Model updating for the simulation of surface strains on printed circuit boards considering parameter uncertainty

The efficient and reliable design of power electronic components plays an important role in the development process of electrically driven vehicles. One key aspect is the reliability of solder joints on printed circuit boards (PCB) that greatly depends on the surface strain at the solder joint locations. It is therefore unavoidable to use precise simulations of surface strains to reliably estimate the solder joint lifetime. This work presents a procedure of model updating of a printed circuit board model that considers the variability of board behavior due to uncertainties in the material composition or the manufacturing process. Hierarchical Bayesian model updating is applied to incorporate this variability. The printed circuit board is seen as a multi-level model that is updated in two steps. Experimental data from system and component level are used to sequentially update the printed circuit board and the board mounting. The experimental data combine modal information and measured frequency response functions. The proposed procedure is applied to a test PCB and the updated model is validated with experimental surface strain data.