Development and optimization of a prediction system model for mechanical properties in rotary friction-welded polyamide joints using the SVM approach and GA optimization

Abstract

The objective of this experimental study is to utilize rotary friction welding (FW) for assembling similar polyamide materials. The application of the SVM approach enables the development of a predictive model for estimating mechanical properties in RFW processes. Furthermore, the optimization of RFW parameters through GA proves pivotal in selecting optimal welding conditions, providing a variety of choices. The welding parameters considered in this study included rotation speed at five levels and traverse speed at three levels. The strength of the welded samples was characterized by a tensile test. Additionally, temperature measurements were taken to determine the maximum temperature in the joint area. The results demonstrated the dependence of tensile strength and maximum temperature on the rotation speed. Maximum tensile strength is achieved at an optimal rotation speed. Moreover, analysis of variance (ANOVA) indicates that rotation speed is the parameter most influenced by tensile strength.