Research And Application Of Heating Pipeline Leakage Monitoring Method Based On Infrasoun

Heat supply pipeline is a key infrastructure for urban construction,and is also the artery of "energy flow" and "quality flow" of the heat and power production system,so it is extremely important to ensure the safe and stable operation of heat supply pipeline.At present,leaks in the heating pipeline network occur from time to time,which not only pose a safety threat,but also cause huge economic losses and environmental pollution.Therefore,it is of great significance to carry out online monitoring,leak point location and intelligent alarm for heating pipelines.In recent years,acoustic-based pipeline leak detection methods have become a hot research topic in this field.The acoustic method assesses the running status of pipelines by detecting changes in the characteristics of acoustic signals generated during the pipeline operation,which has stronger fault sensing capability compared with other methods.Aiming at the important and difficult problems in the pipeline leakage monitoring process,this thesis carries out research in the areas of leakage infrasound signal noise reduction,leakage aperture prediction,leakage point location,development of leakage monitoring system and etc.The main work is as follows:(1)The principle of pipeline leak monitoring based on infrasound is introduced,the generation mechanism and propagation characteristics of pipeline leak acoustic signals are elucidated,and the time-frequency characteristics of leak signals and the main noise components are analyzed;based on the characteristics of leak acoustic signals,the joint noise reduction algorithm based on mode-preferred reconstruction and sparse expression(CEEMDAN-SR)is proposed,considering the shortcomings of traditional noise reduction methods in filtering out strong noise interference.The correlation coefficient and Shannon entropy criterion are introduced to filter and reconstruct the components after modal decomposition,and the sparse expression of the reconstructed signal is further performed based on the sparse decomposition algorithm;finally,the proposed method is verified by the measured signal.(2)To solve the problem of quantitative estimation of leakage size,the improved gray wolf algorithm(IGWO)based on chaotic mapping,elliptic convergence parameter and dynamic weight update position is proposed,and the performance of the improved strategy is verified by performing the optimization search test on the benchmark function;further,the hyperparameter optimization of the LSTM neural network is performed by using the improved gray wolf algorithm,and the leakage aperture prediction model(IGWO-LSTM)is established;finally,the proposed prediction model is validated based on the timefrequency characteristics of the leak signal after noise reduction by CEEMDAN-SR algorithm,the waveform characteristics and the pipe wall pressure as the input feature vector.(3)To address the problem of large localization error in time delay estimation by mutual correlation,time delay estimation is performed by using generalized mutual correlation,and the performance of four generalized mutual correlation weighting functions is compared to select the weighting function applicable to the leakage signal;further,noise reduction is performed on the noisy signal based on the CEEMDAN-SR algorithm proposed in this thesis,and time delay estimation is performed on the signal by generalized mutual correlation;finally,the proposed method is verified to have high accuracy and stability by experimental signals.(4)In order to meet the actual field application requirements of the heating pipeline leakage system,a software of heating pipeline leakage monitoring and intelligent localization based on the joint programming of LABVIEW and MATLAB was developed,which includes the basic business modules of data acquisition,data management,signal analysis,feature extraction and intelligent leakage warning.The software system was finally tested and verified in the laboratory and field environments,and achieved good application results.