Sound Source Depth Localization Based On Beamforming

Acoustic fault diagnosis is an effective complement to vibration fault diagnosis technology and plays an important role in the field of mechanical equipment condition monitoring and product noise reduction.The premise of acoustic fault diagnosis is the identification and location of noise source.The noise source location technology based on the combination of the base and the beam of microphone array is simple in measurement process and mature in algorithm.It can accurately locate the medium and long range and medium and high frequency sound sources and has its unique advantages in the field of noise source identification.This paper mainly studies the noise source identification with the combination of microphone array and beamforming algorithm.The specific research contents include:(1)the basic knowledge of acoustic testing is introduced,and the far-field plane wave and near-field spherical wave models of beamforming are given.The principles of beamforming algorithms such as time-domain delay summation,frequency-domain phase shift algorithm,deconvolution DAMAS algorithm,and CLEAN algorithm are introduced.Numerical simulation experiments are conducted to verify the recognition effects of several algorithms on mono-frequency,same-frequency,and heterofrequency sound sources.(2)theoretical and simulation studies are carried out on various parameters that affect the performance of microphone array.According to the gaussian distribution of sound source energy,a random gaussian array is proposed.The array performance is tested by comparing with several common arrays.(3)in order to solve the problem that the traditional beamforming algorithm could not locate the focus depth of sound source,a three-dimensional depth identification method for spatial multiple sound sources was proposed based on the principle of nearfield focused beamforming combined with the method of spatial singular value decomposition,and the spatial search process of sound source depth was simplified by the golden section optimization algorithm.The mathematical model of the adaptive three-dimension sound source recognition method is deduced and the feasibility of the method to the three-dimensional recognition of spatial multi-sources is verified by numerical simulation.(4)the actual positioning effect of the beamforming algorithm in this paper is verified through experiments.