICU) due to their immature physiological development, such as underdeveloped lungs, apnea, inability to maintain body temperature, inability to feed orally. Because they are born at an earlier stage in their gestational development, the birth weight of the neonatal premature infant is often precariously low, sometimes one-fifth to one-sixth the size of normal newborns (Bendel, 1992). However, infant clothing is generally designed for full-term infants or those whose body weight is more than 3 kg. Therefore, finding clothing in the proper sizes for premature infants is a big problem, as few manufactures address their needs. Currently many standards for the anthropometric measurement of human body have been established as the baseline for designing apparel for humans. For infants, BS 7231-2:1990, Body Measurements of Boys and Girls from Birth up to 16.9 Years (British Standards Institute, 1990), and ASTM D 4910-02, “Standard tables of body measurements for infant, sizes 0 to 24” (ASTM D4910-02, 2002), only involve term infants. However, for premature infants, there is currently no established standard for use in premature infant’s clothing design. To identify the syndrome of newborns through observation of abnormal body parts, proportion, and unusual features, related anthropometric measurements were conducted by Merlob et al. (1984) in 87 term and 111 preterm infants whose gestation ages were between 27 and 41 weeks. Each measurement of the infants was used to diagnose relevant syndromes. For the design of clothing for premature infants, there have been no published detailed anthropometric studies of premature infants. Normally, only paediatricans and anatomists take body measurements of premature infants, fetus and cadavers and find that the measurements are not easy to take accurately. The only measurements usually recorded are the weight, head circumference, chest girth and total length. Therefore, in order to develop a outfits which can fit different sizes of premature infants, the anthropometric measurements and their study should be conducted further. 2. Anthropometric measurement design This anthropometric measurement of premature infants is designed to obtain the body size and growth status of premature infants. The data collected from this work will be used for developing a measurement chart of premature infants. 2.1 Measurement items In this study, three types of information were measured and/or recorded. 2.1.1 Hospital information. The hospital information consists of the hospital number and bed number of the premature infants when they stay in the hospitals. 2.1.2 Birth information. General information about the birth consists of the maturity, age, gender, birth weight, and present weight. The details of these items are described as follows: . Maturity. The maturity of the premature infant expressed in gestation weeks and days. . Age. The premature infant’s age counted in days after birth.

. .

.

Gender. Male or female. Birth weight. The weight of the nude infant on a calibrated infant scale immediately after birth. Present weight. The weight of nude infant on a calibrated infant scale at the time of measuring.

2.1.3 Body size information. The immediate use for the body measurements was for the development of a textile nest for premature infants, and 13 physical parameters of body size of premature infants were selected after through discussions between textile product designers and paediatricians. The meaning of each parameter are described below and the positions of the measurement for each parameter are shown in Figures 1-13. . Stature. The distance from the top of the head to the soles of the feet while subject is lying down flat with legs extended and foot positioned at 1.57 rad (908 to the leg) (Figure 1). . Head girth. The maximum horizontal circumference of the head above the ears (Figure 2). . Armscye girth. The circumference around the armhole passing under the arm and over the highest shoulder point at the armscye intersection (Figure 3).

Anthropometric measurement

321

Figure 1. Stature

Figure 2. Head girth

IJCST 19,5

322

Figure 3. Armscye girth

Figure 4. Chest girth

Figure 5. Arm length

Anthropometric measurement

323

Figure 6. Max girth

Figure 7. Abdominal girth .

.

.

. .

Chest girth. The horizontal circumference around the body under the arms and across the fullest part of the chest apex, including the lower portion of the shoulder blades (Figure 4). Arm length. The distance from the top of the shoulder joint along the outside of the arm over the elbow to the prominent wrist bone, with the arm bent at 1.57 radius (908) and placed on the hip (Figure 5). Maximum girth. The horizontal circumference around the body between the chest and the abdomen (Figure 6). Abdominal girth. The maximum circumference of the abdomen (Figure 7). Hand length. The distance from the wrist line to the second finger (Figure 8).

IJCST 19,5

324

Figure 8. Hand length

Figure 9. Thigh girth .

Thigh girth. The maximum circumference of the upper leg close to the crotch (Figure 9).

.

Shoulder width. The distance from the right side of the intersection of the armscye with the shoulder line to the left side at the back (Figure 10).

.

Head to nape length. The distance from the head crown level to cervical level (Figure 11).

.

Inside leg to heel length. The distance from the crotch to the sole of the foot (Figure 12).

.

Foot length. The distance of longest length from toe to heel (Figure 13).

Anthropometric measurement

325

Figure 10. Shoulder width

Figure 11. Head to nape length

Figure 12. Inside leg to heel length

Figure 13. Foot length

2.2 Measurement tools The weights of the premature infants were obtained using a medical electronic digital weight balance, “Detecto,” which incorporated a tray onto which the baby was placed. After taking the weight measurement of each premature infant, the tray was disinfected. The body size measurements were taken using a disposable paper tape measure, 64 cm long, marked in centimeters and millimeters.

IJCST 19,5

326

2.3 Measurement methods Patient consent from their parents and/or guardians was obtained before taking measurements. The measurements were carried out by experienced medical staff members, and each measurement and the procedure used were determined after careful consideration of the condition of the premature infants. In taking the measurements, it was not possible to landmark the positions for taking the measurements on the premature infant’s body. This was because the infants’ skin is too delicate or too immature to have adhesive tapes stuck on their bodies or marked using a marking pen. The surface area of the bodies was extremely small; therefore, the premature infant was lying on a firm flat surface and no ease allowance was given. Two persons were always needed in taking the measurements. While the nurse was stabilizing the premature infant with one hand placed firmly on the trunk and the other hand held the completely extended knee, the paediatrician conducted the body measurements for the infant. It is inevitably that the procedures have to be taken cautiously since the subjects were all extremely delicate. Difficulties have further increased as it has to be done inside the incubators, considering the need to keep the babies warm. 3. Anthropometric measurement results and analysis In this work of study, body measurements were taken on 42 premature infants nursed in the NICU of Queen Mary Hospital, Hong Kong. The range, minimum, maximum, mean, standard deviation and coefficient of variation for each of the anthropometric measurements are summarized in Table I. Among these 42 premature infants, the range of the maturity (gestational age) was from 23 to 34 weeks, the age in days from 3 to 96 days, the birth weight from 0.512 to 1.940 kg, and the present weight varied from 0.640 to 2.140 kg.

Table I. A summary of the anthropometric measurement results obtained from the 42 premature infants

Maturity (weeks) Age (days) Birth weight (kg) Present weight (kg) Stature (cm) Head girth (cm) Armscye girth (cm) Chest girth (cm) Arm length (cm) Max girth (cm) Abdomen girth (cm) Hand length (cm) Thigh girth (cm) Shoulder width (cm) Head to nape length (cm) Inside leg to heel length (cm) Foot length (cm)

Range

Min.

Max.

Mean

SD

Coefficient of variation (CV) (percent)

10.4 93 1.43 1.50 15.00 9.00 4.80 9.70 11.40 9.60 12.80 3.10 6.90 6.10 3.70 5.60 2.60

23.6 3 0.51 0.64 30.50 23.20 6.00 19.60 9.70 21.40 19.50 3.20 7.40 8.50 8.00 11.40 4.70

34.0 96 1.94 2.14 45.50 32.20 10.80 29.30 21.10 31.00 32.30 6.30 14.30 14.60 11.70 17.00 7.30

29.9 24.5 1.22 1.37 39.37 28.64 8.76 25.02 13.46 26.37 24.97 5.25 11.61 11.88 9.47 14.70 6.19

2.85 21.57 0.47 0.39 3.63 2.35 1.31 2.39 2.74 2.29 2.66 0.67 1.87 1.41 0.83 1.47 0.67

9.5 88.0 38.2 28.3 9.2 8.2 15.0 9.6 20.4 8.7 10.7 12.8 16.1 11.9 8.9 10.0 10.8

From these results, it was found that although there was a large coefficient of variation in age (CV ¼ 88.0 percent), there was only a small CV in birth weight (CV ¼ 38.2 percent) and present weight (CV ¼ 28.3 percent). For the body size, the maximum CV is the arm length (CV ¼ 20.4 percent), the minimum CV is head girth (CV ¼ 8.2 percent). Therefore, the results indicated that the premature infants stayed in NICU change very little in size and weight comparing with the change of age in days.

Anthropometric measurement

327

3.1 Size distribution The distribution of the 13 body size measurements are shown in Figure 14. Data include the range of box (from 75 to 25 percentile), the extreme value and outlier of the measurements. The figure reveals that the median size for stature, chest girth, shoulder width and inside leg to heel length is relatively centered as symmetrically distributed. Nevertheless, the median size for head girth, armscye girth, abdomen girth and thigh girth falls toward the top of its box, indicating that the distribution is left skewed. The median size for arm length and max girth falls toward the bottom of its box, indicating that the distribution is right skewed. The range values of hand length, head to nape length and foot length are very narrow, indicating that the central 50 percent of the data value fall within a narrow range, which is more than two or three times less than the others. Amongst the 13 body measurements, the largest range of box is obtained from the stature measurement. Results also indicated that there is no significant variation amongst the 42 premature infants observed in this study in terms of stature, head girth, armscye girth, chest girth, max girth, thigh girth and shoulder width. One extreme value is found for inside leg to heel length, and one outlier is found for abdomen girth, hand length and head to 50.00

40.00

30.00 175 177 176 174 178 170 172

20.00

444

10.00 499 309

Foot length

Head to nape length

Inside leg to heel length

Thigh girth

Measurements

Shoulder width

Hand length

Abdomen girth

Max girth

Chest girth

Arm length

Head girth

Armscye girth

0.00 Stature

Length (cm)

291

Figure 14. Distribution of body size measurements

IJCST 19,5

nape length, while more extreme values and outliers are found for arm length. In this respect, except the arm length measurement, all of the body measurements taken in this work are normally distributed.

328

3.2 Correlation analysis In order to investigate the relationships between pairs of measurements, correlation analysis was carried out. Correlation coefficient between pairs of measurements were calculated based on the following formula (Kwok, 1992): P ðx 2 x Þð y 2 y Þ r ¼ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ P P ðx 2 x Þ2 £ ð y 2 y Þ2 where r is the correlation coefficient, x and y are the individual measurement; x and y are the means of the individual measurements for which the correlation coefficient is found. A correlation coefficient equal to 1 means there would be perfect linear correlation between the pair of measurements, whereas a correlation coefficient close to 0 indicates no linear correlation between the pair of measurements. In practice, perfect correlation is seldom achieved. In this analysis, a correlation coefficient larger than or equal to 0.4 was considered to indicate a significant correlation. Correlations were performed between all the parameters using the Statistical Package for Social Science to determine those giving the most consistent correlations and thus, would be suitable for developing a sizing system. The summary of correlation coefficients for maturity, age, birth weigh and present weight with other measurements are summarized in Table II. Maturity has good correlations with birth weight (r ¼ 0.83) and present weight (r ¼ 0.51), respectively, correlations are significant at the 0.01 level. In view of body size, maturity has good correlations with stature (r ¼ 0.659), head girth (r ¼ 0.561), hand length (r ¼ 0.576), inside to nape length (r ¼ 0.429) and foot length (r ¼ 0.688). Nevertheless, age is not significantly correlated with the body size measurements. Amongst the 42 premature infants, birth weight has good correlation with most of the body girth and length measurements (significant at the 0.01 level), expect the max girth and abdomen girth measurements. Present weight also has good correlations with all girth and length measurements (significant at the 0.01 level). The values of correlation coefficient between each pair of body size measurement are summarized in Table III. Results indicate that most of the body girth, length and width measurements correlate well with each other. Most of the correlations are significant at the 0.01 level (two-tailed) and only some of them at the 0.05 level. The correlations of stature, head girth, chest girth, thigh girth, shoulder width and foot length with the other body measurements are all significant at 0.01 level. Length measurements of stature and foot length have consistently good correlations with all of the other length measurements, whilst girth measurements of head girth, chest girth and thigh girth always give good correlations with others. Amongst the width measurements, the correlations observed from shoulder width are constantly good with others.

Maturity (week) Age (day) Birth weight (kg) Present weight (kg) Maturity (week) Age (day) Birth weight (kg) Present weight (kg) Stature (cm) Head girth (cm) Armscye girth (cm) Chest girth (cm) Arm length (cm) Max girth (cm) Abdomen girth (cm) Hand length (cm) Thigh girth (cm) Shoulder width (cm) Head to nape length (cm) Inside leg to heel length (cm) Foot length (cm)

1 20.773 * * 0.824 * * 0.514 * * 0.659 * * 0.561 * * 0.309 0.347 * 0.308 * 0.245 20.018 0.576 * * 0.251 0.364 * 0.256 0.429 * * 0.688 * *

20.773 * * 1 20.720 * * 20.099 20.382 * 20.076 0.045 0.024 20.130 0.165 0.348 * 20.421 * 0.082 20.028 20.118 20.206 20.364 *

0.824 * * 2 0.720 * * 1 0.681 * * 0.820 * * 0.667 * * 0.496 * * 0.567 * * 0.395 * * 0.391 0.194 0.743 * * 0.547 * * 0.549 * * 0.451 * * 0.604 * * 0.835 * *

0.514 * * 2 0.099 0.681 * * 1 0.817 * * 0.831 * * 0.639 * * 0.901 * * 0.578 * * 0.781 * * 0.713 * * 0.717 * * 0.836 * * 0.811 * * 0.510 * * 0.649 * * 0.851 * *

Notes: * *Correlation is significant at the 0.01 level (two-tailed); *correlation is significant at the 0.05 level (two-tailed)

It is noteworthy that the girth measurements correlate well with each other, partly with the lengths, but only with shoulder width. The length measurements also correlate well with each other and partly with the widths. 3.3 Linear regression analysis Based on the correlation results discussed above, stature, head girth, chest girth, thigh girth and foot length measurements were identified to be used for the construction of a measurement chart. Nevertheless, as concerning the practical clinical environment of taking body measurements, present weight and stature of premature infants would become the most desirable size parameters for developing a measurement chart for premature infants. Some of the major reasons include: . the stature is easy to measure, but the other measurements may be more difficult, such as the finding the exact position; . the medical staff are very familiar with the use of present weight measurement of premature infants to determine their size and the stature correlates very well with present weight; . body weight is regularly taken at hospitals so as to monitor the growth of premature infants; and . the medical doctors use body weight, which is commonly referred to for the growth of premature infant. In this respect, linear regression analysis has been carried out to build up the linear relationships of the present weight and stature with the other 13 size measurements, respectively. These regression equations could also used to estimate the other measurements based on the stature or present weight.

Anthropometric measurement

329

Table II. The correlation coefficients for maturity, age, birth weight and present weight

Table III. The correlation coefficients between each pair of body size measurement

0.746 * * 0.658 * * 0.428 * 0.708 * * 0.376 * 0.695 * * 0.881 * * 0.824 * * 0.695 * * 0.795 * * 0.471 * * 0.733 * *

0.413 * * 0.298

0.497 * * 0.479 * *

0.351 *

0.448 * * 0.653 * * 0.592 * 0.044

0.574 * *

0.772 * * 0.717 * * 0.625 * * 0.803 * * 0.415 * * 0.844 * *

0.404 * 0.457 * * 1 0.337 1 0.422 * 0.697 * *

0.632 * * 1 0.457 * * 0.916 * *

0.459 * * 0.571 * * 0.574 * * 0.806 * * 0.442 * * 0.859 * * 0.835 * * 0.731 * * 0.569 * 0.728 * * 0.475 * * 0.749 * * 0.763 * * 0.738 * * 0.735 * * 0.841 * * 0.485 * 0.787 * *

1 0.632 * * 0.404 * 0.607 *

0.574 * * 0.806 * * 0.442 * * 0.859 * *

0.668 * * 0.798 * * 0.320 * 0.804 * *

0.607 * 0.916 * * 0.337 1

Abdomen girth

0.574 * * 0.787 * * 0.476 * * 0.655 * *

Max girth 0.459 * * 0.571 * *

Arm length

1 0.795 * * 0.574 * * 0.787 * * 0.476 * * 0.655 * * 0.795 * * 1 0.668 * * 0.798 * * 0.320 * 0.804 * *

Chest girth

Thigh girth

Shoulder width

0.735 * * 0.841 * * 0.485 * 0.787 * *

0.625 * * 0.803 * * 0.415 * * 0.844 * *

0.622 * * 0.444 * * 0.695 * * 0.646 * * 0.627 * * 0.813 * * 0.736 * * 0.727 * *

0.374 *

0.680 * * 0.795 * * 1

0.422 * 0.697 * * 0.574 * * 1 0.658 * * 0.680 * * 0.658 * * 1 0.795 * *

0.569 * 0.728 * * 0.475 * * 0.749 * *

0.835 * * 0.763 * * 0.772 * * 0.731 * * 0.738 * * 0.717 * *

Hand length

0.341 * 0.539 * *

1

0.444 * *

0.351 * 0.374 * 0.622 * *

0.592 * * 0.413 * * 0.298 0.044

0.448 * * 0.653 * *

Head to nape length

Notes: * *Correlation is significant at the 0.01 level (two-tailed); *correlation is significant at the 0.05 level (two-tailed)

Stature Head girth Armscye girth Chest girth Arm length Max girth Abdomen girth Hand length Thigh girth Shoulder width Head to nape length Inside leg to heel length Foot length

Armscye girth

1 0.625 * *

0.341 *

0.627 * *

0.497 * * 0.695 * * 0.646 * *

0.428 * 0.708 * * 0.376 * 0.695 * *

0.746 * * 0.658 * *

Inside leg to heel length

330

Stature

Head girth

0.625 * * 1

0.539 * *

0.727 * *

0.479 * * 0.813 * * 0.736 * *

0.695 * * 0.795 * * 0.471 * * 0.733 * *

0.881 * * 0.824 * *

Foot length

IJCST 19,5

Assuming a linear relationship between the variables x (present weight or stature) and y (other dimension), the equation of the line would be as follows:

Anthropometric measurement

y ¼ a þ bx where a is a constant, b is the gradient. The linear regression analysis between the present weight and the respective 13 size measurements of premature infants are summarized in Table IV. Results reveal that the present weight has good linear relationships with all other size measurements (R . 0.57 and P , 0.001), except with the measurement of head to nape length (R ¼ 0.510, P , 0.002). The highest standardized coefficient is found in the linear relationship between foot length and present weight (R ¼ 0.851). The linear regression between stature and the remaining 12 size measurements are summarized in Table V. As shown, stature has good linear relationship with head girth, chest girth, hand length, thigh girth, shoulder width, inside leg to heel length and foot length (R . 0.770 and P , 0.001). The highest standardized coefficient is found in the linear relationship between foot length and stature (R ¼ 0.881). Measurement items Stature Head girth Armscye girth Chest girth Arm length Max girth Abdomen girth Hand length Thigh girth Shoulder width Head to nape length Inside leg to heel length Foot length

Measurement items Head girth Armscye girth Chest girth Arm length Max girth Abdomen girth Hand length Thigh girth Shoulder width Head to nape length Inside leg to heel length Foot length

a

b

R

Significance

SD

28.907 21.614 4.124 17.411 6.574 19.924 18.270 3.466 6.626 7.840 7.975 9.600 4.188

7.636 5.083 3.103 5.554 4.824 4.786 4.891 1.342 3.595 2.946 1.090 3.547 1.461

0.817 0.831 0.639 0.901 0.578 0.781 0.713 0.717 0.836 0.811 0.510 0.649 0.851

P , 0.001 P , 0.001 P , 0.001 P , 0.001 P , 0.001 P , 0.001 P , 0.001 P , 0.001 P , 0.001 P , 0.001 P , 0.002 P , 0.001 P , 0.001

2.962 1.952 1.350 2.154 1.871 1.793 1.897 0.477 1.564 1.143 0.423 1.376 0.567

a

b

R

Significance

SD

8.166 2 5.428 4.592 2 3.561 8.375 11.721 2 0.819 2 4.676 0.067 5.441 2 2.707 2 1.810

0.519 0.351 0.519 0.425 0.454 0.337 0.155 0.413 0.300 0.102 0.436 0.162

0.795 0.574 0.787 0.476 0.655 0.459 0.835 0.763 0.772 0.448 0.746 0.881

P , 0.001 P , 0.003 P , 0.001 P , 0.002 P , 0.002 P , 0.003 P , 0.001 P , 0.001 P , 0.001 P , 0.004 P , 0.001 P , 0.001

1.868 1.214 1.882 1.542 1.503 1.221 0.556 1.427 1.088 0.371 1.581 0.587

331

Table IV. Linear regression analysis of relationships between present weight and other body size measurements

Table V. Linear regression equation of relationships between stature and other body size measurements

IJCST 19,5

332

4. Applications of anthropometric results Over the past years, the numerical data of anthropometric measurements concerning size, shape and other physical characteristics of human beings have been used in the design context successfully. In this respect, the anthropometric results of premature infants obtained from this study could be applied to develop an appropriate sizing system for efficient design and construction of different outfits. For example, as designing a textile “nest” for premature infants, it is expected that the “nest” should provide a womb-like environment, and lead to physical comfort while the premature infants are lying inside an incubator or radiant warmer bed. The textile “nest” should give support to their limbs which normally exhibit little movement as the premature infant tends to lie in a more extended (open) position, making him/her prone to heat loss. To meet these requirements, the anthropometric measurement results are very important in the following aspects: . the design of the length of textile nest should be based on the stature measurements of premature infants; . the design of the width of textile nest should be based on the shoulder width and thigh girth measurement of premature infants; . the size levels of textile nest are determined by the distribution of stature of premature infants; . the design of head supporting in textile nest should be based on the head girth; . the design of trunk supporting and its adjusting range in textile nest should be based on chest girth, thigh girth and shoulder width; and . the design of legs supporting and its adjusting range in textile nest should be based on the inside leg to heel length and stature. 5. Conclusion Anthropometric measurements were taken on 42 premature infants nursed in the NICU of Queen Mary Hospital. For each of the participated infant, with the consent of their parents, the birth information and thirteen body size measurements were obtained. Based on the anthropometric measurement data, their size distribution, correlations between each pair of body size measurements, and the linear regression for present weight and stature with other body size measurements were analyzed. Upon these analysis, the present weight and stature of premature infants were found to be the most desirable and significant size parameters so as to develop a measurement chart for efficient design and construction of different outfits for premature infants. The anthropometric parameters, such as stature, shoulder width, thigh girth, head girth, etc. in relations to the design of a textile “nest” for premature infants were also discussed. References ASTM D4910-02 (2002), “Standard tables of body measurements for infant, sizes 0 to 24”, Annual Book of ASTM Standards, ASTM. Bendel, P. (1992), “Apreemie’s plea”, Sew News, pp. 17-18. British Standards Institute (1990), Body Measurements of Boys and Girls from Birth up to 16.9 Years, BS7231-2, British Standards Institute, West Conshohocken, PA.

Kwok, Y.L. (1992), The Design of Garment for Premature Infant to Wear in a Hospital Environment, University of Leeds, Leeds. Lam, D.J. and Tam, A.Y.C. (1991), A Guide for Parents with Newborn in the Neonatal Intensive Care Unit, University of Hong Kong, Hong Kong. Merlob, P., Sivan, Y. and Reisner, S.H. (1984), “Anthropometric measurements of the newborn infant”, Birth Defects, Original Article Series, Vol. 20 No. 7. Redshaw, M., Rivers, R. and Rosenblatt, D. (1985), Born Too Early: Special Care for Your Preterm Baby, Oxford University Press, New York, NY. Whitelaw, A. and Cooke, R.W.I. (1988), The Very Immature Infant: Less than 28 Weeks Gestation, Churchill Livingstone, Edinburgh. Corresponding author Yi-lin Kwok can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Anthropometric measurement

333

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,5

Prototype garment pattern flattening based on individual 3D virtual dummy

334

Yang Yunchu and Zhang Weiyuan Fashion Institute, Donghua University, Shanghai, China

Received February 2007 Accepted June 2007

Abstract Purpose – In order to mass-customize clothes, it is essential to create prototype pattern according to individual body shape. The purpose of this paper is to present a new method to generate prototype pattern based on individual three-dimensional (3D) virtual dummy for further study on apparel customization. Design/methodology/approach – The symmetrized preprocessing and convex hull method are employed to create a dress-like virtual dummy based on 3D body scanning data. The corresponding structure lines of 2D prototype pattern are defined on the 3D dummy in advance and 3D dummy surface (only half) is cut into ten zones. Based on the characteristics of each surface, further subdivision was made in each zone to create 3D wireframe of garment prototype by calculating the intersection curves between the dummy surface and local planners. Via flattening geometrically 3D wireframe of each zone, final pattern of the prototype is got. Moreover, during the course of flattening of each zone, define constrained lines in advance so as to ensure the position and direction of each cutting pattern beforehand. Findings – The paper finds that 2D cutting patterns of the prototype have been constructed from the computerized 3D dummy. The length of major structure lines for both 3D model and 2D cutting pattern remain the same. The seven out of ten of cutting patterns have area error within ^ 1 cm2 compared to 3D surface. Only two cutting have relatively larger error but controlled within 3 cm2. Originality/value – The most outstanding property of the method developed is the possibility of geometrical transformation of 3D surface to 2D pattern through constructing 3D wireframe of the prototype garment, with no need to define physical-mechanical properties of fabric used. The newly created 2D cutting patterns have the coincident construction and shape with conventional prototype and are of outstanding quality and preciseness. Keywords Computer aided design, Garment industry Paper type Research paper

International Journal of Clothing Science and Technology Vol. 19 No. 5, 2007 pp. 334-348 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710819528

1. Introduction In recent years, the clothing industry has changed due to the industry’s increased awareness of fashion and design. As the industry has changed from mass production to mass customization, it has needed to develop new production techniques that provide to individual preferences, particularly apparel fitting that must be incorporated into pattern-making systems. Therefore, the development of techniques for better fitting clothing for individual body shapes is continuing. However, while technology may enable mass customization efforts, the processes involved are far from automatic. A significant amount of “behind the scenes” effort is still required in order to provide fit of each garment that might be requested by individual customers.

Made-to-measure (MTM) apparel pattern systems are a vital part of mass customization because it can solve how to produce attractive and accurately fitting apparel according to customers’ needs (Turner, 1994; Istook, 2002). In the prior research, the authors have been outlined the technologies of digital pattern developing for MTM pattern systems (Yang et al., 2006). Digital pattern developing process for customized apparel has two paths. One path develops individual based on 2D CAD technology. One of the typical approaches, used in several commercial MTM pattern systems, is individual pattern altering according to traditional grading rules. It is the most practical method because of its simple theory, which pattern masters are familiar with. However, the basic pattern for altering is made based on one-dimension measurements of the human body and it is difficult to reflect three-dimensional (3D) body shape. In order to get over this disadvantage of 2D CAD technology, a lot of studies have been done to solve the problem of digital pattern making based on 3D CAD technology. Several 3D apparel CAD systems have been developed in some company, such as Gerber, Lectra and Pad. Another path is just to develop pattern through flattening directly from individual 3D apparel model. However, the practicability of these systems is not good and there are always some distortions in the 2D pattern flattened from 3D apparel model. It is difficult to construct the real 3D garment model in these systems, especially for the apparel with complex style. Therefore, many improvements should be done if the 3D CAD systems are applied in apparel mass customization. This ongoing research develops a new scheme to customize individual pattern according to customer’s body shape. Figure 1 shows the overview of this scheme. Considering the advantage and disadvantage of the 2D CAD and 3D CAD technology, the process of customizing individual apparel pattern was separated into two steps. The fist step is the mapping process from 3D virtual dummy to 2D patterns of individual prototype garment. In this step, 3D CAD technology shall be adopted in order that prototype pattern can reflect 3D body shape roundly. The second step is the altering process from individual prototype to the cutting patterns of customized garment with the given style, using 2D apparel CAD technology.

Prototype pattern database for different category garments Step 2

Knowledge database for pattern making

335

The final cutting patterns

2D prototype pattern generation by flattening the 3D wireframe

2D prototype pattern generation by flattening the 3D wireframe

3D wireframe of prototype garment

Feature points definition and structural curves construction

3D Virtual dummy

Step 1

Patterns generation of customized garment based on the prototype design theory

Prototype garment pattern flattening

Figure 1. The overview of the individual pattern customization

IJCST 19,5

336

The purpose of this paper is to solve the task of the first step, namely generating individual prototype garment. Our work is not going to replace the current industrial successful 2D pattern system, but try to promote its’ practicability through 3D CAD technology in order to satisfy the demand on making fit and individual garment pattern. 2. Relevant study investigations Surface flattening is applied in many applications, such as in the aircraft, shoe, and garment industries. Owing to its importance, in both theory and practice, research in surface flattening has been active for a number of years. According to the latest theoretic study of the surface flattening, two typical methods for surface fattening are outlined, namely geometry flattening and physical flattening: (1) Geometrical flattening. The geometrically based flattening techniques do not consider the physical properties of materials. The method maps all 3D patches of the surface in a 2D plane with some geometrical constraint conditions. Parida and Mudur (1993) presented the flattening approach of complex surface based on constraint satisfying. Shimada and Tada (1991) stated an approach to transform approximately an arbitrary curved surface into a plane using dynamic programming. Yang et al. (2001) introduced a method for making complex surface developable and developing it. A developable surface is obtained between two curves on the complex surface and approximates to complex surface. Using this method, 3D front garment basic pattern is made developable and developed into plane pattern. Petrak et al. (2006) presented a method to transforming the 3D cutting pattern contour segments of a garment constructed on a computer-generated body model into the 2D segments. A systematic series of complex spatial rotations and translations is used to transform all the 2D segments from the individual planes of projection into the chain-linking plane (x, z) of the segments into the contours of individual 2D cutting patterns. (2) Physical flattening. The physical-based technique usually represents cloth models as triangular or rectangular grids with points of finite mass at the intersections. The various forces and energies are considered in surface fattening. In general, the physical flattening process from 3D garment surface can be summarized by following procedures: . create the energy model for fabric distortion; . map all 3D patches into 2D plane with constraint; and . minimize the energy by diffusion. McCartney et al. (1999) flatten a triangulated surface by minimizing the strain energy in the 2D pattern. The 3D surface is first triangulated using Delaunay triangulation. Then the triangles are transformed onto a plane. Wang et al. (2003) improve McCartney’s algorithm by using a spring-mass system. The system guides the endpoints to approach better positions by the force of springs and the computational speed of the minimization is improved. The accuracy of the flattening can also be controlled by using the spring constant. Wang et al. (2005) presents a surface flattening approach based on fitting a woven-like mesh model on a 3D freeform surface. The fitting algorithm is based on tendon node mapping (TNM) and diagonal node mapping

(DNM), where TNM determines the position of a new node on the surface along the warp or weft direction and DNM locates a node along the diagonal direction. During the 3D fitting process, strain energy of the woven model is released by a diffusion process that minimizes the deformation between the resultant 2D pattern and the given surface. The flattening technology has been adopted to develop pattern by several commercial 3D garment CAD systems, such as Gerber, PAD and DressingSim. However, the practicability of these systems is not good and there are always some distortions in the 2D pattern flattened from 3D apparel model. It is difficult to construct the real garment model in these systems, especially for the apparel with complex style. Therefore, many improvements should be done if the 3D CAD systems are applied in apparel mass customization. In this paper, a new approach is developed to transform 3D surface of the prototype garment to 2D cutting pattern. The symmetrized preprocessing and convex hull method are employed to create a dress-like virtual dummy based on 3D body scanning data. Then the corresponding feature lines with 2D prototype pattern are defined manually using 3D cursor on the model in advance and cut right half bodice surface into ten zones. Further subdivision was made in each zone to create 3D wireframe of garment prototype by calculating the intersection curves between the surface and local planners. The final pattern of the prototype is got by flattening geometrically 3D wireframe of each zone. Details of the method are discussed in Section 3. 3. Method 3.1 Preparing of 3D virtual dummy By using 3D human body data measured by 3D scanning technology, it is possible to make personalized dummies for use in fitted pattern making. Therefore, computerized 3D body models of the human body have recently attracted considerable attention in the clothing industry. For instance, in order to express the curved body surface geometrically, many methods are developed such as cross section model, triangle polygon model, B spline model, super quadrics, the Delaunay triangulation and meta-balls (Cho et al., 2005). The most complex part of the body to be covered by garment is usually the bodice, of which the shapes differ from person to person and the surface curvature variations are large. Therefore, the automatic generation of bodice pattern was focused in this study, as it is very difficult to design accurate bodice patterns by flat pattern processes only. The raw data scanned by (TC)2 body scanner, can be output in a Virtual Reality Modeling Language file. Each body scan data is made up of seven segments: right leg, left leg, torso (including head), right arm, right hand, left arm, left hand. The data of each segment includes two matrixes. One matrix contains the coordinates of the 3D point cloud, another save the coordinate indexes and color values of the triangle patch. In this research, only the point matrix of torso part was used and the data of head was deleted in advance. In order to ensure that each cross-section of the 3D body data has a same common topological structure, the data of the cross-sections were preprocessed with three methods, including re-sampling, symmetrizing and convex hull calculating. Cubic B-spline technology was adopted to create surface model based on the 3D curve

Prototype garment pattern flattening 337

IJCST 19,5

338

network. The data preprocessing and surface modeling (Figure 2) can be outlined with the following steps: (1) Sort the data points in order to fit the needs of B-spline interpolation. (2) Fit horizontal B-spline curve along the cross-section points, and calculate the center point O of the bounding rectangle of the curve. (3) Re-sample each cross-section at uniform angular centered on point O and re-ordering the sample sequence. (4) Symmetrize the points with y axes as symmetry axis to ensure r1 ¼ r2, in order to get the symmetric torso model for 3D garment CAD system. (5) Repeat step 1 and step 2 to get new samples and center point of the cross-section again because the center point O is changed after carrying step 3. (6) Calculate the convex hull (Figure 3) of each cross-section points to create a dress-like 3D body model. The convex hull method is used to mimic the physical tape-measurement of the mannequin for the purpose of bust girth generation, as well as to simplify the body surface complicacy. (7) Build curve network of the torso according to the point matrix after preprocessed and calculate the control points of mesh, then reconstruct the torso with the B-spline surfaces. We may thus first start fitting each row data of the matrix using B-spline curves (Figure 4(a)) and then process each column of the result y

P1

P2 r2

r1 O

Figure 2. Preprocessing of each cross-section data

x

Notes: O is their center point of the bounding rectangle; P1, P2 are the sample point of the cross-section; ρ1 is the distance between the center point O and P1; ρ2 is the distance between the center point O and P2 P12 P13 P16

P14 P15

Figure 3. An example of the convex hulls calculating

P9 P0

P8 P7 P10 P5 P11

P6

P4

P3

P2 P1

Prototype garment pattern flattening 339

(a)

(c)

(b)

Figure 4. The procedure of 3D curve network construction, where (a) is horizontal curves; (b) is vertical curves; and (c) is 3D curve network

matrix (Figure 4(b)). A 3D network with smooth row curves and column curves should be created, as shown in Figure 4(c). A B-spline surface was generated based on the 3D curve network. Figure 4 shows the results of surface modeling, where (a) is the model without the convex hull calculating of each cross-section and (b) is the model with the convex hull calculating. 3.2 Wireframe constructing of the 3D prototype garment As the prototype developed by Japanese Bunka Women’s University is quite comprehensive in theory and applicable in practice, it prevails in both China and Japan. Their newly garment prototype, as shown in Figure 5, is made based on the bust girth and back length. Other detailed measurements of the prototype are calculated by the

(a)

(b)

Figure 5. The results of body surface modeling based on 3D curve network, where (a) is the model without the convex hull calculating of each cross-section; and (b) is the model with the convex hull calculating

IJCST 19,5

340

regression formulas between each of them and bust girth. In light of different individuals, the prototype made by this method maybe has some error because these regression formulas reflect the average relationship between detail measurements and bust girth. In order to tailor-make a personalized prototype, the author defines a set of feature structure lines based on 3D model with reference to the newly Bunka’s garment prototype (Figure 6). In this way, the individual prototype can reflect 3D shape of human body roundly, because the detailed measurements of the prototype can be got directly from the 3D model. Furthermore, it has the corresponding shape and structural feature with the newly Bunka’s prototype as well as its good practicability. The feature points on dummy are set manually using 3D cursor. The detail specifications are shown in Figure 7 and Table I. As for the feature lines, they can be generated through calculating the intersection curves of body surface and the local planes, which are built according to the coordinates and the normal vectors of the corresponding feature points. Regarding the intersection algorithm between NURBS surface and plane, it has undergone a lot of study in the field of computer graphics and implemented by several 3D graph software. Based on the feature structure lines on 3D dummy, the right half bodice’s surface has been divided into ten zones. The front part is inclusive of five zones (A, B, C, D, E) and the back inclusive of other five zones (F, G, H, I, J) as shown in Figure 7. In order to Β + 6.2 8 Β + 3.4 = 24

Β + 7.4 8 + 0.2 1.8

22°

1.5

0.5

B – 0.8 32

0.5

B + 8.3 5

B + 13.7 12 0.5 B –2.5 4 BP 0.7

+ 0.5 + 0.8

1 2 BL

2∼3

1.5

Figure 6. The new prototype pattern developed by Japanese Bunka Women’s University

WL f

+ 0.5

18°

e

d

c B +6 2

b

a

SNP

SNP

A CP1 B

BP1

BP3 BP2

D

D1

F

E

D2 SWP

G

H

SWP D4 D3 BWP

Feature points

Specification

BP FBP

Bust point The intersection point between the hull curve at the level of bust point and the central vertical plane of human body Shoulder point Side neck point, front neck point, back neck point Front waist point, Side waist point, Back waist point The armpit point of the prototype garment Front armpit, Back armpit The intersection point between the center front line and the planner, which crossing the point CP1 and have the orthogonal relation with the center front line The intersection point between the hull curve of the cross-section at the level the back armpit BP1 and the central vertical plane of the human body The mid-point of the center back line segment between BNP and BP1 The intersection point between armhole curves and the planner, which crossing the point BP3 and have orthogonal relation with the line BNP-BP1 Back protruding point on the curve of BP1-BP2 Mapping point of dart A on waist line Mapping point of dart B on waist line Mapping point of dart E on waist line Mapping point of dart D on waist line

SP SNP, FNP, BNP FWP, SWP, BWP XP CP1, BP1 CP2

BP2 BP3 BP4 BP5 D1 D2 D3 D4

341

XP

XP BP

C

FWP

I BP5

BP4

CP2

FBP

J

SP

SP

FNP

BNP

Prototype garment pattern flattening

Figure 7. The feature points and structure curves on dummy

Table I. The feature points’ specification of 3D prototype garment

IJCST 19,5

342

flattening the 2D cutting pattern, it needs to subdivide the surface on the ten regions of the dummy. The subdivision methods differ to the varied features on the each surface. The detailed way is listed as follows: . As to the zone A, the separating curve FNP-CP4 was defined firstly, through calculating the intersection of body surface and the local plane, which crossing the point FNP and keeping orthogonal relation with the line FNP-CP2. Here, the point CP4 is the intersection point between the armhole curve and the local plane. Then the curve SNP-CP5 was make by calculating the intersection of body surface and the local plane, which crossing the point FNP and keeping orthogonal relation with the line FNP-CP4, where CP5 is the intersection point. Hereby, we can take FNP-CP5 and SNP-CP5 as the front neck width and height of the prototype pattern, respectively. The surface FNP-CP4-CP1-CP2 on the chest and the surface FNP-CP5-SNP around the neck can be subdivided by horizontal strips at even distance. As for the surface SNP-CP5-CP4-SP, it can be subdivided by horizontal strips along the two path curve SNP-CP5 and CP4-SP at even distance. In addition, for the sake of reducing the gap when flattening the 2D pattern, the vertically separating curve CP6-SP2 was made, where CP6 and SP2 are the two mid-points of CP4-CP5 and SP-SNP. . As to the zone B, take point CP3 on the curve CP1-CP2, from which CP2-CP3 is of the same length as FBP-BP. Then make a vertical intersection line over both BP and CP3. Make vertical strip subdivision at even distance on the surface of FBP-BP-CP3-CP2 and horizontal strip subdivision at even distance on the surface of BP-CP1-CP3. . As the surface C, G, H is similar to that of the transforming cylinder, some horizontal intersection will be conducted to make subdivision so as to form a series of horizontal intersection line. . As to the surface of zone D, E, F, the part under the bust line will be subdivided by a series of horizontal section, forming the horizontal intersection line. The part above the bust line will be subdivided by vertical section, forming vertical intersection line. . As to the zone I, make vertical strip subdivision at even distance. . As to the zone J, the subdivision is made as that of zone A. First make orthogonal planner of BP3-BP4 over SP, forming intersection curve SNP-BP6 with dummy surface. As to the surface BNP-BP3-BP6-SNP, make horizontal and vertical segmentation, constructing a grid. As to the surface SNP-BP6-BP4-SP, make horizontal strip segmentation along the two paths of SNP-BP6 and BP4-SP at even distance. Then make vertical intersection line of BP7-SP2, where BP7 and SP2 are the mid-points of SNP-SP and BP4-BP6 separately. Figure 8 shows the final subdivision result of the right half part of the dummy’s body. 3.3 Generating of the 2D prototype pattern The flattening process from 3D surface to 2D pattern for apparel shall keep the following principles. First, the length of key structure lines shall be kept the same as much as possible. Second, the area of 3D surface and 2D planner of the prototype should be controlled within certain gap. Third, the shape of pattern and 3D model should be kept the same as much as possible.

SNP SP2

SP2 SP

FNP

Prototype garment pattern flattening

BNP

SNP SP

CP5 CP6 CP2

CP4

BP3 BP4

CP3

BP7

CP1 XP XP1

XP1

XP2

FBP BP

FWP

BP2

BP5 BP1 XP

343

BP6

BP8 XP3

BP9

SWP D1

D2

SWP

D4

D3

BWP

Based on the foregoing method, the wireframe of prototype has been created. Then will generate 2D cutting pattern of the prototype garment by flattening 3D wireframe geometrically in each zone. During the flattening of each zone, it is necessary to define the structure lines, whose direction and position is constrained, so as to preserve the shape of prototype pattern after flattening as the newly Bunka’s prototype. For example, both the center front line and center back line can keep vertical upon flattening. The back and front width line can keep horizontal upon flattening. The principle of flattening 3D wireframe in all zones is concluded as follows: . The central line of both the front and back of the prototype shall keep vertical upon flattening. The flattening of surfaces shall follow the procedure from center front line for the front pattern and center back line for the back pattern to side seam. All the surfaces shall have mapping from 3D to 2D in order to preserve the edges of all the unit cells at the same length. . As to the surface of zone A and B, first flatten chest line CP1-CP2 and keep it horizontal. Then flatten other intersection lines at same length based on CP1-CP2 and center front line. . When flattening zone C, first flatten bust line FBP-BP and center front line FBP-FWP, among which FBP-BP shall keep horizontal and FBP-FWP vertical. Based on these two lines, flatten other intersection lines at same length. . As to the zone D, first flatten bust line BP-XP2 and keep it horizontal. Then get symmetrized the dart line of D100 -BP in zone D from D10 -BP in zone C. Based on the

Figure 8. The final wireframe of the right half part of the dummy’s body

IJCST 19,5 .

344

.

.

.

bust line of BP-XP2 and D100 -BP, flatten other intersection line one-by-one at same length. In zone E and F, the flattening principle is preserving the bust line XP2-XP3 horizontal. The dart line D200 -CP1 in zone E can be got by symmetrizing vertically the dart line D20 -CP1 in zone D. The dart line D400 -BP1 in zone F will be got by symmetrizing vertically the dart line D20 -BP1 in zone G. In zone H and G, flattening is made in the same way as in zone C and D except in the reversed direction. The flattening in zone H and G shall follow the back width line BP12-BP2 and center back line BP2-BWP, in which BP1-BP2 should be kept horizontal. BP2-BWP shall be kept straight. And dart line D30 -BP5 and D300 -BP5 should be kept vertically symmetrized. In zone I, flatten back width line BP3-BP4 and center back line BP2-BP3, in which BP3-BP4 shall be kept horizontal and BP2-BP3 vertical. Based on this, flatten other intersection lines at even length. In zone J, first flatten the back width line BP3-BP4 and center back line BNP-BP3, in which BP3-BP4 should be kept horizontal and BNP-BP3 vertical. Based on this, flatten the intersection line for the surface BNP-BP3-BP6-SNP and BP4-BP6-SNP-SP. Besides, in order to keep back shoulder dart while making round the curve of armhole at the back, create the armhole curve at on 2D planner prior to flattening the surface SP-BP4-BP7-BP8. Then based on the armhole SP-BP4 and back width line BP4-BP7, map other intersection line and finally form the shoulder dart at the back pattern.

Take the flattening of zone C, for example. Please see the detailed explanation as shown in Figure 9. Via calculating the intersection between surface and plane, constitute a series of intersection lines, which will make zone C a series of small units. Each units will have four edges. When flattening, first map the length of on the center front line in the plane, preserving it vertical and get. Then identify the mapping point of Q1 based on , making and keeping the line horizontal. Here, and is just the structure lines constrained in the direction and position for zone C. As for the flattening of other points, follow the order of units and get the mapping points on the plane in turn. (BP) Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8

T1 T2 T3 T4 T5 T6 T7 T8 T9 T10

Figure 9. The flattening of the wireframe in zone C

T11

(D1)

(FBP) P1 P2 P3 P4 P5 P6 P7 P8 P9 P10

(BP) Q'1 Q'2 Q'3 Q'4 Q'5 Q'6 Q'7

I1

T7

Q'8 I2

(FBP) P '1 P'2 P'3 P'4 P'5 P'6 P'7 P'8 P'9 P'10

P11

P'11

P12 (FWP)

P'12 (FWP)

For instance, suppose three points in unit T have has already been flattened and their mapping points are , respectively, now comes to the mapping point from for flattening the unit T. The position of is found by finding the intersection of two circles, which are centered at and and have radii l1 and l2, respectively, where l1 is the length of and l2 is the length of on the original 3D surface T. Based on the above principle, the author generates 2D cutting pattern in DXF document by AutoCAD. The flattening result is shown in Figure 10.

Prototype garment pattern flattening 345

4. Results and discussion The corresponding feature lines of 2D prototype pattern are defined on the 3D model in advance. Based on the characteristics of each zone, cut 3D dummy surface (only half) into ten zones and meanwhile make subdivision for each zone. The composition of feature line is got from computer by crossing the plane with 3D surface. Via preserving the same length of intersection line, flatten geometrically 3D surface of each zone. Final prototype sample is indicated in the graph. Moreover, during the course of flattening of each zone, define a constrained line in advance so as to ensure the position and direction beforehand. In this way, the shape of the prototype sample and the existing 2D prototype can be kept the same upon flattening. As the length of each edge is kept the same during flattening of each unit surface, the length of major structure lines for both 3D model and 2D cutting pattern remain the same. The lengths of structure lines are listed in Table II.

Figure 10. The final cutting pattern of the individual prototype garment

IJCST 19,5

346

Table II. The length of the main structure lines in the prototype garment

Structure lines

Length

Bust girth/2 ¼ (FBP-BP) þ (BP-XP2) þ (XP2-XP1) þ (XP1-XP3) þ (XP3-XP9) þ (XP9-XP8) Center Back length ¼ (BWP-BP2) þ (BP2-BP3) þ (BP3-BNP) Front armhole length ¼ (XP-CP1) þ (CP1-CP4) þ (CP4-SP) Chest width (CP1-CP2) Back width (BP1-BP2) Front Neck girth (FNP-SNP) Back Neck girth (BNP-SNP) Shoulder width (SNP-SP)

42.142 ¼ 9.0470 þ 9.073 þ 4.5420 þ 5.8720 þ 7.560 þ 6.0480 37.20 ¼ 23.2040 þ 6.2530 þ 7.7430 20.655 ¼ 9.2420 þ 6.9360 þ 4.4770 14.964 14.740 11.764 8.384 8.939

Structure lines Waist girth/2 ¼ (FWP-D1) þ (D1-D2) þ (D2-SWP) þ (SWP-D4) þ (D4-D3) þ (D3-BWP) Center Front length ¼ (FWP-FBP) þ (FBP-CP2) þ (CP2-FNP) Back armhole length ¼ (XP-BP1) þ (BP1-BP4) þ (BP4-SP) (D1-BP) (D2-CP1) (SWP-XP) (D4-BP1) (D3-BP5)

Length 32.617 ¼ 7.8030 þ 7.206 þ 4.1740 þ 3.9560 þ 4.8780 þ 4.600 33.875 ¼ 16.3550 þ 10.6360 þ 6.8840 23.057 ¼ 9.6080 þ 5.4010 þ 8.0480 16.857 24.041 17.452 23.513 23.276

In order to analyze the error after flattening, the area of 3D and 2D will be calculated and compared, respectively. Table III gives the areas of ten zones of 3D surface and 2D pattern, respectively. As the table indicated, the seven out of ten segments of the plane have error within ^ 1 cm2 from 3D surface. Segment G has error of 1.1 cm2. Only two segments have relatively large error but controlled within 3 cm2. The main reason for the larger error is that it has comparatively complex shape on the surface and the relatively large subdivision. By summing up the areas of each segment, the total error of the whole sample is less than 1 cm2. It means this method has relatively good accuracy. In addition, looking from Figures 4 and 5, the cutting patterns after flattening keeps the same shape as newly Bunka’s prototype. This helps to apply. The traditional pattern maker can also create fit and personalized sample by modifying on

Zones

Table III. The area comparison of 2D cutting pattern and 3D surface

A B C D E F G H I J S

2D pattern (cm2)

3D surface (cm2)

Error

164.826 126.120 133.662 161.595 80.388 91.052 154.705 128.971 88.352 138.310 1267.981

167.798 125.638 133.736 161.541 79.551 91.296 155.797 129.267 87.700 136.024 1268.348

2 2.972 0.482 2 0.074 0.054 0.837 2 0.244 2 1.092 2 0.296 0.652 2.286 2 0.367

the existing prototype with reference to the traditional prototype design theory. Meanwhile, it ensures the task of pattern digital customized at the second stage. We can adopt more mature techniques of 2D CAD, and also help to modify the design via the personalized 3D technique and traditional 2D pattern making experience. In addition, what deserves your attention is that the sample mentioned in the context is the naked size of human body without any ease value consideration. However, the method is widely applicable. We can change the body model by enlarging some feature parts so as to be more flexible. Then the flattening can be made on the feature structure line got through the foregoing method. 5. Conclusions In this research, the process of customizing individual pattern was separated into two steps: the mapping process from 3D body model to 2D prototype and the altering process from prototype to the final cutting patterns. For the first step, a new approach is developed to transform 3D surface of the prototype garment to 2D cutting pattern. 3D prototype wireframe has been constructed based on a computer-generated body model by calculating the intersection curves between 3D surface and local planners. Half bodice surface was separated into ten zones according to the 2D structure of the new Bunka’s prototype pattern. For each zones, 2D cutting pattern of the prototype has been generated by flattening 3D wireframe geometrically with some constraint. The 2D cutting patterns fattened by this method have a satisfactory shape, e.g. the shape of the 2D segment meets the rules of conventional garment construction and the mapping relation from 3D virtual dummy to 2D prototype pattern is obvious. Another outstanding property of the method developed is the possibility of gradual mapping from 3D wireframe into 2D patterns, with no need to define physical-mechanical properties of the fabric used and no need to introduce fabric drape. In addition, there has relatively good accuracy through comparing the area of 3D surface and 2D segment. Certainly, many improvements should be done in future, e.g. it is necessary to construct general 3D prototype wirefame based on different 3D body model. References Cho, Y., Okada, N., Park, H., Takatera, M., Inui, S. and Shimizu, Y. (2005), “An interactive body model for individual pattern making”, International Journal of Clothing Science & Technology, Vol. 17 No. 2, pp. 91-9. Istook, C.L. (2002), “Enabling mass customization: computer-driven alteration methods”, International Journal of Clothing Science & Technology, Vol. 14 No. 1, pp. 61-76. McCartney, J., Hinds, B.K. and Seow, B.L. (1999), “The flattening of triangulated surfaces incorporating darts and gussets”, Computer-Aided Design, Vol. 31, pp. 249-60. Parida, L. and Mudur, S.I.P. (1993), “Constraint-satisfying planar flattening of complex surface”, Computer Aided Design, Vol. 25 No. 4, pp. 225-32. Petrak, S., Rogale, D. and Vinko, M.B. (2006), “Systematic representation and application of a 3D computer-aided garment construction method – part II: spatial transformation of 3D garment cut segments”, International Journal of Clothing Science & Technology, Vol. 18 No. 3, pp. 188-99.

