Abstract

Adhesive bonding is a widely applied joining method in car body assembly processes. Up to now, the application of adhesive joints is still based on empirical experience and investigation studies. A numerical simulation tool could drastically improve the process understanding and optimize the manufacturing process in an early stage of development. In addition, the simulation tool can serve as a time and cost saving alternative to expensive prototype testing. This requires a robust and efficient simulation method including a valid material model. During the assembly, the adhesive is not cured and behaves like a non-Newtonian fluid. The measured viscosity curves can be highly non-linear, i.e., the viscosity depends strongly on the local shear rate. In the present work, a material model, which can take arbitrary nonlinearities in the viscosity curve into account, is developed. The numerical model is validated by experimental squeeze flow tests which are conducted at different squeezing velocities. The flow behaviour is qualitatively compared by means of cross section pictures from which the velocity profile can be estimated. For comparison, the classical power law relation is considered in the simulations. It is shown that for highly non-linear viscosity curves, this simple approach results in a high error, especially at high shear rates. The validated viscosity model is then applied to an exemplary assembly process of automotive body parts. In this feasibility study, two sheet parts, connected to other car body parts, are bonded together. During this process, the adhesive bead is squeezed into a thin layer. The simulation provides information about the final distribution of the adhesive, the squeezing force, and the extent of distortions in the joined and connected parts. The obtained results can be considered in subsequent joining processes. This is illustrated by the example of an automotive assembly process involving resistance spot welding as a subsequent assembly operation.