A study on fiber laser micro-spot welding of thin stainless steel using response surface methodology and simulated annealing approach

作者：Liao Hsin Te^*; Chen Zhi Wei

来源：International Journal of Advanced Manufacturing Technology, 2013, 67(5-8): 1015-1025.

DOI：10.1007/s00170-012-4544-y

• Response surface methodology
• Back propagation neural network
• Simulated annealing algorithm
• ANOVA
• Optimization

This study analyzed variations of shear strength that depend on the fiber laser process during micro-spot welding of AISI 304 stainless thin sheets. A preliminary study used ANSYS results to obtain initial process conditions. The experimental plan was based on a Taguchi orthogonal array table. A hybrid method that includes the response surface methodology (RSM)- and back propagation neural network (BPNN)- integrated simulated annealing algorithm (SAA) is proposed to search for an optimal parameter setting of the micro-spot welding process. In addition, an analysis of variance was implemented to identify significant factors influencing the micro-spot welding process parameters, which was also used to compare the results of BPNN-integrated SAA with the RSM approach. The results show that the RSM and BPNN/SAA methods are both effective tools for the optimization of micro-spot welding process parameters. A confirmation experiment was also conducted in order to validate the optimal welding process parameter values.

• 出版日期2013-7