Modeling of the mechanical and physical properties of hybrid composites produced by gas pressure infiltration

Abstract

In this study, mechanical and physical properties of hybrid composites were analyzed by using a data acquisition technique like artificial neural networks. Matlab software was used to examine the effects of test parameters on infiltration temperature and pressure. Bending strength, hardness and density of hybrid composite were affected in neural network training and test. Firstly, training set was prepared for the MMCs. In this set, three various training algorithms were used with artificial neural network. At the end of the each learning, the accuracy of the each training algorithm was checked by using test set. As a result, the predicted results which were closest to the experimental results were obtained with the Levenberg–Marquardt training algorithm for bending strength, hardness and density of hybrid composites. These trained values had an average error of 0.376%, 2.969% and 2.648% for bending strength, density and hardness values, respectively.