Prototype garment pattern flattening 347

IJCST 19,5

348

Shimada, T. and Tada, Y. (1991), “Approximate transformation of an arbitrary curved surfaces into a plane using dynamic programming”, Computer Aided Design, Vol. 23 No. 2, pp. 153-9. Turner, J.P. (1994), “Development of a commercial made-to-measure garment pattern system”, International Journal of Clothing Science & Technology, Vol. 6 No. 4, pp. 28-33. Wang, C., Tang, K. and Yeung, B. (2005), “Freeform surface flattening based on fitting a woven mesh model”, Computer-Aided Design, Vol. 37, pp. 799-814. Wang, C.C.L., Wang, Y. and Yuen, M.M.F. (2003), “Feature based 3D garment design through 2D sketches”, Computer-Aided Design, Vol. 35 No. 7, pp. 659-72. Yang, J.X., Liu, J., Xiao, Z.Y. and Wang, Z. (2001), “A new method for marking complex surface developable and its development”, Mechanical Science and Technology, Vol. 20 No. 4, pp. 520-2. Yang, Y., Zhang, W. and Cong, S. (2006), “Research on digital pattern developing for customized apparel”, The Proceeding of 3rd International Textile, Clothing and Design Conference, pp. 946-52. About the authors Yang Yunchu is currently working as a PhD Student in Fashion Institute of Donghua University. His research interests include body shape analysis, sizing survey and human body modeling. Yang Yunchu is the corresponding author and can be contacted at: [email protected] Zhang Weiyuan is a PhD Supervisor in Fashion Institute of Donghua University.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

Assessing insulating fabric performance for extremely cold weather Wayne Seames, Ben Ficek and William Line Department of Chemical Engineering, University of North Dakota, Grand Forks, North Dakota, USA

Assessing insulating fabric performance 349 Received 10 November 2006 Accepted 29 March 2007

Abstract Purpose – The purpose of this paper is to present the quantification of the thermal conductivities and thermal resistances of 12 insulating fabrics extracted from commercial clothing products under static, simulated sweating, and simulated wind chill conditions. Design/methodology/approach – Triplicate coded (blind test) samples of each fabric were tested in a modified ASTM 1518-85 test apparatus enclosed in a cold box capable of temperatures as low as 2858C to determine thermal conductivity and thermal resistance. Sweat and wind chill were also simulated and evaluated. Findings – One fabric, Vaetrex0, was clearly found to be superior under all conditions to the other 11 fabrics tested. The performance of many of the other fabrics varied when exposed to simulated sweat. Practical implications – An objective evaluation of fabrics that can assist manufacturers in fabric selection for cold weather clothing manufacture. Originality/value – The paper provides an extension of the ASTM 1518-85 method to cold conditions and a unique blind comparison test of commercial clothing fabrics under these extreme conditions. Keywords Heat transfer, Thermal insulating properties, Thermal insulation, Fabric testing, Clothing Paper type Research paper

Introduction Work outside in January in the Northern latitudes or go mountain climbing at high elevations and the motivation for selecting coats and insulated pants with the best insulating properties becomes clear. But which fabric materials will keep people warmest under these extremely cold conditions? Currently no standard, objective test exists to assist clothing manufacturers and buyers in answering this question. This paper documents our adaptation of the most applicable standard method available in order to fill this need. A cold weather test chamber that replicates ASTM (2003) standard D 1518-85, with appropriate modifications to allow cold temperature testing was designed and constructed. This chamber was used to perform a series of experiments to determine the thermal conductivity and thermal resistance of a series of 12 unknown clothing fabrics under cold temperature conditions. Static tests were performed as well as tests which included the addition of a wetting device on the skin side of the fabric to simulate performance under sweat release conditions while additional tests simulated wind chill. Background Thermal conductivity is a proportionality factor that quantifies the efficiency of unidirectional heat transfer per unit area through a substance. Following the method

International Journal of Clothing Science and Technology Vol. 19 No. 5, 2007 pp. 349-360 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710819537

IJCST 19,5

outlined in ASTM (2003) standard D 1518-85, thermal conductivity is found by first calculating the thermal transmittance of a material using equation (1): U¼

350

P ADT

ð1Þ

where: U ¼ thermal transmittance (W/(m2 · 8C)), P ¼ heat flux from test plate (W), A ¼ area of the copper test plate (m2), DT ¼ change in temperature across the material (8C). These results are then used to calculate the thermal conductivity using equation (2): k ¼ ðti ÞðU Þ

ð2Þ

where: U ¼ thermal transmittance (W/(m2 · 8C)), k ¼ thermal conductivity (W/(m · 8C)), ti ¼ thickness of the specimen (m). It is important to note that these equations assume that thermal conductivity is linear with respect to temperature. While this assumption is not 100 percent accurate, the relatively thin layers of material tested insure that deviations from this assumption do not affect the conclusions drawn from the results. Thermal resistance (R-value) is a measure of a material’s resistance to heat flow. Mathematically thermal resistance is the thickness of the material divided by the thermal conductivity of the material. In other words: R¼

ti k

ð3Þ

where: k ¼ thermal conductivity (W/(m · 8C)), ti ¼ thickness of the specimen (m), R ¼ thermal resistance ((m2 · 8C)/W). The thermal conductivity and thermal resistance of a fabric are of great importance when determining the applicability of a fabric in cold weather conditions. A low-thermal conductivity means the fabric material has a high resistance to heat flow. This means that if two fabrics are of the same thickness, the material with the lower thermal conductivity will allow less heat to be lost from the body and thus make the wearer more comfortable under extreme cold conditions. Alternately, for a given level of comfort (that is, of heat loss) materials with lower thermal conductivities can be made thinner and thus the clothing will be less bulky. Thin, comfortable clothing products are the ideal, particularly for outdoor active people like workers and athletes. R-value (thermal resistance) measures the overall effectiveness based on the combination of the thermal conductivity of the material and its thickness. Two materials with different thermal conductivities may have the same R-value if their thicknesses are correspondingly different. Thus, both values are important in designing or evaluating clothing fabrics. When wearers of clothing sweat, the water released from the body can have a significant effect on the fabric’s overall thermal conductivity. One way that fabrics reduce heat loss is through the use of micro-sized air pockets encapsulated in the fabric. Wetting these fabrics often replaces the air in these voids with water and since water conducts heat much more efficiently than air, the overall thermal conductivity of the fabric increases. This effect can be limited by: using a hydrophobic (water repelling) inner liner, using a hydrophobic insulating fabric material, and/or by increasing the

porosity of the fabric (making it easier for the water to diffuse through the clothing and evaporating from the outer surface). The third method is the least desirable because it will also increase the overall thermal conductivity of the clothing. Wind chill increases heat loss by continuously removing hot air from the surface of the clothing. This increases the driving force for heat loss from the body to the outside and therefore makes the wearer less comfortable under extremely cold conditions. In general, wind chill should not impact the relative effectiveness of clothing materials. However, quantification of overall heat transfer, using an equivalent global thermal conductivity can still be useful for clothing designers. This information allows the designer to relate static thermal conductivities to overall user comfort by providing insight into how much additional heat loss can be expected under windy conditions. Experimental approach Experimental test chamber An experimental test chamber was designed and constructed according to the specifications provided in ASTM (2003) standard D 1518-85. This chamber, shown in Figure 1, was placed inside of an insulated cold box. The temperature in the box was controlled using an external chiller, cooling the temperature of an air exchange stream with inlet and outlet connections far removed from the fabric surface in the cold box so that convection from this air exchange did not affect the air boundary layer adjacent to the fabric surface. Heating was provided using a 720 W heating blanket mounted to the underside of the test plate. A rheostat controlled the blanket heat input in order to allow the copper plate to slowly heat to the steady state desired value. A data acquisition computer running Labview Visual Instrumentation software (Version 6) was used to collect and store time-dependent temperature data collected during the experiments. Triple replicate samples of 12 code marked fabrics, cut and sewn into the exact same surface area, but with the thickness of the original clothing item, were fabricated by personnel other than those conducting the tests. None of the personnel involved in the tests had access to or were aware of the identification of the 12 fabric samples. The Labview software used for automatic data monitoring collected and stored the temperatures of the seven thermocouples in the system every 4 s. Every 5 or 10 min, an average of 15 points (1 min of data) were sent to a spreadsheet to record the average current temperatures of the copper test plate and the air immediately above the test fabric. Collecting averages ensures precision in the temperature readings. When a trial was begun, data were collected more frequently, every 3 min. For the sweat simulation tests, a small grid-shaped rack was installed on the under side of the fabric and water was trickled uniformly onto the test area of the fabric at a rate of 15 ml/h (Freudenrich, 2006). This flow rate simulates the typical sweat generation rate of a human onto an 8 £ 8 in. square portion of the average person’s torso (Elert, 2006). The “sweat rack” consisted of a series of perforated copper tubes, laid out in a grid arrangement. This arrangement distributed the water evenly over the entire skin side surface of the test region of the fabric. For the wind chill experiments, compressed air was regulated into the test chamber at a rate that corresponds to external wind speeds of 5 and 20 miles/h in the test chamber. The air inlet for this wind was placed right above the fabric and the air blown

Assessing insulating fabric performance 351

IJCST 19,5

352

Recycled Air Outlet

Dry Ice and Liquid Nitrogen Basket

Recycled Air inlet

Thermocouple

24 in High Plexiglass Walls

Wind Speed Inlet Sweat Rack Heat Blanket

Thermocouple Wooden Frame Test Fabric Test Plate Bottom Plate

Insulation 13 by 13 in Square Plexiglass Base (a)

Gaurd Ring

Sweat Rack Test Plate Heat Blanket Wooden Frame

Figure 1. Experimental test system: (a) side view; and (b) top view

(b) Fabric placed under wooden frame on top of sweat rack during trials. Basket and 2 thermocouples hang in the air above the fabric not shown here

straight across the face of the fabric. A velocity meter was installed in the opposite face of the chamber and used to measure the air flow rate. Static test procedure One of the 12 code marked test fabrics, chosen at random, was placed in the test chamber on top of the copper test plate and held down by a wooden frame. The wooden frame did not cover any of the area over the copper test plate. The heat blanket was then energized until it reached the desired temperature to simulate human external skin temperature (30 ^ 0.58C). A block of dry ice (frozen carbon dioxide, CO2) or a beaker of liquid nitrogen (ultra cold tests) was inserted into a basket hanging inside the test chamber to bring the temperature of the test chamber down to the desired level, simulating the external weather condition. The test chamber was then sealed. With the temperatures above and below the fabric at their desired initial and stable values (steady for at least 15 min), the experiment was initiated by deenergizing the

heating blanket. Each trial lasted one hour during which temperatures at seven locations within the test chamber were recorded. During the static tests with dry ice, the temperature of most fabrics decreased between 8 and 138C over the 1 h test period. When using liquid nitrogen for the extreme cold temperatures, the fabrics decreased as much as 238C over the test period. Shorter, bare plate trials were performed as calibration tests over the same temperature decrease range. Simulated sweat test procedure A set of experiments were performed to ascertain differences in fabric performance under conditions that might occur during exertion (i.e. if the wearer of the garment was sweating). The experimental procedure for these tests was identical to that used for the static tests, as outlined above with the following exception. When the heat blanket was deenergized at the beginning of each test (at the simulated body temperature), water flow was initiated at the predetermined flow rate to simulate sweat. The water flow rate was stopped when the trial was stopped (after 1 h) and restarted only when the next trial was begun. Wind chill simulation test procedure A set of experiments was then run to test the fabrics for their ability to withstand wind chill. These tests used nearly the same procedure as described for the static tests with one small difference, when the actual trial data were started, the compressed air was turned to a predetermined setting to give the desired wind speed. This wind speed was held constant for the hour long trial and if necessary the compressed air supply bottle was changed out before the next trial began. To account for the difference in temperature between the wind coming in and the air in the test chamber, the chiller unit was modified so that the air from the supply bottle was routed through the chiller unit before entering the test chamber. Results Static cold temperature test results Static experiments were run for each of the three replicates of the 12 unknown fabrics at 4-5 different temperatures ranging from 0 to 2 358F using dry ice as the cooling medium. Additional experiments were run for each of the three replicates of the 12 unknown fabrics at a single temperature ranging from 2 97 to 2 1158F using nitrogen as the cooling medium. This ultra-cold test was run so that we could correctly quantify the relationship between fabric insulating properties and cold temperatures over the range of 2 100 to 08F. Bare plate test trials were performed to isolate the insulating properties of the fabric from the surrounding test apparatus. These tests, coupled with the thermal conductivity of copper (Perry and Green, 1997) and the thermal conductivity of silicone rubber (Dow Corning Corporation, 2006) (the material of the heating blanket) were used to correct the overall results obtained when testing the fabrics. Equation (1) was solved to find the heat flux, “P” that was then used in subsequent calculations. Equation (1) was used to calculate the thermal transmittance “U” for each of the 12 fabrics based on the experimental test data and the heat flux, “P” found in the bare

Assessing insulating fabric performance 353

IJCST 19,5

354

plate test trials. Equation (2) was then used to find the thermal conductivity “k” and equation (3) to find the thermal resistance “R” of each individual fabric. Figure 2 shows the thermal conductivity of each of the 12 fabrics at their actual test temperatures (similar curves were generated for thermal resistance, not shown). Note, all samples results are shown with their internal number code as designated for the blind experiments. The sample number code is mapped to the actual fabrics in the conclusion section of this paper. Figure 2 results can be grouped into four statistically “equivalent” groups of fabrics. The best group (in terms of insulating properties), group 1, consists solely of fabric 4. The second best group consists of fabrics 3 and 7 which had essentially the same conductivity. The third best group consists solely of fabric 5 while the fabrics with the highest values and which thus represent the worst insulating properties were 2, 1, 12, 10, 6, 8, 9, and 11 which all had similar conductivities (within the statistical accuracy of the tests). To obtain the thermal conductivity and thermal resistances of the 12 unknown fabrics at extremely cold temperatures, liquid nitrogen was used. A low point was taken for each fabric and the conductivities and resistances at these low temperatures were found as described above for the dry ice-chilled data set shown in Figure 2. Typical results are shown in Figures 3 and 4. The groupings of the fabrics, based on thermal conductivity are the same in the ultra cold region as in Figure 2, which tells us there is little or no change in the relative effectiveness of the fabrics over the entire temperature range. Using the ultra-cold values combined with the rest of the data, we fit the best polynomial equation from the region 0 to 2 1008F, which defines the value of thermal conductivity or thermal resistance of a given fabric as a function of static air temperature over this temperature range. Thermal conductivities and resistances calculated from these relationships at various temperatures for all 12 fabrics are summarized in Table I. The density and thickness of each fabric sample was measured and are also included in Table I. Note that fabric no. 5 has the highest thermal resistance even though its thermal 0.4

Figure 2. Thermal conductivities of cold weather fabrics under static conditions as a function of temperature in the range obtained using dry ice cooling

Thermal Conductivity (Btu/fthr°F)

0.375

Fabric 1 Fabric 5 Fabric 9

Fabric 2 Fabric 6 Fabric 10

Fabric 3 Fabric 7 Fabric 11

Fabric 4 Fabric 8 Fabric 12

0.35 0.325 0.3 0.275 0.25 0.225 0.2 0.175 0.15 0.125 0.1 0.075 –35

–30

–25

–20 Temperature(°F)

–15

–10

–5

Assessing insulating fabric performance

Thermal Conductivity (W/m*°C)

0.48 0.46

y = 0.0076x + 0.6353 R2 = 0.9524

0.44 0.42

355

0.4 0.38 0.36 0.34 –36

–34

–32

–30

–28

–26

–24

–22

Temperature˚C

Figure 3. Assessment of random variability in the static tests via multiple trials for three different samples of test fabric 6

Thermal Conductivity (W/m*°C)

0.46

0.44

0.42

0.4

0.38

0.36 –36

–34

–32

–30

–28

–26

–24

–22

Temperature°C

conductivity is at least double that of fabric no. 4 under the entire range of temperature conditions. The manufacturer has compensated for the lower insulating qualities of the fabric material by making the fabric substantially (over 100 percent) thicker – 18 vs 8 mm. The published thermal conductivities of other similar insulating materials are within the same range as those determined in our experiments. Felted cattle hair has a reported thermal conductivity of 0.021 Btu/(ft h 8F) and flex fiber has a value of 0.31 Btu/(ft h 8F) (Weast, 1972). It should be noted that these values were not generated in the same test apparatus or at the same temperatures, so conclusions should be drawn from this comparison with great care. However, these values suggest that the data generated in the present study are in the realistic range. Three replicates of test fabric 6 were used to quantify the variability among different samples of the same fabric under static conditions. The variability is then

Figure 4. Model equation for the thermal conductivity of fabric 6 under static conditions with error bars of uncertainty

Table I. Static test results for 12 fabrics over the temperature range 2100 to 08F

19.2 ^ 0.8 27 ^ 2.0 46 ^ 3.1 39 ^ 1.4 18 ^ 1.3 29 ^ 1.4 33 ^ 1.6 23 ^ 1.0 28 ^ 1.6 14 ^ 2.6 25.0 ^ 0.4 25.7 ^ 0.3

24 ^ 1.0 24 ^ 1.4 12.8 ^ 0.6 8.0 ^ 0.2 18 ^ 1.3 21.8 ^ 0.8 13.3 ^ 0.4 24 ^ 1.2 24.5 ^ 0.78 23 ^ 2.4 26.6 ^ 0.33 21 ^ 0.26

0.16 0.16 0.09 0.05 0.12 0.15 0.08 0.15 0.17 0.15 0.17 0.14

Ka 0.47 0.47 0.48 0.47 0.49 0.47 0.48 0.49 0.46 0.47 0.43 0.48

Rb 0.17 0.18 0.09 0.05 0.12 0.16 0.08 0.16 0.19 0.16 0.20 0.15

k

R 0.42 0.43 0.43 0.43 0.45 0.42 0.45 0.44 0.41 0.43 0.39 0.43

2 808F

0.19 0.20 0.11 0.06 0.14 0.18 0.10 0.19 0.22 0.18 0.23 0.17

k

R 0.37 0.38 0.38 0.38 0.40 0.38 0.40 0.39 0.36 0.38 0.35 0.38

2608F

0.23 0.24 0.13 0.08 0.17 0.21 0.12 0.23 0.25 0.22 0.27 0.21

k

R 0.32 0.33 0.33 0.33 0.35 0.33 0.35 0.34 0.31 0.33 0.31 0.33

2408F

0.29 0.28 0.15 0.10 0.21 0.26 0.16 0.28 0.30 0.28 0.32 0.25

k

R 0.27 0.28 0.28 0.27 0.29 0.27 0.28 0.28 0.27 0.28 0.27 0.28

2 208F

0.39 0.34 0.19 0.13 0.26 0.32 0.21 0.35 0.36 0.34 0.38 0.31

K

08F

0.21 0.22 0.22 0.20 0.22 0.22 0.20 0.21 0.22 0.22 0.2 0.22

R

Notes: ak ¼ thermal conductivity in Btu/(ft -hr- 8F); uncertainty of these values is estimated to be ^1.4 percent; bR ¼ thermal resistance in (ft2-hr-8F)/Btu; uncertainty of these values is estimated to be ^1.4 percent

1 2 3 4 5 6 7 8 9 10 11 12

Thickness (mm)

21008F

356

Fabric (ID)

Density (kg/m3)

IJCST 19,5

applied as a standard variability for all fabrics. Three non-consecutive trials were performed to eliminate any bias in the error. Figure 3 shows all of the trials for test fabrics 6-1, 6-2, and 6-3. A regression line is shown through the data. This line served as the data model of thermal conductivity versus temperature that was used for determining the error in the data set. Using the pooled variance method, an error of ^ 1.4 percent among samples of the same composition was found. Figure 4 shows this variance in the form of error bars to the thermal conductivity versus temperature function. Comparing these error bars to the data groupings described earlier for the data shown in Figure 2 validates these groupings. Fabrics within each group have thermal conductivity values that are statistically equivalent due to the uncertainty of variability of the results.

Assessing insulating fabric performance 357

Simulated sweat test results Figure 5 shows the thermal conductivities determined from tests in which sweat was simulated by a constant very low trickle of water onto the skin side of the fabric. Comparable thermal resistances were also found from the same tests (not shown). Comparing Figures 2 and 5, we see many of the fabrics unaffected by sweat, but a few of them became significantly worse insulators under sweating conditions. Most notably, fabric 4 continues to have the lowest thermal conductivity, even under wet conditions. Fabric 7 is one of the fabrics which seem the most affected by the addition of sweat. Its thermal conductivity increased drastically from one of the best fabrics in the static tests to nearly the worst fabric under sweating conditions. Variance error for the sweat test trials was determined using all three samples of fabric 7. Using the pooled variance method, an error of ^ 0.94 percent among the three trials of fabric 7 was found. Fabrics that were described to be in a group in the sweat tests are considered statistically equal due to this level of uncertainty in the results. 0.4

Thermal Conductivity (Btu/fthr°F)

0.35 0.3

Fabric 4 Fabric 7 Fabric 11 Fabric 12

0.25 0.2 0.15 0.1 0.05 0 –120

–100

– 80

– 60 Temperature(°F)

– 40

–20

0

Figure 5. Test fabrics from each of the groupings (Figure 2) with both dry ice and liquid nitrogen static test data shown

IJCST 19,5

358

Simulated wind chill test results Wind chill increases the severity of cold conditions by using convective heat transfer to remove heat from the outer surface of the fabric much more efficiently than under static conditions. This results in an increased driving force for heat flux through the fabric, thus making it more difficult to keep the body warm. Figure 6 shows the thermal conductivities determined from tests in which compressed air was used to simulate a wind chill of 20 miles/h. Thermal resistances from the same set of tests were also calculated (not shown). Comparing Figures 2-7, we see that there is almost no change in the order of fabrics from the static tests. Fabric 4 is still by far the best fabric with 3 and 7 in the next best grouping. Fabric 5 still constitutes the third group with the remaining fabrics grouped together as the worst insulators. Similar results were obtained for tests performed to simulate a wind chill of 5 miles/h (not shown). All three samples of fabric 4 were tested to quantify variance in our wind chill test results. Using the pooled variance method, an error of ^ 0.38 percent among the three trials of fabric 4 was found. Fabrics that were described to be in a group in the wind chill tests are considered statistically equal due to the uncertainty of the results. Conclusions This study documents an objective method for testing the capabilities of clothing fabric materials to perform under extremely cold weather conditions. Table II shows the relative ranking of the 12 fabrics for the static, the sweat, and the wind chill simulated conditions based on values of thermal conductivity, listed in overall order of performance as insulating materials. From all test results, dry and wet, windy or calm, it is clear that the thermal conductivity of fabric 4, Vaetrex0 is lower than any of the other fabrics tested. From these data we can conclude that clothing made from Vaetrex0 has the potential to have the best insulating properties. For any given thickness of material, a garment made 0.45

Figure 6. Thermal conductivities of cold weather fabrics under simulated 20 mph wind conditions as a function of temperature

Thermal Conductivity (Btu/fthr°F)

0.40 0.35

Fabric 1 Fabric 4 Fabric 7 Fabric 10

Fabric 2 Fabric 5 Fabric 8 Fabric 11

Fabric 3 Fabric 6 Fabric 9 Fabric 12

0.30 0.25 0.20 0.15 0.10 0.05 0.00 –40

–38

–36

–34

–32

–30

–28

Temperature (°F)

–26

–24

–22

–20

Assessing insulating fabric performance

Thermal Conductivity (Btu/fthr°F)

0.32

0.27

359

0.22

0.17 Fabric 1 Fabric 4 Fabric 7 Fabric 10

0.12

0.07 –30

–28

Fabric 2 Fabric 5 Fabric 8 Fabric 11 –26

Fabric 3 Fabric 6 Fabric 9 Fabric 12 –24

–22

–20

–18

–16

Temperature(°F)

Fabric ID Fabric type 4 3 7 5 6 12 8 2 10 9 1 11

Vaetrex0 Cordura/foam/microfiber Vaetrex30 Primaloft Vaetrex60 Vaetrex3/4 Refrigiwear (polyfil) Cordura/Primaloft/microfiber Down Samoco (polyfil) Versatech/Primaloft/Versatech Vaetrex 2 1

Figure 7. Thermal conductivities of cold weather fabrics under simulated sweating conditions as a function of temperature in the range obtained using dry ice cooling

Static test rank Sweat test rank Wind chill test rank 1 3 2 4 6 5 8 9 7 11 10 12

1 2 7 3 4 12 9 8 11 5 11 10

1 3 2 4 7 5 6 9 8 11 8 12

from Vaetrex0 will resist heat loss the best. The next best appears to be fabric 3, a Cordura/foam/microfiber. Fabric 7, Vaetrex30 performed well in dry and wind chill tests but not as well for wet. The next best is fabric 5, Primaloft.

References ASTM (2003), D 1518-85: Standard Test Method for Thermal Transmittance of Textile Materials, ASTM International, West Conshohocken, PA. Dow Corning Corporation (2006), “Rubber physical and chemical properties”, available at: www. dowcorning.com/content/rubber/rubberprop/thermal_expansion.asp (accessed June 26). Elert, G. (2006), “Surface area of human skin”, available at: http://hypertextbook.com/facts/2001/ IgorFridmans.html (accessed August 10).

Table II. Relative rank of 12 commercial fabrics in order of insulating properties for both static and sweat tests, based on thermal conductivities

IJCST 19,5

Freudenrich, C.C. (2006), “How sweat works”, available at: http://health.howstuffworks.com/ sweat.html (accessed August 10). Perry, R.H. and Green, D.W. (1997), Perry’s Chemical Engineers’ Handbook, 7th ed., McGraw-Hill, New York, NY. Weast, R.C. (Ed.) (1972), Handbook of Chemistry and Physics, 53rd ed., Chemical Rubber Co., Cleveland, OH.

360 Further reading Bird, R.B., Stewart, W.E. and Lightfoot, E.N. (2002), Transport Phenomena, 2nd ed., Wiley, New York, NY. McCabe, W.L., Smith, J.C. and Harriott, P. (2005), Unit Operations of Chemical Engineering, 7th ed., McGraw-Hill, New York, NY. Wikimedia Foundation (2006), “Polystyrene”, available at: http://en.wikipedia.org/wiki/ Polystyrene (accessed June 26). Corresponding author Wayne Seames can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

approach to An approach to predicting bra cup Anpredicting bra cup dart quantity dart quantity in the 3D virtual environment Jian Ping Wang and Wei Yuan Zhang

361

Fashion Institute of Donghua University, Shanghai, China Abstract Purpose – The aim of this paper is to explore a method of predicting the amount of personalized bra cup dart in the 3D virtual environment for supporting the made-to-measure research of the optimum fitted brassiere pattern design. Design/methodology/approach – Very useful enhanced FFD (free-form-deformation) techniques used in both computer animation and geometric modeling were skillfully transplanted to the female breast model deformation. Meanwhile, on the basis of 3D scan and surface modeling technologies, the realization approach of the abstract female breast model library focusing on the individual variations of shapes and sizes was presented. Then according to the principle of isometric area and flabellate segments, the personalized bra cup dart quantity and its distributive information were provided by 3D-2D transformation. Findings – The paper finds that personalized female breast shapes and various aesthetic breast forms sculpted by different bras could be interactively simulated. Accordingly, the amount of corresponding individual bra cup dart and its distributive information were provided. The cup darts were mainly distributed below the bust line. Moreover, dart shapes were curvy. Research limitations/implications – The principles of virtual breast library construction and 3D-2D transformation are also suitable for other parts of the human body such as buttocks, abdomen and, etc. for intimate apparel research. Originality/value – The method of predicting the personalized bra cup dart quantity based on the 3D virtual breast model library was delivered for the first time. The novel findings provided an important guideline for designers to improve the well-fitted bra pattern design technique. Furthermore, it would reduce the manufacturing cost without keeping physical dummies. Keywords Clothing, Women, Computer aided design Paper type Research paper

1. Introduction Garment fit is a significant problem for retailers, manufacturers, and consumers in the clothing industry. Bra, a kind of sheathy foundation of women, has tight-fit requirement to comfortably lift body tissue, provide support for body movement and improve the aesthetic appearance of the wearer. However, current surveys showed that about 80 percent of all women were walking around in wrong-size bras or wearing unfitted bras (Spiegel and Sebesta, 2002). And most of fitting failures were cup wrong (Bras, 2004). The reason might be that the female breast shape was one of the most complex configurations of the body. Different people had different breast forms. Even one person had several shapes of breast, because the breast was a pliable object. It could be pushed, lifted, flattened by assorted constructional bras to sculpt one’s bosom. According to The Oxford Dictionary (2002), fit was defined as the ability to be the right shape and size. Most of the fitting problems were created by the bulges of the human body. Rasband (1998) pointed out that bra fitting began with a fair assessment

International Journal of Clothing Science and Technology Vol. 19 No. 5, 2007 pp. 361-373 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710819546

IJCST 19,5

362

of one’s body. Therefore, 3D body models were the foundation and essential terms for well-fitted bra design. Nevertheless, storing enough physical body models was both exorbitant and non-realistic. Constructing a non-objective 3D female breast model library which precisely described breast sizes and shapes was an effective basic approach of improving bra fitting and the global competitiveness of brassiere products. Recently, the emergence of sophisticated 3D body scan technology accelerated the digital development of apparel industry and made it possible to construct realistic 3D female breast models. Human body models had been studied in computer graphics almost since the introduction of the medium. Various modeling methods were explored from stick figure models (Zeltzer, 1982), surface modeling (Fetter, 1982) to volume models (Yoshimoto, 1992) and multi-layered models (Shen and Thalmann, 1995). Some methods were based on passive 2D images and others were found on active 3D surface measurement systems. Much research had dealt with the region of human hand (Magnenat-Thalmann et al., 1988), leg (Lamousin and Waggenspack, 1994), facial animation (Kahler et al., 2003), human organ reconstruction (Yu et al., 2005) and whole-body model (Yesil, 2003). Moreover, there was considerable interest in modeling human body in the clothing industry (Stylios et al., 2001; Xu et al., 2002; Kurokawa, 1997). Each field put its specific requirements for the models. However, very little attention had focused on the breast modeling and simulation especially in its detailed flexible deformation for bra design. One of the purposes of this study was to create a female breast shape library of the 3D virtual environment using a novel method. After having the breast characteristic information, the other motivation was to predict the personalized bra cup dart quantity and have a better understanding of dart distribution in the bra cup pattern design. Because pattern design which was always recognized as the emblems of the highest skill was one of the most important factors of affecting garment fit (Dobson, 1885). 2. Realistic human body modeling 2.1 3D body scanning The 3D body scanning system could generate thousands of data points which corresponded to the shape of the subject yielding a raw image called 3D Point Cloud. For the purpose of forming an abstract 3D female breast shape resources, the initial work was to reconstruct a realistic human body model based on the 3D scanner data points. The posture of the subject in the experiment was assumed that the female body was erect and stood with the feet between the shoulder width apart, and arms held slightly away from the body to aid segmentation. The upper body was nude and lower body was in a close-fitting panty to expose the body shape. The initial data were sampled and part of the points for rebuilding was used to reduce the process complexity. The data of extremities were deleted, namely, not modeled, because the breast shape was only related to the torso. The model was simple and compact enough for communication. The triangle mesh torso model was then represented according to the 3D geometric and texture information. The torso was described by a Qi, j matrix, where i indicated layers and j meant the columns. The data of columns on each layer were end to end to ensure curve closing. Figure 1 showed the face and profile of the triangle mesh breast model.

2.2 Surface fitting NURBS (Non-uniform rational B-splines) had become industry standard tools for the representation and design of geometry that could accurately describe any shape. Therefore, the description of female body surfaces with realistic and smooth was represented by a series of B-spline surfaces. According to the parametric surface theory, a B-spline surface of degree ð p; qÞ was defined by equation (1): Sðu; vÞ ¼

m X n X

An approach to predicting bra cup dart quantity 363

N i;p ðuÞN j;q ðvÞP i;j ;

u; v [ ð0; 1Þ

ð1Þ

i¼0 j¼0

where P i;j were the ðm þ 1Þ £ ðn þ 1Þ control points and N i;p ðuÞ, N j;q ðvÞ were the B-spline basis functions defined, respectively, by the following knot vectors of equations (2) and (3): U ¼ {0; 0; . . . ; 0; upþ1 ; . . . ; um ; 1; 1; . . . ; 1}

ð2Þ

V ¼ {0; 0; . . . ; 0; vpþ1 ; . . . ; vn ; 1; 1; . . . ; 1}

ð3Þ

Fitting a B-splines surface along each regular grid was a standard interpolation problem with linear complexity, which required the solution of a number of tri-diagonal equation systems. The established technique by Piegl and Tiller (1997) was employed. Then the realistic 3D torso model in different points of view was originated as shown in Figure 2. 3. Schemes and techniques of designing 3D virtual personalized female breast environment As previously mentioned, one of the motivations of this research was to find a way of interactively, intuitively, realistically, and efficiently creating a rich variety of female breast shapes. However, the generated 3D body torso based on 3D scan data had only a real naked breast configuration which, in general, was not the ideal form to the fashion of the day. The reason was that the breasts always pointed out towards the arms and usually were saggy due to the gravity. Since, the female breasts were flexible matter, a well-fitted bra could bring the breast tissues back to an attractive position and hold them in place. In order to construct a virtual library including assorted breast shapes of different figures and even various aesthetic or erotic breast forms which could be trained under different bras, the following techniques were required to properly modify

Figure 1. 3D breast mesh models

IJCST 19,5

364

Figure 2. Realistic torso models

the existing model according to given body characteristic parameters and desired breast identity parameters. 3.1 Deformation technology In this study, a suitable deformation method was one of the crucial techniques. Free-form deformation (FFD), abbreviated to FFD, proposed by Sederberg (Sederberg and Parry, 1986) at the earliest, was an important and a powerful modeling tool in both computer animation and geometric modeling. It enabled the object deformation by deforming the space around them. Till now, there were several deformation methods including the original FFDs, extended FFDs (Coquillart, 1990), direct FFDs (Hsu et al., 1992) and NURBS-based FFDs (Lamousin and Waggenspack, 1994). Sometimes using these techniques was difficult due to limitations. The main resulting constraint was that control point boxes had to be rectangular. However, Farin (1990) presented Dirichlet free-form deformation (DFFD), which offered many advantages over traditional FFDs. It was a local deformation technology based on the natural neighbors coordinates (Sukumar, 2001). It removed any constraint on the position and topology of control points. It kept all the capabilities of FFD extensions, such as extended FFDs and direct FFDs. The generalized method for FFDs was employed in this investigation. It was found on natural neighbor coordinates system or Sibson local coordinates combining the traditional FFD model with techniques of scattered data interpolation on the basis of Delaunay and Dirichlet/Voronoi diagrams. Sibson local coordinates was shortly described in 2Ds for simplification and it could be extended to 3Ds straightforward. Consider a set of distinct points P i ði ¼ 1; 2; . . .nÞ, and let them be divided into Delaunay triangulation. The corresponding Voronoi/Dirichlet polygon T i ði ¼ 1; 2; . . .nÞ could be acquired. If the point S was located inside the convex hull of the set, and was taken part in partition, the Dirichlet tile of S point was named for TS. Then there were overlapped areas between TS and corresponding Dirichlet cell of its

neighbor points, which were denominated as Ai ði ¼ 1; 2; . . .nÞ, respectively. TS was fully partitioned by these areas. The natural neighbor coordinates value of S point was defined as equation (4): ui ¼

Ai ; ði ¼ 1; . . . ; nÞ; TS

ð4Þ

365

where: TS ¼

n X

Ai ;

with

i¼1

n X

ui ¼ 1;

ui [ ½0; 1:

i¼1

The point S could be expressed as equation (5), a linear combination of its Delaunay neighbors in the set: S¼

n X

ui P i :

ð5Þ

i21

If one or more points Pi were moved and DP i was individual displacement, the displacement of S was defined by equation (6): X ð6Þ DS ¼ ui DP i where i indicated the points having been displaced. The new position S0 could be described by equation (7): S 0 ¼ S þ DS:

An approach to predicting bra cup dart quantity

ð7Þ

As noted previously, it was clear that there were four main steps to implement the process of deformation: (1) Design of the control point set. Control points could be interactively created anywhere around, on the surface or inside the object to be deformed. The terms control box and lattice were no longer necessary. (2) Embed the object within the control point set. This step consisted of computing Sibson coordinates for all the points of the object. (3) Deformation of the control point set. Control points were moved in space either individually or in a group, using any technique. (4) Deformation of the object. Once the control point set had been deformed, the deformation was passed to the object. In order to increase or decrease the influence of some Sibson neighbors on the 3D space points, the concept of varying weight should be introduced and was assigned to the control points. The formula (5) was then extended into equation (8): Pn ui W i P i : ð8Þ S ¼ Pi¼1 n i¼1 ui W i

IJCST 19,5

366

When the weights were all given a value of 1.0, DFFDs were then the special case, namely, the formula (5). 3.2 Model layers and control point design In order to simulate diverse desired female breast forms by scaling or modifying existing models, a highly effective multi-layered deformation model driven by torso characteristic parameters was designed using DFFD technology. It was based on three layers. Meanwhile, designing control points was very important for the model deformation effect. Based on the functions, control points were plotted into integral control point, transformative constrained and character control points. The body torso to be deformed and all control points were assigned to the following layers. The first layer was composed of integral control points which constituted the peripheral convex hull and other control points inside the convex hull. The height was divided into several sub-layers. Each layer reflected the corresponding high information. The height and primary girth information were controlled by the integral control points, among which there were two points to be used to define the central axis for transforming. The second layer was formed by character control points which could curb the deformation by changing the weight value or moving the position. It would completely determine the deformation of this place when the point was on the surface of the body torso. The third layer consisted of character control points and transformative constrained points, which could confine the deformation of the object or shield from the influence among control points. According to the DFFD principle, the object to be deformed should be located inside the convex hull of the control point set. Then the integral deformation could be controlled using a small quantity of convex hull points. Other control points and the object to be deformed became control target of the convex hull point set. Figure 3 shows the diversiform control points. 3.3 Deformation process and personal breast shape realization The principle of breast deformation process was keeping away from interacting among body characteristic parameters. The primary breast parameters dealt with bust circumference, under bust girth, bust point height, bust point distance and so on. The height information of the body torso was controlled directly by the relative integral control point set. The detailed perimeter and other characteristics were adjusted by character control points and transformative constrained points, such as moving character control points or changing the weight value. The detailed characteristic parameters of the instances to be deformed were listed in Table I. The having been transformed data were highlighted. With an interactive environment, the corresponding female breast model was implemented to simulate various 3D virtual breasts as shown in Figure 4. Image (a) was the naked real breast model based on the 3D scan data. By inputting the different parameters of bust point distances and bust point heights to modify model (a), picture (b), (c) and (d) could be Integral control point Transformative constrained control point

Figure 3. Control point set

w w

Character control point w Character control point determined by weight

(a) Real torso (b) Separate (c) Push (d) Droop (e) Lift (f) Push-up (g) Appointed torso

Body dimensions (cm)

78.6 78.6 78.6 78.6 78.6 78.6

71.5

97.3

63.8

76.8 76.8 76.8 76.8 76.8 76.8

35.6 35.6 35.6 35.6 35.6 35.6

98.2 35.6

90.1 90.1 90.1 90.1 90.1 90.1 42

39.8 39.8 39.8 39.8 39.8 39.8 18.0

23.2 28.8 17.0 23.2 23.2 17.0 112.8

108.5 108.5 108.5 105.6 113.3 113.3 103.5

102.7 102.7 102.7 102.7 102.7 102.7

Under Bust point Bp Under bust bust Waist Hip Neck Shoulder distance height height

92.1 92.1 92.1 92.1 92.1 92.1

Bust

102.5

99.7 99.7 99.7 99.7 99.7 99.7

Waist height

80.8

78.9 78.9 78.9 78.9 78.9 78.9

137.5

136.2 136.2 136.2 136.2 136.2 136.2

Hip Neck height height

128.2

129.6 129.6 129.6 129.6 129.6 129.6

0.8 0.8 0.8 0.8 0.8 0.2

0.7

Waist weight 0.2 0.2 0.2 0.8 0.8

Shoulder Bust height weight

An approach to predicting bra cup dart quantity 367

Table I. Characteristic parameters of body torsos

IJCST 19,5

368

(a) Original

Figure 4. Simulation of multi-breast forms

(b) Separate

(e) Lift

(c) Push

(f) Push-up

(d) Droop

(g) Appointed torso

generated which, respectively, represented the separated, pushed and droopy breast shape. Illustration (e) and (f) showed the lifted and push-up breast shape by changing the corresponding breast parameters and weight value. Example (c), (e) and (f) were also regarded as the aesthetic breast shapes sculpted by different brassieres. The personalized torso model as shown in picture (g) was produced by changing the trunk parameters of model (a). It was clear that the abstract and magic female breast shape library could be developed on the basis of the 3D scan data and deformation techniques. 4. Predicting method of individual bra cup dart quantity 4.1 Related work It was more and more realized that one of the greatest challenges facing apparel companies was to find a cost-effective method of providing quality fit in apparel. In the intimate apparel field, accordingly, some researchers attempted to develop pattern drafting methods to improve the fit of brassiere. Morris (2001) presented a way of tracing pattern from the well-fitted bra sample which was popularly used in the business for training new designers, because little experience and knowledge were required for this process. However, the bra was readily distorted in shape when being traced. Yu and So (2001) once launched a typical approach of direct drafting bra pattern which was to develop a 2D pattern directly from body measurements and fashion sketches by free hand drawing with the aids of straight and curve rulers.

Although direct drafting might be suitable for the business of made-to-measure, the procedures of different styles and sizes were distinct and complicated. Furthermore, there were many assumed numbers that waited to be verified. Bray (1986), Armstrong (1987), Haggar (1990) and other researchers all analyzed the courses from basic bodice block to bra structure. This drafting operation flow was simplified and regular owing to the base of antitypes, although the top and lower edges of bra block could be changed for the style. However, none of them explained the fit relations or reasons of the steps and involved quantity such as the dart quantity, distribution of darts and some fixed parameters. Bra pattern technicians admitted that cup dart quantity designing was crucial for well-fitted bra cup pattern design. Until now, there was no standard guideline or model in both academic and industrial fields. “Trial and error” method was commonly used in the pattern adaptation process, yet how the body shape and pattern geometry affect the bra fitting performance was not clearly understood. In this section, a new approach for intuitively predicting the amount of personalized bra cup dart and better understanding dart distribution in a bra cup from 3D to 2D based on the above magic breast library was presented.

An approach to predicting bra cup dart quantity 369

4.2 Principle and approach of 3D-2D cup development Using 3D CAD system, the method of breast surface unrolling was approximate since the complex surface of human body actually almost could not be spread out. The principle of garment development from 3D surface to 2D plane was to keep isometric surface. According to this rule, every part of body surface could be unfolded from 3D to 2D by conducting the overlapping and separating regions. Then the dart angle would be approximately obtained. The efficiency and precision were much better than those of the traditional estimated value by experiences and “trial and error” method. The detail was as follows. Firstly, the generated 3D body surface was divided into triangular meshes. The accuracy of triangulation could be artificially controlled. The approximate error could be mastered to a desired scope by increasing triangulation. Secondly, the obtained triangles with space construction could be unrolled along a path to be connected with many grid points under keeping isometric area. For instance, there were a series of space triangles 123, 134, 345. . . in Figure 5(a). In the beginning, 3D triangle 123 was unfolded on to the plane with the same length of each edge as shown in Figure 5(b). Then 3D triangle 134 was spread out along 13 border on the plane with the isometric line or isometric area. Finally, the rest could be done in the same manner. The choice of flattening datum line was very important. The shape and position of darts might be different along assorted paths. On the basis of the breast configuration

1

W

4

4

1

5

5 2 3 (a) Space triangles

2

3

(b) Plane triangles

Figure 5. 3D-2D transformation of body surface

IJCST 19,5

370

characteristic, it was better for results and visual comprehension to develop breast surface by flabellate mode. The fan-shaped center was on the bust point. The triangle meshes were connected into a continuous larger piece. Among triangle meshes, there were gaps which were the dart angles as shown in Figure 6. 4.3 Application results and discussions Figure 6 shows the 3D breast virtual model and the corresponding 2D unrolled cup, as well as the described interface of the bra cup dart quantity and distribution in a cup automatically reported by the system. The subject in portrayal (a) was the same as the one in Figure 4(e). From Figure 6(b), it could be seen clearly that the cup was divided into 18 flabellate segments. Each sector was 208. Take AC segment as an example. The triangle meshes were connected along the datum line A to bust point. Then BC crevice was considered as the dart in this sector. The line connected with each other by every node till to the bust point was dart trace. Obviously, dart shapes were curvy. It was reported that the total amount of bra cup dart was 80.4918218 in chart (c). Other detailed data consisting of 18 rows (18 sectors) and 11 columns (11 nodes) were dart angles. Each angle was corresponding to each sector and every node in graph (b). The parameters of segments and isometric nodes could be artificially controlled and adjusted according to the requirement of dart precision and breast curve length. The negative quantities hinted that there were a little bit of overlapped areas when the

A B C

(a) 3D breast virtual model

Figure 6. Prediction of bra cup dart quantity and distribution

(b) Corresponding 2D unrolled breast cup

(c) Bra cup dart quantity information

triangle pieces were spread out. Their dart quantities were zero in practice. Furthermore, diagram (b) even provided the dart distributive information in a bra cup that darts were mainly allotted below the bust line, which novel finding was consistent with the results of systematic breast flattening experiments. According to the program flow chart of Figure 7, predicting the amount of individual bra cup dart in the 3D virtual environment could be realized.

An approach to predicting bra cup dart quantity 371

5. Conclusions A novel approach of personalized bra cup dart quantization in the 3D virtual environment was developed. During the investigating courses, a hybrid method of FFD model based on Delaunay and Dirichlet/Voronoi diagrams and weight concept was employed. Meanwhile the interactive multi-layered deformations with intercrossed control points were designed to enhance the realistic and desired female breast configuration effect. The realization of the virtual magic female breast shape library provided a favorable 3D environment for bra pattern design. On the basis of the principles of isometric area and flabellate segments, the 3D breast surface was

Realistic 3D naked breast

Yes Breast confirmation No Input deformation parameters

Multi-layered model

Desired breast shape

No Breast confirmation Yes

Bra cup 3D - 2D transformation

Bra cup dart quantization

Figure 7. Program pipeline

IJCST 19,5

372

successfully transformed into a plane. Consequently, the personalized the amount of bra cup dart and dart distribution were predicted intuitively and efficiently. The results provided a theoretic foundation for bra pattern design and made it possible to meet the customer demand of made-to-measure well-fitted bra efficiently.

References Armstrong, H.J. (1987), Patternmaking for Fashion Design, Harper & Row, New York, NY. Bras (2004), “Brace yourselves”, Bras, September, pp. 16-19. Bray, N. (1986), More Dress Pattern Designing, Collins Professional and Technical Books, London. Coquillart, S. (1990), “Extended free-form deformation: a sculpting tools for 3D geometric modeling”, Proceedings Siggraph’90. Dobson, W.L. (1885), Twenty Pound Prize Essay on Cutting by Block Patterns. Farin, G. (1990), “Surface over Dirichlet tessellations”, Computer Aided Geometric Design, Vol. 7, pp. 281-92. Fetter, W.A. (1982), “A progression of human figures simulated by computer graphics”, IEEE Computer Graphics and Applications, Vol. 2, pp. 9-13. Haggar, A. (1990), Pattern Cutting for Lingerie, Beachwear and Leisurewear, BSP Professional Books, Oxford. Hsu, W., Hugues, J.F. and Kaufman, H. (1992), “Direct manipulation of free-form deformations”, Proceedings Siggraph’92. Kahler, K., Haber, J. and Seidel, H.P. (2003), “Reanimating the dead: reconstruction of expressive faces from skull data”, ACM TOG (SIGGRAPH Conference Proceedings). Kurokawa, T. (1997), “Measurement and description of human body shape and their applications”, Journal of the Society of Instrument and Control Engineers, Vol. 36 No. 2, pp. 77-83. Lamousin, H.J. and Waggenspack, W.N. (1994), “Nurbs-based free-form deformations”, IEEE Computer Graphics and Applications, Vol. 14 No. 6, pp. 59-65. Magnenat-Thalmann, N., Laperriere, R. and Thalmann, D. (1988), “Joint-dependent local deformations for hand animation and object grasping”, Proceedings Graphics Interface’88, pp. 26-33. Morris, K. (2001), Sewing Lingerie that Fits, The Taunton Press, Newtown, CT. (The) Oxford Dictionary (2002), Oxford University Press, Oxford. Piegl, L. and Tiller, W. (1997), The NURBS Book, Springer, New York, NY. Rasband, J. (1998), Fabulous Fit, Fairchild Publications, New York, NY. Sederberg, T.W. and Parry, S.R. (1986), “Free-form deformation of solid geometric models”, Proceedings Siggraph’86. Shen, J. and Thalmann, D. (1995), “Interactive shape design using metaballs and splines”, Proceedings of Eurographics Workshop on Implicit Surfaces 95, Grenoble, France. Spiegel, M. and Sebesta, L. (2002), The Breast Book, Workman Publishing, New York, NY. Stylios, G.K., Han, F. and Wan, T.R. (2001), “A remote online 3-D human measurement and reconstruction approach for virtual wearer trials in global retailing”, International Journal of Clothing Science & Technology, Vol. 13 No. 1, pp. 65-75.

