Imaging Location Method Of Gas Leakage Source Based On Acoustic Array

In order to reliably ensure that gas storage and transportation equipment operates normally and safely,and avoid catastrophic hazard caused by gas leak as much as possible,how to effectively detect and locate gas leak is the problem that people pay more and more attention to today.Acoustic array methods have gradually become one of the hot research directions in the field of gas leak detection and location due to their excellent characteristics such as good stability,strong anti-interference,and high accuracy.In this paper,we first improve the existing typical beamforming algorithms,to propose the beamforming algorithm based on symbol coherence and feature space decomposition,and then verify that the algorithm can significantly reduce the sidelobe level from-30 d B to-100 d B,which enhances the beam directivity.At the same time,capturing the audio and image data collected by the multi-arm spiral acoustic array with camera designed in this paper in the compressor station,acoustic imaging and angle of arrival estimation of the gas leakage source are carried out.The results show that the algorithm proposed can improve the imaging resolution of the leak point,and its location accuracy is increased by about 4times compared with the delay multiply and sum algorithm.Then,on the basis of the audio imaging location method proposed in this paper,the method of natural gas station equipment operation and leakage diagnosis based on beamforming acoustic spectrum is proposed.In this paper,a linearly constrained minimum variance beamformer is designed.Compared with the delay and sum beamformer,it can obtain the acoustic spectrum of the target device more accurately according to the angle of arrival.By extracting the Mel frequency feature from the beamforming spectrum,the audio feature data set of the station equipment is established.Using the support vector machine model can completely distinguish the audio data of different devices in the data set,and it can accurately diagnose the valve leakage in different degrees.Finally,the array outlier detection method is used to further improve the imaging location and spatial filtering effect of beamforming for device audio.By detecting and filtering out the outliers in the array,the resolution and contrast of audio imaging can be improved,and the maximum deviation of the azimuth angle estimation can be reduced from 19° to 1° when there are 8 outliers in the array.Summary,combining beamforming and machine learning technology,a method for monitoring the condition of natural gas station equipment based on acoustic array is proposed in this paper.This method can accurately image and locate gas leak,and can efficiently and intelligently monitor station equipment,which provides a new idea for the design and application of acoustic gas leak monitoring systems in the future.