Study On Rotating Mechanical Noise Source Localization Technique Based On Beamforming Techniques

In this paper, combined with the National Natural Science Fund project,according to the need of detection and diagnosis of acoustic fault of large equipment,and starting from the actual engineering application, to improve the beamformingrobustness and resolution are studied by theory and experiment. In this paper, themain contents include(1) Chapter one introduces the origin and significance of the topic in the firstplace, and then briefly reviews the development history of beamforming and theresearch status at home and abroad. According to the research object of this thesis,analyzed the noise characteristics of rotating machinery.(2) In this chapter, beamforming theory and several typical algorithms areanalyzed Systematically. And constracting the performance of the conventionalbeamforming (CBF), minimum variance distortionless response (MVDR)beamforming and multiple signal classification (MUSIC) beamforming by thesimulation analysis.(3) The impact of array disturbance to the beamforming is analyzed, and Errormodels were established. After that, effect of disturbance on array, when only theamplitude error or phase error exists, are respectively discussed. The results show thatboth the amplitude and phase errors affect the output of array, but the phase error playa dominant role.(4) In view of the influence of error, the null widening and diagonal loading areused to improve the robustness of array. And by analyzing the data structure of nullwidening, the contact come to light. In allusion to the difficulty in determining thediagonal loading level, this chapter puts forward an iterative method based onquadratic constraints to solve the optimal loading level. The method can be used in the situation unknown the prior information of signal.Simulation results show that thismethod can adaptively adjust the value of load according to disturbance error.Compared with the fixed loading method, the method can guarantee the stability ofthe beam forming,and make the array gain maintain at a high level at the same time.(5) This chapter introduces the principle of beam space high resolutionrobustbeamforming algorithm.Then the phase errors is imported, and the mathematicalmodel of beam-space beamforming is derived under the condition of error. Simulationexperiment display that it will seriously influence the performance of beamspacealgorithm, when the array phase errors are big enough. Aiming at the effects of arraydisturbance on beamspace method, an improved algorithm is proposed based ongeneralized eigendecomposition. The simulation results show that the improvedmethod puts up outstanding robustness.(6) The experiment is conducted according to the above theory study,experimental results show that the sound source localization method in this paper caneffectively.