Parameter optimization using a regression model and fitness function in laser welding of aluminum alloys for car bodies

The objective of this paper is to determine the optimal welding conditions in terms of the productivity and weldability for laser welding of aluminum alloy AA5182 using filler wire AA 5356. The experiments were performed with laser power, welding speed, and wire feed rate as control factors. Tensile tests were carried out in order to evaluate the weldability under each welding condition. In order to estimate the tensile strength, three regression models are proposed. One is a multiple linear regression model, another is a second order polynomial regression model, and the last is a multiple nonlinear regression model. Of the three models, the second order polynomial regression model had the best estimation performance with respect to ANOVA (analysis of variation) and average error rate. Also, this study defines objective functions for tensile strength, which represents weldability, and for the welding speed and wire feed rate, which represent productivity. In addition, fitness functions are obtained using the objective functions and a weight matrix which shows the importance of each objective function. The steepest descent method is used to find the optimal point where the fitness function was maximized. Optimal welding conditions were found at a filler wire feed rate of 2.7 m/min, laser power of 4 kW, and welding speed of 7.95 m/min.