Research On Pipeline Leakage Detection Based On Wavelet And Depth Neural Network

With the rapid development of pipeline transportation industry,it has become one of the five transportation modes with its characteristics of economy and convenience,which plays a role in promoting the development of national economy.The number of pipeline laying in our country has also increased greatly every year,which has formed a pipeline network throughout the country.However,many pipeline networks in our country have to face the aging problem after decades of service,and illegal elements steal national resources from time to time,which leads to frequent pipeline leakage accidents,resulting in the loss of national resources and property and damage to the ecological environment.Therefore,how to find the leakage and find the leakage location is very important to protect the national property security and reduce the loss.However,the pipeline operation environment is complex and contains many uncertain factors.When the pipeline leaks,it is often accompanied by a large number of noise information interference,so that the existing pipeline leak detection accuracy is reduced.In this paper,a small pipe network system is built to collect data,and a set of oil and gas pipeline leakage detection platform based on Wavelet and depth neural network is designed.This paper analyzes the noise interference that may exist in the pipeline operation,improves and optimizes the wavelet denoising method to retain the real information of the pipeline network leakage well,and uses the negative pressure wave method to locate it,solves the problem that the negative pressure wave velocity is difficult to measure by using simple mathematical knowledge,and the detection effect of small leakage is also very good.The actual pipeline leakage detection often leads to the occurrence of false alarm due to some working conditions interference,which wastes human resources and financial resources.In this paper,the improved deep belief network is applied to the anomaly detection and classification of pipe network and improved.The performance of deep belief network fault identification and classification is good,which can avoid the problem of pipeline anomaly false alarm.