In this paper, an attempt is made to determine input-output relationships of the Electron Beam Welding (EBW) process using regression analysis based on the data collected as per central composite design of experiments. The process performance characteristics directly measured include bead width (BW), depth of penetration (BP) and bead reinforcement (BH). Other relevant bead geometry parameters may be calculated using these outputs. To model EBW process, several regression functions have been fitted to the experimental data to develop mathematical models establishing the relationships between EBW process variables and its response characteristics. The best set of models is then selected based on the statistical analysis. In the final stage of this research, Simulated Annealing (SA) algorithm has been employed to find the best set of process parameter levels in order to obtain desired weld bead geometry. Computational results demonstrate that the proposed SA method is quite effective and efficient in optimizing process parameter values for any desired weld bead geometry in EBW.

Keywords