Sukumar, N. (2001), “Sibson and non-sibsonian interpolants for elliptic partial differential equations”, paper presented at First MIT Conference on Computational Fluid and Solid Mechanics. Xu, B., Huang, Y., Yu, W. and Chen, T. (2002), “Body scanning and modeling for custom fit garments”, Journal of Textile and Apparel Technology and Management, Vol. 2 No. 2, pp. 1-11. Yesil, M.S. (2003), “Realistic rendering of a multi-layered human body model”, MS thesis, The Institute of Engineering and Science of Bilkent University, Ankara. Yoshimoto, S. (1992), “Ballerinas generated by a personal computer”, The Journal of Visualization and Computer Animation, Vol. 3, pp. 85-90. Yu, W. and So, Y.K. (2001), “Special techniques for making push-up bra”, ATA Journal, Vol. 12 No. 3, pp. 69-70. Yu, Z.Y., Zheng, S.S., Chen, L.T., He, X.Q. and Wang, J.J. (2005), “Dynamic concision for three-dimensional reconstruction of human organ built with virtual reality modelling language (VRML)”, Journal of Zhejiang University SCIENCE, Vol. 6 No. 7, pp. 611-6. Zeltzer, D. (1982), “Motor control techniques for figure animation”, IEEE Computer Graphics and Applications, Vol. 2, pp. 53-60. Corresponding author Jian Ping Wang can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

An approach to predicting bra cup dart quantity 373

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0955-6222.htm

IJCST 19,5

COMMUNICATIONS

Comparison of body shape between USA and Korean women

374

Jeong Yim Lee Paichai University, Daejeon, South Korea

Cynthia L. Istook North Carolina State University, Raleigh, North Carolina, USA, and

Yun Ja Nam and Sun Mi Park Seoul National University, Seoul, South Korea Abstract Purpose – The purpose of this paper is to compare body shape between USA and Korean women. It aims to analyze the distribution and proportion of body shapes of two countries and compare the differences of body shape according to age. Design/methodology/approach – SizeUSA and SizeKorea measurement data were evaluated using the Female Figure Identification Technique for apparel system developed at North Carolina State University. Once the samples were defined by shape, comparisons were made of the distribution according to age and country through statistical analysis. Findings – The paper finds that the largest shape category was the rectangle shape in both countries, but the distribution within each shape category for Korean women was different from that of USA women. More body shape categories were found in the USA women than in Korean women. In addition, most body shape categories had different body proportions when comparing the USA women and Korean women. The USA women had the higher measurements in the waist, high hip, and hips height and the larger measurements in the bust, waist, high hip, and hips circumference. Research limitations/implications – Of the over 6,300 US female subjects in this study, only five failed to be identified by the seven shapes identified. These subjects had over 50.2 in. of hip circumference, over 10 in. larger hips than bust circumference, and over 15.5 in. larger hips than waist circumference. Further refinement of the mathematical definitions or a second group of criteria may be required for sorting the women that have no shape as defined by this study. Originality/value – The opportunity to compare the body shapes between two very different countries, using national anthropometric survey data, is very rare, indeed. This comparison allows the opportunity to discover ways to improve the sizing systems of each country, as well as impact the development of international sizing standards that could have a significant impact on brands producing product for a variety of international consumers. Keywords Women, Measurement characteristics, United States of America, South Korea Paper type Research paper

International Journal of Clothing Science and Technology Vol. 19 No. 5, 2007 pp. 374-391 q Emerald Group Publishing Limited 0955-6222 DOI 10.1108/09556220710819555

This work was supported by the KATS, Ministry of Commerce, Industry and Energy, Under Size Korea and by (TC2), located in Cary, NC, the non-profit organization developed to serve the sewn product industries through technology development and supply chain improvement that conducted the SizeUSA National Sizing Survey.

Introduction Most people want clothing that fits well. Although each person’s definition of that may well be subjective, the satisfaction with clothing fit will be higher if the body type of the wearer can be considered when the clothing is designed and manufactured. New developments within the apparel industry can now enable the customization of clothing fit through automatic measuring with 3D body scanning, in addition to enabling a better understanding of the adjustments that need to be made that might allow better fit for a mass market. The demand for manufacturing clothing that considers the traits of various body types is getting higher. Body types differ between races. The body sizes, shapes, and proportions of each are different. There has been a great deal of research regarding fit preferences (LaBat and Delong, 1990; Alexander et al., 2005), or satisfaction with clothing and body cathexis among races or countries (Rabolt et al., 2004; Newcomb and Istook, 2005). But, there has been little research conducted that compared the body type among races. The major factor that has an influence on fit and satisfaction with clothing is body type. It is very important to find the characteristics of each body type among races or countries. As the economic, social, and cultural interchanges among countries increase, many global apparel brands like Banana Republic, Gap, and Crew are becoming more popular in many countries. Body type comparisons between countries allow the opportunity to discover ways to improve the sizing systems of each country, as well as impact the development of international sizing standards that could have a significant impact on brands producing product for a variety of international consumers that have different sizes and shapes. Simple mean comparison of body size between countries is not meaningful to understand the body size and shapes because there are various body types within each country. In this study, we compared the body shapes between USA and Korean women. The distribution and traits of body shapes between two countries were analyzed and the differences according to age were compared. The body proportion of each body type between two countries was compared through statistical analysis. Purpose of the study The purpose of this study was to compare the body shape and body proportion between USA women and Korean women by analyzing the distribution and traits of body shapes and comparing the differences of body shape according to age. Review of literature The national anthropometric survey in the USA Historically, almost all US sizing surveys conducted were for the US military for the purpose of sizing equipment and military apparel. In the mid-1990s, a project was processed to provide civilian measurement information, in addition to military information using a full body 3D scanner. This project, known as Civilian American and European Surface Anthropometry Resource (CAESAR), was put together jointly by the Society of Automotive Engineers and the US Air Force, and was designed around a new 3D laser body scanning technology. A Cyberware scanner was used in the US States to gather data and a Techmath (now known as Human Solutions) scanner was used in Europe. In addition to the scanned measurements that were taken, quite a number of physical measurements were also taken. This study was designed to

Comparison of body shape

375

IJCST 19,5

376

generate information that could serve multiple industries including automobile, aerospace, furniture, apparel, and others (TC2, 2004). The SizeUSA project, completed in September 2003, was anthropometric research developed to gather US sizing data that would be representative of the US adult population (based on an NHANES sampling protocol) with the use of the TC2 3D body scanner, which uses white light to define the 3D image of a subject. The image data were used in combination with the body measurement system measurement extraction software to obtain more than 130 measurements for each subject (TC2, 2006). Over 10,000 subjects were scanned, evaluated and surveyed during the research project. About 65 percent of the subjects scanned were women and 35 percent men. The objective of the SizeUSA National Sizing Survey was to measure the body dimensions of a representative sample of the US population, so the variety of age, ethnicity, gender, and geographic area was considered when determining the sampling methodology (TC2, 2004). The national anthropometric survey in Korea A National Anthropometric Survey in Korea has been carried out every 5-6 years since 1979 and size and shape information about human bodies have been provided to the various industrial areas. The first National Anthropometric Survey in Korea was conducted in 1979. The 18,013 Koreans from 0 years to 45 years were measured and there were 117 total measurements. With the result of first survey, a standard for 46 kinds of industrial products such as clothing, shoes and school instruments was developed. The second anthropometric survey was in 1986 and 21,648 Koreans from 0 years to 51 years were measured on 80 items. The third survey was in 1992 and 8,886 Korean from 6 years to 50 years were measured on 84 items on the basis of the Korean Standard KS A 7003 and KS A 7004. With the result of third survey, 41 standards were developed or revised. The fourth survey of 13,062 Koreans from 0 years to 70 years was measured in 1997. About 120 items were selected by KS A 7003 and measured by KS A 7004. For the first time, the elderly people from 50 years to 70 years were measured. Based on the results of the four national anthropometric surveys, more accurate body data of Koreans were provided to industries and more suitable clothing, shoes, furniture and the other equipment were manufactured Korean Agency for Technology and Standards (KATS, 2006). The demands of industry are varied and constantly changing making it difficult to meet the needs of all from the four previous studies. It became important to find a way to collect body information that could provide data to meet the wide variety of needs for different industries. The SizeKorea project was developed as the fifth National Anthropometric Survey to satisfy the changing demands of each industrial area. This survey was conducted from March 2003 to November 2004, and included not only the traditional physical anthropometric measuring techniques, but also 3D body scanning and body movement. The 19,700 Koreans from 0 years to 90 years were measured on 119 traditional items, 35 action items and 205 3D items. The purpose of SizeKorea was to: . gather size and shape information of the Korean body which corresponded to international criteria; . obtain 3D Korean body data which has importance to many industrial areas; and . build a database of body measurements for the development and revision of industrial products and system standards (KATS, 2004).

Body type classification in the US Sizing system The body types of the apparel sizing systems in USA have been classified by body proportions as might be influenced by age (Glock and Kunz, 2005). Commonly used size ranges classified by body types for women are Junior, Petite, Misses, Women, and Plus sizes. Junior sizes were designed for the more youthful (less well developed), short, and slim figure. The torso is cut shorter and the bustline higher than in Misses. They are 1.62-1.67 m (5.4-5.6 in.) in height. Sizes are odd numbers and range from size 3 to 15. The Petite size range has the general circumferential proportions of the Misses sizes, but their height ranges from 1.52 to 1.62 m (5-5.4 in.) and the sizes are even numbers from 2 to 14. Misses sizes were designed for the woman who has average proportions and height. They have a longer back waist length than Juniors and a more mature female figure with a height from 1.62 to 1.73 m (5.4-5.8 in.). Sizes are even numbers and range from 2 to 20. Women’s sizes generally range from 14 to 24. The torso has larger proportions and a fuller figure than Misses. Height ranges from 1.62 to 1.75 m (5.4-5.9 in.). Plus Sizes are larger figure types corresponding to Misses 16 and larger. The Height is over 1.62 m (5.4 in.) and sizes range from 12 to 32 (Stamper et al., 1991; Ashdown, 1998; Glock and Kunz, 2005). ASTM (formerly the American Society for Testing and Materials, 2001) updated the standard for Misses sizes, designation D5585-95, in 2001. All of the standards related to body sizing are slated for update every five years. This timeline has no relationship to measurement surveys that have been conducted. Generally, any changes to the standards have been based on industry (military, apparel manufacturers, and retailers) practice, rather than representative measurement data. Body type classification in Korean Sizing system Body type has been the fundamental basis of sizing systems in Korea. KATS suggested the apparel sizing system according to body type based on the ISO sizing system in 1998 and improved it in 2004. The sizing system for women’s apparel includes three basic body types. The upper body can be classified into an A, H, or N type, according to the difference between Hips circumference and Bust circumference. The A type has a small bust and large hip, with a difference between Hips and Bust ranging from 9 to 21 cm (3.54-8.27 in.). The H type has large bust and hips, and the difference between Hips and Bust ranges from 2 14 to 2 3 cm (2 5.51 to 2 1.18 in.). The N type is the average type with a difference ranging from 3 to 9 cm (1.18-3.54 in.). The lower body is classified into three types, as well: the average type, the small waist type, and the large waist type. The criterion for classification is the difference between hips circumference and waist circumference. The average type has the difference between hips and waist from 14 to 22 cm (5.51-8.66 in.) and the small waist type with a difference from 22 to 38 cm (8.66-14.96 in.). The difference range of the large waist type is from 2 4 to 2 14 cm (2 1.57 to 2 5.51 in.) (Korean Standard Association, 2004). FFIT software The software Female Figure Identification Technique (FFIT) for Apparelq was developed as a starting point for the representation of female figure shapes in a mathematical way (Simmons, 2002). The software was developed as a part of a doctoral dissertation, which aimed at creating “a methodology for characterization of body types/forms that would more appropriately replicate the diverse shapes of the American

Comparison of body shape

377

IJCST 19,5

378

population” (Simmons, 2002). The aim of the research was “to utilize software that can take 3D data and ‘sort’ it into congruous and related shape categories (body or shape sort) based on measurements, proportion, and shape” (Simmons, 2002, p. 3). The study resulted in development of preliminary subgroups for the female population that may aid in better fit of clothing, after careful review of all previous studies done on somatotyping, body typing (no study was conducted before FFIT for Apparelq using 3D bodyscan data) and by expert knowledge of researchers who worked in the apparel field for more than thirty years. Six body measurements taken from the body scan data of women were used to predict the body shapes of women. Further research was conducted to validate the FFIT methodology. Discriminant analysis was performed to assess the classification accuracy of the FFITq software and compare it to the Discriminant function developed from the results of visual analysis. The classification accuracy of the Discriminant function developed from the training data set is 66.01 percent for the test dataset, while the classification accuracy of the FFITq software is 89.27 percent for the test dataset. The predictive validity of the Discriminant function is higher by itself and the performance of the FFITq software is better than the Discriminant function (Devarajan and Istook, 2004). An additional objective of this research was to identify whether the nine shape groups are statistically different. The results obtained from the MANOVA analysis found that the nine shape groups proposed in the software are significantly different and the five variables used by the software significantly influence the group membership (body shape). The software could be used on a larger random sample set, representing the general US population and expected to produce good classification accuracies (Devarajan and Istook, 2004). Method Data The 3D data of 6,310 women of SizeUSA and 1,799 women of SizeKorea were analyzed for comparing body shapes. The 3D data of SizeUSA was measured with the scanner of TC2 from 2002 to 2003, and the subjects were women over 18 years of age. The 3D data of SizeKorea was measured with the Cyberware scanner in 2003 and the subjects were women over 18 years of age. The Body Measurement Software (BMS) of TC2 was utilized for extracting measurements from the two 3D data sets and the six measurements that were needed to sort and compare the shapes were extracted. Analysis instrument Shape sorting and shape category information. The criteria for shape sorting was the FFIT for Apparelq discussed previously (Simmons et al., 2004a, b). The reliability of the criteria was scrutinized through preliminary analysis to determine the criteria are effective for sorting not only the USA women but also Korean women. The criteria were defined by mathematical analysis and visual check. Body shapes were sorted with the mathematical description by FFITq for Apparel and visually verified individually to determine that the shape designation given by FFITq was correct. About seven of the nine body shape categories were adapted to classify the shape of subjects (the Oval and Diamond shapes were not used for this study due to the lack of availability of measurement data from SizeUSA). The seven shapes were the hourglass, spoon, bottom hourglass, top hourglass, inverted triangle, triangle, and rectangle shapes.

Hourglass. The hourglass category and the rectangle category were the basis from which many of the other categories were created. The body measurements used to define the hourglass shape were the bust, waist, and hips. The underlying criteria of the hourglass shape says that if a subject has a very small difference in the comparison of the circumferences of the bust and hips AND if the ratios of bust-to-waist and hips-to-waist are about equal and significant, then the shape will be defined as an hourglass. The person with an hourglass shape has the appearance of being proportional in the bust and hips with a defined waistline. The mathematical formula defined for hourglass shape is: If (bust-hips)< ¼ 1 Then If (hips-bust)<3.6 Then If (bust-waist)> ¼ 9 Or (hips-waist)> ¼ 10 Then shape ¼ “Hourglass” Bottom hourglass. This shape category is a subset of the hourglass category. This shape was determined by utilizing the same body measurements of the bust, waist, and hips, as in the hourglass. The bottom hourglass category utilizes the underlying criteria that if a subject has a larger hips circumference than bust circumference AND if the ratios of the bust-to-waist and hips-to-waist are significant enough to produce a definite waistline, then the shape will be defined as a bottom hourglass. The mathematical formula defined for bottom hourglass category is: If (hips-bust)> ¼ 3.6 And (hips-bust)<10 Then If (hips-waist)> ¼ 9 Then If (highhip/waist)<1.193 Then shape ¼ “Bottom Hourglass” Top hourglass. This shape category is also a subset of the hourglass category. The underlying criteria for the top hourglass shape says that if a subject has a larger bust circumference than hips circumference and if the ratios of their bust-to-waist and hips-to-waist measurements are significant enough to produce a definite waistline, then the shape will be defined as a top hourglass. The mathematical formula defined for top hourglass shape is: If (bust-hips)>1 And (bust-hips)<10 Then If (bust-waist)> ¼ 9 Then shape ¼ “Top Hourglass” Spoon. The shape category of Spoon was determined by utilizing the body measurements of the bust, waist, hips, and high hip. The Spoon shape category utilizes the underlying criteria that if a subject has a larger circumferential difference in their hips and bust and if their bust-to-waist ratios is lower than the hourglass shape and the high hip to waist ratio is great, then that shape will be defined as a Spoon. The mathematical formula defined for Spoon shape is: If (hips-bust)>2 Then If (hips-waist)> ¼ 7 Then If (highhip/waist)> ¼ 1.193 Then shape ¼ “Spoon”

Comparison of body shape

379

IJCST 19,5

380

Triangle. The shape category of triangle was determined by utilizing the body measurements of the bust, waist, and hips. The triangle shape category utilizes the underlying criteria that if a subject has a larger hips circumference than their bust and if the ratio of their hips-to-waist is small, then the subject can be identified as having a triangle shape. The person with a triangle shape has the appearance of being larger in the hips than the bust without having a defined waistline. This shape differs from the bottom hourglass because the triangle does not consider the bust-to-waist ratio where the bottom hourglass does. The mathematical formula defined for the triangle shape is: If (hips-bust)> ¼ 3.6 Then If (hips-waist)<9 Then shape ¼ “triangle” Inverted triangle. The shape category of inverted triangle was determined by utilizing the same body measurements of the bust, waist, and hips as in the triangle. The inverted triangle shape category utilizes the underlying criteria that if a subject has a larger bust circumference than their hips and if the ratio of their bust-to-waist is small, then it will fall into the shape category of an inverted triangle. The person with an inverted triangle shape has the appearance of being heavy in the bust as compared to the hips but not having a defined waistline. This shape differs from the top hourglass because the inverted triangle does not consider the hips-to-waist ratio where the top hourglass does. The mathematical formula defined for inverted triangle category is: If (bust-hips)> ¼ 3.6 Then If (bust-waist)<9 Then shape ¼ “Inverted Triangle” Rectangle. The rectangle category was determined by utilizing the bust, waist, and hips circumferences. The underlying criteria for this shape is that if the bust and hips circumferences are fairly equal and bust-to-waist and hips-to-waist ratios are low, then the shape will be defined as a rectangle. The person with a rectangle shape is characterized by not having a clearly discernible waistline. Therefore, the bust, waist, and hips are more inline with each other. The mathematical formula defined for rectangle shape is: If (hips-bust)<3.6 And (bust-hips)<3.6 Then If (bust-waist)<9 And (hips-waist)<10 Then shape ¼ “Rectangle” Results Comparison of distribution of shape categories between USA and Korean women The distribution of shape categories between USA (Table I) and Korean (Table II) women was compared. The distribution according to age group was analyzed and the most frequent shape of each country was determined. Hourglass. The hourglass shape was the third largest shape category of the USA women. It had the distribution with 11.8 percent of the US subjects and getting older, it showed lower distribution (Table I). In Korean women, the hourglass shape was the fifth largest shape category. The distribution of the hourglass shape category was under 0.5 percent and the hourglass shape appeared just under 35-years old group of Korean women (Table II). Figure 1 is an example of a true hourglass shape. She has the appearance of

Hourglass Spoon Bottom hourglass Top hourglass Inverted triangle Triangle Rectangle Not classified Total

(3.8) (3.0) (3.8) (0.6) (0) (1.3) (11.9)

1,537 (24.4)

239 191 237 40 2 79 749

18 , 25 years (percent) (3.3) (4.9) (2.5) (0.7) (0.1) (1.0) (10.4)

1,448 (22.9)

206 311 160 43 8 64 656

26 , 35 years (percent) 154 337 105 51 7 55 631 1 1,341

(2.4) (5.3) (1.7) (0.8) (0.1) (0.9) (10.0) (0) (21.3)

103 269 61 40 11 53 602 3 1,142

(1.6) (4.3) (1.0) (0.6) (0.2) (0.8) (9.5) (0.1) (18.1)

USA women 36 , 45 years 46 , 55 years (percent) (percent) 36 163 21 18 6 34 327 1 606

(0.6) (2.6) (0.3) (0.3) (0.1) (0.5) (5.2) (0) (9.6)

56 , 65 years (percent)

236 (3.7)

7 (0.1) 84 (1.3) 5 (0.1) – – 16 (0.3) 124 (2.0)

66 , years (percent)

745 1,355 589 192 34 301 3,089 5 6,310

(11.8) (21.5) (9.4) (3.0) (0.5) (4.8) (49.0) (0.1) (100)

Total

Comparison of body shape

381

Table I. Distribution of body shape categories in the USA women

Table II. Distribution of body shape categories in Korean women 2 (0.1) 57 (3.2) 20 (1.1) – – 104 (5.8) 295 (16.4) 478 (26.6)

7 (0.4) 81 (4.5)

57 (3.2) –

– 101 (5.6) 186 (10.3) 432 (24.0)

26 , 35 years (percent)

– 49 (2.7) 261 (14.5) 327 (18.2)

–

2 (0.1)

– 15 (0.8)

3 17 225 247

(0.2) (1.0) (12.5) (13.7)

–

–

– 2 (0.1)

Korean women 36 , 45 years 46 , 55 years (percent) (percent)

– 4 (0.2) 171 (9.5) 177 (9.8)

–

–

– 2 (0.1)

56 , 65 years (percent)

– 6 (0.3) 132 (7.3) 138 (7.7)

–

–

– –

66 , years (percent)

382

Hourglass Spoon Bottom hourglass Top hourglass Inverted triangle Triangle Rectangle Total

18 , 25 years (percent)

3 (0.2) 281 (15.6) 1,270 (70.6) 1,799 (100)

0 (0.0)

79 (4.4)

9 (0.5) 157 (8.7)

Total (percent)

IJCST 19,5

Comparison of body shape

Hourglass

383

Korea

USA

Figure 1. The example of the hourglass shape

being proportional in the bust and hips with a defined waistline. The ratios of her bust-to-waist and hips-to-waist are about equal. She has a back line being proportional in the lateral upper and lower body. Bottom hourglass. The bottom hourglass shape was the fourth largest shape category with 9.4 percent of the US subjects. It showed lower distribution as subjects aged like the hourglass shape category (Table I). In Korean women, the bottom hourglass shape distributed with 4.4 percent of subjects and was the fourth largest shape category. The distribution of bottom hourglass appeared just under 45-years old group (Table II). Figure 2 is an example of a true bottom hourglass shape. She has the circumferential measurements that meet the bottom hourglass shape criteria. The ratios of her bust-to-waist and hips-to-waist are significant with her hips measurements being slightly larger than her bust. She also has a defined waistline. Bottom Hourglass

Korea

USA

Figure 2. The example of the bottom hourglass shape

IJCST 19,5

384

Top hourglass. The top hourglass shape was the sixth largest shape category of the USA women. It did not appear in the over 66-years old group. It showed larger distribution under 55-years old group than over 56-years old group of the USA women (Table I). The top hourglass shape did not appear in Korean women (Table II). Figure 3 is an example of the top hourglass shape. Spoon. The Spoon shape was the second largest shape category of the USA women and it showed the distribution with 21.5 percent of the subjects. The 14.5 of 21.5 percent Spoon subjects was in 26 , 55 years old group of the USA women (Table I). The Spoon was the third largest shape category of Korean women with 8.7 percent of the subjects. The distribution of Spoon shape was different according to age. The Spoon shape showed the larger distribution under 45-years old than over 45-years old group of Korean women (Table II). In Figure 4, the subject has the circumferential measurements that meet the Spoon shape criteria and is an example of a true Spoon shape. She has a definite waistline and there is a distinct shelf that protrudes from the hip area in the front view. She has the flatter curve in her lateral body or at least in her upper back. Her ratio of upper body-to-lower body in lateral shape is larger than that of the hourglass shape and usually the back waist point of the Spoon shape is lower than that of the hourglass shape. Triangle. The triangle was the fifth largest shape category of the USA women with 4.8 percent of the subjects. It showed lower distribution when getting older like hourglass and bottom hourglass shape category (Table I). In Korean women, the triangle shape was the second largest shape category. The body shape of 15.6 percent subjects met the definition for triangle shape. The triangle shape category was seen over all age groups but frequently in younger age group than in older group of Korean women (Table II). In Figure 5, the subject has the circumferential measurements that meet the triangle shape criteria and is an example of a true triangle shape. She has the appearance of being larger in the hips than the bust without having a defined waistline.

Top Hourglass

Figure 3. The example of the top hourglass shape

USA

Comparison of body shape

Spoon

385

Korea

USA

Figure 4. The example of the Spoon shape

Triangle

Korea

USA

Some of the triangle shape category have protruding abdomen and they also looked like the triangle in the lateral shape. Inverted triangle. The inverted triangle shape category showed the least distribution with 0.5 percent of subjects of the USA women. The inverted triangle and top hourglass shape category did not appear over 66-years old group of the USA women (Table I). In Korean women, the distribution of the inverted triangle shape category was just 3 subjects (0.2 percent) and it was the sixth largest shape category. The inverted triangle shape appeared just in 46 , 55 years old group of Korean women (Table II). Figure 6 has the circumferential measurements that meet the inverted triangle shape criteria and is an example of a true inverted triangle shape. She has the

Figure 5. The example of the triangle shape

IJCST 19,5

Inverted Triangle

386

Figure 6. The example of the inverted triangle shape

Korea

USA

appearance of being heavy in the bust as compared to the hips but not having a defined waistline. Rectangle. The rectangle shape was the largest shape category of the USA women. The 49.0 percent of subjects of the USA women met the definition for rectangle shape (Table I). In Korean women, the body shape of 70.6 percent subjects met the definition for rectangle shape previously defined. The rectangle shape distributed over all age groups and it was also the most frequent category in each age group of Korean women (Table II). The rectangle shape is the largest shape category in both countries, so it includes several sub-shapes. Some subjects of the rectangle shape category have the bust, waist, and hips that are relatively inline with each other in both front and lateral shape. Other subjects of the rectangle shape category have a discernible waistline in the front shape but the flatter backline in the lateral shape. But the common feature among them is that her bust and hips circumferences are fairly equal and bust-to-waist and hip-to-waist ratios are low. Figure 7 has the circumferential measurements that meet the rectangle shape criteria and is an example of a true rectangle shape. Five subjects of the USA women were not classified into any group. They had over 50.2 in. of hips circumference, over 10 in. larger hips than bust circumference, and over 15.5 in. larger hips than waist circumference. It implicates the possibility that it is difficult to define the body shape of the women that have large hips, especially if her hips circumference is much bigger than her bust and waist circumference. Comparison of body proportion between USA and Korean women The body proportion for each shape category between USA women and Korean women was also compared. The ratios of bust height to cervical height, waist height to cervical height, high-hip height to cervical height, hips height to cervical height, bust circumference to cervical height, waist circumference to cervical height, high-hip circumference to cervical height, and hips circumference to cervical height were utilized for analyzing body proportion. Table III demonstrates the comparison of body proportion.

Comparison of body shape

Rectangle

387

Korea

USA

Hourglass. Korean women had the higher proportion of bust height to cervical height and the lower proportion of hips height to cervical height than the USA women. There were no significant differences in the proportion of waist height and high-hip height to cervical height between two countries. The bust, waist, high hip, and hips circumferences of the USA women were greater than that of Korean women. Bottom hourglass, spoon, and triangle. The proportion of waist, high hip, and hips heights of the USA women were greater than those of Korean women. The proportion of the bust height of the USA women was less than that of Korean women. The bust, waist, high hip, and hips circumferences to cervical height of the USA women were greater than those of Korean women. Inverted triangle. Korean women had less proportion of waist and high-hip height to cervical height than the USA women. There were no significant differences in the proportion of four circumferences to cervical height between USA and Korean women. Rectangle. There was no significant difference in the bust height proportion to cervical height between two countries. The proportion of waist height, high-hip height, hips height, and four circumferences to cervical height were greater in the USA women than in Korean women. Discussions and conclusion The body shape was determined using the FFIT for Apparelq methodology and was visually verified that the shape designation given by FFITq for Apparel was correct for both the USA women and Korean women. The mathematical description of this study was the proper criteria for sorting the body shape. The largest shape category for Korean women was the rectangle shape and the next was the triangle shape, and then spoon, bottom hourglass, hourglass, and inverted triangle shape category were followed (Table I). The top hourglass shape did not appear in Korean women. The younger age groups like 18 , 25 years and 26 , 35 years old group included various body shapes more than other age groups, but getting older,

Figure 7. The example of the rectangle shape

USA (n ¼ 301) 0.780 0.679 0.623 0.570 0.700 0.646 0.740 0.769

Bust height Waist height High-hip height Hips height Bust girth Waist girth High-hip girth Hip girth

0.819 0.704 0.640 0.558 0.611 0.469 0.552 0.659 Bottom hourglass Korea (n ¼ 79) 0.813 0.686 0.623 0.553 0.592 0.508 0.591 0.682 Triangle Korea (n ¼ 281) 0.808 0.665 0.604 0.548 0.607 0.557 0.635 0.689 t-value 2 3.094 * * 5.189 * * * 7.549 * * * 9.506 * * * 13.787 * * * 12.269 * * * 13.544 * * * 11.128 * * *

USA (n ¼ 3,089) 0.802 0.686 0.631 0.572 0.739 0.629 0.720 0.742

USA (n ¼ 34) 0.805 0.672 0.626 0.579 0.865 0.729 0.779 0.772

0.800 0.693 0.632 0.568 0.671 0.565 0.696 0.740

2 2.389 * 2 1.114 2 0.169 2.438 * 3.612 * * * 3.226 * * 3.708 * * * 2.196 * t-value 2 5.869 * * * 4.416 * * * 6.361 * * * 8.004 * * * 8.809 * * * 7.567 * * * 7.610 * * * 6.709 * * *

USA (n ¼ 1,355)

t-value 0.814 0.680 0.616 0.555 0.580 0.489 0.594 0.660 Inverted triangle Korea (n ¼ 3) 0.810 0.632 0.591 0.564 0.795 0.672 0.712 0.713 Rectangle Korea (n ¼ 1,267) 0.801 0.663 0.604 0.554 0.678 0.607 0.673 0.700

Spoon Korea (n ¼ 157)

t-value 1.089 24.914 * * * 32.437 * * * 23.781 * * * 22.527 * * * 7.428 * * * 16.301 * * * 16.945 * * *

t-value 20.547 2.808 * * 3.163 * * 1.782 1.574 1.114 1.463 1.318

210.236 * * * 9.347 * * * 11.052 * * * 9.884 * * * 15.652 * * * 12.963 * * * 14.735 * * * 12.520 * * *

t-value

Notes: Top hourglass shape was excluded because it did not appear in Korean women; the values of all measurements mean the proportions divided by cervical height; *0.01 # p , 0.05; * *0.001 # p , 0.01; and * * *p , 0.001

USA (n ¼ 589) 0.801 0.696 0.638 0.570 0.697 0.602 0.701 0.770

Bust height Waist height High-hip height Hips height Bust girth Waist girth High-hip girth Hip girth

Table III. Comparison of body proportion between USA and Korean women 0.807 0.698 0.639 0.571 0.690 0.539 0.651 0.709

Hourglass Korea (n ¼ 9)

388

Bust height Waist height High-hip height Hips height Bust girth Waist girth High-hip girth Hip girth

USA (n ¼ 745)

IJCST 19,5

the triangle and rectangle shape were the major shape categories in Korean women. In over 56-years old, most subjects belonged to the rectangle shape category. The distribution of each shape category in the USA women was different from that of Korean women. The distribution of the rectangle shape, the most frequent shape category in both countries was 70.6 percent of subjects in Korean women and 49.0 percent of subjects in the USA women. The hourglass shape category was just 0.5 percent of subjects in Korean women, but 11.8 percent of subjects in the USA women. The Spoon shape category showed the greater distribution in the USA women (21.5 percent) than in Korean women (8.7 percent). The triangle shape appeared in 15.6 percent of Korean women, but in just 4.8 percent of the USA women. The top hourglass shape that did not appear in Korean women was shown in 3.0 percent of the USA women. More various body shape categories were found in the USA women than in Korean women. Most of the body shapes had different body proportions between USA women and Korean women. The USA women had a higher proportion in waist, high hip, and hips heights except bust height and the greater proportion in bust, waist, high hip, and hips circumference. Implications, limitations and recommendation Implications Research comparing the body shape between two countries which utilizes the data of national anthropometric surveys is extremely rare. The results of this research may be used as a guideline which might help improve the sizing system of each country, as well as ISO, and also help global brands produce apparel for various international consumers that have different sizes and shapes. Limitations It is very important to know the body shape category for producing more suitable clothing for each individual. However, it is not easy to define the shape of the human body because the human body has a very complex structure and each individual has unique body characteristics. All of the Korean subjects were classified into one of the six body shape categories except the top hourglass. However, five USA women subjects were not classified into any of the seven body shape categories. Women in this group had over 50.2 in. of hip circumference, over 10 in. larger hips than bust circumference, and over 15.5 in. larger hips than waist circumference. Since, two of the shapes (Oval and Diamond) originally defined by the FFITq for Apparel methodology were not included in this study (due to lack of essential measurement data), it may well be that these women would have been classified in one of the two other shapes. The further refinement of the mathematical definitions or second step criteria may be required for sorting the women that are very large. In addition, the variety of human body shapes includes the possibility that some of the females may have body shapes that fall right between the edge of two shapes and might not be classified properly. Recommendations for future research This study focused on the classification and comparison of body shapes between USA and Korean women. The results of this study supply the framework for the

Comparison of body shape

389

IJCST 19,5

development of new sizing systems. It is vital to analyze the size of each body shape category for the development of new sizing systems connected with body shapes. The mass customization system needs a great deal of information between the body shape and garment patterns, in order to have the best success with the ultimate fit of customized garments. Further study needs to include the creation of slopers to better fit the shapes of women.

390 References Alexander, M., Connell, L.J. and Presley, A.B. (2005), “Clothing fit preferences of young female adult consumers”, International Journal of Clothing Science & Technology, Vol. 17 No. 1, pp. 52-64. Ashdown, S.P. (1998), “An investigation of the structure of sizing systems”, International Journal of Clothing Science & Technology, Vol. 10 No. 5, pp. 324-41. ASTM (2001), Standard Table of Body Measurements for Adult Female Misses Figure Type, Sizes 2-20. D5585-95, American Society for Testing and Materials, Philadelphia, PA. Devarajan, P. and Istook, C.L. (2004), “Validation of female figure identification technique (FFIT) for Apparelq Software”, Journal of Textile and Apparel, Technology and Management, Vol. 4 No. 1. Glock, R.E. and Kunz, G.I. (2005), Apparel Manufacturing, 6th ed., Prentice-Hall, Upper Saddle River, NJ. KATS (2004), The 5th Size Korea Report, Korean Agency for Technology and Standards, Seoul. KATS (2006), The National Size Survey, Korean Agency for Technology and Standards, Seoul, available at: http://sizekorea.ats.go.kr/ Korean Standard Association (2004), Sizing Systems for Female Adult’s Garments, KS K 0051-2004, Korean Standard Association, Seoul. LaBat, K.L. and Delong, M.R. (1990), “Body Cathexis and satisfaction with fit of apparel”, Clothing and Textiles Research Journal, Vol. 8 No. 2, pp. 43-8. Newcomb, B. and Istook, C.L. (2005), “Hispanic women: body shapes and sizing”, paper presented at the International Textile and Apparel Association Conference, Alexandria, VA. Rabolt, N.J., Rocha, J. and Teele, M. (2004), “Body satisfaction of Asian, African-American, and Caucasian students”, paper presented at the International Textile and Apparel Association Conference, Portland, OR. Simmons, K. (2002), “Body measurement techniques: a comparison of three-dimensional body scanning and physical anthropometric methods”, doctoral dissertation, North Carolina State University, Raleigh, NC. Simmons, K., Istook, C.L. and Devarajan, P. (2004a), “Female figure identification technique (FFIT) for apparel – part I: describing female shapes”, Journal of Textile and Apparel, Technology and Management, Vol. 4 No. 1. Simmons, K., Istook, C.L. and Devarajan, P. (2004b), “Female figure identification technique (FFIT) for apparel – part II: development of shape sorting software”, Journal of Textile and Apparel, Technology and Management, Vol. 4 No. 1. Stamper, A.A., Sharp, S.H. and Donnell, L.B. (1991), Evaluating Apparel Quality, 2nd ed., Fairchild, New York, NY. TC2 (2004), The National Sizing Survey-body Measurement and Data Analysis Reports on the US Population, TC2, Cary, NC. 2 TC (2006), “The US national size survey”, available at: www.tc2.com/what/sizeusa/index.html

Further reading Al-Haboubi, M.H. (1992), “Anthropometry for a mix of different populations”, Applied Ergonomics, Vol. 23 No. 3, pp. 191-6. Brown, P. and Rice, J. (1998), Ready-to-wear Apparel Analysis, 2nd ed., Merrill, Upper Saddle River, NJ. Chun-Yoon, J. and Jasper, C.R. (1993), “Garment-sizing: an international comparison”, International Journal of Clothing Science & Technology, Vol. 5 No. 5, pp. 28-37. Lee, J. (2001), “The study of standard body type for Korean women-at the age between 18 and 24 years old”, doctoral dissertation, Seoul National University, Seoul. Lee, S., Kim, K., Nam, Y., Noh, H., Jung, M., Choi, K. and Choi, Y. (2002), The Apparel Somatology, Kyohakyunkusa, Seoul. Nam, Y. (1991), “A study on classification of somatotype based on the lateral view of women’s upper body”, doctoral dissertation, Seoul National University, Seoul. Salusso-Deonier, C.J., Delong, M.R., Martin, F.B. and Krohn, K.R. (1985), “A multivariate method of classifying body form variation for sizing women’s apparel”, Clothing and Textile Research Journal, Vol. 4 No. 1, pp. 38-45. Tamburrino, N. (1992a), “Apparel sizing issues – part 1”, Bobbin, Vol. 33 No. 8, pp. 44-6. Tamburrino, N. (1992b), “Apparel sizing issues – part 2”, Bobbin, Vol. 33 No. 9, pp. 52-60. Winks, J.M. (1997), Clothing Sizes: International Standardization, The Textile Institute, Manchester. Workman, J.E. (1991), “Body measurement specifications for fit models as a factor in clothing size variation”, Clothing and Textiles Research Journal, Vol. 10 No. 1, pp. 31-6. Corresponding author Jeong Yim Lee can be contacted at: [email protected]

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Comparison of body shape

391

International Journal of Clothing Science and Technology

ISSN 0955-6222 Volume 19 Number 6 2007

International textile and clothing research register Editor

George K. Stylios

Access this journal online _________________________

2

Editorial advisory board __________________________

3

Editorial _________________________________________

4

In memoriam _____________________________________

6

Research register _________________________________

8

Research index by institution______________________

110

Research index by country ________________________

111

Research index by subject_________________________

112

Research index by principal investigator ___________

114

Awards for Excellence ____________________________

115

Access this journal electronically The current and past volumes of this journal are available at:

www.emeraldinsight.com/0955-6222.htm You can also search more than 150 additional Emerald journals in Emerald Fulltext (www.emeraldinsight.com) See page following contents for full details of what your access includes.

CONTENTS

www.emeraldinsight.com/ijcst.htm As a subscriber to this journal, you can benefit from instant, electronic access to this title via Emerald Fulltext. Your access includes a variety of features that increase the value of your journal subscription.

How to access this journal electronically

To benefit from electronic access to this journal, please contact [email protected] A set of login details will then be provided to you. Should you wish to access via IP, please provide these details in your e-mail. Once registration is completed, your institution will have instant access to all articles through the journal's Table of Contents page at www.emeraldinsight.com/0955-6222.htm More information about the journal is also available at www.emeraldinsight.com/ ijcst.htm Our liberal institution-wide licence allows everyone within your institution to access your journal electronically, making your subscription more cost-effective. Our web site has been designed to provide you with a comprehensive, simple system that needs only minimum administration. Access is available via IP authentication or username and password. Emerald online training services Visit www.emeraldinsight.com/training and take an Emerald online tour to help you get the most from your subscription.

Key features of Emerald electronic journals Automatic permission to make up to 25 copies of individual articles This facility can be used for training purposes, course notes, seminars etc. This only applies to articles of which Emerald owns copyright. For further details visit www.emeraldinsight.com/ copyright Online publishing and archiving As well as current volumes of the journal, you can also gain access to past volumes on the internet via Emerald Fulltext. You can browse or search these databases for relevant articles. Non-article content Material in our journals such as product information, industry trends, company news, conferences, etc. is available online and can be accessed by users. Reference linking Direct links from the journal article references to abstracts of the most influential articles cited. Where possible, this link is to the full text of the article. E-mail an article Allows users to e-mail links to relevant and interesting articles to another computer for later use, reference or printing purposes. Structured abstracts Emerald structured abstracts provide consistent, clear and informative summaries of the content of the articles, allowing faster evaluation of papers.

Additional complimentary services available

Your access includes a variety of features that add to the functionality and value of your journal subscription: Xtra resources and collections When you register your journal subscription online, you will gain access to Xtra resources for Librarians, Faculty, Authors, Researchers, Deans and Managers. In addition you can access Emerald Collections, which include case studies, book reviews, guru interviews and literature reviews. E-mail alert services These services allow you to be kept up to date with the latest additions to the journal via e-mail, as soon as new material enters the database. Further information about the services available can be found at www.emeraldinsight.com/alerts Emerald Research Connections An online meeting place for the research community where researchers present their own work and interests and seek other researchers for future projects. Register yourself or search our database of researchers at www.emeraldinsight.com/ connections

Choice of access

Electronic access to this journal is available via a number of channels. Our web site www.emeraldinsight.com is the recommended means of electronic access, as it provides fully searchable and value added access to the complete content of the journal. However, you can also access and search the article content of this journal through the following journal delivery services: EBSCOHost Electronic Journals Service ejournals.ebsco.com Informatics J-Gate www.j-gate.informindia.co.in Ingenta www.ingenta.com Minerva Electronic Online Services www.minerva.at OCLC FirstSearch www.oclc.org/firstsearch SilverLinker www.ovid.com SwetsWise www.swetswise.com

Emerald Customer Support

For customer support and technical help contact: E-mail [email protected] Web www.emeraldinsight.com/customercharter Tel +44 (0) 1274 785278 Fax +44 (0) 1274 785201

EDITORIAL ADVISORY BOARD Professor Mario De Araujo Minho University, Portugal Professor H.J. Barndt Philadelphia College of Textiles & Science, Philadelphia, USA Professor Dexiu Fan China Textile University, Shanghai, China Professor Carl A. Lawrence University of Leeds, Leeds, UK Professor Gerald A.V. Leaf Heriot-Watt University (Hon), UK Professor P. Grosberg Shankar College of Textile Technology and Fashion, Israel Professor Trevor J. Little North Carolina State University, USA

Professor David Lloyd University of Bradford, Bradford, UK Professor Masako Niwa Nara Women’s University, Nara, Japan Professor Issac Porat School of Textiles, UMIST, UK

Editorial advisory board

3

Professor Ron Postle The University of New South Wales, Australia Professor Rosham Shishoo Swedish Institute for Fibre and Polymer Research, Mo¨lndal, Sweden Professor Paul Taylor University of Newcastle, Newcastle upon Tyne, UK Professor Witold Zurek Ło´dz´ Technical University, Poland

International Journal of Clothing Science and Technology Vol. 19 No. 6, 2007 p. 3 # Emerald Group Publishing Limited 0955-6222

IJCST 19,6

4

International Journal of Clothing Science and Technology Vol. 19 No. 6, 2007 pp. 4-5 Emerald Group Publishing Limited 0955-6222

Editorial ITCRR – championing the research efforts of the community The international textile and clothing research register (ITCRR) is in its 13th year of publishing the research efforts of our community. It provides a breadth of activity in the field of textile and clothing research and it encourages participation and dissemination to those working in this discipline and further afield. Again as you will see in this new edition, textile and clothing research and practice is increasing in volume, in quality and in diversity – all good news for all of us involved in it. Research, development and innovation can, without doubt, give us more wisdom, enable our industries to become more competitive, and contribute to our quality of life. I believe that registering research projects will provide the due credit to originators of the research and contribute much more to the future development of this field. Groups of expertise can be identified in this manner, repetition and re-invention can be avoided leading to best utilisation of time and funding for faster and better directed research in the international arena. Important since globalisation is on everybody’s agenda. The ITCRR was set up with all these things in mind. Textiles and clothing originate from the physiological need to protect ourselves from the environment. This has made necessary the art of hand knitting and weaving, and cut and sew processes which have been evolving for many centuries. Although, the original need for clothing has somewhat changed, the mechanisation of this process started after the industrial revolution and has continued this century with automation developments on a massive scale. If you consider the upstream part of the whole textile and clothing production chain, yarn-making is the most highly automated area, followed by fabric, with high speed knitting and weaving machines. The downstream part of garment making, however, still remains probably the less developed connection in this chain; one that no doubt many of us have our eyes on as the candidate for development into the new century. With massive computerisation over the last 20 years, logistics and sales have also changed dramatically from pen and paper to electronic data interchange and electronic point of sale. New challenges are already upon us with nanotextiles, nanofibres, nanocoatings with multifunctional and smart textiles and clothing, and with wearable electronics. A special issue on smart textiles and clothing is at an advanced stage of development and we will continue to welcome contributions from textile and clothing aesthetics, design and fashion highlighting our belief that design and technology go hand to hand. We predict that more exciting projects will come from this synergy. Consistent and extensive research and development in textile and clothing design and science and technology underpin all these developments by the international research community whether in educational establishments, in research trade organisations, or in companies. IJCST was been set up 19 years ago as a platform for the promotion of scientific and technical research at an international level. The manufacture of clothing in particular needs to change to more technologically advanced forms of manufacturing and retailing – IJCST continues to support the community in these and other efforts. q Professor George K. Stylios.

The journal, now fully indexed in SCI, continues with its authoritative style to accredit original technical research and by adhering to our refereeing processes however difficult these may prove at times. IJCST will be instrumental in continuing to support conferences and meetings from around the world in its effort to promote the science and technology of clothing. The 2007 volume saw the publication of a double issue from “3rd International Textile, Clothing and Design Conference” Dubrovnik, October 2006. I finally praise the enthusiasm of our research community and those authors that have made IJCST an invaluable resource to all involved with textiles and clothing. I thank our editorial board for their continuous support and our colleagues who have acted in a refereeing capacity and have given us their free time and expertise to progress our research efforts. I take the liberty of listing some of those names below (apologies in advance if anyone has accidentally been omitted from this list): . Professor Paul Taylor, University of Newcastle . Professor Haruki Imaoka, Nara Women’s University . Professor Jachym Novak, Vysoka Skola Sronjni a Tectilffi . Professor Isaac Porat, UMIST . Professor Ron Postle, The University of New South Wales . Dr Taoruan Wan, University of Bradford . Professor David Lloyd, University of Bradford . Professor G.A.V. Leaf, Heriot-Watt University . Dr David Brook, University of Leeds . Dr Jaffer Amirbayat, UMIST . Dr David Tyler, Manchester Metropolitan University . Professor Jintu Fan, Hong Kong Polytechnic University . Dr Jelka Gersak, University of Maribor . Dr Hua Lin, Heriot Watt University . Dr Sharon Lam Po Tang, Heriot Watt University . Dr Lisa McIntyre, Heriot Watt University . Dr D Sun, Heriot Watt University Thank you all subscribers, authors, editorial board members, referees, publishing team, colleagues, students for your support and especially Dr S Lam Po Tang for her contribution. For further information, please contact: Heriot-Watt University, School of Textiles, Netherdale, Galashiels, Selkirkshire, TD1 3HF, Scotland, UK. E-mail: [email protected]; [email protected] George K. Stylios Editor-in-Chief

Editorial

5

IJCST 19,6

In memoriam Professor Blazˇ KNEZ, PhD 1928-2007

6

International Journal of Clothing Science and Technology Vol. 19 No. 6, 2007 pp. 6-7 q Emerald Group Publishing Limited 0955-6222

Professor Blazˇ Knez, PhD, from the Faculty of Textile Technology, University of Zagreb, Croatia, a long-time member of the International Journal of Clothing Science and Technology Editorial Advisory Board, sadly died on 17 June 2007 after a long fight with illness. He was 79. Born in Libelicˇe, Slovenia in 1928 Professor Knez gained his bachelor’s degree at the Polytechnic of Belgrade and continued his graduate education at the Faculty of Technology and Metallurgy in Belgrade, graduating in 1968. From 1952 to 1959 he attended numerous specialisation courses in Germany and the USA and this rich experience enabled him to become chief engineer of the biggest Yugoslav clothing factories (BEKO Belgrade and Kamensko, Zagreb). After experience in industrial practice he moved into the education of experts in all the branches of garment technology and engineering. From 1973 to 1976 he was a Professor and Dean of the Professional College of Garment Technology in Zagreb, and from 1977 to retirement in 1990 a Professor at the Faculty of Technology and the Director of the Institute of Textile and Clothing, a part of that Faculty. He obtained his PhD from the Faculty of Technology, University of Zagreb, Croatia in 1980. To promote higher education in the field of garment engineering together with his collaborators he founded the Department of Clothing Technology and Engineering. This facility has two research laboratories and is used as a centre for the university education of experts in garment technology in Croatia. He was the supervisor on over 400 graduate theses, 11 master’s theses and four PhD theses at the Faculty including the first PhD theses in garment engineering in Croatia. Professor Knez was the author of over 130 original scientific papers and took part in 43 international conferences with his contributions. It should be noted that his rich practical experience enabled him to lead 37 technological projects for founding garment manufacturing plants all over Croatia. As an educator, he published six university textbooks in the field and was the Chief Researcher of six scientific research projects funded by the Croatian Government. Professor Knez actively participated in the workings of scientific research magazines, and was a member of the editorial boards of the Tekstilna Industrija (Belgrade), Tekstil (Zagreb) and International Journal of Clothing Science and Technology (Bradford, UK). He was a member of professional associations in the field, as a board member mostly, and was the President of the organisation committees of four scientific/professional conferences. He was also a guest lecturer at the universities of Budapest, Lodz, Sofia and Moscow. His activity did not stop with his retirement – he was chief researcher of research projects and took part in constructing modern curricula for university courses in garment engineering.

