Simulated transfer path accuracy vs. sound perception

International audienceThe use of numerical methods is becoming more and more common in industry design processes. However, to finely represent the physics of the problem, numerical models are always more complex and time consuming. In the meantime, sound perception is now a very important topic for car manufacturers, who try to predict sound quality during the development phase of the vehicle. It is therefore useful to get some ideas about how accurate a simulation should be in order to be used in sound quality applications. The goal of this paper is to understand the minimum accuracy of simulated transfer paths (TPs) required to give meaningful psychoacoustic results. A physical model is build and the associated Finite Element (FE) model is setup, validated and taken as reference. Then, some deteriorations of the FE model are introduced and a sensitivity analysis checks how those changes affect objective metrics associated to the TPs. Laslty, hearing tests are performed in order to understand how people percieve the differencies in the various simulated sounds