Abstract:

In this paper, an intelligent model of heat and moisture propagation in light nonwoven fabrics was designed by conversion of a numerical propagation model of a partial differential equation to a feed forward propagation neural network. Propagation coefficients of heat and moisture transfers were estimated from the intelligent model for nonwoven samples containing hydrophilic natural and hydrophobic synthetic fibres. The results presented that the error of the model is acceptable less than 4.7% and 7.9% for estimation of heat and moisture diffusivity coefficients, respectively. The Anova test revealed that while fibre type and fabric thickness affected heat and moisture transfer through the fabric, factors such as surface mass density, heat and the moisture transfer constant were not significant. Also it was found that the optimum transfer rate was observed in the case of samples containing viscose or a viscose and polypropylene blend in the ratio of 60%:40%, respectively.