The work and efforts by Professor Knez have significantly contributed to the development of textile and garment university courses in Croatia, as well as to shaping the profession and scientific field in Croatia and the whole of former Yugoslavia. His public lectures, projects on new garment manufacturing plants and his other work contributed immensely to the development of textile and garment industries in Croatia, while his participation in foreign publications and conferences contributed to the affirmation of garment engineering globally. He never hesitated to transfer his accomplishments and rich experience to generations of his students who are today top managers of garment manufacturers, university teachers and renowned experts not only in Croatia but also further afield. George K. Stylios Editor-in-Chief

In memoriam

7

Research register

IJCST 19,6

Athens, Greece 8

TEI Piraeus, P. Ralli & Thivon 250, GR-12244 Athens, Greece; Tel: +30 210 5381224; Fax: +30 210 5450965; E-mail: [email protected] Principal Investigator(s): Lect. Savvas Vassiliadis, Prof. Christos Kotsios, Prof. Anthony Primentas Research Staff: Prof. Thanos Peppas, Argyro Kallivretaki

Strengthening the University-Industry Links in Uzbekistan Other Partners: Academic

Industrial

TechMinho (P), Logotech S.A. (GR) Tashkent Institute of Textile and Light Industry (UZ), Namangan Engineering and Economic Institute (UZ) Project start date: 1 September 2005 Project end date: 31 August 2008 Project budget: e300,000 Source of support: European Commission Keywords: Industial liaison office, distance learning During the three years long project the situation of the University – Industry links in Uzbekistan will be analysed. In parallel the need of the distance learning will be investigated taking in acount the local conditions. After the first definition phase, the project foresees the development of structures serving the University – Industry links, like the liaison offices in the Universities and related actions. The distance learning activities will result in the operation of the necessary infrastructure and the creation of many modules for use between the Universities and the Industries.

Project aims and objectives The project aim is to analyse the current conditions in Uzbekistan and to develop structures and materials related to the strengthening of the university – Industry links and distance learing courses. The objective is to enhance the interaction between universiteis and industries using modern technological tools and methods.

Research deliverables (academic and industrial) Establishment of structures supporting the university – industry links in Uzbekistan and development of distance learning courses. Publications Not available

Bolton, UK University of Bolton, Centre for Materials Research and Innovation, University of Bolton, Deane Road, Bolton BL3 5AB, UK; Tel: +44 1204 903559; Fax: +44 1204 399074; E-mail: [email protected] Principal Investigator(s): Dr S. Rajendran, Principal Investigator, Prof. S.C. Anand, Co-investigator Research Staff: Dr Alister Rigby

Design and Development of Novel Compression Therapy Regimes for the Treatment of Venous Leg Ulcers Other Partners: Academic None

Industrial Vernon-Carus Ltd, Rossendale Combining Company Ltd Project end date: 23 October 2008

Project start date: 24 October 2005 Project budget: £152,142 Source of support: EPSRC Keywords: Compression therapy, Leg ulcers, Bandages

Venous leg ulcers are the most common type of ulcers and their prevalence increases with age. In the UK alone about 1 per cent of the adult population suffers from active ulceration during their life time. The total cost to the National Health Service in the UK for venous leg ulcers treatment is about 650 million per annum, which is 1-2 per cent of the total healthcare expenditure. Costs per patient have recently been estimated to be between £1,200 and £1,400. Venous leg ulcers are chronic and there is no medication or surgery to cure the disease other than the compression therapy. A sustained graduated compression mainly enhances the flow of blood back to the heart, improves the functioning of valves and calf muscle pumps, reduces oedema and prevents the swelling of veins. In the UK four layer bandaging system is widely used whilst in Europe and Australia the non-elastic two layer short stretch bandage regime is the standard treatment. Both the two layer and four layer systems require padding bandage that is applied next to the skin and underneath the short stretch or compression bandages. It is generally agreed by the clinicians that four layer bandages are too bulky for patients and the cost involved is high. A wide range of compression bandages is available in the Drug Tariff but each of them has different structure and properties and this influences the variation in performance and properties of bandages. The research carried out at the University of Bolton showed that there are significant variations in properties of commercial padding bandages, more importantly the commercial bandages did not distribute the pressure evenly at the ankle as well as the calf region. When pressure is applied using compression bandages, the structure of the nonwoven padding bandages collapsed and the bandages could not impart cushioning effect to the limb. In view of the above mentioned limitations and problems, it is vital that research and development

Research register

9

IJCST 19,6

10

work should be carried out to design, develop and characterise novel single layer bandages that would effectively fulfil the requirements of both padding and compression bandages. It is recognised that spacer is the right technology to produce novel compression bandages that meet the prerequisites of both ideal padding and compression bandages. In three-dimensional (3D) spacer fabrics, two separate fabric layers are combined with an inner spacer yarn or yarns using either warp knitting or weft knitting route. It is possible to produce low modulus spacer fabrics by making use of elastic yarns. Elastic compression could be achieved by altering the structures. It should be mentioned that 3D structure allows greater control over elasticity and these structures can be engineered to be uni-directional, bi-directional and multi-directional. Uni-directional elasticity is one of the desired properties for compression bandages. The three-dimensional nature of spacer fabrics makes an ideal application next to the skin because they have desirable properties that are ideal for the human body. 3D fabrics are soft, have good resilience that provides cushioning effect to the body, breathable, ability to control heat and moisture transfer. For venous leg ulcer applications, such attributes together with improved elasticity and recovery promote faster healing.

Project aims and objectives .

To study in-depth the current practices and problems in managing venous leg ulcers.

.

To test and characterise the currently available commercial bandaging systems.

.

To design and develop single layer bandage that would replace the conventional multiplayer padding and compression bandages using warp knitting technology. To design and develop single layer bandage that would replace the conventional multiplayer padding and compression bandages using weft knitting technology. To study the feasibility of using environmentally and human skin friendly biodegradable fibres in designing the single layer bandage. To test and characterise the properties of novel compression bandages and bandaging regimes. To mathematically model and verify the performance and properties of the developed structures.

.

.

.

.

.

To quantify, predict and optimise the characteristics of the novel compression therapy regimes.

Publication “Venous Leg Ulcer Treatment and Practice”, submitted to Journal of Wound Care.

Bucharest, Romania The Research Development National Institute for Textile and Leather, Lucretiu Patrascanu Street, No. 16, sector 3, 030508; Tel: 004-021-3404928; Fax: 004-021-3405515; E-mail: [email protected]

Principal Investigator(s): 289 Research Staff: 70

Research register

ITA TEXCONF Institutional construction development Project start date: 3 October 2006 Project end date: 30 November 2007 Project budget: 381 819 EURO Source of support: Public funds and own cofinancing Keywords: Technological transfer, Textile and garment industry, Innovative products, Technologies, Research results Textile field in Romania is an important sector within national economy: .

employing an important workforce, mainly female;

.

contributing to social stability through its representation across all the areas in the;

.

having a significant share in national economy export; and

.

contributing with a positive balance to foreign trade.

In Romania, structure of the enterprises in the textile field, in terms of employees’ number, presents as follows: total number of enterprises in the textile field: over 8,000. As easily seen above the number of textile SMEs represents around 96 per cent of the total number of enterprises in Romania. The Romanian Government is aware and appreciates the importance of these companies. The policy in the field of SMEs in Romania has become a part of the Regional Development Policy, due to the great role of the SMEs – “locomotives” of a sustainable economic growth, creating jobs and having an important role in fighting the unemployment. At present, the majority of Romanian enterprises in the textile field are not sufficiently prepared for responding to the opportunities and challenges that the accession to enlarged knowledge-based EU supposes. The innovation policy has as center the research-industry relation. Research creates markets, and industry’s needs generate new researches. ITA TEXCONF is created with the purpose of valorizing the results of technological research and development and, especially for the realization of technological transfer to SMEs. The institutional construction of a technologic and business incubator is a complex theme, this character being determined by the multitude of activities that must be carried out for acquisitions consisting of the endowing with the equipment necessary for carrying out the activity specific for the incubator centre, in order to provide a study to select the SMEs team which are going to be incubed. Relations between research and industry have to be dealt with based on a coherent, long term and comprehensive policy, which should incorporate as core element researchindustry connection through: .

marketing research results;

.

promotion of start-ups by researchers in the new technologies field;

.

creation of conditions favoring activities and partnerships transfer between research and industry;

11

IJCST 19,6

12

.

fostering the enterprises in creating and developing innovative competences;

.

launching joint research programs, based on direct collaboration between research and industry;

.

enhancing the access to information regarding internal and foreign networks of opportunities, informing sources, knowledge and technologies;

.

realization of connections between portals of national and International information; and

.

fostering SMEs participation in research networks and technologies markets, directly or through multi-level innovative networks.

Project aims and objectives Aims: ITA TEXCONF is an entity from the TT infrastructure for the textile-garment field, of which activity is mainly directed towards: facilitating the initiation and development of new innovative private enterprises based on advanced technology; stimulation of innovation and technologic transfer for the purpose of introducing the research results in the economic circuit, the results being turned into new or improved products, processes and services. Short term objectives: Institutional construction; textile-garment field SMEs incubation, logistics service, supplying; training and improving experts in technologic transfer, intellectual property ensuring, entrepreneurship, human resources involved in innovative actions; partnership creating, developing groups of interest for transferring innovative products/technologies towards the industry, especially towards SMEs; ITA TEXCONF integrating into the National Network of Innovation and Technologic Transfer (RENITT). Long term objectives: Know-how transfer between INCDTP and SMEs; increasing the degree of turning into good account the research results from the textile field and the patents obtained by INCDTP, by disseminating the results, by mediating technologic transfers, license selling, etc.; stimulating the financing of the research projects; creating data bases in the textile field; integrating the Romanian innovative structures into EU.

Research deliverables (academic and industrial) ITA TEXCONF institutional construction of the technologic and business incubator Specialists trained in the field of technological transfer, intellectual property, entrepreneurship, enterprise evaluation. ITA TEXCONF will aim at businesses in the textile field by: . .

finding commercial partners especially in the field of SMEs; developing the interest groups in the field;

.

determining R&D Units and Universities interest to offer their technical services to the SMEs;

.

interchanging partner ideas with these enterprises;

.

aiming to establish international collaborations through Technology Transfer; and

.

encouraging the SMEs to become exporters.

Publications Leaflets of presenting and disseminating of the project, distributed with the occasion of participating to the .

Open forum for innovation and technological transfer, the 6th edition, Bucharest, Romania, 20-21.03.2007.

.

International Industrial Fair HANOVVER -MESSE, Germany, 16-20.04.2007.

.

The National Fair of Clothing-Footwear Manufacturers TINIMTEX Mamaia, Romania, 16-20.05.2007.

.

The European Reseach and Innovation Fair, Paris, France, 07 -09.06.2007.

.

3rd INSME Association General Assembly – “Reinvigorating our global Partnership for Innovation”, Rome, Italy, 12 -13.07.2007.

We have to highlight that INCDTP- ITA TEXCONF is an observer member within INSME – International Network for SMEs, which is an international non profit association with headquarters in Italy, having as purpose cross-national cooperation between partners in the public and private sector in the field of innovation and technology transfer (TT) to Small and Medium Enterprises. The INSME Association includes 77 Members: 29 governmental bodies, 8 International Organizations, 8 International NGOs and representatives of 32 networks and intermediaries acting in the field of innovation and technology transfer to SMEs.

Budapest, Hungary Budapest University of Technology and Economics, Budapest XI. Mu˝egyetem rkp. 3, Postal A., H-1521 Budapest, Hungary. Tel: +36-1463-1376; Fax: +36-1-463-1376; E-mail: [email protected] Keywords: Cellulose, Cotton, Hemp, Swelling, Chemical modification, Carboxymethylcellulose, Functional textile, Antimicrobial textile, Textile for hospital use

Modification of cellulose fiber for extension of its application Department or Laboratory: Department of Plastics and Rubber Technology, Department of Physical Chemistry, Department of Chemical Technology Principal Investigator(s): Prof. Judit Borsa Other Partners: Academic Johan Be´la National Center for Epidemiology, Inst. for Isotops and Surface Chemistry of the Hungarian Academy of Sciences, Johannes Kepler University, Linz, Austria, Dr Habil. Ildiko Tanczos

Industrial None

Research register

13

IJCST 19,6

Source of support: Hungarian National Research Fund (OTKA), Governmental Fund (GVOP) Project start date: 1 January 2005 Project end date: 31 December 2008 Cellulosic fibers are modified by physical and chemical methods:

14

.

Interaction of cotton cellulose with quaternary ammonium hydroxide (tetramethylammonium hydroxide) is studied in comparison with sodium hydroxide.

.

Cotton fiber is modified by slight carboxymethylation. Effect of technology parameters on the properties of the modified fiber, theoretical aspects of modification and some possible application of modified fiber (e.g. antibacterial textile for hospital use) are investigated. Delignification and refinement of various kinds of Hungarian hemp are studied.

.

Publication Borsa, J., La´za´r, K., La´szlo´, K., “Antibacterial effect of carboxymethylated cotton fiber”, The Fiber Society Spring Conference, StGallen, Switzerland, May 2005.

Budapest, Hungary Budapest University of Technology and Economics, Mu˝egyetem rkp. 3-9., H-1111 Budapest, Hungary; Tel: +36 1 4632495; Fax: +36 1 4631689; E-mail: [email protected] Principal Investigator(s): prof. Dr Ja´nos Somlo´; Dr Marianna Hala´sz; Dr Pe´ter Tama´s; Ga´bor Ambrus; Dr Be´la Bo˝di Research Staff: Dr La´szlo´ Monostori, Dr Petra Aradi, Dr Gyo¨rgy Lipovszki, Dr Kla´ra Zalatnaidr.Judit Barna, Lajos Szabo´, Bertalan To´th, Jekatyerina Kuzmina

RuhaRobot (ClothRobot) Other Partners: Academic

Industrial

Budapest University of Technology and Economics, Budapest Tech Rejto˝ Sa´ndor Faculty of Light Industry, Institute of Leather, Textile and Garment Technology

MAGA Informatics Ltd., National Osteoporosis Foundation

Project start date: 1 January 2005 Project budget: e327600

Project end date: 31 December 2007

Source of support: National Development Plan – Structural Funds and the Cohesion Fund EU Keywords: 3D dress design, Model of human body, Visual robot, Wearing simulation, 3D drape tester There is a 3D dress designing system in the focus of the project. A parameterized body model is evaluated. The model is appropriate not only for the symmetric and asymmetric body modelling but for motion simulation too. The real sizes of the human body can be defined by a photo based on reduced measuring processes or by robotized 3D scanning using re-engineering. We are developing an integrated traditional scanner and a 6D KUKA robot for scanning the visible and masked body parts. Shape and sizes of clothparts are derived from the geometry of the human body model. Designers modify the model and shape of parts in space. The shape of patterns is calculated by direct laying out. The mechanical model of clothes is a multi mass-point flexible mechanical structure. A virtual mannequin wears the model dress. Draping and texture behaviour is real-time simulated by the mechanical model. There is a newly evaluated drape tester equipment. Tester scans the 3D shape of parts and material properties are defined by the same simulation as used for the virtual mannequin.

Project aims and objectives .

Real 3D Computer aided design for ready made clothes and for handicapped people.

.

Usage of re-engineering technics in body size measuring processes.

.

Development of a visual robot system and application in reg-trade.

.

Material parameter measuring for clothes simulation.

.

Special methods based on material simulation and the newly developed equipment.

Research deliverables (academic and industrial) .

Computer aided 3D dress design system for ready made clothes and handicapped people.

.

3D drape tester equipment.

.

Measuring results.

.

Visual robot system.

Publications Dr Tama´s Pe´ter, Dr Hala´sz Marianna (2003), “3D body modelling in clothing design”, IMCEP 2003, 4th International Conference, 9-11. Oktober, Maribor, Slovenia, ISBN 86-435-0575-7, pp. 64-8. L. Kokas Palicska; J. Gersak; M. Hala´sz, (2005) “The impact of fabric structure and finishing on the drape behavior of textiles”, AUTEX 2005, 5th World Textile Conference, Portorozˇ, Slovenia, 27-29 June, ISBN 86-435-0709-1, pp. 891-7. P. Tama´s; M. Hala´sz; J. Gra¨ff (2005), “3D dress design”, AUTEX 2005, 5th World Textile Conference, Portorozˇ, Slovenia, 27-29 June, ISBN 86-435-0709-1, pp. 436-1.

Research register

15

IJCST 19,6

16

J. Kuzmina; P. Tama´s; M. Hala´sz, Gy. Gro´f (2005), “Image-based cloth capture and cloth simulation used for estimation cloth draping parameters”, AUTEX 2005, 5th World Textile Conference, Portorozˇ, Slovenia, 27-29 June, ISBN 86-435-0709-1, pp. 904-9. L. Szabo´; M. Hala´sz (2005), “Automatic determination of body surface data”, AUTEX 2005, 5th World Textile Conference, Portorozˇ, Slovenia, 27-29 June, ISBN 86-435-0709-1, pp. 715-20. L. Kokas Palicska; M. Hala´sz (2005), “Analysing of draping properties of textiles”, at IN-TECH-ED’05, 5th International Conference, 8-9 September, Budapest, ISBN 963 9397 06 7, pp. 133-8. O. Nagy Szabo´; P. Tama´s; M. Hala´sz, (2005) “Garment construction with a 3 dimension designing system”, at IN-TECH-ED’05, 5th International Conference, 8-9 September, Budapest, ISBN 963 9397 06 7, pp. 248-357. J. Kuzmina; P. Tama´s; M. Hala´sz; Gy. Gro´f (2005), “Image-based cloth capture and cloth simulation used for estimation cloth draping parameters”, paper presented at IN-TECH-ED’05, 5th International Conference, 8-9. September, Budapest, ISBN 963 9397 06 7, pp. 358-65.

Budapest, Hungary Budapest University of Technology and Economics, Department of Polymer Engineering, 9 Muegyetem rkp. Budapest H-1111 Hungary; Tel: 36-1-463-1529; Fax: 36-1-463-1527; E-mail: [email protected] Principal Investigator(s): La´szlo´ M. Vas Research Staff: Pe´ter Tama´s, Pe´ter Nagy, Veronika Nagy, Zsolt Ra´cz, Zolta´n L. Simon, Zolta´n Gombos

Structural and strength modelling and complex testing of fibrous structures and fibre reinforced composites Other Partners: Academic

Industrial

None BUTE Department of Information Engineering Project start date: 2005 Project end date: 2008 Project budget: 43000 euro Source of support: Hungarian Scientific Research Fund – OTKA T049069 Keywords: Modelling, Testing, Fiber bundles, Fibrous structures, Reinforcements, Composites Besides the fibers, fiber groups called fiber bundles as intermediate structural elements play an important role in the macro-scale behavior and properties of the fibrous structures such as slivers, rovings, yarns, webs, fleeces, fabrics, and reinforcements of polymeric composites. The Department of Polymer Engineering has dealt with testing textiles and fibrous reinforcing systems using image processing and with the theoretical fundamentals of their computational structural-mechanical modeling essentially for more than one decade. In the framework of some earlier projects (OTKA I/3 821, 1991-1994; OTKA I/5 T7652, 1993-1994) a novel modeling method [1-4] was developed applying the so

called idealized fiber-bundle-cells as structured statistical mechanical model elements to modeling fiber flows such as slivers or fiber bundles to flat bundle test or twisted fiber structures such as filament or spun yarns. In addition, in cooperation with the Research Institute of Technical Physics and Materials Science Budapest some geometrical and mechanical testing techniques [5-8] for measuring the diameter of fibers or yarns and the fiber orientation on webs or yarn surface as well as the local deformation of fabrics during tensile test were elaborated using image processing technology. The idealized fiber-bundle- cells [1,2,12] are statistical because the strength parameters of the fibers (tensile strength, breaking strain, initial tensile stiffness), and the bundle properties of the fiber cells (pretension, order), as well as the parameters for describing the grip of the fibers or fiber chains (slippage load, slippage length) are stochastic variables. From the sub-bundles combined, so called composite bundles or fiber bundle chains can be formed. According to the modeling concept a finite composite fiber bundle or fiber bundle chains or network built up of randomly or deterministicly oriented fibers ungripped or gripped at one or two ends can be the structural-mechanical model of a real textile considered concerning not only the increasing load but the damage process as well [1-4, 12-13, 16-17]. Another research work (OTKA T022077, 1997-1999) extended to the fiber reinforced composites focused on modeling the structure of glass fiber mats and testing the structural-mechanical properties of the one layer glass mat reinforced unsaturated polyester (UP) resin composite sheets. Besides the structural-geometrical model of short fiber structures developed [9-10] a testing method based on image processing was worked out for studying the local deformation and crack propagation of trans-illuminated composite specimens during tensile test. Using the mobile CCD camera image processing system developed in the framework of this project [5-6] the cracks coming into being during the tensile test were detected as fibers or edges and by that the fiber deformation and crack propagation processes could be described numerically by evaluating the fiber orientation histograms recorded [11]. The next project in this topic (OTKA T038220, 2002-2004) aimed mainly at testing and modeling the damage and fracture process of unidirectional carbon fiber reinforced epoxy resin composites during flexural test as well as analyzing the effect of the geometry. According to experiences the modeling method using statistical fiber-bundlecells can be describe the expected value of the bending load in an acceptable way even if the simplest fiber-bundle-cell was used and the breakage of the stretched layers were modeled [12-13]. The fiber-bundle-cells method was also applied to modeling the mechanical behavior of braided carbon fiber reinforced composite tubes subjected to tensile load [15]. Two PhD Theses made use of the results. The current research project (OTKA T049069, 2005-2008) undertook to summarize and complete the results of the structural-mechanical testing and modeling methods and work out some new applications, as well as – in collaboration with the Department of Information Engineering BUTE – develop a software package for modeling and a complex testing method including geometrical and mechanical tests, acoustic measurements and image processing [14,18]. This is the topic of a PhD Thesis and a DSc Thesis.

Research register

17

IJCST 19,6

18

Project aims and objectives The aim of the research is, based on the theory of the fiber flows and fiber bundles as well as the results of experimental and FEM examinations, to develope a modelling method and software suitable for estimating and characterizing statistically the structuralgeometrical and strength properties of different fibrous structures (fiber bundles, slivers, rovings, yarns, fiber mats, and fabrics), unidirectional composites and specimens cut from laminated composite sheets in different directions as well as their damage processes during tensile and flexural tests. On the other hand, it is planned to develope a complex testing and avaluating system based on a universal tensile tester, some CCD camera image processing equipment and an acoustic emission device suitable for studying, analysing the statistical behaviour, the damage process, and the scale effects of single fibers, rovings, fiber reinforced composite specimens. Objectives and tasks: .

Developing image and AE processing aided methods for the complex testing of the strength properties of unidirectional and laminated composites specimens.

.

Experimental and FEM examinations of the relationships between the structural and the strength properties of unidirectional and laminated composite specimens cut in different directions.

.

Developing statistical structural-strength basic models and modelling methods for describing the tensile and flexural strength properties of the unidirectional and laminated composite specimens.

.

Developing some applications of the structural-strength models.

.

Developing a statistical structural-strength modelling software package based on the results obtained above.

Research deliverables (academic and industrial) . .

The realized results of the project have been expected as follows: Testing and evaluation methods using image processing

.

Windows based modelling software realized in Delphi Code

.

2-3 papers published in periodicals annually

.

2-3 papers published at conferences annually

.

2-3 MSc Theses annually

.

In the framework of the research three PhD Theses have been elaborated as well as the results have been expected to be used by undergraduate and PhD students for preparing papers, MSc Theses and other PhD Theses.

Publications Earlier publications: 1. Vas, L.M. (1990). “The statistical fiber bundle strength and its application in testing fibers and yarns (in Hungarian)”, Magyar Textiltechnika, Vol. 43 No. 4, pp. 165-85. 2. Vas, L.M. (1992). “Latest results in the tensile test theory of fiber and yarn bundles ordered into plane (in Hungarian)”, Magyar Textiltechnika, Vol. 45 No. 3, pp. 71-5, (5-6), 137-42, (7-8), 187-191.

3. Vas, L.M. and Csa´szi, F. (1993). “Use of composite-bundle theory to predict tensile properties of yarns”, Journal of the Textile Institute, Vol. 84 No. 3, pp. 448-63. 4. Vas, L.M. and Hala´sz, G. (1994). “Modelling the breaking process of twisted fiber bundles and yarns”, Periodica Polytechnica Ser. Mech. Eng., Vol. 38 No. 4, pp. 325-50. 5. Vas L.M., Hala´sz G., Taka´cs M., Eo¨rdo¨gh I.,- Sza´sz K. (1994) “Measurement of yarn diameter and twist angle with image processing system”, Periodica Polytechnica Mech. Eng., Vol. 38 No. 4., pp. 277-96 6. Hala´sz G., Taka´cs M., Vas L.M. (1994) “Image processing system for measuring geometrical properties of fibres and yarns”, Fibres & Textiles in Eastern Europe, January/February, pp. 30-33 7. Vas L.M., Hala´sz G. (1994) “Untersuchung der Vera¨nderungen in Fadendiameter und DrehungsWinkel bei der Zug- und Drehbeanspruchung”, Periodica Polytechnica Mech. Eng., Vol. 38 No. 4, pp. 297-324. 8. Vas L.M., Hala´sz G., Nagy P., Eo¨rdo¨gh I., Juha´sz Gy., Sza´sz K. (1997): “Testing the deformation of textile fabrics using image processing system (in Hungarian)”, Magyar Textiltechnika, Vol. 50, 1.sz., pp. 19-25. 9. Vas L.M., Balogh K., Nagy P., and Gaa´l J. (1998): “Structural modeling and testing of Glass Fiber Mats (in Hungarian)”, Magyar Textiltechnika, Vol. 51. (In-Tech-Ed’98 Conference Budapest Oct. 21-22. 1998, special issue) (pp. 67-71). 10. Vas, L.M., Balogh K. (2000) “Testing fiber orientation and its effect on glass mats by using image processing system”, VI. International Conference IMTEX’2000. Lodz, June 5-6, (Zeszty Naukowe No.845. Wlo´kiennictwo No.58.) Proceedings (p 69-78). 11. Vas, L.M. – Balogh, K. (2002) “Investigating damage processes of glass fiber reinforced composites using image processing”, Journal of Macromolecular Sciences Part B – Physics, Vol. B41. Nos. 4/6, pp. 977-989. 12. Vas, L.M. – Ra´cz, Zs. (2004) “Modeling and testing the fracture process of impregnated carbon fiber roving specimens during bending, Part I. Fiber Bundle Model”, Journal of Composite Materials, Vol. 38 No.20, pp. 1757-1785. 13. Vas, L.M., Ra´cz, Zs. – Nagy, P. (2004) “Modeling and testing the fracture process of impregnated carbon fiber roving specimens during bending Part II. Experimental studies”, Journal of Composite Materials, Vol. 38 No.20, pp. 1787-1801. 14. Ra´cz Zs., Simon Z.L., and Vas L.M.: “Analysing the Flexural Strength Properties of Unidirectional Carbon/Epoxy Composites”, COMPTEST 2004, 2nd International Conference on Composites Testing and Model Identification, 21-23 September 2004. University Bristol, UK. (Poster paper, CD)

Publications of the current research: 15. Simon Z. and Vas L.M. (2005): “Relationship between the flexural properties and specimen aspect ratio in laminated composites”, 22nd DANUBIA-ADRIA Symposium on Experimental Methods in Solid Mechanics. Monticelli Terme/Parma, Italy, September 28-October 1. (Poster) 16. Ra´cz Zs. and Vas L.M. (2005): “Relationship between flexural strength and size effects in unidirectional carbon/epoxy”, Composite Interfaces, Vol. 12. pp. 325-39. 17. Nagy V., Vas L.M.: “Testing polyester yarns with overtwisting and estimating the pore sizes (in Hungarian)”, Magyar Textiltechnika, Vol. LVIII. e´vf. Part I: 2005/1. 1-3, Part II: 2005/2. 26-27. 18. Gombos Z., Vas L.M., Gaa´l J. (2005) “Testing and evaluation of resin take-up processes (in Hungarian)”, Anyagvizsga´lo´k Lapja, Vol. 15 No. 3, pp. 97-9. 19. Gombos Z., Vas L.M. (2005) “Determining the pore size of glass fiber mats on the basis of the statistical fiber mats model (in Hungarian)”, Magyar Textiltechnika, LVIII, pp. 89-91. 20. Simon Z.L., Vas L.M. (2005): “Analyzing the flexular properties of laminated composites produced by RTM technology (in Hungarian)”, Anyagvizsga´lo´k Lapja, Vol. 15 No. 4, pp. 119-121. 21. Zsigmond B., Vas L.M. (2005) “Tensile testing of carbon fiber reinforced braided tubes ˝ anyag e´s Gumi, Vol. 42 No. 12, pp. 488-491. (in Hungarian)”, MU 22. Nagy V. and Vas L.M. (2005) “Pore characteristic determination with mercury porosimetry in staple yarns”, Fibers and Textiles in Eastern Europe, Vol. 13 No. 3, July/September, pp. 21-26.

Research register

19

IJCST 19,6

20

23. Gombos Z., Nagy V.,Vas L.M., Gaa´l J. (2005) “Investigation of pore size and resin absorbency in chopped strand mats”, Periodica Polytechnica, Ser. Mech. Eng., Vol. 49 No. 2, pp. 131–148. 24. Vas L.M. (2006) “Strength of unidirectional short fiber structures as a function of fiber length”, Journal of Composite Materials, Vol. 40 No. 19, pp. 1695-1734. 25. Vas L.M. and Cziga´ny T. “Strength modeling of two-component hybrid fiber composites in case of simultaneous fiber failures”, Journal of Composites Materials, Vol. 40 No. 19. 1735-1762. (IF ¼ 0,671) 26. Zsigmond B. and Vas L.M. (2006) “Examination of the tensile state of fibers in braided fiber reinforced composite tubes”, Periodica Polytechnica Ser. Mech. Eng., Vol.50 No.1, pp. 67-76. 27. To¨ro¨k P., Gombos Z. (2006) “Determination of the gel point from the exotherm effect of UP resin (in Hungarian)”, Anyagvizsga´lo´k Lapja, Vol. 16 No. 1, pp. 14-18. 28. Gyivicsa´n P., Gombos Z., Vas L.M. (2006) “Examination of pore size distribution of chopped strand mats with image analysis system (in Hungarian)”, Magyar Textiltechnika, Vol. 59 No. 5, pp. 146-49. 29. Vas L. M. and Tama´s P. (2006) “Fiber-bundle-cells method and its application to modeling fibrous structures”,.GE´PE´SZET 2006, 5th Conf. on Mech. Eng. Budapest, May 25-26, Proceedings (CD – Fulltext) ISBN 963 593 465 3. 30. Vas L.M. (2006) “Idealized fiber bundles and their application to modeling fibrous structures (in Hungarian)”, Scientific Session of Committee of Fiber and Composite Technology Hungarian Academy of Sciences, Budapest, 28 February. 31. Vas L. M. and Tama´s P. (2006) “Fiber-bundle-cells method and its application to modeling fibrous structures”, GE´PE´SZET 2006, 5th Conf. on Mech. Eng. Budapest, May 25-26, Proceedings (CD – Fulltext) ISBN 963 593 465 3. 32. Gombos Z. and Vas L. M. (2006): “Temperature dependence of resin absorption of various chopped strand mats”, GE´PE´SZET 2006, 5th Conf. on Mech. Eng. Budapest, May 25-26, Proceedings (CD – Full-text), pages: 6. ISBN 963 593 465 3. 33. Meggyes G., Gombos Z., and Vas L. M. (2006) “Analysing the orientation and size effects in composite sheets in tensile tests”, GE´PE´SZET 2006, 5th Conf. on Mech. Eng. Budapest, May 25-26, Proceedings (CD – Full-text), pages:6. ISBN 963 593 465 3. 34. Simon Z., Szabo´ L., and Vas L. M. (2006) “Determination of flexular modulus by image processing”, GE´PE´SZET 2006, 5th Conf. on Mech. Eng. Budapest, May 25-26, Proceedings (CD – Full-text) ISBN 963 593 465 3. 35. To¨ro¨k P., Gombos Z., and Vas L. M. (2006) “The effect of curing temperature upon the mechanical properties of unsaturated polyester resin”, GE´PE´SZET 2006, 5th Conf. on Mech. Eng. Budapest, May 25-26, Proceedings (CD – Full-text), pages: 6. ISBN 963 593 465 3. 36. Nagy V. (2006) “Examination and modeling of porosity in polyester twisted fibrous structures”, PhD thesis, BUTE Budapest. 37. Ra´cz Zs. (2006) “Analyzing the bending characteristics and damage processes of unidirectional composite beams (in Hungarian)”, PhD thesis, BUTE Budapest. 38. Vas L.M. (2006) “Statistical modeling of unidirectional fiber structures. macromolecular symposia”, Special Issue: Advanced Polymer Composites and Technologies, Vol. 239 No. 1, pp. 159-175. 39. Gombos Z., Nagy V., Kosˇta´kova´ E., Vas L.M. (2006) “Absorbency behaviour of vertically positioned nonwoven glass fiber mats in case of two different resin viscosities”, Macromolecular Symposia, Vol. 239 No. 1, pp. 227–231. 40. Nagy P. and Vas L.M. (2006) “Investigating the time dependent behavior of thermoplastic polymers under tensile load”, Macromolecular Symposia. Special Issue: Advanced Polymer Composites and Technologies, Vol. 239 No. 1, June, pp. 176-181. 41. Simon Z., Vas L.M. (2007) “Relationship between bending modulus and test sizes of laminated glass/polyester composites”, Materials Science Forum, 537-538, pp. 71-9. 42. Vas L.M., Poloskei K., Felhos D., Deak T., Czigany T. (2007) “Theoretical and experimental study of the effect of fiber heads on the mechanical properties of non-continuous basalt fiber reinforced composites”, Express Polymer Letters, Vol. 1 No. 2, pp. 109-121-91.

43. Vas L.M. (2007) “Modelling polymer composites by using fiber bundle theory based methods (in Hungarian)”, Day of Materials Science. Scientific Session of Committee of Materials Science and Technology Hungarian Academy of Sciences, Budapest, 11 May. 44. Vas L.M., Tama´s P. (2007) “Modeling method based on idealized fiber bundles”, 3rd Chine-Europe Symposium on Processing and Properties of Reinforced Polymers, 11-15 June, Budapest (Poster). 45. Gombos Z., Vas L.M., Hruza J. (2007) “The effect of layer number on air permeability in case of glass fiber mats”, 3rd Chine-Europe Symposium on Processing and Properties of Reinforced Polymers, 11-15 June, Budapest (Poster).

Dundee, Scotland, United Kingdom Duncan of Jordanstone College of Art and Design, University of Dundee, School of Design, Duncan of Jordanstone College of Art and Design, University of Dundee, Dundee DD1 4HT; Tel: 00447973621932; Fax: 00441382 201378; E-mail: [email protected] Principal Investigator(s): Sara Keith

PhD working title: Can Electrodeposition of metals on a textile substrate enhance performance and appearance, causing the evolution of a new material? Other Partners: Academic

Industrial

Project start date: October 2006 Project end date: 2009 Source of support: Arts and Humanities Research Council Keywords: Electroform, Electroplate, Textile, Fibre, Silver Research for this PhD will examine a wide variety of inert and conductive fibres to determine the optimum textile substrate for silver and copper electroforming. Conductive performance and aesthetic appearance of various textile structures will also be examined, analysing the most favourable mandrel configurations.

Project aims and objectives This body of research takes the form of a Full Time PhD. Initial aims are to identify the most successful fibre and structure combinations for electroforming. The overall objective of this research is to determine the potential for this medium within the fields of design, medicine, defence, transport as a technical and apparel textile.

Research deliverables (academic and industrial) The final thesis will be presented as an exhibition and CD, supported by documentation of the tests and consultation process reflecting on results. Potential applications for this medium will also be outlined.

Research register

21

IJCST 19,6

22

Publications The Cutting Edge National Museums of Scotland ISBN: 978-1-905267-07-1 Touring exhibition 2007-8 and book 2007 New Craft Future Voices Conference University of Dundee, United Kingdom Exhibit 2007 The Art of Research Symposium University of Art and Design Helsinki, Finland Paper and exhibit

Edinburgh, Scotland, UK Heriot-Watt University, School of Engineering and Physical Sciences, Riccarton, Edinburgh, Scotland, UK EH14 4AS; Tel: +44 131 451 3034; Fax: +44 131 451 3473; E-mail: [email protected], [email protected] Principal Investigator(s): Prof. J.I.B. Wilson and Dr R.R. Mather

Solar cells in textiles Other Partners: Academic

Industrial

None More Being sought Project start date: 2001 Project end date: Ongoing Keywords: Thin film silicon, solar energy, photovoltaics We are developing thin-film silicon solar cells on low cost textile substrates, using chemical vapour deposition (CVD) technology, based on previous thin-film diamond expertise. The CVD technology employs a proprietary microwave plasma system (developed at Heriot-Watt University) with silane/hydrogen/dopant gas mixtures to produce the sequence of layers that forms the active part of these cells. We have shown that relatively low deposition temperatures of 200 C and the active plasma conditions of the process do not affect our textile substrates, whether of woven or non-woven construction. In addition, solutions have been determined to the problem of providing reliable electrical contacts over fibrous, flexible substrates, together with a conventional transparent conducting oxide as the top contact in the cell “sandwich” structure. Effective “first barrier” encapsulation may also use our deposition technology.

Project aims and objectives Flexible solar cells for a variety of applications: e.g. building facades, use in remote areas, emergency use in disaster relief, camping/leisure industry, portable chargers.

Research deliverables (academic and industrial) Working prototype

Publications “Textiles make solar cells that are flexible and lightweight”, Technical Textiles International, December 2002, pp 5-6. “Solar textiles: production and distribution of electricity coming from solar radiation. Applications” in Intelligent Textiles and Clothing, ed. H. Mattila, Woodhead Publishing Limited, Cambridge, 2006, pp 202-217.

Research register

23

Galashiels, Scotland Heriot Watt University, Scottish Borders Campus, Netherdale, Galashiels TD1 3HF, Tel: +01896 892245; Fax: +01896 758965; E-mail: [email protected] Principal Investigator(s): Dr Alex. Fotheringham Research Staff: Basel Younes

Optimisation of Weaving Process for Biopolymers Other Partners: Academic

Industrial

None None Project start date: October 2006 Project end date: September 2010 Source of support: British Council Keywords: Fibre extrusion, Yarn production, Weaving This is part of a general research programme in the production and use of biopolymers. Current research investigates the optimisation of biopolymer fibre production using experimental design techniques to statistically map processing parameters to mechanical properties. Having produced a model of the fibre and yarn production, the relationship between fibre/yarn characteristics, weaving parameters and fabric properties can then be established. Related work is researching into the use of gas plasma for the pre-treatment of polylactic acid fabric, the printing of knitted biopolymer fabrics and dyeing of such materials in fibre/yarn form.

Project aims and objectives .

To establish the relationships which exist between processing and final properties.

.

To identify the value of gas plasma treatment on biopolymer fabrics for e.g. dyeing, coating etc.

.

To optimise printing on knitted biopolymer fabrics.

Objectives: .

To use experimental designs to create statistical models.

IJCST 19,6

.

To use statistical models as forecasting tools to establish the relationships between processing and properties.

Research deliverables (academic and industrial) .

24

.

A statistical model of fire, yarn and fabric (weaving) processing. To create similar models for fibre dyeing and printing.

Publications Not available

Galashiels, UK Heriot-Watt University, Heriot-Watt University, Netherdale, Galashiels TD1 3HF, Tel: +44 1896 892234; Fax: +44 1896 75 8965; E-mail: [email protected] Principal Investigator(s): Dr Lisa MacIntyre Research Staff: Pippa Bell

The measurement and control of mean pressure delivery exerted by pressure garments for the treatment of hypertrophic scars Other Partners: Academic

Industrial

None None Project start date: September 2004 Project end date: September 2007 Source of support: EPSRC and Drapers Company Keywords: Pressure garment, Pressure measurement, Hypertrophic burn scars, Powernet fabric Pressure garments are the predominant treatment method for hypertrophic scars. They are constructed from elastic fabrics and are custom-made to suit individual treatment specifications. The properties of the elastic fabrics have an impact on the pressure delivery of the garments and ultimately determine treatment success. During this investigation 12 new pressure garment fabrics have been knitted using the powernet structure. Each fabric has a slightly different specification and by assessing certain fabric properties, the impact on pressure delivery of these fabrics can be established. The reduction factor, fabric tension and model circumference are the main properties that contribute to the pressure delivery of the fabrics; the relationships of which are being thoroughly investigated. The measurement of pressure delivery exerted by pressure garments to the wound site is very important. A mean pressure of 25mmHg has been regularly quoted as the optimum pressure required for treating hypertrophic scars. It has been established that when assessing the extent of pressure delivery to the patient, there is no scientific

method currently used by medical practitioners. Based on this information two sensor measurement systems are being developed and investigated for this project. The sensors are being used to measure the mean pressures exerted by pressure garment samples on cylinder models and have been incorporated into the construction of a pressure sensing mannequin which is a simulation of the male upper torso and left arm. Further measurements are to be taken from human volunteers and compared with the model analysis results. Throughout the investigation development work on a predictive equation for calculating pressure delivery of 25mmHg to small circumference limbs from pressure garments is being carried out.

Project aims and objectives .

.

Evaluate the pressures exerted to small circumference limbs and complex body shapes. Develop a pressure-sensing mannequin incorporating an accurate and reliable pressure measurement system to record pressures exerted by pressure garments.

.

Evaluate the effectiveness of 3D modelling for investigation between fabric specifications and applied pressure.

.

Establish relationships between powernet fabric construction variables and pressure delivery potential.

Publications Not available

Galashiels, Scotland, UK Heriot-Watt University, RIFleX, School of Textiles and Design, Netherdale, Galashiels TD1 3HF, UK, Tel: +44 1896 89 2135; Fax: +44 1896 75 8965; E-mail: [email protected] Principal Investigator(s): Prof. George K. Stylios Research Staff: Mohamed Basel Bazbouz

Investigating the Spinning of Yarn from Electro-spun Nanofibres Other Partners: Academic

Industrial

None None Project start date: November 2004 Project end date: October 2008 Project budget: None Source of support: None Keywords: Electrospinning, Polymers, Nanofibre, Alignment, Yarn, Composite

Research register

25

IJCST 19,6

26

Our laboratory is using a process called electrospinning which has the ability to produce a wide variety of polymeric fibres with diameters from a couple of micrometer down to the nanometer scale. In this case different structures can be made from electrospun fibers to suit the needs of various industries. Electrospinning, a fibre spinning technique that relies on electrostatic forces to produce fibres in the nanometer to micron diameter range, has been extensively explored as a simple method to prepare fibres from polymer solutions or melts. Under the influence of the electrostatic field, a pendant droplet of the polymer solution at the spinneret is deformed into a conical shape (Taylor cone). If the voltage surpasses a threshold value, electrostatic forces overcome the surface tension, and a fine charged jet is ejected. As these electric static forces increase, the jet will elongate and accelerate by the electric forces. The jet undergoes a variety of instabilities, dries, and deposits on a substrate as a random nanofibre mat. In our work, nonwoven electrospun mats of nylon 6 produced from solutions with formic acid with different concentrations are examined. Each nonwoven mat with average fibre diameters from 200 to 1300 nm was prepared under controlled electrospinning process parameters. Effects of electric field and tip-to-collection plate distances of various nylon 6 concentrations in formic acid on fibre uniformity, morphology and diameters were measured. Processing parameters effects on the morphology such as fibre diameter and its uniformity of electrospun polymer nanofibres was investigated. A process optimization summarized the effects of solutions properties and processing parameters on the electrospun nanofibre morphology was issued. In our work, we Control the electrospinning process to move away from just collecting random fibre mesh to enabling the accurate deposit of fibres at any predetermined position. This will be by using a simple method of getting a fibres bundle made of aligned nanofibres between two known points. This collection process has been termed as the ‘gap method of alignment’ involves grounding two circular disks from the spinneret. We have demonstrated that it is possible to produce continuous fiber yarn made out of electrospun nanofibres. The current process has the potential to spin nanofibre at a commercially viable rate.

Project aims and objectives .

Understanding of the electrospinning process.

.

Optimizing its process parameters to electrospin polymers into nanofibres with desired morphology.

.

Discussing the models proposed for jet forming, jet travel, processing instabilities.

.

Tensile testing of polymeric nanofibres.

.

Controlling the spatial alignment of electrospun fibres.

.

A technique for making continuous fibre bundle yarns from electrospun fibres.

.

Investigating the other methods for producing continuous fibre bundle yarns, Core-yarn and laminate composite consisting of aligned fibres in different Directions.

Research deliverables (academic and industrial) .

Nonwoven electrospun mats of nylon 6 produced from solutions with formic acid with different concentrations are examined.

.

Each nonwoven mat with average fibre diameters from 200 to 1300 nm was prepared under controlled electrospinning process parameters.

.

Effects of electric field and tip-to-collection plate distances of various nylon 6 concentrations in formic acid on fibre uniformity, morphology and diameters were measured.

.

Processing parameters effects on the morphology such as fibre diameter and its uniformity of electrospun polymer nanofibres were investigated.

.

A process optimization summarized the effects of solutions properties and processing parameters on the electrospun nanofibre morphology was issued.

.

The well-known Chauchy’s inequality is applied to prediction the velocity of the end of the jet in electrospinning.

.

A critical relationship between radius r of jet and the axial distance z from nozzle is obtained for the straight jet. Draw ratio between the jet and the final fibres was pridicted theoritically.

. .

Manufacturing of aligned fibres array was easily achievable.

.

Processing parameters effects on the aligned fibres such as gab distance and collection time were investigated.

Publications In preparation: .

Systematic parameter study for ultra- fine nylon 6 fibre produced by electrospinning technique.

.

Electrospinning of aligned nanofibres with cotrolled deposition.

Conferences Paper and Poster: .

June 2006 (Electrospinning of nanofibres: potential scaffolds for medical applications) a presentation presented in Research in Support of Medicine, Health and Safety Conference, Edinburgh, UK.

.

June 2006 (Systematic parameter study for ultra- fine nylon 6 fibre produced by electrospinning technique), Poster presentation, Research in Support of Medicine, Health and Safety Conference, Edinburgh, UK.

Galashiels, Scotland, UK Heriot-Watt University, RIFleX, School of Textiles and Design, Netherdale, Galashiels TD1 3HF, UK, Tel: +44 1896 89 2135; Fax: +44 1896 75 8965; E-mail: [email protected] Principal Investigator(s): Prof. George K. Stylios Research Staff: Mohammad Mahfuzur Rahman Chowdhury

Research register

27

IJCST 19,6

Investigating Nano Fibre Production by the Electrospinning Process Other Partners: Academic

28

Industrial

None None Project start date: July 2004 Project end date: June 2008 Keywords: Electrospinning, Electrospinning process, Parameters, Polymer, Nanofibre application Electrospinning is a unique way to produce novel polymer nanofibres with diameter typically in the range of 10 nm to 500 nm. Using this process, the polymer nanofibres can be made from a variety of polymer solutions or melt to produce fibres for a wide range of applications. Electrospinning occurs when the electrical force at the surface of a polymer solution or melt overcomes the surface tension and causes an electrically charged jet to be ejected. When the jet dries or solidifies, an electrically charged fibre remains. This charged fibre can be directed or collected or accelerated by electrical forces, then collected in sheets or other geometrical forms. This research project is an investigation of the electrospinning process and the effect of process variables on orientation, crystallinity, microstructure and mechanical properties of the nanofibres produced. Some of the polymeric parameters investigated are polymer type, solvent type, molecular weight, solution properties, viscosity, conductivity and surface tension. In the case of process parameters, the electric potential, flow rate, concentration, distance between capillary and collection screen, ambient parameters are important

Project aims and objectives .

To investigate process-structure-property relationships in polymer fibres with nanosize diameters produced by electrospinning

.

To innvestigate the morphology and properties of the polymer nanofibres

.

To produce fibres at uniform diameters

Research deliverables (academic and industrial) .

Nanofibres of uniform diameter

.

Defined mechanical and physical properties

.

Process-structure-property relationships

.

Detailed understanding of the electrospinning process

.

Nanofibres suitable for applications such as air filtration, protective clothing, fibre reinforced support, and Biomedical.

Publication “Nano fibre and its medical application”, Poster presentation in Research in support of Medicine, Health and Safety” Conference, Heriot-Watt University, Scotland, UK.

Galashiels, Scotland, UK Heriot-Watt University, RIFleX, School of Textiles and Design, Netherdale, Galashiels TD1 3HF, UK, Tel: +44 1896 89 2135; Fax: +44 1896 75 8965; E-mail: [email protected] Principal Investigator(s): Prof. George K. Stylios Research Staff: Dr Taoyu Wan

Novel Micro-Channel Membranes for Controlled Delivery of Biopharmaceuticals Other Partners: Academic None

Industrial Stryker UK Ltd, St James’s University Hospital, Leeds General Infirmary, Camira Fabrics Ltd, Dinsmore Textile Solutions Ltd Project end date: March 2009

Project start date: April 2006 Project budget: £650,000 Source of support: DTI Technology Programme Keywords: Micro-channel, Micro-porous, Membrane, Controlled delivery, Drug release

This project, which stems out of research findings of an EPSRC-funded research, aims at developing micro channel structures (coatings, membranes, foams, etc.) with encapsulated biopharmaceuticals capable of controlled release by changes in temperature, pH, magnetic field or voltage. A technique for encapsulating biopharmaceuticals into micro-channels of a polymer matrix structure and controlling their subsequent release will be developed. Driven by the consortium, the technologies will be veered towards various pay-load bearing applications, e.g. in self-supporting materials for delivering bone growth hormones (INN eptotermin alpha) in bone fractures, or in coated textiles for the personal hygiene, healthcare, treatment and protective clothing industries.

Project aims and objectives The main objectives for the projects are to: .

Establish criteria for channel size and distribution control; engineer the structure and morphology of the porous material to suit the encapsulation of biopharmaceuticals and their slow release. This involves both self-supported materials and coated textiles.

.

Investigate UV-based techniques for the encapsulation of specific biopharmaceuticals, with particular reference to bone growth hormones.

.

Investigate strategies for the controlled release of the biomaterials from the channels within the developed membrane or coating.

