Martin Berggren
Department of Computing Science, UMIT Research Lab, Umeå Universitet
The compression driver is the standard sound source for mid- and high-range acoustic horns used in, for instance, movie theaters and live audio systems. The compression driver + horn combination delivers a system that is about 10 times as efficient as a standard direct-radiating loudspeaker. The main challenge in the design of the compression driver is to avoid resonance and interference phenomena. The complexity of these phenomena makes it difficult to carry out this design task manually. Therefore, we employ a gradient-based shape optimization method in order to design a driver with a frequency response as free from irregularities as possible. We use a level set function to specify the geometry of the devic's interior, and the geometry is imposed in the computations using the CutFEM approach, which means that mesh deformations or remeshings are not needed when the design is updated. A particular modeling challenge here is that that visco-thermal losses cannot be ignored, due to narrow chambers and slits that occur in the driver. An accurate but computationally inexpensive boundary-layer model, recently developed by our research group, is used here to account for these losses within the optimization loop. The computations were carried out in the FeniCS computing platform. The frequency response of the optimized driver very closely matches the ideal response that a compression driver would have if wave effects could be ignored.
Joint work with Anders Bernland, André Massing, Daniel Noreland, and Eddie Wadbro