Effect of Aspect Ratio on Dynamic Fracture Toughness of Particulate Polymer Composite using Artificial Neural Network

Abstract The present study discusses about the effect of the aspect ratio of the fillers on the fracture toughness of the glass-filled epoxy composites under impact loading. Three different kinds of fillers (spheres, flakes and rods) were used with different volume fractions (5%, 10% and 15%). Experimental results for Stress Intensity Factor (SIF) were obtained using a gas gun setup and a high speed camera. Further experimental investigation was done using SEM fractographs. Then the potential of using Artificial Neural Network (ANN) in predicting the effect of filler shape on the fracture behavior is studied. The framework of Multi-Layer Perceptron (MLP) feed forward network was used to predict the SIF history using four input parameters viz. time, dynamic elastic modulus, aspect ratio and volume fraction of the glass fillers. Experimental results of fracture test under impact loading were fed to train the ANN network and later the predicted results were compared with the experimental ones. Owing to the fact that predicted values had an accuracy of 91%, crack initiation toughness was predicted corresponding to the intermediate values of aspect ratio for which the experiments were not performed. Among the four input parameters, aspect ratio (largest/shortest dimension) was found to be the most important parameter in the prediction of SIF after time, followed by the dynamic modulus and volume fraction. The significance of aspect ratio lies in increasing the surface area to volume ratio which is responsible for the interfacial strength between the matrix and the filler and hence affects the fracture toughness of the overall composite material.