A multilayer perceptron artificial neural network model for estimation of ultraviolet protection properties of polyester microfiber fabric

Abstract

In this study, the use of a polyester fabric produced from microfibers as a siding material for construction industry was investigated. In this context, an ultraviolet (UV) absorber was applied to the polyester fabric samples with and without dyeing process. These samples were treated with UV absorber at 1-4% concentration and then an accelerated aging test was carried out. The UV absorbance capacity of the samples was investigated before and after the accelerated aging test. Furthermore, the effects of dyeing process on UV absorbance capacity of samples were analyzed. Afterwards, a multilayer perceptron artificial neural network (MLP-ANN) model was proposed and utilized to predict the UV protection properties which are UV protection factor, UV-A and UV-B in polyester microfiber fabric. The MLP-ANN based results demonstrate that the regression (R) values are almost 1 for all UV protection properties. Accordingly, it was seen that the proposed MLP-ANN is correctly modeled and the prediction of UV protection properties is successfully performed.