Research register

29

IJCST 19,6

30

.

Characterise and test laboratory and pilot-scale samples - study release rates of the biopharmaceuticals, degradation/absorption rates and efficacy of the system.

.

Investigate alternative exploitation routes as coated textiles with encapsulated biopharmaceuticals.

The project will investigate possibilities both for implantable and non-implantable payload bearing materials, both as self-supported materials, and as coatings or membranes supported by a base fabric. Research work on implantable applications will however only be performed in laboratory scale with the aim to prove the concepts.

Research deliverables (academic and industrial) .

. .

.

Pilot-scale production of membranes, coatings, foams, or any other forms of channel-containing materials with engineered, controllable and tailor-made channel sizes, shapes and distributions. Defined characteristics, properties and performance of the as-produced materials A technique to encapsulate biopharmaceuticals within the membrane channels using a UV cross-linking technique, which eliminates the need to use high temperature treatments for encapsulation. This enables the biopharmaceuticals to maintain their effectiveness. The final expected result is the development of a system for activated release, e.g. following a change in stimuli or external conditions, such as temperature, pH, magnetic field or voltage. This will enable the controlled and targeted release of the biopharmaceutical as, when and where required for optimum treatments after fractures.

Publications Stylios, G.K., Giannoudis, P.V., Wan, T., (2005), “Applications of nanotechnologies in medical practice”, Injury, Vol. 36S, pp. S6-S13. Stylios, G.K., Wan, T., Giannoudis, P.V., (2006), “Present status and future potential in the enhancement of bone healing using nanotechnology”, Accepted for publication in Injury.

Galashiels, Scotland, UK Heriot-Watt University, RIFleX, School of Textiles and Design, Netherdale, Galashiels TD1 3HF, UK, Tel: +44 1896 89 2135; Fax: +44 1896 75 8965; E-mail: [email protected] Principal Investigator(s): Prof. George K. Stylios Research Staff: Liang Luo

Interactive Wireless and Smart Fabrics for Textiles and Clothing Other Partners: Academic

Industrial

None

None

Project start date: September 2002 Project end date: September 2008 Project budget: None Source of support: Worshipful Company of Weavers Keywords: Smart, Interactive, Textiles, Garment, Clothing, Sensors, Wireless The last few years have witnessed an increased interest in wearable technologies, smart fabrics and interactive garments. This has come about by certain technological innovations n the areas of sensor-based fabrics, micro devices, wire and wireless networks. In terms of textiles, most of current developments are towards the fashion markets and have resulted in glorifying garments as gimmicky gadgets. However, some efforts are also being directed in using the technology for improving the quality of life, or even for life saving purposes. Examples of such uses can be found in the military, healthcare, fire fighting, etc. This research project investigates new interdisciplinary technologies in fabrics, sensors and wireless computing, for the development of a prototype interactive garment for monitoring various functions of the wearer.

Project aims and objectives The general aim of the project is to develop technologies for use in interactive garments, which can provide monitoring functions for various applications such as the clinical or healthcare sector. More specifically, objectives are: .

Develop suitable wireless sensors for various measurements, including ECG, temperature, breathing, skin conductivity, mobility and movement, humidity, positioning, etc.

.

Develop a Personal Area Network and a Wireless Communication Centre.

.

Optimise suitable wireless technologies such as Bluetooth to enable communication between sensors and a central processing unit.

.

Conceptualise a smart multilayer fabric.

.

Integrate technologies.

Research deliverables (academic and industrial) .

Wireless sensors for physiological and other measurements.

.

Wireless communication centre for relaying information between sensors, wearers, central processing unit and Internet.

.

Conceptual multilayer fabric suitable for interactive garments.

Publications Stylios, G.K., Luo, L., (2003), “Investigating an interactive wireless textile system for SMART clothing”, 1st International Textile Design and Engineering Conference (INTEDEC 2003), Fibrous Assemblies at the Design and Engineering Interface, Edinburgh, UK, 22-24 September. Stylios, G.K., Luo, L., (2003), “The concept of interactive, wireless, SMART fabrics for textiles and clothing”, 4th International Conference, Innovation and Modelling of Clothing Engineering Processes – IMCEP 2003, Maribor, Slovenia, 9-11 October.

Research register

31

IJCST 19,6

32

Stylios, G.K, Luo, L., (2004), “A SMART wireless vest system for patient rehabilitation”, Wearable Electronic and Smart Textiles Seminar, Leeds, UK, 11 June. Stylios, G.K., Luo, L., Chan, Y.Y.F., Lam Po Tang, S., (2005), “The concept of smart textiles at the design/technology interface”, 5th International Istanbul Textile Conference, Recent Advances and Innovations in Textile and Clothing, Istanbul, Turkey, 19-21 May.

Galashiels, UK Heriot-Watt University, School of Textiles & Design, Netherdale, Galashiels, Scotland TD1 3HF, UK, Tel: +44 1896 892140; Fax: +44 1896 758965; E-mail: [email protected] Principal Investigator(s): R H Wardman and R M Christie Research Staff: S Lam Po Tang

Multifunctional Technical Textiles for Construction, Medical Applications and Protective Clothing Other Partners: Academic

Industrial

9 universities across Europe 13 SMEs across Europe Project start date: 1 May 2006 Project end date: 30 April 2010 Project budget: e12.7million Source of support: European Union Keywords: Protective textiles, Digital printing The project involves a consortium of 13 high tech and less RTD intensive SMEs across the enlarged European Union, together with 3 non-SMEs and 9 universities and research institutes. The project will involve the development of digital printing procedures and equipment to enable the disposing of fluids of various chemical agents to textile fabrics to impart functionalities geared towards protection.

Project aims and objectives The objective of the project is to develop breakthrough technology based on digitally microdisposing fluids on textiles, enabling high-speed protective functionalization, continuous processing and customised production. Digital microdisposal has the ability of exact localisation and patterning of functionalities in multilayer textile substrates, integrating advanced thermo- and hydro-regulation, sensorics, actuating and controlled release functions, based on nanotechnology and multifunctional materials.

Research deliverables (academic and industrial) Creation of new knowledge in mechatronics and micro-fluids, nanotechnology and multifunctional materials. Understanding the behaviour at nano-level of droplets on textile substrates.

Technology breakthrough in textiles, enabling new functionalities; mass customisation and launching of new product services systems integrating the supply chain. New standards in personal protective equipment by enabling functionalities that empower the worker (sensorics and controlled release) to customise functionalities to specific uses. A shift from water-based processes, thereby achieving considerable savings in water, chemicals, and effluent. Publications Not available

Galashiels, Scotland, UK Research Institute for Flexible Materials, School of Textiles & Design, Heriot-Watt University, Galashiels, Selkirkshire TD1 3 HF, Tel: +01896 892135; Fax: +01896 758965; E-mail: [email protected] Principal Investigator(s): Prof. G K Stylios, RIFleX, Heriot Watt University (Academic), Dr K. Lee, Unilever Research (Industrial) Research Staff: L. Luo plus others in partner universities.

Multi-scale Integrated Modelling for High Performance Flexible Materials Other Partners: Academic

Industrial

University of Nottingham and Manchester Unilever, OCF Plc, TechniTex Faraday Ltd, Crode International Plc, ScotWeave University Ltd, Airbags International Ltd, Moxon of Huddersfield Ltd, Carrington Carrer & Workwear Ltd Project start date: 1 January 2007 Project end date: 31 December 2010 Project budget: £1.7 Million Source of support: Department of Trade and Industry DTI Keywords: Modelling, Yarn, Fabric, Garment, Hagh performance This is a flagship proposal for the UK. It is based on integrating micro, meso and macro scale structure/property and deformation models for high performance flexible materials. The outputs will be industry targeted solutions for predicting the properties and behaviour of high performance flexible materials in deformed states during usage including garments. The proposal stems from the modelling achievements of the three academic partners, combining their complimentary work, and after integration, adapting them for industry use, with Unilever, the lead partner, and 12 companies

Research register

33

IJCST 19,6

34

covering diverse applications of the outputs, attempting to represent this sector. OCF and Scotweave will help commercialising the output in the form of a software product or licence. The main aim of the project is to develop the models such that they can be used primarily by the high performance textiles and garment industry, but also by other industries dealing with flexible materials. The key strength of the proposal is that for the first time an attempt is made to model the behaviour of flexible materials in a 3D manner, taking into account the dynamic changes of performance related properties with physical changes during use. .

Bridging the gap between (a) industry needs for predictive and development tools and (b) academic modelling efforts on two levels: structure/property micro/meso scale and macro-scale whole flexible structures such as high performance clothing.

.

Modelling of overall properties and performance of simple deformed textile structures such as draped/creased/folded fabrics, predicting the dynamic changes of performance related to physical deformations. This will be an innovative engineering tool for industries involved in designing, developing and manufacturing flexible materials, initially targeted at fabrics, but with applications to paper and thin films.

.

Modelling simple high performance whole garments based on structural/geometrical parameters of the flexible materials, to predict the changes in whole garment performance caused by yarn and fabric changes. No model that can perform this function for the high performance clothing industry is yet available. This aspect of the work will be a powerful tool for the high performance clothing and garment industry (high performance medical wear, protective clothing, sportswear, etc.)

The project will lead to two way modelling: predicting properties and performance of deformed textile materials and whole garments from structural, mechanical and geometrical parameters, or vice versa, i.e. generating structural, mechanical and geometrical requirements for specific end-product characteristics. Focusing on the high performance clothing sector, the work will provide a first-in-its-kind tool for design, development, engineering and optimisation of high performance flexible materials and exploited by a consortium of diverse companies, three of which are technology providers and the rest users of the technology. Different mechanisms (e.g. interfacial modification) can be exploited in the manufacture, modification, cleaning and care of high performance multi-component textiles. This will permit rational design of these materials and associated products, reducing the need for experimental development and testing. This project will push the boundaries in multi-scale modelling by building on leading edge expertise in fibre to yarn scale modelling at Manchester University, yarn to textile scale modelling at Nottingham University, and textile to whole flexible structures and clothing at Heriot-Watt University. By developing and interfacing these areas closely coupled tools will allow completely predictive models to be developed. It will also be the first time that modelling techniques have been used to understand the effects of interfacial properties on textile performance, leading to increasing innovation in high performance textile products.

Project aims and objectives .

Combining the strong modelling expertise of the four academic partners to create multi-scale integrated models for predicting structural properties and performance of flexible materials.

.

Optimising the integrated models for industry use, focusing in the first instance at high performance clothing and technical textiles.

.

Increasing time and cost efficiency of industry in product design and fabrication, starting with the high performance clothing industry.

.

The objectives address industry needs for new and improved product design and fabrication aids for high performance materials e.g. in view of the London 2012 Olympics (high performance sportswear, anti-terrorism protective clothing, etc.). They will enable the prediction of properties and performance before manufacturing, hence accelerating the development stage, reducing costs and increasing competitiveness.

Research deliverables (academic and industrial) .

.

.

.

A Virtual Testing capability will be developed and transferred into industry, reducing the amount of experimentation required for new product development (new textiles, treatments and laundry products designed and manufactured) and thereby reducing the associated costs and waste generation (improving sustainability of production). Direct beneficiaries of this work will be manufacturers of apparel, textile, technical and allied industries. Indirect beneficiaries will include manufacturers of carpets, non-wovens, composites, paper and structural materials, as well as retailing and software (after commercialising). Bridging the gap between (a) industry needs for predictive and development tools and (b) academic modelling efforts on two levels: structure/property micro/meso scale and macro-scale whole flexible structures such as high performance clothing. Modelling of overall properties and performance of simple deformed textile structures such as draped/creased/folded fabrics, predicting the dynamic changes of performance related to physical deformations. This will be an innovative engineering tool for industries involved in designing, developing and manufacturing flexible materials, initially targeted at fabrics, but with applications to paper and thin films. Modelling simple high performance whole garments based on structural/geometrical parameters of the flexible materials, to predict the changes in whole garment performance caused by yarn and fabric changes. No model that can perform this function for the high performance clothing industry is yet available. This aspect of the work will be a powerful tool for the high performance clothing and garment industry (high performance medical wear, protective clothing, sportswear, etc.)

Publications Too early.

Research register

35

IJCST 19,6

36

Galashiels, Scotland, UK Research Institute for Flexibel Materials, School of Textiles & Design, Heriot Watt University, Galashiels, Selkirkshire TD1 3 HF, Tel: +01896 892135; Fax: +01896 758965; E-mail: [email protected] Principal Investigator(s): Prof. G. K. Stylios Research Staff: X. Zhao

Integration of CFD and CAE for Design and Performance Assessement of Protective Clothing Other Partners: Academic None

Industrial Tilsatec Ltd, TechniTex Faraday Ltd, Camira Fabrics Ltd, St Jame’s University Hospital, Remploy Ltd, Pil Membranes Ltd, Altair Engineering Ltd. Project end date: 31 May 2010

Project start date: 1 June 2007 Project budget: £600,000 Source of support: Engineering and Physical Sience Rerearch Council EPSRC and Department of Trade and Industry DTI Keywords: CFD, CAE, Protective fabric, Garment, Apparel, Modelling high performance This collaborative proposal aims at improving and developing new textiles for protective clothing by integrating Computational Fluid Dynamics (CFD) and Computer Aided Engineering (CAE). A new industrial tool for predicting the diffusion of chemical and/or biological (CB) agents through multilayer, non-homogenous flexible porous materials such as fabrics and whole garments will also be established. Multi-layer textile structures (flat and shaped) and simple garments will be modelled, using equations of mass, heat and momentum balance, integrated with human computational representation. The outcome will be optimisation of current commercial fabrics/garments and developing new protective clothing whilst, at the same time offering a new objective tool for product designers, engineers and developers to predict and evaluate performance of CB protective products.

Project aims and objectives .

Develop and integrate CAE/CFD for modelling of a clothed human to predict the performance of textiles and garments in CB protective applications, linking with fabric design and manufacturing.

.

Establish an objective measure protocole for CB protection end uses.

.

Improve protection performance of (PPE) and (PC) through new product development for extreme conditions.

.

Develop new materials by better understanding of the complex interactions between the flow of CB agents and textile/materials properties.

Research register

Research deliverables (academic and industrial) .

With a worldwide focus on CB agents, the project is timely for industry, considering the legislative demands for public health, safety and security. Growing concerns from government, the public and private sector about threats of terrorist attacks or epidemic outbreaks have led to increased performance and evaluation requirements. Increasing emphases on security for international events (e.g. London 2012) are also drivers for the above innovations. The proposed project addresses these issues by development key technologies for immediate and long-term use. The consortium experience can achieve the project objectives and lead to a major breakthrough in PC.

Publications Too early.

Huddersfield, UK University of Huddersfield, Queensgate, Huddersfield HD1 3DH, Tel: +44 1484 472563; Fax: +44 1484 472826; E-mail: [email protected] Principal Investigator(s): Paul Squires Research Staff: Cari Morton

word4word Other Partners: Academic TEKO Centre - The Danish Academy of Design, Technology, Textile, Denmark, Handelsfagsskolen Business and Retail College, Denmark, Baronie College, Breda, Holland Project start date: September 2004 Project budget: UK Budget: e53,000 Source of support: Leonardo Funds

Industrial Dansk Textil Union, Danish Retailers Organisation, Copenhagen, Denmark, IC Companies, Copenhagen, Denmark, ECCO Shoe, Areal, Portugal Project end date: December 2006

The textile/clothing, fashion and footwear industries in Denmark and in Europe are remarkable in experiencing an unsurpassed globalisation process and technological development and at the same time suffering from an educational backlog especially in terms of language competences and terminological knowledge, which often lead to costly mistakes and cross-cultural misunderstandings. Experience and reports from the industry show that the knowledge society and knowledge-based economy vigorously call for a larger degree of flexibility and readiness for innovation which can only be

37

IJCST 19,6

38

achieved through continuous competence building and the development of a common terminological platform. Consequently companies and institutions of education will join forces in an International/European network to develop a common platform for knowledge-sharing by establishing a combined online terminology database and visual training system. The International/European dimension is needed in national training in companies and for students at textile/design schools who are to take up positions in companies operating worldwide. At production sites the database will provide in annovative teaching tool for competence building. The partners will involve the knowledge and competences of their own institutions in a collaboration where their expert knowledge within the various stages of the value chain from raw material to finished product will complement each other. Given the sheer scope and limited time span of the project and to ensure quality, the project will cover terminology in the stages from raw material to sem-finished product in English, German and Danish. Subsequent projects will extend the database from semimanufactures to finished articles and extend the database into other languages.

Project aims and objectives Aims: .

To develop an online terminology database in Danish, English and German for the textile/clothing, fashion and footwear industry.

.

To strengthen the students’ knowldge of the terminology used in the industry thus enhancing their skills and employability.

Objectives: .

To provide a flexible, visual hands-on teaching tool.

.

To provide companies with a tool to achieve consistency and homogeneity in vocabulary thus eliminating costly mistakes and preventing cultural misunderstandings.

.

The work in the project will imply the implementation of a terminology bank for teachers, students and companies.

.

To establish a lasting network between the partners and to strengthen and develop this further as the terminology database is to be extended into other languages. When this network has been firmly established, it will open up for all interested parties and countries in the EU.

.

To strengthen the industry’s ties to trade assocations and other organisations.

.

To utilise the project results from the textile/clothing, fashion and footwear industry within other trades needing to obtain the same methodology for teaching and homogeneity in terminology.

Research deliverables (academic and industrial) .

An online database in English, Danish and German to strengthen consistency and homogeneity of the terminology used within the textile/clothing, fashion and footwear industry.

.

A database for visual teaching materials for teachers, thus introducting a new teaching methodology.

.

A web-platform for interactive information and exchange of information for teachers, students and companies.

.

Consequently, the users can contribute to the contents of the database by using the link provided for suggestions.

.

A database and interactive tool for anyone taking an interest in terminology. Students at textile and design schools who are studying to become employees in the textile/clothing, fashion and footwear industry, be it as controllers, designers, purchasers or exporters.

.

.

Teachers at textile and design schools who undertake special subject teaching within textile/clothing and footwear.

.

Teachers who are trainers at production sites. Companies within the industry.

.

Publications The results of the project work will be disseminated in various ways: Internallly in partner institutions: The companies will implement the database in their daily workin the mother company, subsidiaries and with their suppliers worldwide. The institutions of education will integrate the results in the curricula of the courses offered and place the results on their intranets. Teachers will use the templates as teaching materials. Externally: The partners will inform the public about the results by advertising in trade journals, participating in fairs and holding seminars and information meetings natiionally. Local/national and international trade organisations, chambers of commerce and institutions of education will be informed of the results. A printed leaflet and a CD-Rom will be made about the possibilities for using the results of the project, besides information about the www.word4word.dk site will be given. This will ensure fast access to new knowledge for any interested person besides students, teachers and companies. The database will continuously be extended after the project period and expanded into other European languages. Once the project has been completed the database will become subject to commercialisation on a subscription basis to ensure continous development and up-dating of the database.

Huddersfield, England University of Huddersfield, Department of Design, Queensgate Campus, Huddersfield, West Yorkshire HD1 3DH, Tel: +44 1484 472928; Fax: +44 1484 472826; E-mail: [email protected] Principal Investigator(s): Helen Woodget

Research register

39

IJCST 19,6

How to Realise Complex Textile Craft Techniques by Cad and Ink Jet Printing Other Partners: Academic

40

Industrial

None Dr R Annable, Dr J Pearson, Ms P Macbeth Project start date: June 2005 Project end date: April 2007 Keywords: Ink jet printing, Craft, Colour management, Simulation of craft techniques There are many techniques in textile crafts which are virtually impossible to produce commercially, for example hand and computer generated embroidered textiles and hand knotted woven fabrics. With the growing interest in the consumer market in the UK particular for textile craft products, it is becoming increasingly important to develop reproducible techniques to enable these products to be made to a commercially acceptable level. With the capability of ink jet printing being able to print at high resolution and in almost true colour, it is possible to simulate these techniques on fabric. In doing so, accurate colour representation must be achieved; the method of colour management used will be examined. This research describes the steps required to obtain identical images of the textile craft onto a CAD system, with the correct resolution, number of colours and size. This research is also pushing the boundaries of what is possible to print on using textile ink jet technology and pre-manipulated fabrics. Using a variety of fabrics that previously have not been able to print using ink jet technology; looking into custom treating and the possibility of fixing pigment inks.

Project aims and objectives The aim of this research is to achieve an accurate representation of hand-developed craft techniques using CAD software and textile ink jet technology, and to study the problems that may arise with this and to produce high quality printed fabric results. They will need to be commercially viable and in doing so the beginnings of a catalogue will be generated allowing for choice in custom made printed textiles. The process will be documented for future production of simulated craft techniques onto printed fabrics, resulting in the discussion of the use of colour management, CAD software and potential uses.

Research deliverables (academic and industrial) .

.

The outcomes of this research will be introduced into the teaching of these techniques into the under graduate courses. BA/BSc (Hons) Textile Design for Fashion and Interiors.

.

BA (Hons)Textile Crafts.

.

BA (Hons) Surface Design. BA (Hons) Costume with Textiles at the University of Huddersfield, and

. .

MSc in CAD in Textile Technolgy.

Publications The use of Wide Format Ink Jet Printing for Textile Sampling, paper presented at AUTEX Conference, July 2002, Bruges, Belgium. The Use of Embellishment Techniques with Ink Jet Printed Repeated Designs, paper presented at INDETEC Conference Setember 2003, Edinborugh. The Commercial Realistaion of Complex Textile Craft Techniques by Cad and Ink Jet Printing, paper presented at Archtex Conference, September 2005, Krakow, Poland. Craftily using Ink Jet, paper published in ‘Textiles’ Journal Issue, Vol. 33, No. 2, pp. 20-22.

Izmir, Turkey Dokuz Eylul Uni, Deu Muhendislik Fakultesi Tekstil Muhendisligi Bol. Bornova Izmir Turkiye, Tel: +902323882869; Fax: +902323882867; E-mail: [email protected] Principal Investigator(s): Ass. Prof. Fatma Ceken Research Staff: Gulsah Pamuk (MSc in Textile Eng)

Physical Properties of Automotive Seat Fabrics Other Partners: Academic

Industrial

None Various Fabric Producers in Turkey Project start date: September 2005 Project end date: September 2007 Project budget: 10000 USD Source of support: Dokuz Eylul Uni. and Turkish Fabric Producers Keywords: Automotive seat fabrics, warp knitted fabric, woven fabric, velvet fabrics People spend more time in their cars with increased daily commuter distances to and from work, increased traffic density, more people work away from from home and travel long distances by car at the weekends.For this reason the automobile manufacturers not only improve the technical and mechanical primary functions of the automobile but also improve the secondary parts that appeal to the senses, such as seats, door panels, carpets etc. Seat covers is one of the most important factor for interior car design and for passenger’s comfort. The main fabric technologies used for seat covers are woven velour, jacquard woven, warp knit, warp knit pile sinker, warp knit double needle bar raschel velour, circular knit jacquard pile and spacers. The warp knitted fabrics have a structure that can be controlled and spesific level of stretch and recovery. The major limitiaiton of warp knitted fabrics is their relatively limitted capability to produce large area of color and design effects. The spacer fabrics, when used in seat covers, provide temperature control, comfort and support.

Project aims and objectives In this study the aim is to compare the physical properties of automotive seat fabrics which are produced using different manufacturing techniques. Publications Not available

Research register

41

IJCST 19,6

42

Jalandhar, India National Institute of Technology Jalandhar, Department of Textile Technology, Dr B. R Ambedkar National Institute of Technology, Jalandhar, G T Road Bye Pass, Jalandhar-144011, Punjab (India); Tel: +91 0181 2690301, 2690302; Fax: +91 0181 2690932, 2690320; E-mail: [email protected], [email protected] Principal Investigator(s): Vinay Kumar Midha

Development of Needle Punched Nonwoven Blanket Fabric Other Partners: Academic

Industrial

Dr S. Ghosh None Project start date: 1 April 2004 Project end date: 30 September 2007 Project budget: Rs 15 LACS Source of support: Ministry of Human Resource Development, India Blankets are generally woven or knitted and are made from pure wool or blends of wool and acrylic fibres. Woven or Knitted fabric production is a lengthy, laborious and costly process. Whereas the needle punched fabric production is a short process and the manufacturing cost of needle punched nonwoven fabric is much lower. A nonwoven blanket is defined as a blanket produced by bonding or interlocking of fibres or both, accomplished by mechanical, chemical, thermal or solvent means or combination thereof. The important parameters, which affect the structure and properties in a needle punch nonwoven blanket, are the web parameters (type of fibre and fabric weight per unit area) and machine parameters (depth of needle penetration and needle punch density). The important requirements for end use performance of blankets are good handle, softness, good abrasion resistance, good thermal insulation in winter and good air permeability in summer. Fabric abrasion resistance increases linearly with the increase in needle punch density, but as the needle punch density is increased the other important properties like thermal resistance and handle are adversely affected. So a fabric with good thermal and handle properties will have poor abrasion resistance and vice-versa. So to design a blanket with good abrasion resistance without spoiling the softness and thermal properties, the machine parameters and web parameters are to be systematically optimized.

Project aims and objectives The objective of the project is to produce a cost effective nonwoven blanket by optimising the machine parameters to produce good thermal, softness and abrasive properties

Research deliverables (academic and industrial) Publications Not available

Jalandhar, India National Institute of Technology Jalandhar, Depratment of Textile Technology, National Institute of Technology, Jalandhar – 1440111, India; Tel: +91 181 2690301 2, 453; Fax: +91 181 2690320, 2690932; E-mail: [email protected] Principal Investigator(s): Dr Arunangshu Mukhopadhyay Research Staff: Ms. Sunpreet Kaur

Designing Nonwoven Fabric for Pulse-jet Filtration Other Partners: Academic

Industrial

Project start date: 31 March 2005 Project end date: 30 September 2008 Project budget: Rs. 15 lacs Source of support: Ministry of Human Resource Development, Government of India Keywords: Dust particle, Nonwoven, Pulse-jet filtration Primary factors which determine the selection of a fabric for a particular application are state of aerosol medium (thermal and chemical condition, static charge on the particles, abrasive particles, moisture in the gas stream etc.), filtration requirement (particle capturing efficiency, pressure drop and cleaning) and equipment consideration. The particle capturing mechanism by fabric filters based on reverse jet cleaning is well researched. A number of studies reveal the effect of fabric structural parameters on its filtration characteristics. Further a number of attempts have been made to improve filtration efficiency with reduced pressure drop and better cake release performance. However, it is worth mentioning that the above studies becoming rudimentary since pulse jet filter has become an attractive option of particulate collection utilities. Pulse jet cleaning is a technique whereby a short, periodic, high pressure burst of air is fired into the clean side of the fabric. The particles are dislodged and the pressure drop falls to an acceptable level. Pulse jet filtration can meet the stringent particulate emission limits regardless of variation in the operating conditions. The cleaning device is less expensive than other type of mechanism and requires considerable less space. Other merits of pulse jet fabric filters are high collection efficiency, on line cleaning application and outside collection which allows the bag maintenance in a clean and safe environment. The increasing adoption of pulse jet filters for control of process emissions and recovery of utility dusts has stimulated research on many aspects of their operation. Despite their wide applications, the functioning of fabric filters is poorly understood. The operating condition is usually specified by the manufacturer at the time of commission. In practice, however, the filtration process frequently undergoes changes in the operating condition caused by the disturbances due to the variation in the concentration of dirty air, variation in particle size etc. Due to the complex characteristics inherent in the filtration and cleaning process of pulse jet bag houses, the proper understanding of the role of fabric construction on filtration is very important.

Research register

43

IJCST 19,6

Generation of such knowledge would represent an important first step towards designing fabric with improved barrier properties.

Project aims and objectives

44

The separation of solids from fluids by textile filter media is an essential part of countless industrial process, contributing to purity of product, saving in energy, improvement in process efficiency, recovery of precious materials and general improvements in pollution control. The dust may create environmental pollution problems or other control difficulties caused by their toxicity, flammability and possible risk of explosion. The particles in question may simply require removal and be of no intrinsic value or alternatively may constitute part of a saleable product, for example, sugar or cement. Among several techniques, the most efficient and versatile is the fabric collector, especially when processing very fine particles. Fabric filtration under pulse jet situation becoming very common in industrial bag house filtration. However, there is lack of study on the role of fabric under the said circumstances. It is also important to develop the fabric suitable for pulse-jet filtration. Therefore, we are having the following objectives: . .

.

Understanding the mechanism of filtration under pulse-jet situation. Studing the effect of geometrical parameter of nonwoven fabric on filtration performance. Design and development of suitable nonwoven fabric under pulse-jet situation.

Research deliverables (academic and industrial) Initially the work will be concentrated on studing the impact of structural parameter of nonwoven fabric on its filtration performance in case of cement dust. This will lead to the development of suitable nonwoven fabric under pulse-jet situation in cement industry. At a later stage the work will be extended for optimising fabric design for other dust particles. Publications Not available

Lodz, Poland Technical Univerity of Lodz, Technical University of Lodz, Department of Clothing Technology, Zeromskiego 116 Str, 90-924 Lodz; Tel: +48 042 631 33 21; Fax: +48 042 631 33 20; E-mail: [email protected] Principal Investigator(s): Iwona Frydrych Research Staff: Renata Krasowska

Influence of work conditions of the disc take-up on characteristics of lockstitch Other Partners: Academic

Research register

Industrial

None None Project start date: 20.05.2005 Project end date: 19.05.2008 Project budget: 49100 PLN Source of support: Ministry of Science and Higher Education Keywords: Lockstitch, Lockstitch machine formation zone (STS), Thread demand in STS, Thread feed by the take-up The project presents an elaboration of original model simulating a thread movement in the zone of stitch creaction, on the basis of which a mechanism of take-up disc will be built. It will be evaluated in experimental research. The regulating point at this type of take-up will be presented as a novelty. According to this, reactions of the machine operator on the changes of sewing thread properties will be enabled and useful properties of lockstitch will be created. It requires elaborating the set of criteria assessing an action of this type of take-up disc.

Project aims and objectives The aim of the project is testing the working conditions of the disc take-up disc of the sewing thread taking into consideration the development of its construction of regulation points in the relation to existing solutions.

Research deliverables (academic and industrial) New construction of mechanism of take-up disc should create possibilities to adjust the parameters of its construction to thread sewing parameters and useful properties of the lockstitch Publications 1. R. Krasowska, I. Frydrych: “Possibilities of modelling the control conditions of thread by the disc takeup in the lockstitch machine”. Fibres & Textiles in Eastern Europe, Vol. 1, 2006. 2. R. Krasowska, I. Frydrych: “Formation of the thread control curve by the disc take-up in the lockstitch machine”. Research Journal of textile and Apparel, 2006.

Loughborough, UK Loughborough University, Environmental Ergonomics Research Centre; Dept. Human Sciences, Loughborough University, Loughborough LE11 3TU; Tel: +44 (0) 1509 223031; Fax: +44 (0) 1509 223940; E-mail: [email protected]

45

IJCST 19,6

Principal Investigator(s): Prof. George Havenith Research Staff: Lucy Dorman

Project title: ‘Thermprotect’, Assessment of thermal properties of Protective clothing and their use

46

Other Partners: Academic

Industrial W.L.Gore

Lund University Sweden; EMPA, Switzerland; TNO, Netherlands; INRS, France; TU-Tampere, Finland; Ifado, Germany Project start date: 2003 Project end date: 2007 (extension) Project budget: 700ke Source of support: European Union Keywords: Clothing heat & mass transfer; Radiation, Wet clothing, Energy use due to clothing, Wind This project focuses on effects of radiation, wind and wetting of clothing layers on the thermal strain experienced in Personal Protective Clothing (PPC) (Work Package 1 and WP2), on special issues with cold weather protective clothing (WP3), and on the impact of protective clothing on the energy consumption of workers (WP4). Heat transfer through PPC with different radiant and moisture transport properties is studied. First on material samples, followed by thermal manikin measurements and finally in a select number of conditions, in human wear trials. Data will be used to derive general models of either analytical (WP1, WP4) or conceptual (WP2) nature of heat transfer through PPC in relation to clothing material properties, climate etc. The last stages of the project focus on the effect of protective clothing on energy use. Most protective clothing increases energy use disproportionally more than the weight would suggest. This effects wearers in terms of work capacity, time to exhaustion and e.g. for firefighters in the time available before the compressed air runs out.

Project aims and objectives .

To improve the representation of clothing in ISO and EN heat and cold stress assessment standards.

.

To get a better understanding of the interaction of moisture and heat transfer. To understand which factors influence the energy requirement increase caused by clothing.

.

Research deliverables (academic and industrial) .

.

Data that will be used by standard writers to update the available standards on thermal stress and strain in the workplace. Publications will disseminate information to modellers of heat transfer and industry.

Publications 1. Bro¨de, P., Candas, V., Kuklane, K., den Hartog, E., Havenith, G., Bro¨de, P., Candas, V., den Hartog, E., Griefahn, B., Havenith, G., Holme´r, I., Meinander, H., Nocker, W. and Richards, M., “Effects of heat radiation on the heat exchange with protective clothing – a thermal manikin study”, Central Institute for Labour Protection – National Research Institute, 3rd European Conference on Protective Clothing (ECPC), Poland, May 2006, ISBN: 83-7373-097-4. 2. Bro¨de, P., Kuklane, K., den Hartog, E. and Havenith, G., “Infrared radiation effects on heat loss measured by a thermal manikin wearing protective clothing”, Environmental Ergonomics XI, Proceedings of the 11th International Conference, Ystad, Sweden, May 2005, pp. 74-79. 3. Dorman, L. and Havenith, G., “The Influence of clothing weight and bulk on metabolic rate when wearing protective clothing”, The Third International Conference on Human-Environmental System ICHES’ 05, Tokyo, Japan, September 2005, pp. 439-443. 4. Dorman, L., Havenith, G., Bro¨de, P., Candas, V., den Hartog, E., Havenith, G., Holme´r, I., Meinander, H., Nocker, W. and Richards, M., “Modelling the metabolic effects of protective clothing”, Central Institute for Labour Protection – National Research Institute, 3rd European Conference on Protective Clothing (ECPC), Poland, May 2006, ISBN: 83-7373-097-4. 5. Fukazawa, T., den Hartog, E.A., Daanen, H.A.M., Penders-van-Elk, N., Tochihara, Y. and Havenith, G., “radiant heat transfer network in the simulated protective clothing system under high heat flux”, The Third International Conference on Human-Environment System ICHES’ 05, Tokyo, Japan, September 2005, pp. 435-438. 6. Fukazawa, T., den Hartog, E.A., Daanen, H.A.M., Tochihara, Y. and Havenith, G., “Water vapour transfer in the simulated protective clothing system with exposure to intensive solar radiation”, The Third International Conference on Human-Environment System ICHES’ 05, Tokyo, Japan, 2005, pp. 202-205. 7. Gao, C., Holme´r, I., Fan, J., Wan, X., Wu, J.Y.S. and Havenith, G., “The Comparison of thermal properties of protective clothing using dry and sweating manikins”, Central Institute for Labour Protection – National Research Institute, 3rd European Conference on Protective Clothing, Poland, May 2006, ISBN: 83-7373-097-4. 8. Havenith G. “Clothing heat exchange models for research and application”, The 11th International Conference on Environmental Ergonomics, Ystad, Sweden; Proceedings published as Environmental Ergonomics 2005. Editors I Holme´r, K Kuklane and C Gao. Lund University, Lund, Sweden, ISBN 91631-7062-0, pp. 66-73. 9. Havenith, G., “Clothing heat exchange models for research and application”, Environmental Ergonomics XI, Proceedings of the 11th International Conference, Ystad, Sweden, May 2005, pp 66-73, ISSN 1650-9773. 10. Havenith, G., Wang, X, den Hartog, V.E., Griefahn, B., Holme´r, I., Meinander, H. and Richards, M., “Interaction effects of radiation and convection measured by a thermal manikin wearing protective clothing with different radiant properties”, The Third International Conference on HumanEnvironmental System ICHES 05, Tokyo, Japan, September 2005, pp. 47-50. 11. Havenith, G., Wang, X., Richards, M., Bro¨de, P., Candas, V., den Hartog, E., Holme´r, I., Meinander, H. and Nocker, W., “Evaporative cooling in protective clothing”, Central Institute for Labour Protection – National Research Institute, 3rd European Conference on Protective Clothing (ECPC), Poland, May 2006, ISBN: 83-7373-097-4. 12. Kuklane, K., Gao, C., Holme´r, I., Broede, P., Candas, V., den Hartog, E., Havenith, G., Holme´r, I., Meinander, H. and Richards, M., “Effects of natural solar radiation on manikin heat exchange”, Central Institute for Labour Protection – National Research Institute, 3rd European Conference on Protective Clothing (ECPC), Poland, May 2006, ISBN: 83-7373-097-4. 13. Dorman L., Havenith G. and Thermprotect Network, “The effects of protective clothing on metabolic rate”, The 11th International Conference On Environmental Ergonomics, Ystad, Sweden; Proceedings published as Environmental Ergonomics 2005. Editors Holme´r I., Kuklane K. and Gao C. Lund University, Lund, Sweden, 2005, ISBN 91-631-7062-0, pp. 82-85.

Research register

47

IJCST 19,6

48

14. Bro¨de P., Kuklane K., den Hartog E., Havenith G. and Thermprotect Network, “Infrared radiation effects on heat loss measured by a thermal manikin wearing protective clothing”, The 11th International Conference On Environmental Ergonomics, Ystad, Sweden; Proceedings published as Environmental Ergonomics 2005. Editors Holme´r I., Kuklane K. and Gao C. Lund University, Lund, Sweden, 2005.ISBN 91-631-7062-0, pp. 74-78. 15. Richards M., Rossi R., Havenith G., Richards M., Candas V., Meinander H., Broede P., Candas V., den Hartog E., Havenith G., Holme´r I., Meinander H. and Nocker W., “Dry and wet heat transfer through protective clothing dependent on the clothing properties and climatic conditions”, Central Institute for Labour Protection – National Research Institute, 3rd European Conference on Protective Clothing (ECPC), Poland, May 2006, ISBN: 83-7373-097-4. 16. van Es E.M., den Hartog E.A., Broede P., Candas V., Heus R., Havenith G., Holme´r I., Meinander H., Nocker W. and Richards M., “Effects of short wave radiation and radiation area on human heat strain in reflective and non- reflective protective clothing”, Central Institute for Labour Protection – National Research Institute, 3rd European Conference on Protective Clothing (ECPC), Poland, May 2006, ISBN: 83-7373-097-4. 17. Kuklane K., Gao C., Holme´r I., Giedraityte L., Bro¨de P., Candas V., den Hartog E., Meinander H., Richards M. and Havenith G., “Calculation of clothing insulation by serial and parallel methods: effects on clothing choice by IREQ and thermal responses in the cold”, International Journal of Occupational Safety and Ergonomics (JOSE) 2007, Vol. 13, No. 2, pp. 103–116. 18. Dorman L. and Havenith G., “Protective clothing and its effects on range of movement”, Proceedings International Conference on Environmental Ergonomics, 2007, Mekjavic S. and Taylor (eds). 19. Bro¨de P., Havenith G., Wang X. and Thermprotect network, “Non-evaporative effects of a wet mid layer on heat transfer through protective clothing”, Proceedings International Conference on Environmental Ergonomics, 2007, Mekjavic S. and Taylor (eds). 20. den Hartog E.A., Broede P., Candas V. and Havenith G., “Effect of clothing insulation on attenuation of radiative heat gain”, Proceedings International Conference on Environmental Ergonomics, 2007, Mekjavic S. and Taylor (eds). 21. Kuklane K., Gao C., Holme´r I., Bro¨de P., Candas V., den Hartog E., Havenith G., Meinander H. and Richards M., “Physiological responses at 10 and 25 c in wet and dry underwear in permeable and impermeable coveralls”, Proceedings International Conference on Environmental Ergonomics, 2007, Mekjavic S. and Taylor (eds). 22. Kuklane K., Gao C., Holme´r I., Broede P., Candas V., den Hartog E., Havenith G., Meinander H. and Richards M., “Validation of insulation and evaporative resistance calculations based on manikin and human tests”, Proceedings International Conference on Environmental Ergonomics, 2007, Mekjavic S. and Taylor (eds).

Manchester, UK Manchester Metropolitan University, Department of Clothing Design and Technology, Hollings Campus, Old Hall Lane, Manchester M14 6HR; Tel: +44 161 246 2676; Fax: +44 161 247 6329; E-mail: [email protected] Principal Investigator(s): Dr Jess Power and Dr Rose Otieno Research Staff: T.B.C.

Anthropometrical Data in Knitted garments Other Partners: Academic

Industrial

MIRIAD

T.B.C.

Project start date: 30 September 2006 Project end date: 30 August 2007 Project budget: £36732 Source of support: MIRIAD / Department (internal funding) Keywords: Knitted garments Anthropometrical Knitted garments and the relating technologies have advanced immensely during the last decade especially in the areas of fully fashioning (integral shaping) and complete garment production. Various new techniques have become available in modern knitting machinery including microprocessor-controlled mechatronic systems for needle selection and fabric take down systems, as well as automated loop transferring. These technologies have enabled knitwear to be shaped three-dimensionally as not previously achievable through the sophisticated CAD software that generates pattern and control during knitting. Despite this advancement in programming software and electronic needle selection, garment development is still heavily reliant on the skill of the operator or designer to develop new shapes and manipulate this advanced technology into innovative knitwear. This advance in technology has opened up a large knowledge gap with regards to fit and styling within knitwear, and many questions have arisen; Do we really understand the relationship between anthropometrical data and knitwear sizing, styling and fit and do we utilise the knowledge effectively to enable predictions of fit to occur? In previous research it was found that many new knitwear developments are based on empirical knowledge utilising the trial and error approach to develop styles with good fit which conform to the human figure and the theoretical application is very much lacking. This suggests that there is a void caused by the lacking of theoretical data and a knowledge gap between the relationship of anthropometrics and knitwear development. It is vital to assist the industry by providing a sound theoretical understand in order to minimise time wasted utilising the trial and error approach which appears to be considered the current industry norm. In addition to the strong technical justification for the project investigation regarding the use of anthropometrical data in knitwear development, there is evidence that during resent years knitwear has become a key fashion item, a greater percentage of a clothing range in retail outlets are knitwear based. It has also been acknowledged that the individual knitwear manufacturers are placing much more emphasis on the development of new shapes that conform to the human form. However, there is an uncertainty regarding how much of this is derived from empirical development and the quantity that has a sound theoretical framework grounded in anthropometrical studies.

Project aims and objectives .

A thorough literature search to determine the key players; and establish knowledge gaps in relation to the use of anthropometrical data in knitwear sizing.

.

Intense market research collating knitwear sizing and anthropometrical data.

.

Through visits, interviews will be conducted with five key manufacturers to evaluate the current industry standards in relation to knitwear sizing and the usage of anthropometrical data.

Research register

49

IJCST 19,6

.

.

50

Using laboratory based procedures analysis will occur by scientific evaluation into the current grading procedures in knitted garments, to determine the current industry standards. Devise a database of anthropometrical information related to knitwear styles and sizing.

Research deliverables (academic and industrial) It is thus proposed to conduct investigation through a research project into the use and understanding of anthropometrical data in knitted garments. This will occur in the form of a study into the development and production of current knitted goods by primary research involving extensive practical observation by scientific means. This will result in the analysis of four basic stylelines in a variety of sizing codes. In addition it is proposed that five UK knitwear manufacturers will contribute information regarding their internal knowledge of the use of anthropometrical data within their current knitwear developments. The second part of the study will involve the collating of information gained from the scientific observation and the primary research from the knitwear manufacturers. A Research Assistant funded entirely by the project will conduct the extensive laboratory observational research to enable the research outputs to be achieved. Publications

This research is grounded in previous postgraduate work and will expand on the department’s current research portfolio within anthropometrics and knitwear. The output will be in two phases: Phase 1 . .

Publication to outline the initial findings of the background study. A conference presentation based on the findings of the primary investigation.

Phase 2 . A framework for further research in the form of scholarly activity. .

Development of a short course to relate the findings directly back to the industry.

Maribor, Slovenia University of Maribor, Faculty of Mechanical Engineering, Smetanova ulica 16, SI-2000 Maribor, Slovenia, Tel: +386 2 220 7960; Fax: +386 2 220 7990; E-mail: [email protected] Principal Investigator(s): Prof. Dr sc. Jelka Gersˇak Research Staff: Research Unit Clothing Engineering

Clothing Engineering and Textile Materials Other Partners: Academic

Industrial

None

None

Project start date: 01 January 2004 Project end date: 11 December 2008 Project budget: 75.000 ECU for 2006 Source of support: Slovenian Research Agency Keywords: Clothing, Fabric, Fabric mechanics, Behaviour, Comfort, Prediction The research programme is based on complex research of woven fabric mechanics, matter properties and the relation of matter properties and the parameters of thermalphysiological comfort. It is subdivided into three thematically connected parts, which included: a) basic investigations of the mechanics of woven fabrics as complex textile structures, b) modellling the behaviour of complex textile structures, and c) characterising the parameters of thermal-physiological comfort. The first part of activities is based on the knowledge of mechanical laws of fabric behaviour at low loads and is aimed at defining the quality properties of garment form appearance. Fabric behaviour from the point of view of mechanics in processing from a plane into a 3D garment shape is determined, the importance of individual parameters of mechanical and physical properties is evaluated and a knowledge basis used to predict garment appearance is constructed. Based on the laws defined and appearance quality factors, we have developed the “Intelligent system for predicting garment appearance quality”, called InSyPrGAQ, or InSiNaKVO by Slovenian acronym. InSiNaKVO system is designed using the method of machine learning and represents an innovation – the first effective tool for objective evaluation and prediction of the degree of garment appearance quality. The system is based on the ratio among the parameters of mechanical and physical properties and the degree of garment appearance quality, and it is aimed at predicting the degree of garment appearance quality on the basis of the parameters of mechanical and physical properties of the fabrics built into the garment evaluated. It offers: . objective evaluation of the garment appearance quality degree; .

.

predicting five key factors of garment appearance quality degree: garment drape or fall, 3D form, fitting, the quality of the seams and garment appearance quality as a whole; planning the target degree of garment appearance quality;

.

predicting garment appearance quality degree for particular fabrics in a direct manufacturing process; and

.

planning the parameters of fabric mechanical and physical properties for the target degree of garment appearance quality.

The second part of activities involving modelling complex textile structures deals with continuation of the investigations of complex fabric deformations, with the accent on studying rheological models. Fabric relaxation is here studies from the point of view of mechanical multi-component models, which are able to describe in a satisfactory manner deformation and relaxation processes in fabrics, this being the starting point for numerical modelling of fabric behaviour, fabrics being complex textile structures. The third part of investigations deals with the characterisation of thermalphysiological comfort parameters. Comprehensive research of heat transfer and

Research register

51

IJCST 19,6

52

flow, as well as thermoregulation, together with the investigations of textile material properties have been performed. Investigations of matter properties and specific requirements imposed on textile materials of a particular target group have been performed for convetional and functional fabrics, as well as for the combinations, aimed at planning garment systems of business suits. The research is organised at three levels. The first includes investigations of thermal-physiological parameters on so called sweating cylinder for the selected fabrics and combinations of the planned garment systems (This research was performed in SmartWearLab of the Tampere University of Technology). The sweating cylinder simulates the functioning of human body allowing heat flow through the textiles or combination of textiles tested, as well as water vapour flow (the process of sweating), employing numerous sweat glands on the cylinder surface, which provide, i.e. transport particular amount of water to the surface. The purpose of the experiment is to establish the best combinastion of textile fabrics, such that could grant the required thermalphysiological comfort to the wearer. The second part of the research is aimed at evaluating and testing the business garment system by using a sweating thermal manikin, that simulates heat and moisture production in a similar way to the human body and measures the influence of garment system on heat exchange in different environmental and sweating conditions. The third level of the investigations includes evaluating and testing garment systems, empoloying testing persons under artificially designed weather conditions in a computer-controlled chamber. The results to be obtained will influence the characterisation of the thermal-physiological comfort parameters. It will be used to design a numeric simulation of heat exchange between the body and the garment, e.g. other textile products and the environment.

Project aims and objectives The main objectives of this project are as follows: .

To define the relationships between fabric mechanical properties and quality of a produced garment.

.

To develop a model for prediction of garment appearance quality.

.

To set-up a model of fabric behaviour as shell. To systematically investigate the parameters of thermophysiological comfort.

. .

To develop a numerical simulation of heat exchange between the human body and garment system or other textiles products and environment.

Research deliverables (academic and industrial) The research programme is systematically designed so as to result in the knowledge that could be used in establishing and engineering concept of garment and/or other textile product planning particularly the so called “knowledge-based products”. The results have partially been made applicable through the development of the InSiNaKVO system, the main characteristics of which are as follows: (a) knowledgebased designing and predicting garment appearance quality degree, using known fabric mechanical and physical properties (b) planning fabric mechanical and physical

properties so as to get the target garment quality level, and (c) tools to simulate the degree of garment quality. Publications Gersˇak, J. (2004), “Study of relationship between fabric elastic potential and garment appearance quality”, International Journal of Clothing Science and Technology, Vol. 16, Nos. 1/2, pp. 238-251. Zavec Pavlinic´, D., Gersˇak, J. (2004), “Vrednovanje kakvoc´e izgleda odjec´e”, Tekstil, Vol. 53, No. 10, pp. 497-509. Jevsˇnik, S., Gersˇak, J., Gubensˇek, I. (2005), “ The advance engineering methods to plan the behaviour of fused panel”, International Journal of Clothing Science and Technology, Vol. 17, Nos 3/4, pp. 161-170. Gersˇak, J., Sˇajn, D., Bukosˇek, V. (2005), “A study of the relaxation phenomena in the fabrics containing elastane yarns”, International Journal of Clothing Science and Technology, Vol. 17 Nos 3/4, pp. 188-199. Tama´s, P., Gersˇak, J., Hala´sz, M. (2006), “Sylview 3D Drape Tester – New system for measuring fabric drape”, Tekstil, Vol. 55, No. 10, pp. 497-509. Zavec Pavlinic´, D., Gersˇak, J., Demsˇar, J., Bratko, I. (2006), “Predicting seam appearance quality”, Textile Research Journal, Vol. 76 No. 3, pp. 235-242. Sˇajn, D., Gersˇak, J., Flajs, R. (2006), “Prediction of stress relaxation of fabrics with increased elasticity”, Textile Research Journal, Vol. 76, No. 10, pp. 742-750. Gersˇak, J., Marcˇicˇ, M. (2006), “Development of a mathematical model for the heat transfer of the system man – clothing – environment”, International Journal of Clothing Science and Technology, Vol. 19, Nos 3/4, pp. 234-241. Gersˇak, J., Marcˇicˇ, M. (2006), “Thermophysiological comfort”, Book of Proceedings of the 1st International workshop Design – Innovation – Development, Iasi, 24-27 July, pp. 143-147. Gersˇak, J., Zavec Pavlinic´, D. (2006), “Garment appearance quality as aesthetic function of clothes”, Annals of DAAAM for 2006 & Proceedings of the 17th International DAAAM symposium “Intelligent manufacturing & Automation: “Focus on mechatronics and robotics”, Vienna, Austria 8-11-November. Gersˇak, J. (2006), “Thermophysiological comfort of the driver”, Book of Papers of 2nd International Material Conference TEXCO, Ruzˇomberok, Slovakia, 17-18 August. Gersˇak, J., Marcˇicˇ, M. (2006), “Modeling the heat transfer of the system man – clothing – environment”, Book of Proceedings of 3rd International Textile, Clothing & Design Conference ITC&DC, Dubrovnik, Croatia, 8-11 October. Zavec Pavlinic´, D., Gersˇak, J. (2006), “Garment appearance quality: from subjective estimation to objective evaluation”, Book of Proceedings of 3rd International Textile, Clothing & Design Conference ITC&DC, Dubrovnik, Croatia, 8-11 October. Zavec Pavlinic´, D., Gersˇak, J. (2007), “The advanced method for the evaluation of garment appearance quality grade”, Proceedings of 7th Annual Textile Conference by AUTEX, Tampere, Finland, 26-28 June.

