Finite element modeling and optimization of hybrid laser–TIG welding of type 316L(N) austenitic stainless steel

Abstract

An integrated approach was followed for determining the optimized process parameters for hybrid laser-tungsten inert gas (TIG) welding of 316L(N) stainless steel which involve FEM followed by optimization. FEM-based model was developed to predict the thermal distribution for a various combination of welding process parameters. The hybrid heat source combining conical and double-ellipsoidal heat sources was used for generating the temperature distribution during welding. The penetration depth produced by welding was estimated from the thermal model for various set of process parameters by running the model several times. The generated dataset consisting of process parameters and the depth of penetration (DOP) was used for developing adaptive neuro-fuzzy information system (ANFIS) model for correlating the set of input process parameters with the output DOP. This ANFIS model was then used in the objective function of a genetic algorithm (GA) for identifying the optimum process parameters to achieve the maximum DOP during autogenous hybrid laser–TIG welding. The optimized solutions were validated by performing bead-on-plate experiments on 5.6-mm-thick type 316L(N) stainless steel plates. The above-integrated approach was found to accurately determine the optimum hybrid laser–TIG welding process parameters for achieving the maximum DOP with minimum experimentation.