Study Of Sound Source Identification Using Inverse Boundary Element Method

Mechanical noise has a greater harm on human health. The most effective method of reducing noise is to control the sound source. For this reason, the simulation and experiment of the sound source identification are carried out based on the inverse boundary element method. The main contents of the research are as follows:The acoustic transfer matrix is established based on the basic theory of acoustics and the Helmholtz boundary integral equation. The singular integral problem in the boundary integral equation is dealt with polar coordinate transformation. The transfer matrix is verified by a pulsating sphere. The sound pressure of the radiated sound field is successfully predicted by this method.The ill-posed problem of the transfer matrix for the reconstruction of the surface vibration velocity is suppressed by the direct regularization method. The principle and characteristics of Tikhonov regularization,TSVD regularization,GCV method and L-curve criterion of regularization parameter selection method are discussed. Meanwhile, field point selection influencing factors are studied from three aspects: the number of the measurement field points, the size of the measurement surface, the distance between the measurement field points and the sound source. Based on the radiation characteristics of the pulsating sphere and the vibration characteristics of the thin plate, the vibration velocity of the surface is reconstructed. The results show that Tikhonov regularization and GCV regularization parameter selection are more effective than the other method. The selection principles of field point selection are given.To solve the problem of sound source identification in mixed sound field, the noise separation is executed based on the boundary element method using the sound pressure and particle velocity on the monolayer measurement surface. Using the vibration radiated sound field of the thin plate, the simulation of the noise separation and the reconstruction of the surface vibration velocity is carried out by interference of the monopole and the plane wave.The results show that the interference noise can be separated effectively by this technique and the reconstruction accuracy of the surface vibration velocity is improved.The mono sound source and multi sound source identification experiments are studied in an anechoic chamber. The single frequency noise signal and white noise signal are input to the speakers. The sound source identification results verify the validity and accuracy of the theory.