The predictive development of the electroconductive textile using artificial neural network

This paper presents several aspects concerning the predictive development of the electroconductive textiles based on the artificial neural network. The electroconductive textiles have been obtained using thin film coating with metal micro/nanoparticles and printing. For these textiles, we have made a mathematical model and we optimized dependent parameters, such as pH, dispersion conductivity, mass, thickness, air permeability, and surface resistance, by neural networks based on optimization algorithm Levenberg-Marquardt back propagation. The methodology consists in developing three feed-forward neural networks that are capable of estimating the surface conductivity of the textile materials, based on six parameters obtained in the laboratory. Moreover, this work is significant in the development of surfaces with conductivity directed to obtain the antistatic, electroconductive textiles or insulators, starting from input matrix values.