Vehicle Interior Audio System Noise Prediction Based on Inverse Acoustic Approach and Directivity Pattern

Automotive audio components must meet high quality expectations with ever-decreasing development costs. Predictive methods for the performance of sound systems in view of the optimal locations of loudspeakers in a car can help to overcome this challenge. Use of simulation methods would enable this process to be brought up front and get integrated in the vehicle design process. The main objective of this work is to develop a virtual auralization model of a vehicle interior with audio system. The application of inverse numerical acoustics [INA] to source detection in a speaker is discussed. The method is based on truncated singular value decomposition and acoustic transfer vectors The arrays of transfer functions between the acoustic pressure and surface normal velocity at response sites are known as acoustic transfer vectors. In addition to traditional nearfield pressure measurements, the approach can also include velocity data on the boundary surface to improve the confidence of the source identification. The surface vibration pattern over the surface of the virtual speaker is first extracted based on measured sound pressure data. The acoustic response in a free field generated by the virtual speaker is validated by comparing the sound pressure level from direct measurements and from numerical prediction. The validated virtual speaker with vibration pattern is then applied in a full vehicle model to predict interior sound field. Investigated the interior noise due to speaker with its directivity considered. Inverse numerical acoustics used to retrieve the surface normal velocities on the acoustic model. The technique allows to back calculate the operational vibrations based on operational near field pressure measurements. Near field pressure measurements are required to capture all acoustic waves (radiated waves + evanescent waves). More mid-field pressure measurements were taken to verify the correctness of the suggested method. A good agreement is discovered when the measurements are compared to the re-computed field.