Applying the artifical neural network to predict the thermal properties of knitted fabrics

Abstract

Fabric thermal properties have been of great interest and importance for textile researchers, since they are among the major characteristics that determine wearing comfort. In this study, thermal conductivity and thermal contact properties of a large number of knitted fabrics were measured instrumentally and a learning group was proposed to train the artificial neural network (ANN) algorithm and then prediction was achieved with a strong regression coefficient. Therefore it is concluded that it was possible to predict thermal properties of knitted fabrics by using basic fabric properties as input.