Maribor, Slovenia University of Maribor, Faculty of Mechanical Engineering, Smetanova ulica 16, SI-2000 Maribor, Slovenia, Tel: +386 2 220 7960; Fax: +386 2 220 7990; E-mail: [email protected] Principal Investigator(s): Prof. Dr sc. Jelka Gersˇak Research Staff: Prof. Dr Vili Bukosˇek, Dr Dunja Sˇajn, MSc. Rozalija Blekacˇ

Research register

53

IJCST 19,6

54

Study of the relationship between deformation and relaxation of fabrics containing elastane yarns in the spreading process Other Partners: Academic

Industrial

ELKROJ, modna oblacˇila d.d., Nazarje University od Ljubljana, Faculty of Natural Sciences and Engineering, Department of Textiles Project start date: July 2004 Project end date: June 2007 Project budget: 17.000 ECU for 2006 Source of support: Slovenian Research Agency Keywords: Fabrics containing elastane, Relaxation phenomena, Stress relaxation, Deformation, Spreading process In textile materials subjected to the impact of tensile force deformations appear in the form of increasing the length in the direction of force impact. After the relaxation because of a typical viscoelastic properties one part of deformation disappears immediately (elastic deformation), the other part disappears with time (elastic deformation retained) and the remaining part stays behind as a durable and irreversible deformation. The aim of the research project is to study the elastic behaviour of fabrics with elastane yarns taking into account the elastic potential, elastic deformations and relaxation ability based on the study of fabric mechanics, with the purpose to gain new cognition about the relationship between elastic deformations and relaxation phenomena in such fabric type. The main part of the research is focused on the study of elastic deformations and relaxation phenomena in fabrics with elastane yarns, where the influence of the parameters of tensional-elastic properties of such fabrics on elastic deformations and relaxation level will be investigated taking into account loading intensity. Special attention is given to the study of time dependence of relaxation phenomena in fabrics with elastane yarns and special relaxation conditions of such fabrics during spreading and laying into fabric layers, i.e., influence of the length, surface friction between fabric layers and number of layers. For this purpose the stress relaxation during witholding under constant deformations and the relaxation of fabrics after manually unwinding and spreading were studied. The relaxation of fabrics represents as elastic recovery and response of fabric as the result of loading during winding and spreading to form by lay. Based on given results of relaxation of fabric as the consequence of winding process was found that with higher percentage of containing elastane in the yarn grows the degree of deformation. The investigation results show also an important influence of the elastane yarn construction on relaxation phenomena in fabrics.

Project aims and objectives The main objectives od this research are as follows: .

To find the correlation between the elastic deformation and relaxation level in fabrics with elastane yarns.

.

To set up the model for predicting the relaxation of fabrics taking into account the parameters of fabric’s tensional-elastic properties, parameters of the spreading process, as well as the length and number of fabric layers.

Research register

Research deliverables (academic and industrial) . .

New knowledge of the elastic behaviour of fabrics with elastane yarns. The model for predicting the relaxation of fabrics with elastane yarns.

Publications Gersˇak, J., Sˇajn, D., Bukosˇek, V. (2005), “A study of the relaxation phenomena in the fabrics containing elastane yarns”, International Journal of Clothing Science and Technology, Vol. 17, Nos 3/4, pp. 188-199. Sˇajn, D., Gersˇak, J., Bukosˇek, V., Nikolic´, M. (2005), “Utjecaj konstrukcije prede s elastanom na relaksacijska svojstva tkanina”, Tekstil, Vol. 54, No. 10, pp. 489-496. Sˇajn, D., Gersˇak, J., Flajs, R. (2006), “Prediction of stress relaxation of fabrics with increased elasticity”, Textile Research Journal, Vol. 76 No. 10, pp. 742-750.

Nottingham, UK University of Nottingham, School of Mechanical, Materials & Manufacturing Engineering, University Park, Nottingham NG7 2RD, Tel: 0115 9513779; Fax: 0115 9513800; E-mail: [email protected] Principal Investigator(s): C. D. Rudd, A. C. Long, R. Brooks, I. A. Jones, S. J. Pickering, N. A. Warrior, M. J. Clifford, C. A. Scotchford, G. S. Walker Research Staff: H. Lin, L. Harper, R. R. H. Naqasha

Platform Grant: Processing and Performance of Textile Composites Other Partners: Academic

Industrial

None None Project start date: 1 February 2005 Project end date: 31 January 2009 Project budget: £445k Source of support: EPSRC Keywords: Textile composites, Unit cell analysis, TexGen Our current research portfolio is centred on the processing of polymer matrix composites with a growing emphasis on modelling and simulation. Given our high level of interest in textile-based composites and their growing importance in the field, we wish to introduce a common platform for our modelling studies based on our formalised textile

55

IJCST 19,6

56

generator (TexGen). Textile modelling provides a launchpad for downstream simulation of processing, damage mechanics and environmental performance. The functionality of our existing TexGen software will be extended and coupled to materials models for simulation of each of the above aspects of physical behaviour. Simulation of each of the physical processes will be enhanced by a common, interchangeable geometric definition of the textile structure within the rigid composite. This will enable a rapid understanding of fabric architecture effects to be built and the approach has excellent potential for application to other physical problems which relate to rigid and flexible composites or technical textiles. The platform grant application seeks continuity of support for key workers during the period of this development. Further details are available at: www.textiles.nottingham.ac.uk

Project aims and objectives .

Implement an approach based on textile modelling throughout our research portfolio, integrating the multiple streams of processing, energy management, biomedical applications and textile modelling.

.

Develop a series of downstream models relating to: three-dimensional permeability, formability (including shear compliance), static mechanical properties, damage mechanics and residual property estimation, diffusion and environmental degradation. Exploit the potential of the platform grant to raise our international profile, develop strategic links with other leading groups, and enhance our technology transfer activities.

.

Publications Not available

Ohtsu-City, Japan SCI-TEX, 12-15, hanazono-cho, Ohtsu-city, 520-0222 Japan, Tel: + 81 77 572 3332; Fax: + 81 77 572 3332; E-mail: [email protected] Principal Investigator(s): Tatsuki Matsuo

Propagation of knowledge on new textile science and technology Other Partners: Academic None Project start date: Project budget: Source of support:

Industrial None Project end date:

Keywords: Advanced technical textiles, Knowledge propagation The importance of advanced technical textiles has increased in the textile industry of developed countries. In addition, R & D on nano-technologies and electric-textiles are now intensively carried out. In this situation, propagation of knowledge on new textile science and technology must be meaningful. This project is being conducted individually by T. Matsuo through symposium lectures, journal articles, monographic books. Publications Not available

Patras, Greece University of Patras, University of Patras, Mechanical Engineering & Aeronautics Department, Robotics Group, 26500, Rio, Patras, Tel: +302610997268; Fax: +302610997212; E-mail: [email protected] Principal Investigator(s): Prof. Nikos Aspragathos (Project Coordinator), Evangelos Dermatas (Associate Professor), Emmanouil Psarakis (Assistant Professor) Research Staff: Panagiotis N. Koustoumpardis (Technical Researcher); George Zoumponos (PhD Student); Paraskevi Zacharia (PhD Student); Ioannis Mariolis (PhD Student); Ioannis Chatzis (PhD Student); George Evangelidis (PhD Student)

Handling of non-rigid materials with robots. Application in robotic sewing (XROMA) Other Partners: Academic (1) Mechanical Engineering & Aeronautics Department, University of Patras; (2) Department of Electrical & Computer Engineering, University of Patras; (3) Department of Computer Engineering & Informatics, University of Patras

Industrial (1) ELKEDE – Technology & Design Centre, Athens, Greece

Project start date: 26 May 2003 Project end date: 25 May 2007 Project budget: e220,103 Source of support: General Secretariat for Research and Technology, Hellenic Republic Ministry of Development

Research register

57

IJCST 19,6

58

Keywords: Fabric handling, Artificial intelligence, Intelligent control, Robotic handling for sewing, Robotic sewing, Fabric properties, Quality control, Seam quality, Color matching The main concept and purpose of this project is the development of a robotic workcell for sewing fabrics. The research work deals with the handling and the quality control of fabrics and cloths in garment assembly. The project is focused in the fabric handling tasks: ply separation, translation, placement, folding, feeding and orientation for sewing. Sensor fusion, fuzzy logic, neural networks and machine learning techniques are used for controlling the robotic grippers and the applied forces for intelligent fabric handling. The quality control system is based on machine vision for detecting the fabric’s defects and color matching as well as for inspecting the seams’ quality. Each of the experimental stage sub-system devices are individually tested and at the end would be integrated to form a workcell. In addition, the project provides a new researchers training program, concerning the technology of automated handling of non-rigid materials and the technology of cloth making. The results will be demonstrated and disseminated to national cloth making industries – SME’s, and published to international journals. This is a joint project, where three Departments of the University of Patras and the ELKEDE – Technology & Design Centre are involved. Mechanical, Electrical and Computer Engineers are cooperated to integrate the automated handling and quality control system.

Project aims and objectives The main aim of this project is to develop new intelligent methods for the robotic handling of non-rigid materials, seam quality control and fabric color matching. The innovative approaches are based on sensor fusion and artificial intelligent techniques: Fuzzy Logic, Neural Networks and Genetic Algorithms. The objectives: . Training new researchers in the technology of automated handling of non-rigid materials and the technology of the cloth making industry. . Development of intelligent algorithms for fabric handling strategies, color matching and seam quality. .

.

.

Development of experimental intelligent mechatronic devices, for separatinghandling-translating-inspecting fabrics, in order to test the intelligent algorithms in laboratory conditions. Development and test an integrated experimental system for the robotic sewing and the fabric quality control. To demonstrate the results (acquired knowledge and experimental devices) in garment assembly.

Research deliverables (academic and industrial) Five PhD dissertations A laboratory prototype system of the robotic garment assembly Two Technical reports

Publications Journals: Panagiotis N. Koustoumpardis, Nikos A. Aspragathos, (2003), “Fuzzy Logic Decision Mechanism Combined with a Neuro-Controller for Fabric Tension in Robotized Sewing Process”, Journal of Intelligent and Robotics Systems, Vol. 36, No. 1, pp. 65-88. Zacharia, P.Th., Aspragathos, N.A. (2005), “Optimal task scheduling based on Genetic Algorithms”, Robotics and Computer-Integrated Manufacturing, Vol. 21, No. 1, pp. 67-79. Panagiotis, N., Koustoumpardis, J.S., Fourkiotis, Nikos, A., Aspragathos, “Robotized Measurement and Intelligent Evaluation of Fabrics’ Extensibility – Tensile Test”, submitted for publication to the International Journal of Clothing Science and Technology. Karybali, I.G., Berberidis, K., Psarakis, E.Z., Evangelidis, G.D., “An Efficient Spatial Domain Technique for Subpixel Image Registration, submitted for publication to the IEEE Transactions on Image Processing. Zoumponos, G.T., Aspragathos, N.A., “Fuzzy Logic Path Planning for the Robotic Placement of Fabrics on a Work Table” submitted for publication to the Robotics and Computer-Integrated Manufacturing Journal. Furthermore, five papers are under preparation and will be submitted to international journals soon. Conferences: Zacharia, P.Th., Aspragathos, N.A. (2004), “Optimization of industrial Manipulators Cycle Time Based on Genetic Algorithms”, pp. 517-522, 2nd IEEE International Conference on Industrial Informatics INDIN’04, 24-26 June, Berlin, Germany. Koustoumpardis, P.N., Aspragathos, N.A., (2004), “A Review of Gripping Devices for Fabric Handling”, International Conference on Intelligent Manipulation and Grasping IMG04, Genova, Italy, ISBN 88 900 426-1-3, pp. 229-234, July. Zoumponos, G.T., Aspragathos, N.A. (2004) “Design of a robotic air-jet gripper for destacking fabrics”, IMG ‘04, pp. 241-246, Genova, Italy, July. Ioannis C., Evangelos, D., (2004), “Semi blind gamma correction”, 1st International Conference “From Scientific Computing to Computational Engineering”, Athens, 8-10 September. Mariolis, I., Dermatas, E., (2004), “Robust Detection of Seam Lines using the Radon Transform”, 1st International Conference “From Scientific Computing to Computational Engineering”, Athens, 8-10 September. Zoumponos, G.T., Aspragathos, N.A. (2005), “A fuzzy robot controller for the placement of fabrics on a work table”, IFAC World Congress, Prague, 4-8 July. Zacharia, P.Th., Mariolis, I.G., Aspragathos, N.A., Dermatas, E.S., (2005), “Visual servoing of a robotic manipulator based on fuzzy logic control for handling fabric lying on a table”, 1st I*PROMS Virtual International Conference on Intelligent Production Machines and Systems, IPROMS 2005, 4-15 July, pp.411-416. Psarakis, E.Z., Evangelidis, G.D., (2005), “An enhanced Correlation-Based Method for Stereo Correspondence with Subpixel Accuracy’’ 10th IEEE Int. Conf. On Computer Vision, October, Beijing, China Karybali, I.G., Psarakis, E.Z., Berberidis, K., Evangelidis, G.D., (2006), “Spatial Domain for Image Registration with Subpixel Accuracy, submitted to European Signal Processing Conf. (EUSIPCO), (accepted). Ioannis, S.C., Dermatas, E.S., (2006), “Non-parametric Estimation of camera response function”, 13th IEEE MELECON, 16-19 May 2006 (accepted in journal preselection). Koustoumpardis, P.N., Zoumponos, G.T., Zacharia, P.Th., Mariolis, I.G., Xatzis, I., Evagelidis, G., Zabetakis, A., (2006), “Handling of Non-Rigid Materials with Robots (XROMA), Application in Robotic Sewing”, 37th International Symposium on Novelties in Textiles, Ljubljana, Slovenia, 15-17 June. Mariolis, I., Dermatas, E., (2006) “Automated quality control of textile seams based on puckering evaluation”, 37th International Symposium on Novelties in Textiles, Ljubljana, Slovenia, 15-17 June.

Research register

59

IJCST 19,6

60

Zacharia, P.Th., Mariolis, I.G., Aspragathos, N.A., Dermatas, E.S., (2006), “Visual servoing controller for robot handling fabrics of curved edges”, I*PROMS NoE Virtual International Conference on Intelligent Production Machines and Systems, 3-14 July. Ioannis, C., Gavrilis, D., Dermatas, E. (2006), “Spectral characterization of digital cameras using genetic algorithms”, IPROMS 2006 3-14 July. Ioannis, S.C., Kappatos, V.A., Dermatas, E.S., (2006), “Filter selection for multi-spectral image acquisition using the feature vector analysis method”, IPROMS 2006 3-14 July. Dimitris, G., Ioannis, S.C., Evangelos, D., (2006), “Detection of Web Denial-of-Service Attacks using decoy hyperlinks”, Communication Systems, Networks And Digital Signal Processing (CSNDSP’06), 19-21 July 2006, Patras Univ. Conference Centre, Patras, Greece. Zacharia, P.Th., Mariolis, I.G., Aspragathos, N.A., Dermatas, E.S., (2006), “Polygonal approximation of fabrics with curved edges based on Genetic Algorithms for robot handling towards sewing”, to be presented in CIRP, 25-28 July, Ischia, Italy. Zacharia, P., Zoumponos, G., Koustoumpardis, P., Zampetakis, A., Aspragathos, N., (2006), “Robot handling of non-rigid materials for the sewing process”, to be presented in ITCDC 2006, 8-11 October 2006, Dubrovnik, Croatia.

Pisa, Italy University of Pisa, Via Diotisalvi, 2, 56126 Pisa, Italy, Tel: +39 0502217053; Fax: +39 0502217051; E-mail: [email protected] Principal Investigator(s): Prof. Danilo De Rossi Research Staff: Post doc researchers, PhD students, post graduate students.

FLEXIFUNBAR Multifunctional barrier for flexible structure (textile, leather and paper) Other Partners: Academic Brunel University, Pisa University, Centro Di Cultura per l’ingegneria delle Materie Pastiche, Ghent University, Queen’s University of Belfast, DWI – Aachen, Institut Pasteur de Lille, Institute of Natural Fibres – Poznan

Industrial Alan, Amkey Management, Annebergs, Arjo Wiggins, Basilius, Calsta, CEI, Centexbel, Centro Tecnologico do Calcado, Clotefi, Clubtex, CREPIM, Curtumes Aveneda, Devan, DG Tec, Duflot, ECCO, Gleittechnik, Europlasma, IFTH, IMP Comfort, INCA, IQAP, Lauffenmu¨ hle, Linificio, Nabaltec, Nylstar, Patraiki, Procotex, Siamidis, Sinterama, Subrenat, Sveriges Provnings, Telice, Thrakika Ekkokistiria, Traitex, VTT, Wellman International.

Project start date: 1 October 2004 Project end date: 30 November 2008 Project budget: Total amount: e6,438,995 (University of Pisa: e343,000) Source of support: European Commission Keywords: Multifunctional, Barrier, Textile, Fiber, Flexible, Leather, Paper

All citizens are permanently protected by flexible structures with barrier noise and thermal insulation, shield against electrostatic or electromagnetic phenomena, filtration of dust or insects.. The application of flexible structure is very large thanks to their easy adapting properties and shape. Nevertheless, they will maximise the level of safety in building, transportation and to ensure the well-being of European citizens, The flexible structures, generally based on paper, leather or textile are usually treated to serve only one barrier effect.Ires a whole re-design of flexible structure functions that is the main purpose of FLEXIFUNBAR. For instance to prevent from all external aggressions in hazardous atmosphere, flexible structure must provide at least barrier effects. The ultimate goal of flexifunbar initiative is to develop innovative generation of hybrid multi barrier-effects materials, based on multi layer complex structures and funcionnalisation of micro and nanostructures.The development of such materials covers a large range of applications: . Transportation: filter, thermal and acoustic insulation panel, pollutant detectors. .

Home and building: wall covering, home furniture, carbon monoxide detectors, antibacterial mattress, electromagnetic insulation panels.

.

Health: protection of people against insects, protective clothing for military and defense, hygiene mask, operation area.

Project aims and objectives The innovation of Flexifunbar lies in the principle of associating in one same material several functionalities: Heat insulation, acoustic insulation, shielding against electromagnetic waves, anti odours, anti bacterial, flame retardancy. . . The objective of Flexifunbar is to develop and promote multi-functional flexible structure for use in many multisectorial industrial applications in the health field as well as in the building construction and transportation industrie..

Research deliverables (academic and industrial) New fiber, textile, leather and paper samples with multifunctional barrier properties Publications Not available

Pisa, Italy “E. Piaggio” Centre – University of Pisa, Via Diotisalvi, 2, 56126 Pisa, Italy, Tel: +39 0502217053; Fax: +39 0502217051; E-mail: [email protected] Principal Investigator(s): Prof. Danilo De Rossi

Research register

61

IJCST 19,6

62

Research Staff: Prof. Roger Fuoco; Prof. Arti Ahluwalia; Dr Fabio DI Francesco; Dr Thoas Schafer

BIOTEX: Bio-sensing textiles to support health management Other Partners: Academic Dublin City University(Ireland), University of Pisa (Italy)

Industrial CSEM Centre Suisse d’Electronique et de Microtechnique SA (Switzerland), CEA Commissariat a` l’Energie Atomique (France), Smartex s.r.l. (Italy), Thuasne (France), Penelope SpA (Italy), Sofileta (France)

Project start date: 01 September 2005 Project end date: 29 September 2008 Project budget: Total amount: e3.108.029 (eligible cost) – Requested EC Contribution: e1.900.000 University of Pisa quote: 255.750,00 Source of support: European Commission Integration of health monitoring tools into textiles brings the benefits of safety and comfort to the users. Instrumented clothes will provide remote monitoring of vitals signs, diagnostics to improve early illness detection and metabolic disorder and benefits to the reduction on medical social costs to the citizen. Ambulatory healthcare, isolated people, convalescent people and patients with chronic diseases are addressed. To date, developments in that field are mainly focused on physiological measurements (body temperature, electro-cardiogram, electromyogram, breath rhythm, etc.) with first applications targeting sport monitoring and prevention of cardiovascular risk. Biochemical measurements on body fluids will be needed to tackle very important health and safety issues.

Project aims and objectives The BIOTEX project aims at developing dedicated biochemical-sensing techniques compatible with integration into textile. This goal represents a complete breakthrough, which allows for the first time the monitoring of body fluids via sensors distributed on a textile substrate and performing biochemical measurements. BIOTEX is addressing the sensing part and its electrical or optical connection to a signal processor. The approach aims at developing sensing patches, adapted to different targeted body fluids and biological species to be monitored, where the textile itself is the sensor. The extension to whole garment and the integration with physiological monitors is part of the roadmap of the consortium. Textiles for applications in health monitoring are becoming a major theme for the citizen’s healthcare and safety since they allow:

.

simultaneously comfort and monitoring (for safety and/or health);

.

non-invasive measurements, no laboratory sampling; continuous monitoring during daily activity of the person;

. .

Research register

possible multi-parameter analysis and monitoring; and

distributed sensing thanks to access to 90 per cent of the body surface if integrated on clothing. Societal issues such as ambulatory healthcare, isolated elderly or disabled patient’s care may be tackled by these techniques. Moreover, non-invasive and continuous monitoring of people in a critical state is more and more needed e.g. in emergency services, for heavily burnt patients and safety, e.g. exposed personnel like fire-fighters. At the present stage, health-monitoring systems using electronic textiles are mainly targeting applications based upon physiological parameter measurements, such as body movements or electro-cardiogram. To open a dramatically wider field of applications, biochemical measurements on body fluids (blood, sweat, urine) will be needed. At the present time, biochemical analysis systems compatible with integration into clothing are unfortunately lacking. This is a major drawback for instance in the case of sweat analysis which is potentially very rich in health related information. However, such analysis is hardly performed today because of the difficulty to sample sweat in sufficient quantity. Only a real textile sensor embedded in a garment through textile techniques will allow direct collection of sweat and a large body surface; moreover lower fabrication costs are expected. For blood analysis, the main interest will be to avoid invasive sampling and to allow continuous analysis. BIOTEX aims at the development of technologies to fulfil these needs. .

Research deliverables (academic and industrial) New textile integrated systems for biochemical parameters detection. Publications Not available

Pisa, Italy University of Pisa, Via Diotisalvi, 2, 56126 Pisa, Italy, Tel: +39 0502217053; Fax: +39 0502217051; E-mail: [email protected] Principal Investigator(s): Prof. Danilo De Rossi Research Staff: Prof. Luigi Landini, Ing. Enzo Pasquale Scilingo, Dr Federico Lorussi

MY HEART Cardiovascular Disease by preventive lifestyle & early diagnosis

63

IJCST 19,6

64

Other Partners: Academic

Industrial

University of Pisa, Universidad Philips GmbH Forschungslaboratorien Politecnica de Madrid, Consorzio di (Philips Research Labs. – Aachen), Bioingegneria e Informatica Philips Electronics UK Limited (Philips Medica (CBIM), Politecnico di Milano, Research Labs. – Redhill and Philips University of Padova, University of Firenze,Design – London), Philips International University Clinic Aachen, Hospital Clinico B.V. (Design – Eindhoven), Philips San Carlos de Madrid Insalud, Fondazione Innovative Technology Solutions NV Centro San Raffaele del Monte Tabor, (Philips DSL), Medtronic Iberia SA, Nokia University Pavia IRCCS/FSM, Corporation, Fundacion Vodafone, Nylstar Universidad Politecnica de Valencia CD S.p.A., Manifatture Filati Riunite SPA (Technical University of (Lineapiu Group), Milior – S.P.A., Smartex Valencia – Sports Center) Facludade de – S.R.L, Dr Hein GMBH, Mind Media B.V., ciencias e technologia da universidade de Medgate AG, CSEM Centre Suisse Coimbra, Hospital de Unisersidate D’Electronique et de Microtechnique SA, de Coimbra Commissariat a l’energie Atomique (CEALETI), Eidgenossische Technische Hochschule Zurich (ETH Zurich), Instituto de Aplicaciones de las Tecnologı´as de la Informacio´n y de las Comunicaciones Avanzadas – Asociacio´n (ITACA), Mayo Clinic Rochester, Philips Electronics Nederland b.v. (Philips Research Labs -Eindhoven), Lineapiu S.P.A. Project start date: December 2003 Project end date: September 2007 Project budget: University of Pisa: e976,168 Source of support: European Commission Cardio-vascular diseases (CVD) are the leading cause of death in the west. In Europe over 20 per cent of all citizens suffer from a chronic CVD and 45 per cent of all deaths are due to CVD. Europe spends annually hundreds billion Euros on CVD. With the upcoming aging population, it is a challenge for Europe to deliver its citizens healthcare at affordable costs. It is commonly accepted, that a healthy and preventive lifestyle as well as early diagnosis could systematically fight the origin of CVD and save millions of live-years. The MyHeart mission is to empower citizen to fight cardio-vascular diseases by preventive lifestyle and early diagnosis. The starting point is to gain knowledge on a citizen’s actual health status. To gain this info continuous monitoring of vital signs is mandatory. The approach is therefore to integrate system solutions into functional clothes with integrated textile sensors. The combination of functional clothes and integrated electronics and process them on-body, we define as intelligent biomedical clothes. The processing consists of making diagnoses, detecting trends and react on it. Together with feedback devices, able to

interact with the user as well as with professional services, the MyHeart system is formed. This system is suitable for supporting citizens to fight major CVD risk factors and help to avoid heart attack, other acute events by personalized guidelines and giving feedback. It provides the necessary motivation the new life styles. MyHeart will demonstrate technical solutions. The outcome will open up a l new mass market for the European industry and it will help prevent the development of CVD, meanwhile reduce the overall EU healthcare costs. The consortium consists of 33 partners from 11 countries. It is a research effort of industrial, research institutes, academics and medical hospitals, covering the whole value chain from textile research, via fashion and electronic design, towards medical and home-based applications.

Project aims and objectives The MyHeart approach for solving the key challenges is based on the development of intelligent biomedical clothes for preventive care application tailored to specific user groups. In order to focus on the user motivation and the individual benefit, we define the main objectives along 5 different application areas. These application areas reflect the main risks for developing a CVD and address the user need for early diagnose to limit the severity of an acute event.

Research deliverables (academic and industrial) Electronic systems embedded into functional clothes, functional clothes with integrated textile and non-textile sensors.Combination of functional clothes and integrated electronics will be intelligent biomedical clothes (IBC). Publications Not available

Pisa, Italy University of Pisa, Via Diotisalvi, 2, 56126 Pisa, Italy, Tel: +39 0502217053; Fax: +39 0502217051; E-mail: [email protected] Principal Investigator(s): Prof. Danilo De Rossi Research Staff: Prof. Bruno Neri; Ing. Alessandro Tognetti; Ing. Enzo Pasquale Scilingo; Ing. Federico Carpi; Ing. Antonio Lanata`

PROETEX: Protection e-Textiles: MicroNanoStructured fibre systems for Emergency-Disaster Wear

Research register

65

IJCST 19,6

66

Other Partners: Academic Consiglio Nazionale delle Ricerche – INFM, Technical University of Lodz, Ghent University – Department of Textiles, University of PISA, Dublin City University, Institut National des Sciences Applique´es de Lyon

Industrial Smartex srl, Milior, Sofileta SAS, Thuasne,Commissariat a` l’Energie Atomique -”CEA”,CSEM Centre Suisse d’Electronique et de Microtechnique SA, Sensor Technology and Devices Ltd, Steiger S.A.,Philips GmbH, Zweigniederlassung Forschungslaboratorien, Ciba Spezialita¨tenchemie AG, Diadora Invicta SpA, iXscient Ltd, Zarlink Semiconductor Limited, Brunet-Lion SAS, Brigade de Sappeurs Pompiers de Paris, European Centre for Research and Training in Earthquake Engineering, Direction de la De´fense et de la Se´curite´ Civiles

Project start date: February 2006 Project end date: January 2010 Project budget: University of Pisa: e780,443; Total: e 12,792,242 (Requested: e 8.100.000) Source of support: European Commission

ProeTEX will develop integrated smart wearables for emergency disaster intervention personnel, improving their safety, coordination and efficiency and for injured civilians, optimising their survival management. This core application area, which is of significant societal importance in itself, will drive a wide range of key technology developments, building on current and past EU and national projects and the commercial activities of partners, to create micro-nano-engineering smart textile systems – integrated systems (fabrics, wearable garments) using specifically fibrebased micronano technologies. These are capable of being combined into diverse products addressing this core application area but also a wide range of other markets from extreme sports, through healthcare to transportation maintenance and building workers. The industrial partners can address these markets. Fiber systems can integrate sensors, actuators, conductors, power management, and the emergency disaster personnel smart garment will, within a wireless ambient planning and managing environment, progressively enhance and integrate fiber systems for: .

continuous monitoring of life signs (biopotentials, breathing movement, cardiac sounds);

.

continuous monitoring biosensors (sweat, dehydration, electrolytes, stress indicators);

.

pose and activity monitoring;

.

low power local wireless communications, including integrated fiber antennae;

.

active visibility enhancement, light emitting fibers;

.

internal temperature monitoring using fiber sensors;

.

external chemical detection, including toxic gases and vapours; and power generation – photovoltaic and thermoelectric and power storage.

.

The technological base developed will concentrate on smart fibers/e-textiles, but the IP will combine these where appropriate with ‘conventional’ microsystems (such as accelerometers, gyros, microcontrollers and wireless chips).

Project aims and objectives The central IP goal is to develop an integrated set of functional garments for emergency disaster personnel, such as firefighters and paramedics, plus systems for injured civilians. These will be produced using both enhanced and novel fibre based micronanosystems, whose development will extend the state of the art in this area. The project will roll out a sequence of progressively more capable integrated wearable systems for emergency disaster intervention personnel and injured civilians. Thus overall the IP will: .

Progress the fundamentals of fibre-based sensor, processing, communications and power management systems.

.

Integrate these fibre-based capabilities into functional knitted or woven wearable garments.

.

Produce fully capable integrated communicating, sensor wearables, using additional ‘conventional’ systems where necessary.

.

Test their match of user needs and requirements in a lab-based setting. Demonstrate their function in a real-world application in a number of field trials.

.

Scientific objectives .

Develop a multifunctional garment integrating an increasingly ambitious set of sensors and energy harvesting and storage which is reliable, robust, easy to wear and capable of manufacture.

.

Into this garment: Design, test and integrate a bioelectrical heart rate monitor into whole skin contact garment interface; Design, test and Integrate a cardiac sound monitor; Integrate sensor breathing monitor and ensure that signal conditioning and processing results in successful way. Develop fibre and new textile based technological solutions, with reliable functionality, capable of integration into wearable garments covering the following set of technological area capabilities: . Monitor bioelectrical potential.

.

.

Sensing breath movements.

.

Sensing posture and movement. Biochemical sensing, specifically determination of dehydration status.

. . .

Sensing core temperature. Acting as local communications antennae.

Research register

67

IJCST 19,6

68

.

Sensing external toxic gases/chemicals, including CO.

.

Generating local energy using thermoelectric generation.

.

Generating local energy using photovoltaic processes and

.

Storing energy using Li-Ion textile batteries.

Technical objectives .

Develop and adapt textile manufacturing processes to these new active fibres and layers (weaving, knitting, coating, laminating) but also innovate in terms of clothes conception to optimise the assembly step regarding interconnection needs for e-textile garment.

.

Develop and test a multifunctional (inner and outer) garment integrating an increasingly ambitious set of sensors and energy harvesting and storage which is reliable, robust, easy to wear and capable of manufacture for both intervention people and injured civilians.

.

The inner and out garment will include an adapted set of functionalities based on the developed technologies. As example first inner garment could integrate bioelectrical heart rate monitor, cardiac sound monitor, strain sensor breathing monitor inner temperature measurement and ensure that signal conditioning and processing results in successful and robust physiological monitoring. Energy generated by the heat (thermoelectricity) and the movement (piezoelectricity) of the of the wearer. Outer garment will typically include toxic gas measurement, external temperature; motion and position monitoring, data transmission system, energy could be provided by photovoltaic external layer and textile Li-Ion batteries.

.

Realize field trial of the instrumented garment for technological validation.

Research deliverables (academic and industrial) The key deliverables will be: .

An inner garment for emergency disaster personnel, monitoring the health of the user through vital signs, biochemical parameters, activity and posture, generating and storing own power and communicating locally with other wearables and relaying through (D).

.

An outer garment for emergency disaster personnel, measuring potential environmental insults (temperature, CO, other toxic gases), sensing posture and movement of the wearer and offering improved visibility, generating and storing its own power communicating locally with other wearables and relaying through (D).

.

An under-garment jerkin or chest band for injured civilians (closely related to (A)) monitoring their health, generating and storing its own power and communicating locally, relaying information via (E).

Victims monitoring measures will include: body temperature; cardiac pulse; respiration rate; ECG; percutaneous CO saturation; percutaneous O2.

.

.

A portable unit for the emergency disaster personnel, communicating with A, B & C, but offering additional ‘conventional’ microsystem, providing both local and long range communication (acting as a relay for A, B and C), including some specific sensors not easily integrated into (B), plus accelerometers, gyros and GPS to enable high accuracy position and movement determination. This device should allow data entry and displays/alarms. A portable unit for injured civilians, to include data relay capability and INS/GPS but no data entry or display. Some kind of integrated alarm or indicator to give the overall civilians health status. A simple user input, such as panic button, may be required.

Publications Not available

Shrewsbury, UK Rapra Technology Ltd, Shawbury, Shrewsbury, Shropshire SY4 4NR, Tel: +44 1939250383; Fax: +44 1939 251118; E-mail: [email protected] Principal Investigator(S): Chris Hare Research Staff: R Venables, S Wallace

New Classes of Composite Engineering Materials from Renewable Sources Other Partners: Academic Upper Austria Research, Lulea University of Technology, Wroclaw University of Technology

Industrial Fraunhofer ICT, Risoe, Gaiker, Celabor, VTT, APC Composites, BAFA, Transfurans chemicals, PJH, Chemont,Tecnaro, MERL, Net composites, Tehnos, Griffner, MEDOP, Haidlmair, Fiedler, Ekotex, National Institute of wood

Project start date: 1 April 05 Project end date: 1 October 08 Project budget: e6.5 million Source of support: EU, Integrated Project, FP6 Future product design requires sustainable processes and eco-innovation in material development for engineering applications. The innovative approaches use new engineering materials - biocomposites and their development has to be knowledgebased, whereas predominant issues are resource saving, variability in properties and functionality, light weight, low costs and eco-efficiency in all stages of the product life cycle.

Research register

69

IJCST 19,6

70

The main objective of this project is to obtain a breakthrough for SMEs on the development and use of engineering thermoplastic and thermosetting materials mainly from natural resources, like lignin from the paper industry and from the High Presure Hydrothermolyses (HPH) process, other biopolymers (here referred as biopolymers: e.g. Polylactide, Polyhydroxy-butyrate, Starch), furan resins, woven and non-woven cellulose fibres and fibre mats to final model products. The technical work programme will comprise the complete technical path from the input of natural raw materials (fibres, polymers and natural additives) to the output of final top quality engineering composite materials and model products (e.g. housings for electronic equipment, car front end interiors, glass frames etc.) with an environmentally friendly life cycle. In parallel, there are activities concerning standardisation of characterisation and test procedures and quality control. Demonstration by the model products supporting the dissemination and the exploitation of results will exhibit the benefits of the materials and deliver a first input to material databases. An integrated concept of sustained skill and education of staff and students will provide routines and access to the material data. It includes the most interesting approaches of all current developments for engineering biocomposites. Innovative additives will provide flame retardancy and colouring.

Project aims and objectives To produce a range of ‘hi tech’ composite panels using a variety of natural fibres and natural resins.

Research deliverables Many deliverables in the project, some of which are: Reports on: Raw material characterization, data and tolerances; Compounding of materials; Test data; Safety and emissions; Environmental Benefits; Economic Evaluation; Sample tools and parts; Demonstration Parts; Training, exploitation and dissemination. Publications Dr Tama´s Pe´ter - Dr Hala´sz Marianna (2003), “3D body modelling in clothing design”, IMCEP 2003, 4th International Conference, 9-11. October 2003, Maribor, Slovenia, ISBN 86-435-0575-7, pp. 64-68. L. Kokas Palicska; J. Gersak; M. Hala´sz (2005), “The impact of fabric structure and finishing on the drape behavior of textiles”, AUTEX 2005, 5th World Textile Conference, Portorozˇ, Slovenia, 27-29 June 2005, ISBN 86-435-0709-1, pp. 891-897. P. Tama´s; M. Hala´sz; J. Gra¨ff (2005), “3D dress design”, AUTEX 2005, 5th World Textile Conference, Portorozˇ, Slovenia, 27-29 June 2005, ISBN 86-435-0709-1, pp. 436-441. J. Kuzmina; P. Tama´s; M. Hala´sz, Gy. Gro´f (2005), “Image-based cloth capture and cloth simulation used for estimation cloth draping parameters”, AUTEX 2005, 5th World Textile Conference, Portorozˇ, Slovenia, 27-29 June 2005, ISBN 86-435-0709-1, pp. 904-909. L. Szabo´; M. Hala´sz (2005), “Automatic determination of body surface data”, AUTEX 2005, 5th World Textile Conference, Portorozˇ, Slovenia, 27-29 June 2005, ISBN 86-435-0709-1, pp. 715-720. L. Kokas Palicska; M. Hala´sz (2005), “Analysing of draping properties of textiles”, IN-TECH-ED’05, 5th International Conference, 8-9. September 2005, Budapest, ISBN 963 9397 06 7, pp. 133-138. O. Nagy Szabo´; P. Tama´s; M. Hala´sz (2005), “Garment construction with a 3 dimension designing system”, IN-TECH-ED’05, 5th International Conference, 8-9. September 2005, Budapest, ISBN 963 9397 06 7, pp. 348-257.

J. Kuzmina; P. Tama´s; M. Hala´sz; Gy. Gro´f (2005), “Image-Based Cloth Capture and Cloth Simulation Used for Estimation Cloth Draping Parameters”, IN-TECH-ED’05, 5th International Conference, 8-9. September 2005, Budapest, ISBN 963 9397 06 7, pp. 358-365.

Research register

Zagreb, Croatia Faculty of Textile Technology, University of Zagreb, Prilaz baruna Filipovic´a 30, HR-10 000 Zagreb, Croatia, Tel: +385 1 4877 360; Fax: +385 1 4877 355; E-mail: [email protected] Principal Investigator(s): Prof. Ana Marija Grancaric´, Ph.D. Research Staff: Assoc. Prof. Tanja Pusˇic´, Ph.D.; Assist. Prof. Zˇeljko Penava, Ph.D.; Anita Tarbuk, M. Sc., Lea Markovic´, B. Sc., Assist. Prof. Jasenka Bisˇc´an, Ph.D.; Sonja Besˇenski, M. Sc., Ivancˇica Kovacˇek, Ph.D., D. Med., Prof. Djamal Akbarov, Ph.D., Prof. Emil Chibowski, Ph.D., Prof. Rybicki Edward, Ph.D., Prof. Eckhard Schollmeyer, Ph.D., Prof. M.M.C.G. Warmoeskerken, Ph.D.

Interface Phenomena of Active Multifunctional Textile Materials Other Partners: Academic partners:

Industrial partners

Croatian National Institute of Pamucˇna industrija Duga Resa, Duga Resa Public Health, Zagreb; Tashkent Institute of Textile and Light Industry, Uzbekistan; Maria Curie-Skłodowska University, Lublin, Poland; Technical University of Lodz, Poland; Deutsches Textilforschungsinstitut Nord-West eV; Institut der Universitat Duisburg Essen; University of Twente, Netherlands Project start date: 1 January 2007 Project end date: 31 December 2011 Project budget: None Source of support: Ministry of Science, Education and Sports, Republic of Croatia Keywords: Textile material, Interface phenomena, Surface modification and finishing, Multifunctionality The goal of the project is synergistic effects of some compounds on modified textile surfaces for achieving multifunctionality of textiles. Interface phenomena of textile surfaces with special accent on surface free energy, zeta potential, electroconductivity, adsorption and desorption of surfactants and other compounds usually used in textile finishing will give a great contribution to multifunctionality of textile. The mechanism of adsorption and desorption of surfactants and other finishing agents on modified textile surfaces is expected to be clarified in the present project.

71

IJCST 19,6

72

Different surface modifications, pretreatment and finishing of textile, especially cotton and polyester, will be performed according to European Technology Platform for the future of textile and clothing. For such purpose advance processes like mercerization, cationization, alkali, EDTA, other compounds and enzymes for surface hydrolysis of PET fabric, optical bleaching, implementation of nano antimicrobial active silver ions and mineral delivery mechanism, zeolite and others will be performed. Aminofunctional and other compounds will be added to azalides for the synergistic high antimicrobial effects. In cotton pretreatment enzymatic scouring will be applied using enzymes pectinase and the newest cutinase, for removal of pectins and bioplymers from cotton impurities with lipophylic character, instead of ecologically unfavorable alkali scouring. The goal of the project is synergistic effects of some compounds on modified textile surface. Interface phenomena of the new textile materials produced from electroconductive, low electro resistance fibers will be investigated for the purpose of static electricity and electromagnetic protection and for its implementation as sensors or other electronic devices in intelligent textiles. Traditional protection and aesthetic role of textile will be spread in active textile multifunctionality. Project will deal with elektrokinetic phenomena (zeta potential, isoelectric point, IEP, point of zero charge, PZC, surface electrical charge, surface free energy), hydrophility and hydrophobilicity, whiteness, fluorescence and phosphorescence, friction, fabric cover factor, elasticity, air and water vapor permeability of textile materials and their protection on UV radiation, microbes and fungi, coldness, heat and flame, static electricity and electromagnetic field.

Project aims and objectives Project will continue researching on assignments from previous project (0117012). Purpose of these investigations is based on lightening of interface phenomena on textile which effect directly to its adsorption and interaction intensity between textile fibers and chemical compounds. Almost all possibilities in modification during manufacturing high performance synthetic fibers are used, therefore nowadays attention and research is on textile surface modification. Procedures and compounds for that modification varies, as their effect varies, but the purpose and aim are directed to synergism of two or more components for accomplishing hydrophob or hydrophil textile, textile highly resistant to atmospheric condition, bacteria, microbe and fungi, UV radiation and open flame. Furthermore important aim of the project is cotton high level of purity by unconventional agents and material pretreatment procedures for mercerization and cationization. Pectinase in previous project investigation showed good elimination of pectine from primary cotton layer, but hydrophility was not so high like alkali scoured cotton. Chemical composition of cotton cuticula has lypophilic polymers, biopolyesters, which can be degraded by cutinase, new enzymes for degradation of waxes for better hydrophility. Cotton cationization during mercerization is the most important innovation of previous project and the patent for it was asked. Electronegative cotton surface charge, of which anionic substances adsorption depends, is lower after cationization in harsh mercerization conditions. The aim of this project is antibacterial, UV and flame protection by nanoparticle implementation (Ag) using mineral delivery compound (zeolite and others) as well. Electroconductive fibers implementation in yarns

of textile materials should result in static electricity removal, and hopefully other effects. The aim of polyester surface modification, optical bleaching, other compounds treatments is well-known aesthetic, as well as high UV protection, high material elasticity as a result of changes in fiber microstructure. Interface phenomena research on wide range possible fabric knitted and woven construction will givethe solution of problems of fabric construction influence to high effect in this project.

Research register

73 Research deliverables Interface phenomena of textile materials surface in wet medium results in textile electric surface charge cognition and surface free energy as well on which adsorption depends. Important application of this project results is in ecological enzymatic scouring with pectinases.Enzymatic scouring with new enzymes, cutinase, will remove biopolyester cuticula and improve cotton hydrophility, and therefore replace harsh conventional alkali scouring entirely. Important application will have, patent requested cotton cationization during mercerization. By this pretreatment electropositive cotton is achieved, with great anion adsorption on its surface in all textile finishing processes. These anions enclose all low and high molecular compounds for textile finishing and all pricondensates. Implementation of nanoparticles (Ag and others) is predicted during mercerization and cationization processes, therefore it is important to emphasize rational component of these procedures which gives cotton multifunctionality in all textile usage. The next important application is antibacterial textile accomplished with azalide treatment especially in synergism with aminofunctional and other compounds and systems. It is well-known that fluorescence of optically bleached increases whitening of textiles. Optical brighteners and other compounds researching will be of great importance in UV protection with textile material. Heavy metals are toxic and their research is of great importance in human health protection. Furthermore, in nowadays growing demands on life safety from external influences especially UV radiation, research of differently structured textile material interface phenomena will find application in textile for summer clothing. Publications Grancaric´, Anamarija; Pusˇic´, Tanja; Tarbuk, Anita (2006), Enzymatic Scouring for Better Textile Properties of Knitted Cotton Fabrics // Biotechnology in Textile Processing / Guebitz, Georg; CavacoPaulo, Artur; Kozlowski, Rysard (ED.). New York: The Haworth Press, Inc., 2006. Grancaric´, Ana Marija; Tarbuk, Anita; Dumitrescu, Iuliana; Bisˇc´an, Jasenka (2006), “UV Protection of Pretreated Cotton - Influence of FWA’s Fluorescence”. // AATCC Review 6 (2006) 4, 2–6. Anita Tarbuk, Ana Marija Grancaric´ & Volker Ribitsch (2007), Electrokinetic Phenomena of Textile Fibers//Book of abstracts XX.Croatian meeting of Chemists and Chemical engineers 2007,301. Ana Marija Grancaric´, Lea Markovic´, Anita Tarbuk, Eckhard Schollmeyer (2007), Properties of Multifunctional Cotton in Accordance with International Standards//Conference of Textile days, Zagreb 2007. Ana Marija Grancaric´, Anita Tarbuk, Ivancˇica Kovacˇek (2007), “Micro and nanoparticles of Zeolite for the protective Textiles”, Book of Proceedings of 7th Annual AUTEX Conference, AUTEX 2007, Tampere, Finland, p. 1123. Anita Tarbuk, Ana Marija Grancaric´, Mirela Leskovac (2007), “Surface Free Energy of Pretreated and Modified Cotton Woven Fabric”, Book of Proceedings of 7th Annual AUTEX Conference, AUTEX 2007, Tampere, Finland, p. 1104.

IJCST 19,6

74

Zagreb, Croatia Faculty of Textile Technology, Prilaz Baruna Filipovica 30, HR-10 000 Zagreb, Croatia, Tel: +385 1 4877 360; Fax: +385 1 4877 355; E-mail: [email protected] Principal Investigator(s): Prof. Ana Marija Grancaric, Ph.D., Prof. Ilse Steffan, Ph.D. Research Staff: Iva Rezic, M.Sc., Michaela Zeiner, Ph.D.

Impact of Heavy Metals from Textiles on Human Health Other Partners: Academic

Industrial

Department of Analytical and Food Chemistry, University of Vienna, Austria Project start date: 1 January 2006 Project end date: 31 December 2007 ¨ Source of support: Osterreichischer Austauschdienst (O¨AD, Austria) and Ministry of Science, Education and Sports, Republic of Croatia Keywords: Textiles, Heavy metals, Human health Many textiles may contain various chemicals that represent a health hazard to consumers and an environment when discharged into wastewater from households, laundries or textile industries. The risk involved due to chemical pollutants in respect to health hazard depends on the toxicity of certain chemicals, their quantity present in the products and the duration of exposure to them. One important type of substances mentioned above is heavy metals. Systems in which toxic metal elements can induce impairment and dysfunction include the blood and cardiovascular system, eliminative pathways (colon, liver, kidneys, skin), endocrine (hormonal) pathways, energy production pathways, enzymatic, gastrointestinal, immune, nervous (central and peripheral), reproductive and urinary system. This is the reason why the heavy metals which are present in the textile material represent a source of pollution which must be controlled to protect both the environment and human health.The most important standards in Europe for evaluation of the textiles ¨ ko Tex Standard and M. S. T. (Markenzeichen schadstoffgepru¨fter Textilien). are O According to its requirements, the textiles are controlled for extractable harmful heavy metals (As, Pb, Cd, Cr, Co, Cu, Ni, and Hg).

Project aims and objectives The objective of this project was the investigation of the impact of heavy metals from ¨ ko Tex standard. textiles on human health which will be examined according to the O This means a study of the extractable harmful heavy metals such as As, Pb, Cd, Cr, Co, Cu, Ni, and Hg from Croatian and Austrian textile products made of cotton, flax, wool, silk, viscose, and other similar materials which are manufactured and used daily. Since a majority of problems arise in the field of heavy metal complex dyes which could be extracted from the fabrics by perspiration or saliva solutions, the primary target of

interest was the determination of extracted heavy metals from different textile materials according to the O¨ko Tex standard. The quantity of extractable harmful metals in the textiles should not exceed the given limits. Leaching of heavy metals from textiles was carried out in saliva and perspiration solution (artificial acidic sweat solution, according to the ISO 105 – E04) using ultrasonic as well as microwave extraction methods. The metals will be separated simultaneously, identified and quantified by a rapid screening method (thin layer chromatography and video densitometry). UV/VIS spectrometry was applied as an additional comparable method for determination of particular extractable heavy metals. Main advantages of thin layer chromatography are the low costs, its ease of operation and rapidity.

