Application of artificial neural network and full factorial method to predict the Poisson's ratio of double core helical auxetic yarn

In this work, an experiment according to the full factorial method (FFM) to investigate the effect of initial helical angle of wrap component, diameter ratio of components and modulus ratio of components on Poisson's ratio of double core helical auxetic yarn (DC-HAY) is designed. Then, after statistically confirming the similarity of maximum NPR in both main directions to reduce the calculations, a regression model based on the FFM equation and an artificial neural network (ANN)-based model to predict the maximum NPR of DC-HAY are developed. The results of the analysis of variance indicated that all terms of the regression model are statistically significant. Also, the Student-Newman-Keuls showed that all of the considered levels for parameters are in separated groups. Furthermore, the optimum architecture for the ANN model is selected based on a goodness function constructed by the coefficient of determination and mean squared error. The results of comparison between the models revealed that the ANN model has better performance than the regression model. In addition, the sensitivity analysis of ANN models introduced that the modulus ratio of components is the most sensitive parameter affecting the output of the network.

Automated summary

In this work, an experiment was conducted to investigate the effect of initial helical angle of wrap component, diameter ratio of components and modulus ratio of components on the Poisson's ratio of double core helical auxetic yarn (DC-HAY) using the full factorial method (FFM). Regression and artificial neural network (ANN) models were developed to predict the maximum NPR of DC-HAY, with the ANN model performing better. The sensitivity analysis showed that the modulus ratio of components is the most sensitive parameter affecting the output of the network.