Investigation of temperature distribution in friction stir welding with experimental and regression algorithms

In this current study, the joining process performed by friction stir welding, which has become very popular recently and has started to be used in many sectors, and then the temperature analysis of this welding process was performed. The temperature values obtained during the process carried out in traditional friction stir welding and the temperature values reached in the welding process using insulation material were compared with real experiments and analyzed with Taguchi and ANOVA methods by applying experimental designs in the Minitab program. In addition, the empirical formulas obtained by performing PACE regression analysis and the temperature values obtained from experimental studies, their results and error rates were found. By using these methods, more realistic results were obtained with fewer experiments in order to reduce the cost of their implementation, and the results were evaluated accordingly. It also helped to create a less costly roadmap for subsequent experiments. In the experiments, the desired welding process was carried out at a 15% lower temperature using insulation material; with the formulas obtained as a result of PACE regression analysis, it was seen that the experiment with the lowest error rate was the experiment with 1200 RPM and insulation material.