Crash simulation of mechanical joints with automatically determined model parameters based on test results and prediction algorithms

The increasing usage of innovative light weight concepts in automobile production leads to the application of different mechanical joining techniques like self-pierce riveting- semi tubular (SPR-ST) and -solid (SPR-S), flow drilling screwing (FDS) and high speed bolt joining (HSB) for multi-material constructions. These mechanical joints are used at positions of car bodies which show high stresses under impact loading. For the prediction of the load-bearing capacity, the failure behavior and the energy absorption in crash simulations complete and reliable models are needed. Therefore experimental results on single joint specimens and simulation of these specimen tests are necessary to determine the model parameters. If this had to be done for all existing sheet metal combinations of all mechanical joints in a body-in-white it would result in a very time and cost intensive process. The aim of the research project CraSiFue[1] was to reduce these efforts by developing a forecast algorithm and implementing it in a software. The developed software JoiningLab predicts the joint properties and model parameters of the *CONSTRAINED_ INTERPOLATION_ SPOTWELD (Model 2, SPR4) [2], [3] in LS-DYNA® for untested i.e. unknown mechanical joints. This results in saving real tests and accelerates the crash safety investigations especially in the concept phase of construction, where materials, sheet thicknesses and joints are not definitely specified yet.