CAE technologies for efficient vibro-acoustic vehicle design modification and optimization

In automotive industry, Computer-Aided Engineering (CAE) methods are increasingly used to predict the various functional performance attributes (noise and vibration, crashworthiness, ...) and adapt the design based on the outcome of virtual simulations, such as numerical Finite Element (FE) models. This reduces the need for expensive physical prototypes, so that design cost and time-to-market can be reduced. The last step is still the final validation on a single physical prototype. If issues are still detected in this late stage, CAE techniques are very useful for efficient diagnosis and refinement engineering, to derive countermeasures and perform optimization. At present, the Finite Element Method (FEM) has become the dominant method for modal analysis and interior acoustics in the low and medium frequency range. For accurate predictions up to higher frequency, large FE model sizes are required, which limits the number of design iterations in a given time. Furthermore, at increasing frequency, the effect of uncertainty and variability (in geometric dimensions, material properties ....) becomes ever more important, so that for accurate performance (range) predictions, one has to include their effect in the modeling & simulation process. Many non-deterministic methods are based on using a deterministic “core” method to assess the effect of uncertainty and variability on the vehicle response. Methods for faster deterministic design iterations are useful to address both these issues, since they enable the evaluation of more design variants / a larger-scale optimization / a more extensive uncertainty assessment in a given time. This paper reports on two methodologies for efficient vibro-acoustic design, namely Wave-Based Substructuring (WBS) for efficient local re-design and Modal Modification for fast predictions of the effect of panel thickness and damping variations. Finally, the latter approach is used to speed up a non-deterministic analysis using the Fuzzy Finite Element approach, aimed at assessing the effect of panel uncertainty on the NVH response.