Virtual testing of self-piercing rivet connections

• Main Authors: Andersson, Daniel, Saliba, Fredrik
• Format: Others
• Language: English
• Published: Luleå tekniska universitet, Institutionen för teknikvetenskap och matematik 2020
• Subjects: FEM; SPR; self-piercing rivets; self piercing rivets; LS-dyna; Explicit FEM; lap-shear; cross-tension; KS2; r-adaptivity; 2D axisymmetric; axisymmetric; Applied Mechanics; Teknisk mekanik
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-79405

The automotive industry is currently trying to replace the conventional steels to lightweight materials such as aluminum or carbon fiber to meet all stricter emission targets. When using such materials, traditional joining methods, such as spot welds, could be difficult to use. Therefore, more focus has been put on self-piercing rivets (SPR).In whole car models used in crash simulations, substitution models are used to model SPR joints. It is important to calibrate these models for different load cases. Volvo Cars Corporation (VCC) are currently calibrating using time-consuming physical tests where the SPR joint is subjected to loads in different directions. To save time, a way of virtually evaluating the SPR joint strength is therefore sought after. To do this, a method was developed using non-linear FEM in LS-DYNA. The method was then used to perform sensitivity studies concerning friction, sheet thickness and rivet geometry.The method developed can be divided into three parts. The process simulation, where the rivet insertion was simulated. A springback analysis, where the material is allowed to springback, closer resembling the real behaviour. Finally, the three destructive tests, lap-shear, cross-tension and KS2, were built using the geometry and initial values from the springback.For the process simulation, an explicit solution was used. To handle the large deformations present during the event, r-adaptivity was used together with a kill-element-method to describe failure, based on CrachFEM or Gissmo. The following springback analysis was then performed using one implicit step.For the destructive tests, a solid element representation of the SPR joint was created using the geometry and initial values from the springback. A shell-solid hybrid model was used to keep the computational time low.Using the method, a good correlation was found both for the process- and the destructive test simulations when compared to experiments. Furthermore, it could be concluded that friction, sheet thickness and rivet geometry affects the SPR joint strength and characteristics.