An ANN and RSM Integrated Approach for Predict the Response in Welding of Dissimilar Metal by Pulsed Nd:YAG Laser

Laser welding is an advance non-traditional and high energy beam welding technique for dissimilar material, which is being increasingly used in industries like automobile, nuclear reactors, aerospace etc. Again it is a very tough task to joining dissimilar materials due to their properties variation. In this present work two dissimilar metals such as copper and AISI 304 stainless steel are taken into consideration. Laser weldings are performed using a pulsed Nd: YAG laser welding machine. Experiments are carried out by taking three process parameters such as laser power, welding speed and pulse duration with three levels each. A statistical design of experiment (DOE) technique i.e. Response Surface Methodology (RSM) is adopted for analysis of maximizing the tensile strength. The back propagation artificial neural network (ANN) technique is used to predict the strength of the welded area. The predicted data are compared with the experimental results and is found to be in agreement.