Global Optimization of Welding Parameters Based on Improved EGO Algorithm

Residual stress generated in welding process lead to fatigue failure. Therefore, it is important to optimize welding process for improving fatigue life of weld parts. As the traditional optimization and simulation results are liable to fall into local optimum or more iterations about model. In this paper, improving the global optimization of kriging-based is proposed to find the optimum welding parameters for maximum residual stress. Based on the initial sample, Kriging model is implemented to approximate the objective function regarding the design variables. the introduce of calibration points about MAXEI and maximum variance MSE, Kriging model is updated iteratively by adding two new points to the original sample till the global optimum is obtained. The results show that the proposed method has faster convergence, reducer the iterations and it can search for a more global optimal solution. Applying the method to design optimization of residual stress generated in welding process can better guide designer of welding process planning.