Research And System Implementation Of Acoustic Imaging And Sound Source Location Technology Of Power Equipment

With the development of power system,the real-time monitoring and diagnosis of power equipment fault has become a hotspot in this field.Acoustic imaging is a kind of sound field visualization technology,using this technology for fault diagnosis of power equipment,we can intuitively understand the distribution of sound field inside the equipment,and assist technicians to understand the fault location and formulate corresponding fault solutions.This technology improves the efficiency of fault detection and location,which is of great significance for the construction of intelligent power system.This thesis mainly studies the acoustic imaging and source location algorithm of power equipment,proposes an improved acoustic imaging and source location algorithm,and builds acoustic imaging and sound source location system for practical testing.The main research contents and achievements of this thesis are as follows:(1)Some common algorithms of acoustic imaging are reviewed,including several beamforming algorithms and near-field acoustic holography algorithms.The characteristics,advantages,disadvantages and applicable scenarios of different algorithms are compared.According to the actual application scenarios in this thesis,the research focuses on DAMAS(Deconvolution approach for the mapping of acoustic sources)algorithm,and the main factors which affect the performance of the array beamforming algorithm are simulated and analyzed.(2)The existing noise suppression algorithm and DAMAS algorithm are improved.A noise suppression algorithm based on the improved spectral subtraction and de-spectrum method is proposed.According to the characteristics of noise and sound source signal,the ratio of time domain energy to frequency domain spectral entropy is constructed as a characteristic parameter,and the noise spectrum estimation is carried out by using this parameter.Then,combined with spectral subtraction and self-spectrum elimination,the irrelevant noise and self-spectrum interference of acoustic signal are suppressed.In view of the large time cost of existing DAMAS2 algorithm,this paper proposes a deconvolution beamforming algorithm based on focus grid filtering,which greatly reduces the time cost.(3)Finally,an acoustic imaging and sound source localization system is built,and the main work includes hardware selection,microphone array parameters determination,real-time acquisition program design,camera calibration algorithm design,acoustic imaging and sound source localization algorithm design.A new method is proposed to determine the number of sound sources by using the focus plane sub region division and screening method,which realizes the sound source counting and multi-source localization.The speaker is used to simulate the transformer sound source for acoustic imaging and sound source location test.The results show that compared with the DAS algorithm,the improved noise suppression deconvolution beamforming algorithm can significantly improve the spatial resolution performance and the effective dynamic range in the case of single sound source or dual sound source Compared with DAMAS2 algorithm,the positioning error is improved by 16.9% to36.8% in different arrays.Compared with DAMAS2 algorithm,the time cost of the improved noise suppression deconvolution beamforming algorithm is reduced by 35.6% to 1.03 s,which basically meets the system design requirements.