Abstract
ABSTRACT The main aim of the present study is to perform a comparative investigation, experimental analysis, and optimisation of weld reinforcement in the specific domain of powder-fed gas metal arc welding in relation to AISI 1023 steel. The study conducted a comprehensive analysis of several parameters, encompassing statistical methodologies (i.e. Genetic Algorithm and artificial neural networks) and response surface methodology (RSM). The study shows that the advance gas metal arc welding provides maximum of 95% more reinforcement and current and powder feed rate has emerged as the most influential parameter for bead reinforcement. Based on the results obtained, it can be concluded that the Levenberg-Marquardt (5-14-1) model demonstrates the highest level of suitability as an ANN model for the purpose of forecasting output data. As per R value positive correlation of 0.9674 between the empirical data and the forecasts produced by the LM (5-14-1) model. Genetic algorithm suggested that the optimal point is situated within the predetermined range of process variables. According to validation experimentation the ANN model exhibited superior performance compared to the RSM model and then come GA. ANN has shown the maximum error of 0.06 when compared to experimental results.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call