Research On Cotton Fabric Handle Determination With Artificial Neural Networks

Fabric handle is a representative character of materials or products made of various fibers, and it is objective, which can be used to assess the compatibility between fabrics and human beings, that is, subjective judgment is ranked by sense. In recent years, artificial neural network not only develops towards comprehensiveness, but also has been widely used in other aspects, almost covering every fields of the society. Its market scale is rapidly exposing. In the paper the application of artificial neural network on fabric handle assessment is investigated.Twelve kinds of blended and pure cotton fabrics are employed and tested to settle the problem of sensory evaluation of model recognition and comprehensive judgment for fabrics handle. The sensory property of fabrics handle is obtained by surveying among a special crowd and evaluation of experts. The standards of evaluation on fabrics handle rank are established by comparing the subjective recognition of fabrics using hand and eye with the memory stored in brain and analyzing factors affecting sensory evaluation and the relationships between handle feeling and fabric style. In the investigation of fabrics handle, there are many indistinctive or obscure conceptions. The study shows that softness, flexibility, lubricities are effective words in describing the handle feeling of fabrics.Different physical mechanics properties of fabrics endow various fabrics handle style. By evaluating the style of fabrics with KES-FB measuring system and analyzing effects of every physical mechanics factor on fabrics handle, a predictable artificial neural network model is constructed. And the method of predicting handle feeling of cotton fabrics is put forward by artificial neural network. The objective assessment of fabrics handle is based on eighteen physical index of KES-FB. The results prove that it is feasible to establish a predictive model of the connection between the subjective assessment and objective assessment by artificial neural network, and the artificial neural network predictable model can exactly predict the fabric handle to a certain degree. The artificial neural network overmatches traditional methods in predicting cotton fabric handle.