In the product design of shell-like housing structures for vibrating bodies, acoustic emission plays a significant role in usability and customer comfort. Consequently, the prediction of the emitted sound, e.g., in the form of the radiated sound power, is desired in the earliest possible stage of the development cycle. In industry, performing extensive acoustic testing using microphone arrays or laser scanning vibrometry is often time-consuming or infeasible, as it requires an acoustically treated measurement chamber and specific measurement equipment. On the other hand, accelerometer data is more straightforward to obtain due to the lower cost of the sensors and relaxed constraints on the measurement chamber compared to acoustic measurements. Hence, this contribution utilizes available surface velocity data in an acoustic simulation of the free-field Helmholtz problem. In doing so, it compares a conventional element-based and a mesh-less data-driven mapping technique to investigate the reduction of the number of provided sensor positions for the radiation problem. The open-source boundary element framework NiHu is used to give accurate predictions of acoustic emissions over a wide frequency range. The approach is applied to an industrial problem and validated against microphone measurements.
Read full abstract