Contribution
Exploring the Use of Masking Effects in a Multi-Level Simulation Framework for Efficient Prediction of Perceived Sound
* Presenting author
Abstract:
The ability to assess acoustic properties in the early stages of product development is a major challenge, as the simulation of complex source and propagation mechanisms is costly. An example of this is the prediction of cabin noise in future transport aircraft, which are expected to have a high degree of integration of propulsion and fuselage for reasons of propulsion efficiency. To reduce the computational effort, a multi-level method based on human perception, here, loudness, is proposed. The included masking effect suggests that certain low-amplitude, high-frequency sounds become inaudible in the presence of high-amplitude, low-frequency sounds. Based on the assumption that the simulation effort can be reduced without compromising accuracy if the masking effect is taken into account, high-fidelity models are applied to simulate these high peaks, while low peaks can be handled with low-fidelity models. This contribution validates this concept within a multi-level framework, applied to a cavity model. As an initial step towards utilizing masking, a possible methodology is presented for identifying low-peak regions by assuming that the peak amplitude and width can be estimated prior to completing the full simulation.