Research deliverables The results obtained in Croatia with UV/VIS and TLC were compared to those achieved in Austria by graphite furnace - atomic absorption spectrometry (GF-AAS) and inductively coupled plasma – optical emission spectrometry (ICP-OES). The metal concentrations was be determined by GF-AAS and ICP-OES after preparation by microwave digestion oven in nitric acid. GF-AAS and ICP-OES are widely used analytical methods for the determination of trace and ultratrace elements in all kinds of samples. Whereas ICP-OES is simultaneous multielement method with high sample throughput, GF-AAS is a single element method requiring smaller sample amounts and enabling lower limits of detection especially for the toxic metals As, Cd and Pb. Publications M. Zeiner, I. Rezic´ & I. Steffan (2006), Overview of Analytical Methods for the Determination of Heavy Metals Present on Textile Materials, 3rd International Textile, Clothing & Design conference, 8th – 11th October 2006, Dubrovnik, Croatia. I. Rezic´, M. Zeiner, I. Steffan (2007), UV/VIS Comparison of some Textile Dyes, XX Croatian Meeting of Chemists and Chemical Engineers, Zagreb, February 26 – March 1 2007. I. Rezic´, M. Zeiner & I. Steffan (2007), Risk by Chromium Tanned Leather for Human Skin, Second Conference of Croatian Scientists, 7th – 10th May 2007 Split, Croatia.

Zagreb, Croatia Faculty of Textile Technology, University of Zagreb, Prilaz baruna Filipovica 30, HR-10000 Zagreb, Croatia, Tel: +385 1 37 12 557; Fax: +385 1 37 12 591; E-mail: [email protected] Principal Investigator(s): Prof. Budimir Mijovic, Ph.D. Research Staff: Prof. Miroslav Skoko Ph.D., in retirement; Prof. Dragutin Taborsak Ph.D., Professor Emeritus; Prof. Salah-Eldien Omer, Ph.D.; Prof. Jovan Vucinic, Ph. D.; Nenad Mustapic, Mr. Sc.; Jasenka Pivac, Mr. Sc.; Zlatko Jurac, Mr. Sc.

Ergonomic design of the worker-furniture-environment system

Research register

75

IJCST 19,6

Other Partners: Academic

76

Faculty of Forestry, University of Zagreb Tvin, Virovitica, Croatia Project start date: 1 January 2007 Project end date: 31 December 2009 Source of support: Ministry of Science, Education and Sports, Republic of Croatia Keywords: Sitting, Furniture, 3D model of workplace, Work environment, Virtual reality

Industrial

Sitting furniture should enable the worker to take an optimal bodily sitting posture, ensuring active and dynamic sitting. A long-lasting and non-ergonomic bodily posture in this position causes uncomfortable sitting. Defining optimal working postures and strains makes a contribution to the reduction of necessary energy and facilitates working and circulation functions. The total work space should be designed in compliance with all criteria of the working posture and technical requirements. It is necessary to know the worker well, his working capabilities, work place and work methods to ensure an optimal working environment. Furniture dimensions and workplace, surrounding the furniture, regarding its optimal utilization, should be harmonized with the worker’s anthropometric sizes. Research methods are experimental, theoretical and numerical. Functional dependences of the worker-furniture-environment system will be investigated, based on ergonomic postulates in order to find optimal conditions between work humanization and productivity. Investigations are determined by measuring and recording typical working postures as well as conditions of excessive workload. Using digitally scanned 3D anthropometric characteristics of the human body, a digital 3D biomechanical model is obtained, taking account of the appropriate kinematicdynamic motion rules and the construction of the inner skeleton. 3D program applications with advanced automated defined anthropometric and ergonomic features of biomechanical models and digital figures will be recorded. A 3D visualization of the workplace by using a computer-based 3D model of furniture and computer-based character animations of workers will be performed. By using computer 3D program solutions, the prototype is substituted by 3D models on which all necessary designs and changes in real time have been carried out interactively. Computer visualization will be used to perform a biomechanical analysis of movements based on the real correlation within the space of the interaction of workers and belonging working environment on the obtained 3D models of workers and workspace. It is necessary to analyze the workspace and time studies of motion accurately. The 3D virtual model will enable a detailed biomechanical analysis of motion, speed and acceleration and more designer’s solutions of furniture with biomechanical and ergonomic parameters. Detrimental impact of too a high noise on workers as well as efficient procedures of noise reduction will be investigated. Special attention will be focused on detrimental action of microclimatic conditions regarding technological requirements of the industry. The optimization of work energy during the performance of the work by the worker will be performed to lessen fatigue and to remove excessive workload and to reduce sick-leaves. The performance of these investigations will result in ergonomic technical-economic design of the interactive work-furniture-environment system which is of great importance for the development of the Republic of Croatia and elsewhere in the world.

Project aims and objectives The purpose of these investigations is to achieve the reduction of work energy when workers work in sitting position and to enable the reduction of fatigue and falling ill. When performing work, a better economic performance with a lower energy consumption and operator’s fatigue should be obtained. In testing the level of detrimental sound effects their efficacious reduction could be achieved. The action of noise and vibrations could be prevented by investigating efficient noise and vibration dampers, damping the transfer of vibrations on workers by using pads and personal protective equipment. The aim of these investigations is to optimize work movements in sitting position with less operator’s fatige and reduction of sick-leaves in the industry. Thereby, working conditions and environment as well as safety in technological processes are to be especially stressed. They should be used to improve economic characteristics substantially. Special purposes of the investigation are the optimization of working postures and the confirmation of the knowledge about defining new criteria for the right ergonomic design of working pieces of furniture.

Research deliverables The application of investigations means work simplification, investigation and determination of manufacture time from the point of view of ergonomic starting points. When performing work, a better economic effect is to be achieved with a less energy consumption and operator’s fatigue. Based on the investigated causes of the increase of sound pressure level of machines, it is necessary to find out qualitative solutions to reduce the level into permitted and tolerable limits. To obtain satisfactory microclimatic conditions, work spaces, machinery and air-conditioning instruments with better economic and other characteristics will be defined, whereby current energy consumption will be reduced. Furniture dimensions are determined by computer based visualization of virtual 3D character in interaction with digitized and really designed furniture and environment. Virtual simulation offers the opportunity to create optimal constructive solutions of furniture which enables active and dynamic sitting. It is a relaxation bodily posture that is thereby obtained when sitting. Ergonomic, anthropometric, functional and technical requirements of workers are thereby satisfied. A 3D simulation model of comfortable and safe bodily posture in sitting position is created, together with all biomechanical characteristics, and interactive effects with existing and new materials used for making sitting furniture are attained. Thereby, design and construction of sitting furniture is supplemented, which, because of its designer and constructive solutions and design observance, reduces the possibilities of low-quality designs and problems of body illness and causing inability to work.

Publications A. Agic, V. Nikolic, B. Mijovic (2006), “Foot Anthropometry and Morphology Phenomena”, Collegium antropologicum. Vol. 30 (2006), No. 4; 815-821. A. Agic, B. Mijovic (2006), “Planar Model of the Deformation Behaviour of Electrospun Fibrous Nanocomposites”, Tekstil, Vol. 55 (2006), 606-612. A. Agic, B. Mijovic, T. Nikolic (2007), “Blood Flow Multiscale Phenomena”, Collegium antropologicum. Vol. 31 (2007), No. 2; 523-529.

Research register

77

IJCST 19,6

78

U. Reischl, V. Nandikolla, C. Colby, B. Mijovic, H.C. Wei (2007), “Ergonomics Consequence of Chinese Footbinding: A case study, Ergonomics 2007”, Budimir Mijovic (Ed.) Croatian Society of Ergonomics, 2007. 1-6. D. Novak, B. Mijovic (2007), “Applying Cognitive Rrgonomics to teaching Math, Ergonomics 2007”, Budimir Mijovic (Ed.) Croatian Society of Ergonomics, 15-20. N. Mustapic, B. Mijovic (2007), “Assessing the Slip resistence of Flooring, Ergonomics 2007”, Budimir Mijovic (Ed.) Croatian Society of Ergonomics, 2007. 155-163.

Zagreb, Croatia Faculty of Textile Technology University of Zagreb, Prilaz baruna Filipovic´a 30, HR-10 000 Zagreb, Croatia, Tel: +38513712500; Fax: +38513712599; E-mail: [email protected] Principal Investigator(s): Prof. Darko Ujevic´, Ph.D. Research Staff: Jadranka Akalovic´, B.Sc.; Prof. Jadranka Bacˇic´; Prof. Zoltan Baracˇkai, Ph.D.; Vinko Barisˇic´, B.Sc.; Ing. Iva Berket; Bajro Bolic´, B.Sc.; Blazˇenka Brlobasˇic´ Sˇajatovic´, B.Sc.; Ksenija Dolezˇal, B.Sc.; Mirko Drenovac, Ph.D.; Prof. Milan Galovic´, Ph.D.; Marijan Hrastinski, B.Sc.; Renata Hrzˇenjak, B.Sc.; M.D. Natasˇa Kaleboti; Prof. Isak Karabegovic´, Ph.D.; Ivan Klanac, B.Sc.; M.D. Irena Kos-Topic´; Prof. Tonc´i Lazibat, Ph.D.; Nikol Margetic´, B.Sc.; Prof. Zlatka MenclBajs; M.D. Zˇeljko Mimica; Prof. Gojko Nikolic´, Ph.D.; Alem Orlic´, B.Sc.; PhD.M.D. Vedrana Petrovecˇki; B.Sc.M.E. Zˇeljko Petrovic´; Prof. Dubravko Rogale, Ph.D.; Prof. Andrea Russo; Igor Sutlovic´, Ph.D.; Prof. Vlasta Szirovicza, Ph.D.; Irena Sˇabaric´, B.Sc.; MSc.M.D. Nadica Sˇkreb-Rakijasˇic´, Marija Sˇutina, B.Sc.; Prof. Larry C. Wadsworth, Ph.D.

Anthropometric Measurements and Adaptation of Garment Size System Other Partners: Academic

Industrial

Project start date: 1 January 2007 Project end date: 31 December 2011 Source of support: Ministry of Science, Education and Sports, Republic of Croatia Keywords: Anthropometric Measurements, Garment Size System Systematic anthropometric surveys have been conducted since 1901 with the aim of developing and improving systems for clothing and footwear sizes. The measurement results show how a national population changes over a period of several decades in physical build and size due to a series of factors (food habits, sports development, genetic predispositions, population migrations, climatic conditions etc.). Based on the results of anthropometric measurements in the Republic of Croatia (2004/05) on the sample of 30,866 test persons aged between 1 and 82 a statistical analysis of body measurements was performed, a database including 5 basic studies of sex and age as well as a new standard for clothing and footwear was built. These results enable a significant and stimulating continuation of scientific research and a comparison

to other national standards and their contributions to the creation of systems for clothing and footwear sizes. Elements common for national standards of garment sizing by an exact approach will be investigated and analyzed, in particular because the presumptions of national systems and starting elements respectively are not universally founded like intersize intervals which differ in sizes since the conformity of individual starting places is missing. Data will be provided for a common base with methods of body measuring and size designation of clothes according to the recommendations of the Technical Committee TC133 within ISO and EN standards as well as the design and development of a sophisticated computer system (DOV-KO) for unifying all body measurements and basic garment construction based on one or all other sizes. Within the scope of this project and based on experience, a very important cycle of anthropometric measurements of the sporting population in football, water polo, rowing, basketball and handball will be performed. 4,000 test persons from Zagreb, Osijek, Rijeka, Split and Dubrovnik will be measured, whereby specific body differences and deformations of muscles caused by longstanding training will be analyzed. A comparative analysis of the representative sample of the anthropometric measurements of sportsmen and other population as well as the investigation of other trends of body measurements will be performed. This will enable an exceptional insight into the anthropometric dimensions which reflect body shape, proportionality, composition and elements of success in sports respectively. Stadiometar or a new measuring instrument for continuous measuring body height, foot length and width will be designed too.

Project aims and objectives Problems of garment sizing and fit affect the market globally, and a consequence of bad predictions of the quantity of necessary stocks for manufacturers and dealers poses a risk of high costs. In the case of domestic manufacturers samples may be additionally divided. Particular solutions may be considered more efficiently by interpreting the data from the anthropometric database connecting five studies according to sex and age and the system of sizing. Therefore, using the results of the anthropometric measurements taken and the basic projection of the new standard for clothing and footwear, one of the directives of the project is to investigate other national standards of Europe and the world and to create size intervals and a new Croatian standard. The study of body differences and specific deformities of the body muscles during the longstanding practice of athletes such as leg circumference, chest circumference, torso, shoulder width, arm and leg length, body height, palm length is an additional aim of this project which will show the body elements affecting success level in sport. Besides a greater adaptation of clothing and footwear to the home market, it would be advisable to ensure the continuity of investigating the national size standard by creating a sustainable Croatian system of clothing and footwear sizes in conformity with anthropometric surveys that are conducted periodically and systematically in developed countries in which a change in the morphology of the human body occurred over the last decades. By way of proof, systematic anthropometric measurements and sizing in

Research register

79

IJCST 19,6

80

France showed a tendency of average height growth. In Great Britain in female population a growth of bust circumference was recognized, whereas in USA studies point to the tendency of an increase in obesity (besides an aesthetic also a health problem). The interest of the scientific and professional public, manufacturers, tradespeople and consumers in sizing will continue to grow, since a faster change in established body proportions may be expected thanks to changes in living and food habits of the population, an unavoidable mingling of ethnic groups, increase in the number of older consumers of clothing and footwear which will be doubled in the next two decades etc. Thus, it is additionally stressed how much it essential at the moment to ensure a valid starting point or a Croatian standard to pursue next movements of body dimensions in order to avoid a discrepancy and imposition of the specificities of domicile consumers to home and foreign manufacturers.

Research deliverables Anthropometry is the study of the measurement of the human body, but Pheasant has expanded it as “applied anthropometry” including quantitative data of size, forms and other physical characteristics of people that can be used in garment design. Since the form of the human body changed through time, the problem of ageing proved to make a contribution to perceived changes in body shape and size more than any other individual factor, such as for example improved nutrition and prolonged life, in particular the knowledge that the number of older consumers will be doubled by 2003. Therefore, the systems of sizing shall be updated periodically to ensure a correct fit of ready-made clothing. On the other hand, the home industry of clothing, fashion wear and footwear disposes of modest and aged data based on the out-of-date anthropometric measurements from 1962. It was therefore necessary to conduct a new cycle of anthropometric measurements and to use the obtained results. World fashion industry shows a special interest in measuring anthropometric characteristics of the population so that it gathers such data permanently, motivated by the wish for designing articles of clothing for all population groups, including the persons with pronounced specificities (higher stature, higher body weight etc.). The use of investigations will contribute to creating a new and modern Croatian standard for clothing and footwear harmonized with ISO and EN standards. Besides the clothing and footwear industry, pediatricians, specialists of occupational and sports medicine, experts in wood processing industry, automotive industry, in the army and police will benefit from the investigation results. Teachers and students in undergraduate and graduate studies as well as teachers and pupils at technical schools will benefit from the development results of the computer system based on the selection of garment sizes. By using the investigations of the sporting population, one can get an insight into tendencies of diversities of body measurements and changes in muscles as a result of longstanding practice. Various specialists of sports medicine, orthopedists, garment and footwear designers will benefit form the results of this investigation because based on previous experience it is evident that mass customization is necessary for athletes. Knowing dimensional characteristics, this method would be considerably promoted and improved.

Publications Ujevic´, D., Rogale, D., Hrastinski, M., Drenovac, M., Szirovicza, L., Lazibat, T., Bacˇic´, J., Prebeg, Zˇ., MenclBajs, Z., Mujkic´, A., Sˇutina, M., Klanac, I., Brlobasˇic´ Sˇajatovic´, B., Dolezˇal, K., Hrzˇenjak, R. (2006), “Normizacija, antropometrijski pregledi i Hrvatski antropometrijski sustav”, Tekstil, Vol. 55 No. 10, pp. 516-526. Ujevic´, D., Firsˇt-Rogale, S., Nikolic´, G., Rogale, D. (2006), “Pregled razvojnih dostignuc´a u tehnologiji sˇivanja - IMB 2006.”, Tekstil, Vol. 55 No. 12, pp. 624-631. Ujevic´, D., Dolezˇal, K., Lesˇina, M. (2007), “Analiza antropometrijskih izmjera za obuc´arsku industriju”, Poslovna izvrsnost, Vol. 1 No. 1, pp. 171-183. Ujevic´, D., Hrzˇenjak, R., Dolezˇal, K., Brlobasˇic´ Sˇajatovic´, B. (2007), “Hrvatski antropometrijski sustav jucˇer, danas, sutra”, HZN Glasilo, Vol. 3 No. 1, pp. 5-10. Hrzˇenjak, R., Ujevic´, D., Dolezˇal, K., Brlobasˇic´ Sˇajatovic´, B. (2007), “Investigation of Anthropometric Characteristics and Body Proportions in the Republic of Croatia”, Proceedings of 7th Annual Textile Conference by Autex, Tampere, Finland, 25-28 June, pp. 1191-1198. Ujevic´, D., Brlobasˇic´ Sˇajatovic´, B., Dolezˇal, K., Hrzˇenjak, R., Mujkic´, A. (2007), “Rezultati prvog antropometrijskog mjerenja stanovnisˇtva Republike Hrvatske”, Drugi kongres hrvatskih znanstvenika iz domovine i inozemstva, Split, Croatia, 5-10 May. Nikolic´, G., Ujevic´, D. (2007), “Protractor for Measuring Shoulder Slope”, Patent. Ujevic´, D. (2007), “One-arm and/or two-arm anthropometer”, Patent.

Zagreb, Croatia University of Zagreb Faculty of Textile Technology, Prilaz bazuna Filipovic´a 30, HR-10000 Zagreb, Croatia, Tel: +385 1 48 77 359; Fax: +385 1 48 77 355; E-mail: [email protected] ˚ urdica Parac-Osterman, Ph.D. Principal Investigator(s): Prof. A Research Staff: Martinia Ira Glogar, Ph.D.; Assist.Prof. Darko Golob, Ph.D.; Assoc. Prof. Marija Gorensˇek, Ph.D.; Assoc. Prof. Darko Grundler, Ph.D.; Assoc. Prof. Nina Knesˇaurek, Ph.D.; Prof. Nina Rezˇek-Wilson; Assist.Prof.Tomislav ˚ urasˇevic´ B.Sc. Rolich, Ph.D.; Ana Sutlovic´, M.Sc; Vedran A

Colour and Dyestuff in Processes of Ecologically acceptable sustainable development Other Partners: Academic

Industrial

Jadran Stockings Factory University of Ljubljana & Maribor, Slovenia Project start date: 1 January 2007 Project end date: 31 December 2011 Project budget: Source of support: Ministry of Science, Education and Sports, Republic of Croatia Keywords: Dyestuff selection, Nano-technology, Optimizing dyeing process, Purifying and decolouring wastewaters, Colour management, Fuzzy logic Scientific contribution to sustainable development relies on unlimited support of basic, developing and employable research. Therefore, selection of multi-functional dyes

Research register

81

IJCST 19,6

82

(UV protection, antibacterial protection, micro capsules of multi-functional performance), applying nano-technology in the dyeing processes with the aim of preventing water contamination, development of new methods as well as purifying dyed wastewaters contribute to sustainable development. Both input and output parameters of water will be controlled throughout the entire dyeing process: amount of residual dye in dyebath using Lamber-Beer absorption model; X, Y, Z standard spectral characteristics of colour defined by specific absorption coefficient SAC and water quality defined by BOD5, COD, TOC, AOX, electrical conductivity and other defining values. System of control comprising advance models of control such as fuzzy logic (model based on rules) and model based on physical and chemical processes will be developed and applied. Capital area of research will involve models of dyeing processes, colour control and its correlation to dye as well as the interactive system of dye control. Models should describe and predict kinetics, reactivity, affinity, exhaustion, fixation and interaction of solutions containing various dyestuffs. Prediction of output process result as well as definition of both physical and chemical parameters crucial for controlling the process will be conducted based on afore-mentioned models. These models encompass kinetic models (according to Nernst and Langmuir) modified for interactions between dyestuff on fibre and in the solution. Interdisciplinarity of dye within the system of sustainable development is based on spectral characteristics of colour as the fundament model dependent of the employment conditions. Instrumental measurement of colour is involved in all industrial production processes: textile technology, design, graphic industry and elsewhere which enables implementing control and colour harmonization. Application of evolutionary algorithms for modeling computer aided design of textiles based on principals of examinee’s subjective evaluation. Methods of descriptive statistics as well as methods of statistic reasoning will be applied within the frame of statistic analysis. Scientific affirmation of research results will be computer simulation as well as in vivo confirmation.

Project aims and objectives Contribution to sustainable development relies on unlimited support of basic, developing and employable research. Aim of the project is to contribute to humane ecology (regarding UV, antibacterial and other protective properties), through use of multifunctional dyes and selection of appropriate waste water discoloration and purifying methods, all in order of obtaining biological quality of water (free of toxic, aromatic components which may form in the process of dye degradation). Cognition of structure and use of thermo sensible nano sized dyes will enable their use on fibres for special purposes. Advance dyeing technologies, with the overview on pretreatment of textile substrates (enzymatic, plasmatic and other) in order of preserving environment and saving energy, will be applied. Base of the project is application of dyeing process control, including advance models of control; fuzzy logic and model based on physico-chemical processes. Colour used as constant value will be applied for formation of fuzzy logic model, used for complex colour designing, automatic dye selection, direct transfer of colour coordinates data into the dyeing recipe setup system, advance recipe correction, as well as control and colour matching. Project research will enable use of new instrumental methods and development of researcher’s creativity, while graduate students and

potential PhD students will be given a chance to get acquainted with scientific methodology, development of experimental skills and writing scientific papers.

Research register

Research deliverables Influence of dye’s chemical constitution and mode of dye-fibre bond onto antibacterial (e.g. Staphiloccoci, Escherichia), UVA and UVB protection properties. Influence of additives (electrolytes and surfactants) on dyeing process and degree of water pollution. Control of, in dye-bath and wastewaters, present electrolytes – elaboration of mathematical model. Further results considering influence of dye onto protection properties are expected. Application of thermo sensible dyes on children clothing. Influence of textile substrate’s pre-treatment (enzymatic, plasmatic pre-treatment etc.) on dyeing kinetics and energy saving. Results of wastewater purifying and discoloration methods, with the emphasis on salt removal using physico-chemical methods, nano filtration and reverse osmosis. Colour as constant value of monitoring process, dye properties and colour matching in design applying evolutionary algorithms. Application of nano size particles. Influence of surfactants onto reactivity, affinity, exhaustion and fixation degree of reactive dyes. Advantages and disadvantages of physico-chemical decolouring methods. Dye degradation products and their toxicity (considering aromatic components) in wastewaters. Selection of dyestuff and its interaction with in the dye-bath present additives. Mathematical model based on measured values will be elaborated, while control system including a model based on physico-chemical processes will be applied. Fuzzy logic model, based on colour as constant value within control system will be worked out. Application of capsulated dyestuff and nano particles of zink and silver for special use (medical textiles). Pre-treatment of hydrophobic, synthetic fibres in the aim of increasing hydrophility and applicability of, in water soluble, dyes. From the economical and ecological aspect, a more acceptable system of purifying and decolouring wastewaters ii expected. Mathematical model based on the analyses of input and output measured values, considering coloured waters, will be elaborated, while a control system using advance models, such as fuzzy logic (model based on rules) and based on physicochemical processes model. Evolutionary algorithms for modelling computer design of fabrics based on principals of subjective asessment. Model must comply with standards, flexible, stabile, precise, and easily applicable. It includes complete process modelling: dyeing, colour control and its relation to dye. These models involve kinetics models (Nernst, Langmuir) modified for dye - dissolved dye. Applying CCM (computer colour matching) methods based on Kubelka-Munk theory, spectral characteristics and colour parameters according to CIEL*a*b* system, a model of fuzzy logic for complex design by colour, automatic dye selection, direct transfer of colour coordinates data in to the dyeing recipe setup system, advance recipe correction, as well as control and colour matching, will be elaborated. Evolutionary algorithms for modelling computer design of fabrics based on principals of subjective asessment. Publications ˚ urdica Parac-Ostreman, Ana Sutlovic´, Vedran A ˚ urasˇevic´ and Tjasa Griessler Bulc, “Use of wetland A for dye-house waste waters purifying purposes”, Asian Journal of Water, Environment and Pollution, Vol. 4, No. 1, pp. 101-106.

83

IJCST 19,6

84

˚ urdica Parac-Osterman, Vedran A ˚ urasˇevic´, Ana Sutlovic´, “Comparison of some chemical and A physical-chemical waste water discoloring methods”, Chemistry in Industry, in publication. ˚ urdica Parac-Osterman, Tomislav Rolich, “Fuzzy logic based Martinia Ira Glogar, Darko Grundler, A approach to textile surface structure influence in colour matching”, AATCC Rewiev, in publication. ˚ urasˇevic´ (2007), “Dyeing Properties of Vesna Tralic´-Kulenovic´, Livio Racane, Ana Sutlovic´, Vedran A New Benzothiazol Disperse Dyestuff”, XX. Croatian Meeting of Chemists and Chemical Engineers, February 26 – March 1, 2007, Zagreb, Croatia. ˚ urdica Parac-Ostreman, Nevenka Tkalec Makovec, Ana Sutlovic´, Ljerka Dugan (2007), A “Staphylococcus Aureus and Escherichia Coli Behavior on Undyed and Dyed Wool”, XX. Croatian Meeting of Chemists and Chemical Engineers, February 26 – March 1, 2007, Zagreb, Croatia. ˚ urdica Parac-Osterman, Ana Sutlovic´, Vedran A ˚ urasˇevic´ (2007), “Application of Wetland System”, A Textile Dyes Zagreb 2007, March 9th 2007, Zagreb, Croatia. ˚ urdica Parac-Osterman, Ana Sutlovic´, Martinia Ira Glogar (2007), “Dyeing Wool With Natural Dyes in A Light of the Technological Heritage”, 7th annual Textile Conference by Autex: “From Emerging Innovations to Global Business”, 26-28 June 2007, Tampere, Finland. ˚ urdica Parac-Osterman, Ana Sutlovic´, Vedran A ˚ urasˇevic´: Application of Wool, CA and PP Fibers as A Filters in Wetland Pretreatment Media Formation, University of Zagreb, Faculty of Textile technology Intrenational Conference on Multi Functions of Wetland Systems, 26-29 June, Legnaro (Padova), Italy.

Zagreb, Croatia Faculty of Textile Technology, University of Zagreb, Prilaz baruna Filipovic´a 30, HR-10 000 Zagreb, Croatia, Tel: +385 1 48 77 352; Fax: +385 1 48 77 352; E-mail: [email protected] Principal Investigator(s): Prof. Drago Katovic´, Ph.D. Research Staff: Asoc. Prof. Sandra Bischof Vukusˇic´, Ph.D.; Prof. emeritus Ivo Soljacˇic´, Ph.D.; Dubravka Dosˇen Sˇver, Ph.D.; Sandra Flincˇec Grgac, B.Sc.; Asoc. Prof. Radovan Despot, Ph.D.; Asist. Prof. Jelena Trajkovic´, Ph.D.; Asist. Prof. Branka Lozo, Ph.D.; Luka Cˇavara, M.Sc.; Bozˇo Tomic´, M.c.; Prof. Charles Yang, Ph.D.; Prof. Christian Schram, Ph.D.

Alternative eco-friendly processing & methods of cellulose chemical modification Other Partners: Academic

Industrial

Cˇateks, d.d., www.cateks.hr Faculty of Forestry, Croatia; Faculty of Graphic Art, Croatia; University of Georgia, USA; University of Innsbruck, Austria Project start date: 1 January 2007 Project end date: 31 December 2011. Source of support: Ministy of Science, Education and Sports, Republic of Croatia Keywords: Multifunctional eco-friendly textile finishing, Polycarboxylic acids, Protective functionalities, Chemical modification of cellulose, Microvawe treatment of cellulose materials

One of the requests of European Union for higher competiteveness of european market is rebuilding and reconstruction of traditional industrial sectors, especialy textile and wood industry. According to the strategical goals of the Republic of Croatia the project emphasizes the use of highly sofisticated production processes and treatments of cellulose materials i.e. obtaining additional and improved characteristics of wooden and paper matherials which can be acchieved by using high-tech processes and by introduction of nano- micro- and bio-technologies. One of the alternative methods for replaciong the conventional reactants containing formaldehyde which were used in textile and wood treatments so far, would be the modification with eco-friendly agents such as polycarboxylic acids. Efficiency of these treatments will be determined quantitatively by ester crosslinking analytical methods or by means of isocratic HPLC and spectrophotometric FTIR method. Standard methods of textile, wood and paper material testing would be used for examining their protective performance and resistance to weathering conditions. Part of the proposed project will be development of optional multifunctional treatment that would provide better protection of cellulose materials against microorganisms, UV, electromagnetic rays, flame, oil or water. Therefore a particular attention will be payed to development and application of the agents which will not only improve the characteristics of textile matherials but also give it permanent freshness and provide additional care and protection i.e. medical characteristics. Optimisation of alternative processing and methods will provide ecologically and economicaly favorable characteristics of treated matherials. Further process optimisation in order to improve processing quality could be obtained with new alternative method using microwave energy. Improved characteristics obtained with this method in our previous research confirm its usability in textile finishing processes as well as in chemical modification of wood. Previous research in this field represent worlwide novelty which should be by all means continued.

Project aims and objectives The purpose and aim of the proposed project is to obtain highly valuable and multifunctional treated textile materials that will acquire analogous price on the demanding market. This is the basic condition for the survival of Croatian textile, wood and paper industry on EU market. In textile area experiments will be conducted to obtain multifunctional environmentally friendly textile material which will simultaneously offer dimensional stability, flame retardancy, crease and antimicrobial resistance and will have no effects on human health. Further goal is to obtain chemicaly modified wood that will have reduced shrinking and water absorption as well as to obtain flame retardancy on wood and paper products. One of the equally important goals is construction of a semi industrial microwave device for continuous planar treatment of cellulose materials. The results obtained would be presented in the world best known papers in the relevant field. The most important goal of the project is affirmation of Croatian science in Europe and rest of the World, by presenting the results in international papers so as on International Conferences. It is important to stress that established cooperation with EU and USA experts, so as with their scientific institutions will be continued and expanded.

Research register

85

IJCST 19,6

In this project, where will scientists from abroad have an active contribution with their work, further contribution to development of high quality products will be added. We certainly hope it will affect development of Croatian industry and economy.

Research deliverables

86

Publications Katovic´, D., Bischof Vukusˇic´, S., Flincˇec Grgac, S. (2007), “Crosslinking cotton with citric acid and organophosphorus agent for the purpose of flame retardant finishing”, 85th Textile Institute Conference, Colombo, Sri Lanka, pp. 820-824. Bischof Vukusˇic´, S., Flincˇec Grgac, S., Katovic´, D. (2007), “Catalyst influence in low formaldehyde flame retardant finishing system”, 7th AUTEX Conference, Tampere, pp. 60-61. Flincˇec Grgac, S., Katovic´, D., Bischof Vukusˇic´, S. (2007), “Combination of organophosphorus agent and Citric acid in Durable Press Finishing of Cellulose Fabrics”, XX. Croatian Society of Chemical Engineers, Zagreb, Croatia, p. 281. Bischof Vukusˇic´, S., Flinecˇ Grgac, S., Katovic´, D. (2007), “Antimicrobial Textile Treatment and Problems of Testing Methods”, Tekstil, Vol. 56, accepted for publication.

Zagreb, Croatia Faculty of Textile Technology, University of Zagreb, Prilaz baruna Filipovica, HR-10 000 Zagreb, Croatia, Tel: + 385 1 3712 540; Fax: + 385 1 3712 599; E-mail: [email protected] Principal Investigator(s): Prof. Dubravko Rogale, Ph.D. Research Staff: Prof. Zvonko Drgacˇevic´, Ph.D.; Prof. Gojko Nikolic´, Ph.D.; Prof. Maja Vinkovic´; Snjezˇana Firsˇt Rogale, Ph.D.; Slavenka Petrak, Ph.D.; Goran Cˇubric´, B. Sc.; Bojan Mauser, B. Sc.

Intelligent garment and environment Other Partners: Academic

Industrial

Project start date: 1 January 2007 Project end date: 31 December 2011 Source of support: Ministry of Science, Education and Sports, Republic of Croatia Keywords: Intelligent garment, Active thermal protection, Micropneumatic system, Intelligent sick bed, Adaptable ironing machine, Multiaxial testing of textiles and their joints Investigations, construction and development of intelligent article of clothing related to its direct environment by developing an adaptable bed, adaptable ironing machine and measuring instrument for multiaxial testing physical-mechanical properties of technical textile and joined parts. The purpose of the project is that a research team makes researches resulting in a construction and realization of the first intelligent garment whose basic function is

active thermal protection. It contains a sensor system for monitoring the values of air temperature inside and outside of the garment, data bus for data transfer, microcomputer and micro controller, and execution devices for the automatic regulation of thermal protection value. Controlling conduction and convection of the heat of the human body regulates thermal protection in such a way that based on anthropometric measurements several types of various air thermo insulation elastic chambers are constructed which are integrated into the construction of the garment between the outer shell and lining. Thermoinsulation chambers consist of several segments and have a twofold function so that by inflating sealing properties are assumed, and the heat loss of the human body by convection can be regulated and the thickness of the air chambers can be changed by program, whereby the heat loss of the human body by conduction can to be regulated. Micropneumatic elements and the chambers would be equipped with sensors of air pressure integrated into them, because depending on air pressure values in the chambers there will be defined chamber forms, their sealing properties and thickness on which thermal resistance depends. Investigations would prove that the integration and efficient joint operation of the integrated sensors, microcomputers with associated algorithms of intelligent behavior and actuators so that an independent action of the garment is realized with the aim of thermal protection whereby the garment would have the attribute of active, adaptable and intelligent behavior in variable temperature conditions. Communication possibilities of intelligent garment with the environment would be examined and an intelligent sick bed, adaptable ironing-machine and an instrument for testing load would be developed. They would practically use the same or very similar sensor, computer and micropneumatic actuator systems, connection techniques, constructions and design as well as intelligent garment.

Project aims and objectives The basic aim of the proposed research project is to investigate the possible construction and practical realization of an intelligent article of clothing with thermal protection, adaptable bed, ironing machine for the technological manufacturing process and necessary measuring instruments. The purpose of the investigation is to investigate characteristics of all elements of the system and behavior of the system as a whole and the communication between intelligent garment and environment. In addition to the basic aim of all investigations it is necessary to point out other aims too emerging as the result of the said investigation. Establishment of the leading European and world scientific role in investigations and development of intelligent garment. Writing 4 doctoral dissertations in the mentioned field (two dissertations in the mentioned field have been registered and approved by the Senate of the University of Zagreb. To prove that the Croatian clothing industry possesses a strong scientific research basis that guaranties it a technological excellence on demanding foreign markets.

Research deliverables After 1st year: Selection and investigation of properties of installation materials, sensors, microcontroller systems and actuators as well as materials for making

Research register

87

IJCST 19,6

88

thermoinsulation chambers. Shapes and constructions of thermoinsulation chambers adjusted to dynamic anthropometric properties and necessary degree of freedom of movements, parameters of ultrasonic welding of chambers and influence of geometry, arrangement of chamber elements on thermoinsulation characteristics. Manufacture of microcontroller systems, sensor amplifiers, actuator drive and power supply system. Collection of technical literature and familiarizing with problems of beds for patients, ethimology and decubitus. Collection of technical documentation on the achievements of the technique of the intelligent ironing machine. Investigations of optimal ironing parameters: specific pressure, humidity, temperature, duration of humidification and drying. After 2nd year: Computer digitalization of cutting parts of the garment, modeling, seam allowance and cutting pattern production with NC units. Installation of shell elements and thermoinsulation chambers together with integration of buses of the microcontroller system. Investigations and check of the characteristics of the operation of the sensor and actuator system and the operation of the power supply system. Actuation and investigation of the operation of the program languages and tools for the thermoinsulaiton chambers and 3D construction of all installation elements. Implementation of the microcontroller program. Manufacture of a functional prototype. Experimentation with various forms of the bladder for beds for patients on segments, various materials, procedure of bladder elongation without stretching, method of joining several segments. The process of the controlled generation of wet air with specific humidity and the accurate control of the intelligent ironing machine will be implemented. After 3rd year: Design and examination of the complex program for measuring by means of integrated sensors, data interpretation and making decisions on the optimal combination of thermal protection, actuation of thermoinsulation chambers. Creation of an algorithm of intelligent behavior. The application of the PWM technology and communication protocol with display. Functional examination of the whole system of intelligent garment. Measurements and final check of the characteristics of the 3D construction, thermoinsulation chambers and all integrated components. Design of the sensor system for measuring the pressure in the bladders of the bed for patients and segments respectively. Determination of the optimum air pressure. Devising a proportional pneumatic system that can independently and quickly change the pressure level in each segment and bladder. System design of measuring the ironing surface of the adaptable ironing machine by tactile sensors or camera system with computer program. After 4th year: Investigations of autonomous operation and characteristics of an intelligent article of clothing in variable climatic conditions in the air conditioning equipment and in real atmospheric conditions. Improvement and conformation of the algorithm of intelligent behavior. Investigations and improvement of the communication channel between intelligent article of clothing and its wearer. Devising and experimentation with operation algorithms and control program of the system of the bed for patients, and the use of the touch screen monitors.

Examination of the sensor system for accurate measuring fabric moisture, and a system for data acquisition about material type and its thickness by means of the integrated chip or bar code in the intelligent ironing machine. After 5th year: Investigations of the communication channels between the intelligent garment and its environment by using suitable wireless communication. Unifying all systems of the bed for patients and designing a prototype with examination in real conditions. Unifying all partial systems into the operation system of the adaptable ironing machine that, depending on sensor data, makes own decisions on the most suitable operation parameters using fuzzy logic for making decisions and integrated operation algorithms. The system should be examined on prototype. Publications Rogale, Dubravko; Firsˇt Rogale, Snjezˇana; Dragcˇevic´, Zvonko; Nikolic´, Gojko. Intelligent Article of Clothing with an Active Thermal Protection, PK20030727, (patent). Nikolic´, Gojko; Ujevic´, Darko. Protractor for Measurement of Shoulder Slope (patent claim). Dragcˇevic´, Zvonko; Rogale, Dubravko. Pneumatski ulozˇak za sprecˇavanje deformacija perive obuc´e (patent claim). Zˇeljko Sˇomodi, Anica Hursa, Dubravko Rogale (2007), “A minimisation algorithm with application to optimal design of reinforcements in textiles and garments”, International Journal of Clothing Science and Technology, 2007, 19, 3/4, 159 - 166 (original scientific paper). Nikolic´, Gojko; Cˇubric´, Goran (2007), “Investigating the positioning edge accuracy of sensors in textile and clothing manufacture”, International Journal of Clothing Science and Technology, 2007, 19, 3/4, 178 - 185 (original scientific paper). Edita Vujasinovic, Zeljka Jankovic, Zvonko Dragcevic, Igor Petrunic, Dubravko Rogale (2007), “Investigation of the strength of ultrasonically welded sails”, International Journal of Clothing Science and Technology, 2007, 19, 3/4, 204-214 (original scientific paper). Snjezˇana Firsˇt Rogale, Dubravko Rogale, Zvonko Dragcevic, Gojko Nikolic, Milivoj Bartosˇ (2007), “Technical systems in intelligent clothing with active thermal protection”, International Journal of Clothing Science and Technology, 2007, 19, 3/4, 222-233 (original scientific paper). Petrak, Slavenka (2007), The method of 3D garment construction and cutting pattern transformation models / Doctoral thesis. Zagreb: Tekstilno-tehnolosˇki fakultet, 17.05. 2007., 220 str. Voditelj: Rogale, Dubravko. Snjezˇana Firsˇt Rogale (2007), Intelligent clothing with active thermal protection / Doctoral thesis. Zagreb: Tekstilno-tehnolosˇki fakultet, 19.03 2007., 224 str. Voditelj: Dragcˇevic´, Zvonko. Hursa, Anica; Rogale, Dubravko; Sˇomodi, Zˇeljko (2006), “Application of Numerical Methods in the Textile and Clothing Technology”// Tekstil. Vol. 55 (2006), 12; 613-623 (review paper). Ujevic´, Darko; Firsˇt-Rogale, Snjezˇana; Nikolic´, Gojko; Rogale, Dubravko (2006), “Survey of Development Achievements in the Sewing Technology - IMB 2006”. // Tekstil, cˇasopis za tekstilnu tehnologiju i konfekciju. Vol. 55 (2006.), 12; 624-631 (presentation on symposium). Ujevic´, Darko; Rogale, Dubravko; Hrastinski, Marijan; Drenovac, Mirko; Szirovicza, Lajos; Lazibat, Tonc´i; Bacˇic´, Jadranka; Prebeg, Zˇivka; Mencl-Bajs, Zlatka; Mujkic´, Aida; Sˇutina, Marija; Klanac, Ivan; Brlobasˇic´ Sˇajatovic´, Blazˇenka; Dolezˇal, Ksenija; Hrzˇenjak, Renata (2006), “Standardization, Anthropometric Surveys and Croatian Anthropometric System” // Tekstil. Vol. 55 (2006), 10; 516526 (review paper). Sˇomodi, Zˇeljko; Hursa, Anica; Rolich, Tomislav; Rogale, Dubravko (2007), “Numerical analysis and optimisation of mechanical reinforcements on clothes” // Zbornik radova Prvoga susreta Hrvatskog drusˇtva za mehaniku / Cˇanadija, Marko (ur.). Rijeka: Hrvatsko drusˇtvo za mehaniku, 2007. 173-178 (professional paper).

Research register

89

IJCST 19,6

90

Zagreb, Croatia University of Zagreb, Faculty of Textile Technology, Prilaz baruna Filipovica 30, HR-10000 Zagreb, Croatia, Tel: +385 1 37 12 567; Fax: +385 1 37 12 535; E-mail: [email protected] Principal Investigator(s): Assist.prof. Edita Vujasinovic, Ph.D. Research Staff: Prof. Maja Andrassy, Ph.D.; Prof. Zvonko Dragcevic, Ph.D.; Prof. Dubravko Rogale, Ph.D.; Igor Petrunic, M.Sc.; Ruzica Surina, B.Sc.

Implementation of new materials objective measurement and evaluation system in the processes of technical and intelligent textile design Other Partners: Academic

Industrial

University of Maribor, Faculty of Mechanical Engineering, Maribor, Slovenia Project start date: 1 January 2006 Project end date: 31 December 2007 Source of support: Ministry of science, education and sport, Republic of Croatia Keywords: Objective evaluation, Technical textiles, Intelligent textiles, Design The intention of this project is to investigate and to establish objective evaluation systems for unconventional textile materials (composites, laminates, PCM) according to the specific requirements on particular type of technical and intelligent textiles, through systematic approach and application of conventional and/or modern methods and procedures of textile testing.

Project aims and objectives Exact characterization of new materials will enable comparison of such materials (regardless of their structure and composition), so that expert data basis for the prediction of their behavior during use e.g. in real condition will be possible.

Research deliverables On the basis of such investigations multi-component textile system will be designed and optimized, respecting the high and targeted property of such materials in use. Publications E. Vujasinovic, Z. Jankovic, Z. Dragcevic, I. Petrunic, D. Rogale (2007), “Investigation of the strength of ultrasonically welded sails”, International Journal of Clothing Science and Technology, Vol. 19(2007) No. 3/4, 204-214, ISSN: 0955-6222. E. Vujasinovic, Z. Dragcevic, Z. Bezic (2007), “Descriptors For The Objective Evaluation Of Sailcloth Weather Resistance”, Proceedings of 7th Autex Conference 2007, Tampere 26-28th June 2007, Finland; ISBN: 978-952-15-1794-5. Ruzˇica Surina & Maja Somogyi (2006), “Biodegradable polymers for biomedical purpose”, Tekstil Vol. 55 (2006), No. 12; 642-645.

First Rogale, Snjezana; Rogale, Dubravko; Dragcevic, Zvonko; Nikolic, Gojko; Bartos, Milivoj. (2006), “Technical Systems in Intelligent Clothing with Active Thermal Protection” // Book of Proceedings of the 3rd International Textile, Clothing and Design Conference –; Magic World of Textiles / Dragcevic, Zvonko (ur.).Zagreb: Faculty of Textile Technology University of Zagreb, 2006. 413-419. First Rogale, Snjezana; Rogale, Dubravko; Dragcevic, Zvonko; Nikolic, Gojko; Bartos, Milivoj. (2006), “The algorithm of the intelligent behavior of the article of clothing” Annals of DAAAM for 2006 & Proceedings of the 17th International DAAAM Symposium “Intelligent Manufacturing & Automation: Focus on Mechatronics and Robotics” / Katalinic, Branko (ur.).Bec: DAAAM Intaernational Vienna, 2006. 125-126. Vujasinovic E., Dragcevic Z., Gersak. (2006), “The impact of enzyme finish on the quality and hend of denim fabric”, Proceedings of the 37th International Symposium on Novelties in textiles, Ljubljana 2006, 78-82.

Zagreb, Croatia University of Zagreb, Faculty of Textile Technology, Prilaz baruna Filipovic´a 30, HR-10000 Zagreb, Croatia, Tel: +385 1 37 12 521; Fax: +385 1 37 12 599; E-mail: [email protected] Principal Investigator(s): Assoc. Prof. Emira Pezelj, Ph.D. Research Staff: Prof. Ruzˇica Cˇunko, Ph.D.; Prof. Maja Andrassy, Ph.D.; Assist. Prof. Edita Vujasinovic, Prof. Vili Bukosˇek, Ph.D.; Antoneta Tomljenovic´, Ph. D.; Sanja Ercegovic´, M. Sc.; Maja Somogyi, B. Sc.; Dubravka Gordosˇ, M. Sc.

Mulifunctional Human Protective Textile Materials Other Partners: Academic

Industrial

Project start date: 1 January 2007 Project end date: 31 December 2011 Source of support: Croatian Ministry of Science, Education and Sport, Republic of Croatia Keywords: Protective textiles, Multifunctionality, Smart textiles, Ceramic coatings, Sol-gel process The investigations proposed have been motivated by the fact that people are more and more exposed to various influences from the environment, which can be harmful to their health. Such harmful influences are, for example, UV irradiation, electromagnetic smog, high temperature, fire, etc. Contemporary textile materials for personal protection are required to offer high efficiency, in most cases multifunctionality, as well as a necessary level of comfort. The fabrics used are high-performance ones and interdisciplinary approach is necessary in research dealing with their development and manufacture. The thesis we propose is that the application of contemporary research results in the field of materials can be used to offer a new contribution to the development of multifunctional protective textile materials. The accent will be given to a purposeful surface modification of fabrics, using environmentally friendly agents and processes, which is in accordance with contemporary European trends of research in the field of materials. Special attention will be paid to investigating modifications using the new sol/gel process, combined with preceding ultrasound, laser and plasma treatment of textile surfaces.

Research register

91

IJCST 19,6

92

New possibilities of manufacturing efficient protective layers will be investigated, using various inorganic substances, including functional layers of nano-dimension made of hybrid inorganic-organic polymers. The aim is to optimise modification parameters of achieving efficient protection from UV and EM irradiation, as well as to increase resistance to abrasion, cutting and heat in particular materials, establishing antimicrobial properties at the same time. Adequate testing procedures will be established to evaluate the newly created materials. New levels of knowledge is expected to be achieved regarding correlation of protective properties and textile fabric composition, as well as the development of practical processes of obtaining aimed fabric modifications and the development of the methods of new material evaluation. New knowledge will contribute to the quality of education in the field of textile materials. Transfer of knowledge into actual industrial production is also expected. The results will be presented on international conferences and will be published in relevant international publications. The obtained results to be obtained could be used to stimulate manufacture of new high-performance textile materials for special purposes in Croatia.

Project aims and objectives The purpose of the investigations is to obtain new knowledge in the field of material development, especially regarding the new composites with textiles as a basic component. The knowledge should be directly applicable in practice, and simultaneously used to improve the quality of education, of both students, young researchers and experts from the industry. The new knowledge is expected to further the development of the Department of textile materials, where the investigations are organised. Based on the knowledge of high-performance materials, that has resulted in the development of the composites, and the role of textile component in them, the possibility will be investigated of obtaining high-performance composites for protection, in which textiles are the basic component. These are new textile materials to be used as protection from harmful influences of the general and working environment in high-risk industrial processes and other activities where people are exposed to risks of mechanical, thermal or chemical injuries, of infection by micro-organisms and even fatal risks from the causes. This is why protective materials are expected to offer high efficiency under various conditions, while the best solutions are aimed at obtaining multi-functional protection by a single material. The purpose of the research is to investigate the solutions that could be applied in textile industry, which could stimulate the introduction of knowledge-based and new-technology-based production in the industry, through adapting the industry to manufacture high-performance composite materials for special purposes. The aim of the investigation is to determine the procedures of obtaining multi-functional textiles for personal protection, simple to manufacture and use. The protective properties will be obtained by modifying the surfaces of the fabrics of various constructions, with the aim to establish optimal modification procedures and processing parameters which could offer efficient protection from individual influences, or, otherwise, protection from more influences. The investigations are supposed to result in solutions for objective evaluation of the effect achieved and the durability of protection as well, but also in the evaluation of the adequacy of the materials for a particular purpose. Adequate testing methods and procedures will be developed, appropriate indicators defined and the correlation of the modification parameters and properties achieved established.

