Prediction of Elasto-plastic Behaviour of Short Fiber Reinforced Polymer Composites Using Artificial Neural Network

Abstract

Short fiber reinforced polymer (SFRP) composites have evolved to be a material of interest in recent years due to the high strength-to-weight ratio, durability, and cost-effective manufacturing of complex parts. However, modeling of an SFRP demands intricate micro-mechanical considerations and complex algorithmic implementations. For analyzing mechanical properties, the stress-strain response provides the best framework, which can give important insights like elastic modulus, ultimate strength, and toughness. Considering plasticity models for the material and complex material modeling makes it computationally intensive using numerical techniques like finite element (FE). In the present study, an artificial neural network (ANN) model is developed to predict the elasto-plastic response of an SFRP. The dataset required for training the model has been obtained from simulations using the Mori-Tanaka method. The prediction accuracy of the ANN model is tested with experimental data and FE analysis results reported in the literature.