Estimation of Critical Buckling Loads of Hybrid Composites in Different Environments using Artificial Neural Networks

Abstract

The primary objective of this study is to estimate the buckling behaviors of hybrid composite plate using test data on the effects of different environmental conditions. These estimations were made using an Artificial Neural Network (ANN). The MATLAB software was used to develop the artificial neural network. The data from the buckling tests were used to train the ANN model that used a multilayer, feedforward and backprop algorithm. The input parameters for the ANN modeling were determined as waiting times of the samples, ambient temperatures, ambient conditions, and material arrangement angles. This modeling was used to estimate critical buckling loads. The values obtained after the training and testing of ANN were examined by performing statistical analyses commonly used in ANN models, revealing that the designed model was applied successfully and the results were very close to the real test results.