Adaptive Modeling of Non-rigid Assembly Orientation Processes in a Statistical CAT Simulation

One of the most common uses for Computer Aided Tolerancing (CAT) in the passenger car industry is body design, which universalized, consists of non-rigid assembly processes. There are a number of theoretical approaches to model the joining behavior of assemblies with elastic FEM approaches combined with statistical computation. On the other hand the rising complexity of today's body-in-white production chain makes simplifications inevitable when fast results for tolerance optimization are required.Multiple use-cases representing different scenarios are analyzed. The scenarios vary in the stiffness of the involved components, the joining environment and the general geometrical set-up. Comparing mass production measurement data with simulation output reveals the limits of CAT simulations in certain cases. One of those cases is where so-called geo-stations improve the dimensional output quality of assemblies compared to the single parts. To account for this phenomenon an adaptive approach to model geo-stations in different ways is presented, i.e. as an entity either provoking an additional influence of deviation or supplying geometrical correction. A matrix is used to suggest the appropriate kind of modeling depending on the deduced parameters.