Modelling and interaction analysis of the self-pierce riveting process using regression analysis and FEA

Self-pierce riveting (SPR) is a major joining method used in the automotive industry. However, there still lacks fast and easy-to-use joint quality prediction tool available for engineers. In this study, simple but effective regression analysis was applied to quickly predict SPR quality. Two models were developed of interlock minimum remaining bottom sheet thickness (Tmin). The accuracy validated by comparing with experimental results. Under studied configurations, mean absolute errors (MAE) Tmin 0.047 mm 0.053 mm, respectively, corresponding percentage (MAPE) 10.4% 12.3%. With models, interaction effects between rivet die parameters on also systematically analysed. results revealed that demonstrated significant influences not Tmin. These further examined analysing deformations substrate materials. Moreover, die-to-rivet volume ratio (R) found be critical formation interlock, larger more likely achieved when R close 1.0.