Comparison of RSM with ANN in predicting fatigue and impact behavior of MIG welded AA6061 Aluminum alloy joints

AA6061 aluminum alloy (AI-Mg-Si alloy) has found wide application in the fabrication of light weight structures requiring a high strength-to weight ratio and good corrosion resistance. MIG welding parameters are the most important factors affecting the quality, productivity and cost of welded joints. The effect of MIG weld parameters on fatigue life and impact energy of AA6061 joints was analyzed in the present study.Two methods, response surface methodology (RSM) and artificial neural network (ANN) were used to predict the fatigue life and impact energy of MIG welded AA6061 aluminum alloy joints. The experiments were conducted based on three factors, three-level, and central composite face centered designwith full replications technique, and mathematical model were developed. The results obtained through response surface methodology were compared with those through artificial neural networks. The comparison shows that ANN model is more accurate than the RSM model

Keywords:

fatigue life , impact energy , response surface methodology , artificial neural networks