Objective Hand Evaluation of Stretch Fabrics Using Artificial Neural Network and Computational Model

ABSTRACT

Stretch woven textiles are widely employed because of their excellent elongation and recovery properties. The stretch fabrics studied in this research can be used as power stretch and action stretch sportswear fabrics. The purpose of this paper is to investigate an approach to predict the total hand value by translating the senses into numbers. Computational and artificial neural network models were developed. Five primary hand attributes softness, smoothness, fullness, stiffness, and stretchability were shortlisted that influence the fabric handle. Computational methods were used to generate primary and total hand equations based on basic mechanical parameters. To forecast primary hand values, stretch percent was used with low stress mechanical properties. The association between subjective, computational, and artificial neural network total hand values was investigated using a statistical technique. The subjective and computational hand values have a high correlation of 0.84. The subjective and artificial neural network hand values were shown to have a 0.82 correlation. The accuracy of both models’ prediction of fabric hand was found to be very high. The study finds that both models can forecast the total hand value of stretch materials with a tolerable level of accuracy.

摘要

弹性机织物因其优异的延伸率和回复性能而被广泛应用. 本研究所研究的弹力织物可以用作强力弹力和动作弹力运动服面料. 本文旨在研究一种通过将感官转换为数字来预测手部总价值的方法. 开发了计算和人工神经网络模型. 影响织物手感的五个主要手部属性: 柔软度、平滑度、丰满度、刚度和可拉伸性. 基于基本力学参数, 使用计算方法生成主要和全部手部方程. 为了预测主要手部值, 将拉伸百分比用于低应力机械性能. 使用统计技术研究主观、计算和人工神经网络总手值之间的关联. 主观和计算手值具有0.84的高度相关性. 主观和人工神经网络手动值的相关性为0.82. 两种模型对织物手感的预测精度都很高. 研究发现, 这两种模型都可8以预测拉伸材料的总手部价值, 并且具有一定的准确性.

KEYWORDS:

• Stretch fabrics
• sportswear fabrics
• fabric hand
• artificial neural network
• multiple regression
• computational model

关键词:

• 弹性面料
• 运动服装面料
• 织物手
• 人工神经网络
• 多元回归
• 计算模型

Research highlights

• The kind of stresses acting on stretch fabrics are different from those acting on apparel fabrics. Therefore, the already existing Kawabata hand equations cannot be applied to the stretch fabrics and a set of new hand equations were developed.

• Along with low-stress mechanical properties the stretch percentage was also used.

• Computational method including multiple stepwise block regression method was used to develop primary and total hand equations. And Artificial Neural Network (ANN) was also used to determine the total hand values.

• Five primary hand attributes softness, smoothness, fullness, stiffness, and stretchability were defined and considered for stretch fabrics.

• It was observed that the accuracy of both models’ prediction of fabric hand was found to be very high. The study finds that both models can forecast the total hand value of stretch materials with a tolerable level of accuracy.

• This computational tool can be used by fabric manufacturers, in product development etc, who are desperately looking for such a tool to know the resulting fabric hand of stretch fabrics.

Abbreviations

PHV: Primary Hand ValueTHV: Total Hand Value

Disclosure statement

No potential conflict of interest was reported by the author(s).