Research On Sound Source Localization Technology Based On Microphone Array

In acoustic fault diagnosis,it is very important to find the source of noise accurately.Microphone arrays are widely used in the location detection of sound sources.The geometric rules of the spatial layout of the microphone array play a decisive role in the results of sound source location.The widely used arrays,such as linear arrays,circular arrays,and spiral arrays,are difficult to accurately locate the desired signal when facing multiple similar sound sources.Therefore,this paper proposes a leaf shaped microphone array designed based on the bionic sunflower spiral arrangement theory,aiming to apply it to the location of mechanical noise with a lot of interference,locate the location of the fault sound source in time,and provide help for subsequent fault diagnosis.In this paper,the performance of the leaf shaped array is analyzed and verified in detail through simulation analysis and experimental verification,combined with beamforming and MUSIC sound source location algorithms.The specific research contents are as follows:First,build a simulation platform for the blade microphone array.When the incident signal is a single sound source and multiple similar sound sources,use the beamforming algorithm to simulate and analyze the single sound source and multiple similar sound sources,and analyze the performance of the leaf shaped array from the perspective of the positioning accuracy and resolution of the array.At the same time,uniform circular array,Arcondoulis spiral array and other arrays are simulated to compare and analyze.The results show that combined with beamforming algorithm,leaf array has better resolution and anti-interference ability than uniform circular array and spiral array in the case of single sound source and multiple similar sound sources;Then,the leaf shaped microphone array is combined with MUSIC algorithm for simulation analysis of single source positioning,aiming to compare the positioning accuracy of different positioning algorithms combined with leaf shaped array,and draw a conclusion: under the same other conditions,the positioning accuracy of MUSIC algorithm based on leaf shaped microphone array is higher than that of conventional beamforming algorithm,and the generated positioning cloud image has fewer sidelobes and higher resolution.This paper takes the rotating machinery experimental platform as the research object,analyzes the common fault forms of the mechanical experimental platform and the nearfield propagation model of the sound source,studies three kinds of sound source location algorithms,and selects the beamforming algorithm and MUSIC algorithm to locate the mechanical fault noise source.The microphone arrays of different arrays are used to pick up sound signals,and then the gear fault signals are located through the delay sum beamforming algorithm and MUSIC algorithm.The aim is to verify the ability of the leaf shaped microphone array to locate the mechanical fault sound source through experiments,and draw a conclusion: under the same conditions,the leaf shaped microphone array combined with the location algorithm can achieve the location of the mechanical fault sound source;The positioning accuracy and resolution of the leaf shaped microphone array are superior to the commonly used spiral array.