Auxetic structures exhibit distinctive deformation characteristics that give rise to unconventional mechanical responses, characterized by a negative Poisson's ratio (PR) under deformation. While most designs emphasize auxetic behavior, less attention is directed toward resulting stress distributions. The present study addresses this gap by predicting the PR and stress‐concentration characteristics of an S‐shaped auxetic structure, which is observed to exhibit lower stress concentration than conventional re‐entrant configurations. Two prediction models were developed using machine learning techniques, i.e., regression analysis and an artificial neural network (ANN), based on a full‐factorial design of experiments comprising 27 simulations. The regression model produced high coefficients of determination, with R 2 values of 99.43% for the PR and 92.91% for the von Mises (VM) stress, along with average percentage errors of 1.9023% and 9.6215%, respectively. The ANN model demonstrated stronger performance, achieving an overall correlation of 0.99962 and average percentage errors of 0.6585% for PR and 3.024611% for VM stress. Statistical evaluation confirmed the goodness‐of‐fit for both models at the 95% confidence level, indicating no significant differences between finite element modeling (FEM) and predicted responses. Overall, the the ANN model offers superior predictive accuracy, making it an effective alternative to resource‐intensive simulations/experiments.
Gupta et al. (Fri,) studied this question.