Modeling of multi layered ANN architecture for resistance spot welding process

Abstract

Resistance spot welding process was widely used in the automobile industry since the early nineteenth century and still being used in the production of automobiles. Several works were carried over the several years on spot welding of different materials. The recent researchers are focusing on welding different materials using resistance spot welding process using new techniques. The present study discusses about modelling of ANN architecture. Based on the results of the cross tension failure load of resistance spot welds, the ANN model is created. It is found that the architecture consisting of 4 neurons in the interface layer has least mean square error compared to other architectures. Mean square error of architectures with varying number of neurons in the interface layer is evaluated in the present study.