Research deliverables The purpose of the investigations is to obtain new knowledge in the field of material development, especially regarding the new composites with textiles as a basic component. The knowledge should be directly applicable in practice, and simultaneously used to improve the quality of education, of both students, young researchers and experts from the industry. The new knowledge is expected to further the development of the Department of textile materials, where the investigations are organised. Based on the knowledge of high-performance materials, that has resulted in the development of the composites, and the role of textile component in them, the possibility will be investigated of obtaining high-performance composites for protection, in which textiles are the basic component. These are new textile materials to be used as protection from harmful influences of the general and working environment in high-risk industrial processes and other activities where people are exposed to risks of mechanical, thermal or chemical injuries, of infection by micro-organisms and even fatal risks from the causes. This is why protective materials are expected to offer high efficiency under various conditions, while the best solutions are aimed at obtaining multi-functional protection by a single material. The purpose of the research is to investigate the solutions that could be applied in textile industry, which could stimulate the introduction of knowledge-based and new-technology-based production in the industry, through adapting the industry to manufacture high-performance composite materials for special purposes. The aim of the investigation is to determine the procedures of obtaining multi-functional textiles for personal protection, simple to manufacture and use. The protective properties will be obtained by modifying the surfaces of the fabrics of various constructions, with the aim to establish optimal modification procedures and processing parameters which could offer efficient protection from individual influences, or, otherwise, protection from more influences. The investigations are supposed to result in solutions for objective evaluation of the effect achieved and the durability of protection as well, but also in the evaluation of the adequacy of the materials for a particular purpose. Adequate testing methods and procedures will be developed, appropriate indicators defined and the correlation of the modification parameters and properties achieved established. Publications R.Cˇunko, S.Ercegovic´, D.Gordosˇ, E.Pezelj (2006), “Influence of ultrasound on phisical properties of wool fibres”, Tekstil, Vol. 55 (2006) 1-9. A. Tomljenovic´, E. Pezelj, F. Sluga (2007), “Application of TiO2 nanoparticles for UV protective shade textile materials”, Proceedings of 38th symposium of textile novelity, Ljubljana 21st June 2007, Slovenija. E. Vujasinovic, Z. Jankovic, Z. Dragcevic, I. Petrunic, D. Rogale (2007), “Investigation of the strength of ultrasonically welded sails”, International Journal of Clothing Science and Technology, Vol. 19 (2007) No. 3/4, 204-214, ISSN: 0955-6222. E. Vujasinovic, Z. Dragcevic, Z. Bezic (2007), “Descriptors For The Objective Evaluation Of Sailcloth Weather Resistance”, Proceedings of 7th Autex Conference 2007, Tampere 26-28th June 2007, Finland; ISBN: 978-952-15-1794-5. Ruzˇica Sˇurina i Maja Somogyi (2006), “Biodegradable polymers for biomedical purpose”, Tekstil, Vol. 55 (2006) No. 12; 642-645.

Research register

93

IJCST 19,6

94

Ruzˇica Sˇurina i Maja Andrassy (2007), “Resistance of lignocellulosic fibers to microorganisms”, XX. hrvatski skup kemicˇara i kemijskih inzˇenjera, knjiga sazˇetaka, posvec´en Lavoslavu Ruzˇicˇki i Vladimiru Prelogu, hrvatskim nobelovcima u kemiji, Zagreb, 26. veljacˇa – 01. ozˇujka 2007., 286. Sˇurina Ruzˇica i Andrassy Maja (2007), “Quality of Modified Flax Fibers”, The 18th International DAAAM Symposium, “Intelligent Manufacturing & Automation: Focus on Creativity, Responsibility and Ethics of Engineers”, 24-27th October 2007.

Zagreb, Croatia Faculty of Textile Technology,University of Zagreb, Prilaz baruna Filipivic´a 30, HR-10 000 Zagreb, Croatia, Tel: +38514877351; Fax: +38514877357; E-mail: [email protected] Principal Investigator(s): Prof. emeritus, Ivo Soljacˇic´, Ph.D. Research Staff: Asoc. Prof. Tanja Pusˇic´, Ph.D.; Prof. Ljerka Bokic´, Ph.D.; Asst. Prof. Branka Vojnovic´, Ph.D.; Iva Rezic´, Ph.D.; Prof. Jelena Macan, Ph.D.; Asoc. Prof. Barbara Simoncˇic´, Ph.D., Prof. Sonja Sˇostar-Turk, Ph.D.; Asist.Prof. Sabina Fijan, Ph.D.; Mila Nuber, M.Sc.; Ivan Sˇimic´, M.Sc.; Dinko Pezelj, Ph.D., Versˇec Josip, M.Sc.

Ethics and Ecology in Textile Finishing and Care Other Partners: Academic

Industrial

University of Maribor and University of Ljubljana, Slovenia Labud, d.d. Zagreb and Vodovod, Zagreb Project start date: 1 January 2007 Project end date: 31 December 2011 Source of support: Ministry of Science, Education and Sports, Republic of Croatia Keywords: Wellness finishing of textiles, Determination of harmful substances on textiles, Toxicological and alergenic properties, Environmental protection, Hygiene and effects of textile care, Textile material sample preparation Modern textile finishing processes have to fulfill high demands due to the expectations of new textile materials properties and their persistence during care. Especially interesting in this respect are the new production processes of socks which include implementation of microcapsules that can release active materials for skin moisturizing. Their primal role is prevention of dryness, dandruff and allergenic reactions of the skin. The most suitable analytical methods for determination of durability to washing, friction and sweat will be tested. Durability to washing of products with special properties will be tested with different amounts of anionic and cationic surfactants in liquid detergents. The mechanism of adsorption and desorption, their influence on primary effect of the treatment, and the influence of the pH value and the mechanical way of treatment will be tested. On the ground of the obtained results, analytical methods for determination of micro components in the macro components of textile materials should be proposed, without regards to the specifications of the materials or the method of the treatment. The testing will involve a review of the analytical method of each individual analytical procedure as well as its impact on the obtained information.

The parameters of the analytical procedure will be worked out with the purpose of restoration of historical textile by destructive and non-destructive methods for the preservation of national heritage. European controlling methods of new materials have ethical demands involving the human population health which demands an environmental friendly process. For this purpose the processes of textile finishing and care will be optimized. The possibility of obtaining new preventive properties, which were not previously present on the textile material or improvement of present protection, will be tested. The impact of washing cycles with detergent and UV absorber on pastel colored textile materials made of cotton, polyester and their mixtures on UPF and the shade change will be investigated. The quality control of water and effluents will be based on the determination of micro quantities of potential allergens, heavy metals, pesticides, dyes, and surfactants. The traces of solvents will be controlled on the clothing material and in the air during the chemical cleaning and further treatment processing.

Project aims and objectives The main goal of this investigation is to stimulate ethic ecological demands on the production processes, care processes, and thereby on the utilization properties of the textile materials wherewith it would be possible to get the optimal properties of materials regarding their functional properties by avoiding all possible harmful allergenic and toxicological influences of textile materials to consumers. Elaboration of production and textile finishing processes for the optimal effects (wellness finishing, protection from unwanted changes of utilization properties in texcare, elaboration of pastel dyed textiles laundering in detergent with UV absorber, additional laundering quality – UV protection), formulation of new compositions for laundering for the purpose of avoiding secondary harmful effects in modern conditions with maximal saving of water and energy, more safe treatment with solvents during dry cleaning. By monitoring of harmful inorganic and organic substances that are present in micro quantities on the textile materials, textile accessories, textile wastewaters and finished textile products, new analytical methods would be determined. Sampling procedures, sampling preparation steps, selection of appropriate analytical method and the processing of the obtained result will be optimized. In this investigation the mathematical modes for guiding of analytical procedure will be applied, what is economically justified because the time spend for investigation is much shorter, and the consumption of chemical reagents, energy and emission of harmful substances to the environment reduced. Special contribution will be in development of analytical methods for determination of components present on the historical textile, for the purpose of avoiding the damaging of the textile material during restoration conservation treatments.

Research deliverables The project is scheduled over three years. Eco problems and human ecology, especially presence of heavy metal traces in textile processes and fibres will be investigated and some results will be published. Analytical methods for qualitative and quantitative determination will be developed. The influence of sweat on the heavy metal emission will be tested from colored textile materials.

Research register

95

IJCST 19,6

96

Possibility and durability of wellness finishing effects particularly on PA pantyhose’s as well as methods will be established. UPF and change in shade of white and pastel colored textiles made from cotton, PET, PA and their blend with cotton during laundering with addition fluorescent compounds in detergent will be researched, too. Hygienic laundering with chemothermic and chemical treatments in order to destroy micro-organisms in compliance with existing recommendations will be done. Potentially irritations of the skin caused by textiles, finishing agents and inadequate rinsing during laundering will be studied. Investigation of anionic, cationic and nonionic surfactant adsorption and desorption influenced by different composition of textile fibres, pH and temperature will be performed. The adsorption and desorption will be studied in order to establish a correlation between zeta potential and swelling capacity of textile fibres. Publications Rezic´, Iva; Steffan, Ilse (2007), “ICP-OES determination of metals present in textile materials”, Microchemical Journal. 85 (2007), 1; 46-51 (scientific paper). Fijan, Sabina; Pusˇic´, Tanja; Sˇostar-Turk, Sonja; Neral, Branko (2007), “The influence of industrial laundering of hospital textiles on the properties of cotton fabrics”, Textile Research Journal. (2007) (in publishing). Pusˇic´, Tanja; Jelicˇic´, Jasenka; Nuber, Mila; Soljacˇic´, Ivo (2007), “Istrazˇivanje sredstava za kemijsko bijeljenje u pranju”, Tekstil. Pusˇic´, Tanja; Soljacˇic´, Ivo (2007), “Changes in Shade of Cotton Fabrics during Laundering with Detergents containing Fluorescent Brightening Agent and UV absorber”, AATCC Review. Vojnovic´, Branka; Bokic´, Ljerka; Kozina, Maja; Kozina, Ana (2007), “Optimization of analytical procedure for phosphate determination in detergent powders and in loundry wastewater”, Tekstil.

Zagreb, Croatia University of Zagreb, Faculty of Textile Technology, Prilaz baruna Filipovica 30, HR-10000 Zagreb, Croatia, Tel: +385 1 37 12 566; Fax: +385 1 37 12 599; E-mail: [email protected] Principal Investigator(s): Prof. Maja Andrassy, Ph.D. Research Staff: Prof. Zvonko Dragcevic, Ph.D.; Assoc.Prof. Emira Pezelj, Ph.D.; Prof. Dubravka Raffaelli, Ph.D.; Assist. Prof. Edita Vujasinovic, Ph.D.; Zvonko Orehovec, Ph.D.; Vera Friscic, M.Sc.; Ruzica Surina, B.Sc.; Prof. Majda Sfiligoj Smole, Ph.D.

High performance textile materials and added-value fibres Other Partners: Academic

Industrial

None University of Maribor Faculty of Mechanical Engineering, Maribor, Slovenia Project start date: 1 January 2007 Project end date: 31 December 2011 Source of support: Ministry of Science, Education and Sport, Republic of Croatia

Keywords: HP Materials, Textile Fibers, Textile Design Contemporary global trends of development in the field of textile fibres and fibrous materials have led to their increased use in various fields of industry and technique. Increase in consupmtion of these types of materials has constantly been recorded and by the beginning of the 21st century technical fibres account for half of all the fibres manufactured. The requirements imposed on fibres and materials in particular fields of application and extraordinary high and specific. These requirements have been met through fibre engineering, i.e. development of new generic types of fibres. It can be assumed that innovative manufacturing and finishing processes, applied to conventional fibres, will result in their added value, so that they can be used to design new fabrics of pre-determined end-use properties. Such improvements in fibre properties and their use in fabric manufacture of added market value and broader scope of application are completely in accordance with the intentions of the European technological platform for future textile and garment, but it can also strongly stimulate the development of Croatian textile industry and its comptetitiveness in the global market. This is supported by the fact that there are considerable research, industrial and raw-material potentials in Croatia, necessary to accomplish the goals. Although domestic production is mostly based on imported fibres, clearly defined modifications of fibre structure and propertties, even for domestic fibres, such as wool, flax, textile regenerates and fibres made from recycled PET, that have been used in Croatian textile industry insufficiently until now, could be used as a starting raw material for the manufacture of high-performance textiles. In this manner, domestic fibrous raw materials would cease being waste material and would become strategic Croatian raw material, as well as a basis of future rational management of natural resources and a step in approaching sustainable development trends, recommended by the European Union and United Nations. The investigations proposed aim at establishing the possibilities of modifying conventional fibres, as well as developing the methods and procedures of objective measurement and evaluation of unconventional textile materials, in accordance with specific rules and requirements for individual types of high-performance textiles, including composites reinforced with fibres of modified properties. The results obtained will offer the construction of high-performance materials based on conventional fibres of added value, as well as the design and optimisation in accordance with the properties of the fibres used and pre-determined high end-use properties.

Project aims and objectives The main purpose of the project proposed is to determine possible interventions and modifications of conventional fibres and textiles, so as to obtain added value and to broaden the scope of their application. This immrovement of fibres and their application in the manufacture of new, knowledge-based innovative textiles of added market value is in accordance with the short term (energy and materials) and long-term (nano-science, new materials, constructions and production processes) strategic trends of research in the Republic of Croatia, strategies of the European technological platform for the future of textiles and garment in the XXI century, as well as with the trends of rational management of raw material resources and the concept of sustainable development, as

Research register

97

IJCST 19,6

98

proposed by the EU and UN recommendations. Valuable results and new knowledge are expected, especially regarding ecologically friendly and economically feasible production of high-performance textiles through the usage of domestic raw materials in their manufacture. There is a broad diversity of fibres and constructions present in the area of non-conventional textile materials and structures, which makes objective characterisation of their quality a difficult task, we propose to develop new methods, procedures and equipment for testing, so as to enable higher degree of objectivity in quality evaluation. The investigations are planned to initiate the development and optimising of the manufacture of high-performance textiles, matched with their increased and more and more specified areas of application. The results expected to be obtained will enhance the scope of knowledge in the field of textile fibres, materials and textile testing, also valuable in the education of young researchers, knowledge transfer and preparation of future textile engineers for the European labour market.

Research deliverables The results of the investigation will be directly applicable in Croatian textile industry, since new solutions for designing high-performance materials, based on conventional fibres of added value, will be proposed as based on the results of the investigations proposed. As there are some processing capacities still in Croatia (Regeneracija, Kelteks, Vrbenka, Konoplja, LIO, Feniks), working with imported fibres, the system of objective description and evaluation of domestic fibrous raw materials (especially flax and wool), as well as some instructions regarding their environmentally acceptable manufacture, processing and modifications, with the aim of enhancing end-use properties, are expected to create adequate conditions for econbomically feasible manufacture of textiles. Higher content of domestic fibres and raw materials in manufacture of textile would be a sound basis for new development and growth of the Croatian textile industry and its competitiveness in the European and global markets, as well as for realising the principles of sustainable development and rational raw material resource management. Developments and innovations in testing metodology and objective evaluation of relenat properties of the modified fibres and new high-performance textiles will enhance objectivity of testing and evaluation of high-performance technical textiles in general, which is, not only in Croatia but globally as well, a problem with no acceptable solution on the horizon. Some testing methods are expected to be used in production monitoring and control, which could contribute to more reliable and stable manufacture and realising pre-planned levels of quality. The investigations proposed open the way to scientific and professional collaboration with other institutions and with the industry. Publications Ruzˇica Sˇurina i Maja Somogyi (2006), “Biodegradable polymers for biomedical purpose”, Tekstil Vol. 55 (2006) 12; 642-645. Ruzˇica Sˇurina i Maja Andrassy (2007), “Resistance of lignocellulosic fibers to microorganisms”, XX. hrvatski skup kemicˇara i kemijskih inzˇenjera, knjiga sazˇetaka, posvec´en Lavoslavu Ruzˇicˇki i Vladimiru Prelogu, hrvatskim nobelovcima u kemiji, Zagreb, 26. veljacˇa – 01. ozˇujka 2007., 286. Cindric´, Jasna (2007), Improvements Properties of Flax Fibers, diploma work, Zagreb, Tekstilnotehnolosˇki fakultet, 25.04. 2007., 56 str. Voditelj: Andrassy, Maja. Klasic´, Sanja (2007), Usable Properties of Modified Flax Fabric, diploma work, Zagreb, Tekstilnotehnolosˇki fakultet, 25.04. 2007., 53 str. Voditelj: Andrassy, Maja.

Sˇurina Ruzˇica i Andrassy Maja (2007): “Quality of Modified Flax Fibers”, The 18th International DAAAM Symposium, “Intelligent Manufacturing & Automation: Focus on Creativity, Responsibility and Ethics of Engineers”, 24-27th October 2007 (in press). E. Vujasinovic, Z. Jankovic, Z. Dragcevic, I. Petrunic, D. Rogale (2007), “Investigation of the strength of ultrasonically welded sails”, International Journal of Clothing Science and Technology Vol. 19(2007) No. 3/4, 204-214, ISSN: 0955-6222. E. Vujasinovic, Z. Dragcevic, Z. Bezic (2007), “Descriptors For The Objective Evaluation Of Sailcloth Weather Resistance”, Proceedings of 7th Autex Conference 2007, Tampere 26-28th June 2007, Finland; ISBN: 978-952-15-1794-5.

Zagreb, Croatia Faculty of Textile Technology, University of Zagreb, Prilaz baruna Filipivic´a 30, 10 000 Zagreb, Croatia, Tel: +385 1 48 77 357; Fax: +385 1 48 77 357; E-mail: [email protected] Principal Investigator(s): Assoc. Prof. Sandra Bischof Vukusˇic´, Ph.D. Research Staff: Prof. Drago Katovic´, Ph.D.; Assoc. Prof. Tanja Pusˇic´, Ph.D.; Prof. Emeritus Ivo Soljacˇic´, Ph.D.; Sandra Flincˇec Grgac, B.Sc.

Antimicrobial finishing of textiles Other Partners: Academic

Industrial

University of Maribor, Zagreb Public Health Institute, Croatian National Institute of Public Health

Jadran, Hoisery Factory d.d., www.jadrancarapa.hr, Pamucˇna Industrija Duga Resa, d.d. www.pamucna-industrija.com

Project start date: 1 Janua;ry2005 Project end date: 1 January 2008 Source of support: Ministry of Science, Education and Sports, Republic of Croatia Keywords: Antimicrobial finishing, Antimicrobial agents, Testing methods for antibacterial activity, Testing methods for antifungal activity, Legislative regulations Protection of the environment and human health is now perceived as more important than the profitability and efficiency of a business. Antimicrobial finishing is used on textile fabrics to control bacteria, fungi, mold, mildew and algae present in everyday life, to resolve problems of deterioration, staining, odors and health concerns they cause. It has been long ago recognized that microorganisms, particularly bacteria can negatively influence textile fabrics, especially cotton ones. Most textiles used as clothes, underwear and bedclothes in hospitals and hotels are conducive to cross infection or transmission of diseases caused by microorganisms. During the project time alternative agents for the purpose of antimicrobial treatment will be developed. Two parallel investigations will be performed. Croatian partners will be developing system with polycarboxylic acids. Slovenian partners will be working on grafting of cyclodextrine on textile substrates. Their bacteriostatic activity will be determined using standard methods: JIS L 1902: 2002 and AATCC 147, before and after the washing process. The antimicrobial activity will be tested against gram positive (Staphylococcus aureus) and gram negative

Research register

99

IJCST 19,6

100

(Klebsiella pneumoniae) bacteria. Additional tests would include Escherichia coli, which might cause urinary infection that is of particular interest for underwear knitted fabrics. In order to prevent this negative microbial influence knitted material will be treated with several agents of different producers in order to choose the best one and recommend it to the industrial partners. For efficiency control of the treated products, strict conditions of the microbiological laboratory and qualified staff are necessary. For that reason cooperation with Zagreb Public Health Institute and Croatian National Institute of Public Health will be conducted within the project. Further experiments include development of alternative method for antimicrobial treatment, such as microwave treatment. Microwave effectiveness for the purpose of antimicrobial treatment will be determined.

Project aims and objectives Project aim: .

implementation of optimal antimicrobial treatments to Croatian and Slovenian market.

Project objectives: .

antibacterial and antifungal efficiency control of textiles treated with several antimicrobial agents, present at the market;

.

development of novel antimicrobial systems based on polycarboxylic acids; and

.

investigation of microwave effectiveness for the purpose of antimicrobial functionalisation.

Research deliverables .

Report ordered from SME “Pamucˇna industrija Duga Resa, d.d.”: Antimicrobial hygiene finishes, Usage guidelines.

.

Report ordered from SME “Jadran Hoisery Factory d.d.”.: Antimicrobial functionalisation of socks.

Publications Bischof Vukusˇic´, S., Pusˇic´, T., Katovic´, D., Soljacˇic´, I. (2005), “Antimicrobial finishing of Socks”, 36. Simpozij o novostih v tekstilstvu, Tekstilije za sˇport in prosti cˇas, Ljubljana, Slovenia, ISBN 961-604530-X, pp. 163-177. Flincˇec Grgac, S., Bischof Vukusˇic´, S., Katovic´, D., Matica, B., Dragosˇa, M. (2006a), “Antimicrobial Functionalisation of Knitted Fabrics”, Rewiew 2006, Published papers of Zagreb Public Health Institute, ISBN 953-6998-30-0, pp 147. Flincˇec Grgac, S., Bischof Vukusˇic´, S., Katovic´, D., Matica, B., Dragosˇa, M. (2006b), “Antimicrobial Functionalisation of Knitted Fabrics”, 3rd International Textile, Clothing & Design Conference, Dubrovnik, Croatia, ISBN 953-7105-12-1, pp. 270-275. Bischof Vukusˇic´, S., Flincˇec Grgac, S., Katovic´, D. (2006c), “Antibacterial & Antifungal Protection of Military Socks”, 3rd International Textile, Clothing & Design Conference, Dubrovnik, Croatia, ISBN 953-7105-12-1, pp. 241-246. Bischof Vukusˇic´, S., Flinecˇ Grgac, S., Katovic´, D. (2007), “Antimicrobial Textile Treatment and Problems of Testing Methods”, Tekstil, Vol. 56, accepted for publication.

Zagreb, Croatia Faculty of Textile Technology, University of Zagreb, Prilaz baruna Filipovic´a 30, HR-10000 Zagreb, Croatia, Tel: +385 1 3712575; Fax: +385 1 3712599; E-mail: [email protected] Principal Investigator(s): Assoc. Prof. Stana Kovacˇevic´, Ph.D. Research Staff: Assist. Prof. Zˇeljko Penava, Ph.D.; Josip Hadina, M.Sc.; Ivana Schwarz, B.Sc. Assist. Prof. Andrea Pavetic´, Nikol Margetic´, B.Sc.; Irena Sˇabaric´, B.Sc.; Valent Strmecˇki, M.Sc.; Dubravka Gordosˇ, M.Sc.; Biserka Vuljanic´, M.Sc.; Prof. Vladimir Oresˇkovic´, Ph.D., in retirement; Assoc. Prof. Krste Dimitrovski, Ph.D., Blago Brkic´, Ph.D. in retirement, Diana Franulic´ Sˇaric´, M.Sc.

Advanced Technical Woven Fabrics and Processes Other Partners: Academic

Industrial

Varteks d.d., Varazˇdin; Cˇateks d.d., Cˇ akovec; Kelteks d.o.o., Karlovac; Lola Ribar d.d., Karlovac; Uriho, Zagreb; Peng d.o.o., Zagreb Project start date: 1 January 2007 Project end date: 31 December 2009 Source of support: Croatian Ministry of Science, Education and Sport, Republic of Croatia Keywords: Technical woven fabrics, Composites, Wool, Flax, Tapestry, Ethno heritage

University of Ljubljana, Slovenia

The subject of this project is advanced technical woven fabrics and processes. They are intended for the use in interior decoration, transportation, industrial and medical purposes, tapestry and the like. These fabrics contain raw materials in common, and domestic wool and linen yarn as well as glass and carbon yarn will be preferred, but other natural raw materials will be used too, yarns of chemical fibers from synthetic polymers and “smart” yarns. The aim of this research is to find the most optimal raw material and fabric construction and to make a commercially acceptable, qualitative, healthy, comfortable and smart technical fabric. Basic investigations will include: physical-mechanical, thermal, relaxation and elongation properties, dimensional stability, abrasion, effect of sun rays, inflammability, air permeability, water repellency, degradation and the investigation of these properties depending on fabric application. The scope of investigation will include technical fabrics intended for use in civil engineering, transportation and household (3D fabrics for composites, fabrics for seat covers, furnishing fabrics etc.) on which high requirements are set, such as: safety, resistance, comfort and aesthetics. Technical fabrics for industrial purposes such as filter fabrics and fabrics for composites which are of great importance for better utilization and productivity, and still more important in terms of ecological protection of environment, will be investigated. Healthy fabrics in medical terms from natural raw, and generally fabrics with various properties and applications subjected to additional treatments according to health standards. Part of the project will be directed at the investigation and revival of Croatian eco and ethno heritage, including

Research register

101

IJCST 19,6

102

the manufacture of tapestry, blankets and mats interwoven with art and skill of weaving, using domestic raw materials. The aim of this research is that tapestry authenticity and originality of work of art represent a unique value. The significance of this project is to revive the processing of domestic wool and flax in parallel with the investigation of new constructions and forms of glass technical fabrics, and new materials processed by new technologies. Several technical fabrics replicated in this project will serve as an encouragement for processing domestic wool and flax in smaller batches in karts regions of Croatia.

Project aims and objectives .

.

.

To design and manufacture technical fabrics from natural and new materials, and with their combination, with new methods and manufacturing procedures. To investigate qualitative properties of technical fabrics from natural textile raw materials and to promote the use of primarily domestic wool and linen. Within the basic aim it is necessary to define a numerical model of the manufacturing process and development of new composites so that chemical and physical processes affecting the quality and applicability of composites are investigated and analyzed experimentally. Thereby the influence of particular processes (manufacturing processes, temperatures, outside influences) on the changes of composite properties as well as the interrelationships will be determined.

.

The knowledge gained will be used to define a computer model for the construction of new types of technical fabrics and accordingly new composite materials.

.

To systematize utility and quality values of wool yarn from domestic sheep breeds and to investigate the possibility of the industrial process in blending with other raw materials. To optimize the production of applicable products of domestic raw materials and to promote the revival of the textile ethno heritage in Croatia.

.

.

To investigate the most cost-effective process and the most optimal shares of individual raw materials for the target product with new properties.

.

To include other raw materials with domestic materials with the aim of higherquality, more valuable and attractive products.

.

To manufacture healthy, high-quality, “smart”, comfortable, unique and commercially acceptable textile materials by integrating “smart” materials and new materials dyed with natural dyes.

Research deliverables Manufacture of technical fabrics from different composites with improved properties such as: more healthy (natural fibers), more comfortable (better thermal properties), and more attractive (uneven surface structure made by weaving various densities, fineness, colors and raw materials with different weaving techniques and different finishing treatments). It may be expected that the results obtained will make a

considerable contribution to further development of manufacturing technical fabrics and composites based on them, and that their application range will be expanded. Manufacture of medical fabrics based on natural raw materials and new composites which will be used as part of orthopedic and surgical aids that will dispose of improved properties such as: strength, thermal properties, liquid absorption, elasticity, stability and other properties. Manufacture of tapestries, blankets and mats with natural and new materials dyed with natural dyes whereby modern, high-quality and unique products can be produced. Moreover, by mastering these skills the possibility of a quality restoration of the textile ethno heritage will be provided. By applying the results obtained in various economic branches the development of new construction materials as a substitution for classic materials will be possible which will make a direct contribution environmental protection and entire ecological development of the country. By realizing such results the project will make a further contribution of the development of science in the field of developing new materials which according to long-term strategic directions of research in Croatia. They emphasize the need for the investigation of new materials, constructions and manufacturing processes. At the same time, the project is on the track of short-term strategic directions of scientific researches which direct investigations at natural, glass and organic-inorganic hybrids, making intelligent materials and polymer research. Publications Kovacˇevic´, S.; Schwarz, I. (2007), “Hand weaving – tradition of the future”, 7th Annual Textile Conference by Autex: From Emerging Innovations to Global Business, 26.-28.06.2007., Tampere, Finland. Schwarz, I,; Flincˇec Grgac, S.; Kovacˇevic´, S.; Katovic´, D. & Bischof Vukusˇic´, S. (2007) “The efect of drying methods on sized yarn characteristics”, The 18th International DAAAM symposium - Intelligent Manufacturing & Automation: Focus on Creativity, Responsibility, and Ethics of Engineers, 24-27th October 2007, Zadar, Croatia / editor B. Katalinic. Vienna: DAAAM International, 2007. (in press). Ujevic´, D.; Kovacˇevic´, S,; Schwarz, I.; Brlobasˇic´ Sˇajatovic´, B. (2007), “Novi visˇeslojni tekstilni plosˇni proizvodi”, 6th International Scientific Conference on Production Engineering: Development and modernization of production, October 24th - 26th 2007. / edited by: I.Karabegovic, V. Dolecek, M. Jurkovic, RIM 2007. (in press). Kovacˇevic´, S.; Ujevic´, D.; Schwarz, I.; Brlobasˇic´ Sˇajatovic´, B.; Brnada, S. (2007), “Analysis of Motor Vehicle Fabrics” // Fibres & Textiles in Eastern Europe, 2007 (in press).

Zagreb, Croatia Faculty of Textile Technology University of Zagreb, Prilaz baruna Filipovic´a 30, HR-10000 Zagreb, Croatia, Tel: +385 1 37 12 552; Fax: +385 1 37 12 599; E-mail: [email protected] Principal Investigator(s): Assist. Prof. Zˇeljko Sˇomodi, Ph.D. Research Staff: Assist. Prof. Ana Kunsˇtek, Ph.D.; Slavica Bogovic´, M.Sc.; Anica Hursa, M.Sc.; Igor Petrunic´, M.Sc.; Assist. Prof. Simona Jevsˇnik, Ph.D.; Daniela Zavec-Pavlinic´, Ph.D.

Research register

103

IJCST 19,6

Computational Modelling in Engineering Analysis of Textiles and Garment Other Partners: Academic

Industrial Kamensko d.d., Zagreb

104

University of Maribor, Faculty of Mechanical Engineering, Maribor, Slovenia Project start date: 1 January 2007 Project end date: 31 December 2011 Source of support: Ministry of Science, Education and Sports, Republic of Croatia Keywords: Clothing technology, Numerical methods, Optimisation, Reinforcements on clothing The intention of this project is to give a contribution to advanced application of the methods of engineering analysis in the field of textiles and garment. This goal will be achieved by introduction, adaptation, elaboration and application of up-todate computational methods in the analyses of problems relevant for the field of textile and clothing engineering. Considering the existing experience and an overview of questions and problems actual for the engineering science in the field, the research is to be undertaken in a number of areas, such as: optimal design of structural reinforcements in garment based on the finite element analysis; three parameter model of tensile nonlinearity of textiles; computational evaluation of post-buckling stable state in prediction and simulation of fabric drape; general numerical solution of thin plate bending with application to optimal grip geometry in automated work piece manipulation. Depending on the timing and realization of these researches, there is a possibility of opening further research areas from the field of computational modelling in mechanics of textiles and garment, including the spatial modelling and design of clothing items. The methods of research to be applied primarily consist of derivation and elaboration of numerical models suitable for application in the problems under consideration, and the development and application of computer programmes based on these models. At the same time, the plan is to acquire and apply some of the existing software applicable in the problems to be considered, as well as to prepare and conduct experimental verification of results obtained by computations.

Project aims and objectives Aim and scope of the proposed research is to give a contribution in the improvement of the level of engineering and technological know-how in the field of textile and garment. The research is expected to result in computer programmes or engineering data collected in tables, diagrams etc. that will be useful for the problem solution in the area of expertise covered by the research. The knowledge and methods developed by the research will be on offer for the interested subjects, from Croatia or elsewhere, primarily from the branch of textiles and garment. It can also be expected that after some time the collected knowledge and methods will be included in the teaching process at the Faculty of Textile Technology, primarily as parts of the subjects at the doctoral or diploma levels.

Research deliverables The principal user of the results will be Faculty of Textile Technology, as the institution for production and transfer of knowledge in the field of textile and garment. Further users will be the firms, institutions and individuals from among the designers and manufacturers of textile and garment, who already have the co-operation with the faculty, or shall have that co-operation in the future. The results and findings of the research shall be offered to these subjects by means of the Centre for development and transfer of textile and clothing technology and fashion design, as a unit in the structure of the faculty. The specific applications are expected in the expert analyses related to engineering in preparation of production processes in which the problems from the field of research appear. Publications Hursa, Anica; Rogale, Dubravko; Sˇomodi, Zˇeljko (2006), “Application of Numerical Methods in the Textile and Clothing Technology”, Tekstil, Vol. 55 (2006) No. 12, 613-623. Akrap-Kotevski, Visˇnja; Kunsˇtek, Ana (2007), “Writing Ability after Brain Damage”, Proceedings of 3rd International Ergonomics Conference, Ergonomics 2007, Mijovic´, Budimir (Ed.), Zagreb, Croatian Society of Ergonomics, 2007, 279-285. Sˇomodi, Zˇeljko; Hursa, Anica; Rolich, Tomislav; Rogale, Dubravko (2007), “Numerical analysis and optimisation of mechanical reinforcement on clothing”, Book of Proceedings of the 1st Meeting of Croatian Society of Mechanics, Cˇanadija, Marko (Ed.), Rijeka, Croatian Society of Mechanics, 2007, 173-178. Sˇomodi, Zˇeljko; Hursa, Anica; Rogale, Dubravko (2007), “A minimisation algorithm with application to optimal design of reinforcements in textiles and garments”, Internationl Journal of Clothing Science and Technology, Vol. 19 (2007) No. 3/4, 159-166.

Zagreb, Croatia Faculty of Textile Technology, Prilaz baruna Filipovica 30, 10 000 Zagreb, Croatia, Tel: +385 1 37 12 577; Fax: +385 1 37 12 533; E-mail: [email protected] Principal Investigator(s): Prof. Zenun Skenderi, Ph.D. Research Staff: Prof. Miroslav Srdjak, Ph.D.; Prof. Momir Nikolic´, Ph.D.; Prof. Alka Mihelic´-Bogdanic´, Ph.D.; Bozˇo Tomic´, M.Sc.; Vesna Marija Potocˇic´ Matkovic´, M.Sc.; Ivana Salopek, M.Sc.; Dragana Kopitar, B.Sc.

Multifunctional technical nonwoven and knitted textiles, composites and yarns Other Partners: Academic

Industrial

Cˇateks d.d. Cˇakovec, Croatia, Regeneracija non-woven and carpets j.s.c., Zabok, Croatia Project start date: 1 January 2007 Project end date: 01c2010 Source of support: Ministry of Science, Education and Sport, Republic of Croatia

None

Research register

105

IJCST 19,6

106

Keywords: Yarns, Knitted and nonwoven fabrics, Structures, Properties, Thermal and vater-vapour resistance properties Further dislocation of the textile production from deveoped countries into Asia is a basic characteristic for the world textile industry today. In the field of technical textiles profound resistance is felt against relocation. An increase in the production of technical textiles is recorded due to a permanent expansion of the application range. It is used in: transportation, industry, medicine, hygiene, household, garment industry, agriculture, fishing trade, civil engineering, sport, safety, ecology etc. Nonwovens make the most significant contribution to the development of technical textiles. Over last decades the technology of nonwovens production has experienced a rapid development, and the production of late years has registered an annual increase of approx. 10 per cent. A significant application range for technical textiles or geotextiles is civil engineering, in particular road building. In addition to woven, knitted and similar structures, nonwovens play a predominant role with a share of approx. 75 per cent in 2005. The most important functions of geotextiles are: separation of weak soil, reinforcement of soil or elements of building structures, filtration and drainage. Geotextile properties are: stability, uniform structure, small thickness, high strength and stretching, porosity, small surface mass and water permeability. Various applications require a more or less marked particular structure and characteristic. The first part of the project will deal with various structures and properties of technical textiles based on nonwoven and knitted structures, in particular on geotextiles. Moreover, manufacturing technologies of technical textiles and knitted materials as well as their controlling parameters will be discussed. Conventional technologies such as: spinning, weaving, knitting and clothing technology will probably not withstand the competitiveness coming from Asia. Besides, relocation of the manufacture of man-made fibres into the Far East is taking place. It is undoubtedly the case that only those disposing of raw materials and enough knowledge to produce and sell high-quality products will have the chances of survival on the market. The investigation of possibilities of manufacturing from coarser sorts of wool which have similar fineness as domestic wool and the investigation of their possible use for products such as carpets and several articles of clothing will be within the scope of this project. The limit of fibre spinnability, typical stress-strain curves, yarn behavior in cyclic examinations of elongation properties, surface friction and yarn hairiness.

Project aims and objectives There are two dominant reasons why the field of technical textiles is dealt with in the project: . Intensive development of technical textiles because of an increase in the application (technical textiles on average of approx. 5.5 per cent, nonwoven fabric approx. 10 per cent). . Resistance to the relocation of the manufacture of technical textiles to the Far East. As a result the manufacture of technical textiles (quantity) in the developed countries accounts for more than 40 per cent of the total production of textiles. The state of the Croatian textile industry should be noted where the classic textile industry has

almost vanished. Two companies Regeneracija and Cˇateks manufacture technical textiles which have their markets. Regeneracija and Cˇateks confirmed their participation in the project. Moreover, the use of domestic wool for various products of higher value is a chalange and obligation. However, today domestic wool is sold as a raw material, and in some regions it is not bought off which is an ecological problem. Based on the above mentioned facts, it is reasonable to deal with the subject matter of technical textiles and yarns and products respectively, such as carpets within the scope of the project. In this way production of higher-quality products is promoted which is the purpose of this investigation. The aims of the investigation are as follows: .

Definition of the interdependence, process parameters and physical-mechanical as wel as other relevant properties.

.

technical nonwoven textile. technical textile based on nonwoven fabric coated with polyurethane (PUR).

. .

technical textile based on knitted fabric from PET and PA coated with polyurethane (PUR) and other technical textiles based on knitted fabric.

.

Definition of the interdependence of raw materials, process parameters and physical-mechanical yarn properties, primarily wool, and the behavior of carpets in dynamic investigations (new instrument required for the project).

.

It will be attempted in case of interest of Regeneracija or other interested parties to investigate thermal properties of wool insulation materials.

Research deliverables Obtaining the new understandings of: . .

.

.

.

.

Thermal and vapor resistance properties of the knitted and nonwoven fabrics. Spun yarns, primarily coarser wool yarns: procedures of manufacturing and defined the controlling parameters of the processes. Structure and properties of the coarser, industrially manufactured wool yarns should be emphasized that can be similar to the yarn spun from domestic wool as far as their properties. Influence of fibres parameteres (finenness, length,.) on limit of spinnability and spun yarn properties. Possibilities of manufacturing carpets from coarser wool fibres will be investigate as well as compressibility of carpets on the new instrument purchased from funds of the project. Processes of manufacturing: technical nonwoven fabrics intended for use in civil engineering, technical nonwoven fabric coated with polyurethane (PUR), technical textiles based on knitted fabric coated with polyurethane (PUR), as well as their structures and properties.

Publications Salopek, Ivana; Skenderi, Zenun; Srdjak, Miroslav (2007), “Stoffgriff - ein Aspekt des Tragekomforts von Strickware”, Melliand Textilberichte. Vol. 88 (2007.); 426-428.

Research register

107

IJCST 19,6

108

Salopek, Ivana; Skenderi, Zenun (2007), “Thermophysiological comfort of knitted fabrics in moderate and hot environment“, Proceedings of the 3rd International Ergonomics Conference / Mijovic´, Budimir (ur.). Zagreb: Croatian Society of Ergonomics, 2007. 287-293. Potocˇic´ Matkovic´, Vesna Marija; Salopek, Ivana (2007), “Computer Assisted Study of Knitted Structures”, Proceedings Vol. IV. CE Computers in Education. 2007. 99-102. Kopitar, Dragana; Skenderi, Zenun (2006), “Prsteni i trkacˇi - glavni elementi prstenaste predilice”, Tekstil, Vol. 55 (2006); 543-501.

Zagreb, Croatia Faculty of Textile Technology University of Zagreb, Prilaz baruna Filipovic´a 30, HR - 10 000 Zagreb, Croatia, Tel: + 385 1 37 12 500; Fax: + 385 1 37 12 595; E-mail: [email protected] Principal Investigator(s): Assoc. Prof. Zlatko Vrljicak, Ph.D. Research Staff: Kresimir Hajdarovic, Ph.D.; Valent Strmecki, M.Sc.; Tomislav Koren, M.Sc.; Ivan Basnec, M.Sc.

Design and manufacture of nets for the protection of fruit and vegetables against hail Other Partners: Academic

Industrial

-

Tvornica mreza i ambalaze, Biograd n/m, Croatia Project start date: 1 January 2007 Project end date: 31 December 2009 Source of support: Ministry of Science, Education and Sport, Republic of Croatia Keywords: Fruit production, Nets, Hail, Viticulture, Plant material, Economic profit Approximately 7 per cent of Croatian fruit is present on the Croatian market. The production and sales of Croatian fruit can be quadrupled and sold at present prices on the Croatian market, but as first class fruit. Over the last ten years The Ministry of Finance received damage reports worth more than 100 million kuna which were caused by hail. Fruit, vegetables, plants, flowers, nursery-gardens, and animals, material resources: houses, agricultural machinery, automobiles and the like get damaged. Up to now rockets have been used to ensure hail protection of fruit and vegetables. On account of a rapid increase in the volume of air transport this technique is less used and is substituted by using protection nets. Several more developed and neighboring European countries have started using nets for the protection of fruit and vegetables against hail. Within the scope of this project systems of applying protection nets in European countries and their use in Croatia would be studied. The emphasis here would be on the safety of orchards, new plants or crops and how to pay compensation for damages by hail. Appropriate protection nets would be designed and manufactured for particular agricultural products and then installed on plantations. Net construction depends on the application so that protection nets of various widths, shapes, colors and structures with special emphasis on the raw material for the production of nets and for shadowing the area to be covered. Across Croatia nets would be offered to the registered fruit growers

for use. During the first year of the project a fruit grower would be offered 1000 m2 of nets without charge with the aim that he buys the same quantity (ratio 1:1) and that he should cover only one part of his plantation. By continuous monitoring orchards all changes under the nets would be analyzed and then compared to the results obtained outside the nets. Over the period of five years relevant conclusions about the use of nets for the protection of fruit and vegetables against hail can be made. It is to be emphasized that the nets, which protect agricultural products against hail, can protect against sun, birds, animals etc. By adequate use of the above mentioned nets it is to expect that yield per hectar will be increased as well as fruit quality. Up to now our fruit growers have collected less than 50 per cent of first class fruit, but using nets it can be expected to increase this limit over 80 per cent. In this way, when Croatia enters into the European Union, we can sell our quality fruits on our market and compete on international markets.

Project aims and objectives The purpose of the investigation is to provide assistance to fruit growers to protect their orchards, both trees and fruits against hail. The aim of the project is to design and make nets that will be helpful and commercially acceptable in the protection of fruits and vegetables against hail. Manufactured nets would be offered to fruit growers for use. During the duration of the project several types of nets for the protection of fruit and vegetables would be developed. The application aim of these protection nets is primarily to protect trees and fruits. In this way fruit growers would increase the quality of collected fruit and yield quantity per unit area. In the long term, when Croatia enters into the European Union, our fruit growers will be able to sell their quality fruits on the Croatian market and compete on the international markets. If we do not do it, the European fruit producers will sell their fruits on the Croatian market without competition. The Croatian fruit growers will not be able to enter into this market because of minor fruit quality. This action will prevent the breakthrough of export fruit on the Croatian market

Research deliverables he whole plan and protocol is mentioned in section 9.2. Therefore only basic characteristics are mentioned here. After 1st year more than 20000 m2 of nets will be installed on different fruit plantations across Croatia. We would collect data on necessary properties of nets for the application in a particular sector or for a particular orchard. Based on the gathered data we would design and make new nets with different physical-mechanical properties. We would come to basic data on the influence of sun beams and weathering on the changes in net characteristics. We would publish approximately five papers. During previous three years we would keep records of all important influences on the fruit quantity and quality. We should come to data on the influence of hail, sun, wind, thrash, birds, animals and the like on the fruit quality and quantity. We would publish a newsletter about the application of protection nets in the cultivation of fruit and vegetables. Publications Not available

Research register

109

Research index by institution

IJCST 19,6

110

Institution

Page

Budapest University of Technology and Economics

10-21

Dokuz Eylul University

41

Heriot-Watt University

22-37

Loughborough University

45-48

Manchester Metropolitan University

48-50

National Institute of Technology Jalandhar

42-44

National Institute for Textiles and Leather

10-13

Rapra Technology Ltd

69-71

Sci-Tex, Japan

56-57

Technical University of Lodz

44-45

Technological Education Institute of Piraeus

8

University of Bolton

9-10

University of Dundee

21-22

University of Huddersfield

37-41

University of Maribor

50-55

University of Nottingham

55-56

University of Patras

57-60

University of Pisa

60-69

University of Zagreb

71-109

Research index by country Country

Page

Croatia

71-109

Greece

8, 57-60

Hungary

13-21

India

42-44

Italy

60-69

Japan

56-57

Poland

44-45

Romania

10-13

Slovenia

50-55

Turkey

41

United Kingdom

9-10, 21-41, 45-50, 55-56, 69-71

Index by country

111

Research index by subject

IJCST 19,6

112

Anthropometrics, garment sizing

15, 16, 34, 35, 47, 78

velvet warp knitted woven

41 41 41

Fibre extrusion, weaving, yarn production

23

Heritage

101 40

Bio medical, antimicrobial agents, antimicrobial legislation clothing, embedded electronics, intelligent biomedical clothes IBC

29 99

Ceramic coating, sol gel process

91

Ink jet printing, colour measurement

Chemical modification

13

Medical textiles

Clothing, comfort, prediction engineering, textile materials health monitoring heat and mass transfer, clothing energy, wind fabric lockstitch machine formation, sewing thread, take up medical clothing protective clothing robotic sewing technology, garment modelling, reinforcement, optimisation Colour matching Design Distance learning Environmental protection, harmfull textiles, textile care, textile hygiene, toxicology and wastewaters antifungal activity renewable sources Fabrics, automotive seat knitted mechanics non-woven blanket powernet properties

65

50, 90 50 63 46 45 63, 65, 91 65, 84 57 103 57 37, 90, 96 8 81

99 69 30, 32 41 48, 105 50 42, 43, 105 24 86, 105

antimicrobial textiles antimicrobial finishing bandages bio-sensing compression therapy controlled delivery leg ulcers pressure garment, measurement, hypertrophic burn scars, powernet fabric

13, 29, 32, 74, 86, 94, 99 13 99 9 63 9 29 9 24

Modelling, 34, 55 14, 16 3D design, human body modelling, visual robot, clothing simulation, 3D drape tester 3D modelling, workplace, 75 work environment, VR CFD, CAE, protective fabric, 36 garment, modelling high performance knowledge propagation 56 unit cell analysis 55 Nanotechnology, 81, 65 25, 28, 29 electrospinning, polymers, nanotechnology, nanofibres, nano yarn, alignment, nanocomposite micro-channel, micro-porous, 29 membranes, controlled delivery, drug release

Objective evaluation, fabric mechanics

50, 57, 90

Smart textiles, 91 electroform, electroplate fibre, silver 21 intelligent garments 86 intelligent sick beds 86 intelligent textiles 90 interactive textiles, garments, 30 wireless smart technology protective textiles, digital printing, 32 medical textiles solar energy, thin films, 22 photovoltaics 71 Surface modification and finishing Technical textiles 32, 57, 86, 90, 99, 101, 108 barrier materials 60 composite materials, 16, 71 engineering materials

dust particle, pulse-jet filtration 43 fibres, fibrous structures, 57, 16 reinforcements, composites fruit production nets, anti-hail nets 108 high performance textiles 33, 96 13, 60, 71, multi-functional textiles, 91 chemical modification, microwave treatment of cellulosics powernet fabric 24 Technology transfer

11

Testing, multiaxial testing thermal properties, water-vapour resistance, fabric properties

16 86 105

Wireless communications

30, 63

Index by subject

113

IJCST 19,6

114

Research index by principal investigator Principal investigator Ambrus, G. Anand, S.C. Andrassy, M. Aspragathos, N. Bo¨di, B. Borsa, J. Brooks, R. Ceken, F. Christie, R.M. Clifford, M.J. De Rossi, D. Dermatas, E. Fotheringham, A. Frydrych, I. Gersˇak, J. Ghosh, S Grancaricˇ, A.M. Hala´sz, M. Hare, C. Havenith, G. Jones, I.A. Katovicˇ D. Keith, S. Kotsios, C. Kovacˇevic´, S. Lee, K. Long, A.C. MacIntyre, L Mather, R.R. Matsuo, T. Midha, V.K. Mijovic, B.

Page 14-16 9-10 96-99 57-60 14-16 13-14 55-56 41 32-33 55-56 60-69 57-60 23-24 44-45 50-55 42-43 71-75 14-16 69-71 45-48 55-56 84-86 21-22 8 101-103 33-35 55-56 24-25 22-23 56-57 42-43 75-78

Mukhopadhyay, A. Otieno, R. Parac-Osterman, A.

43-44 48-50 81-84

Pezelj, E. Pickering, S.J. Power, J. Primentas, A. Psarakis, E.

91-94 55-56 48-50 8 57-60

Rahnev, I. Rajendran, S. Rogale, D. Rudd, C.

40-45 9-10 86-89 55-56

Scotchford, C.A. Skenderi, Z. Soljacˇ ic´, I. Somlo´, J. Sˇomodi, Zˇ. Squires, P. Steffan, I. Stylios, G.K. Tama´s, P. Ujevicˇ, D Vas, L. Vassiliadis, S. Vrljicak, Z Vujasinovic, E. Vukusˇic´, S.B. Walker, G.S. Wardman, R.H. Warrior, N.A. Wilson, J.I.B. Woodget, H.

55-56 105-108 94-96 14-16 103-105 37-39 74-75 25-32, 33-37 14-16 78-81 16-21 8 108-109 90-91 99-100 55-56 32-33 55-56 22-23 39-41

Awards for Excellence Outstanding Paper Award International Journal of Clothing Science and Technology

‘‘Comparison of hyperelastic material models in the analysis of fabrics’’ Manuel Julio Garcı´a Ruı´z CAD/CAM/CAE Laboratory, EAFIT University, Medellı´n, Colombia, and

Leidy Yarime Sua´rez Gonza´lez CAD/CAM/CAE Laboratory, EAFIT University, Medellı´n, Colombia Purpose – This work presents a review of the application of hyperelastic models to the analysis of fabrics using finite element analysis (FEA). Design/methodology/approach – In general, a combination of uniaxial tension (compression), biaxial tension, and simple shear is required for the characterization of a hyperelastic material. However, the use of these deformation tests to obtain the mechanical properties of a fabric may be complicated and also expensive. A methodology for characterizing the fabric employing a different experimental test is presented. The methodology consists of a comparison of the results of the fabric characterization with only a tensile test and the combination of shear, biaxial, and tension experimental tests by using FEA. Findings – Numerical results of the fabric behavior contribute to estimate the effects of experimental limitations in the material characterization and to select the best fit material model to modeling fabrics. Finally, a comparison of hyperelastic material models is illustrated through an example of a rigid body in contact with a hyperelastic fabric in 3D. Originality/value – Hyperelastic models are used to characterize textile materials. Keywords Elastic analysis, Finite element analysis, Textile testing www.emeraldinsight.com/10.1108/09556220610685249 This article originally appeared in Volume 18 Number 5, 2006, pp. 314-25 of International Journal of Clothing Science and Technology, Editor: George K. Stylios

www.emeraldinsight.com/authors

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