Experimental Study On Ground Characteristics Of Leakage Signal Of Buried Water Supply Pipeline

Leakage of water supply pipes has always been a major problem for urban water supply systems.Pipeline leakage accidents have a variety of causes,including various natural and human factors,such as soil erosion,geological settlement,and load changes.If it cannot be found and repaired in time,it will cause a lot of waste of resources,and it will also pose a threat to water quality,water quantity,water pressure,and even personal and property safety.Pipeline leak detection methods are various,of which acoustic detection applications are the most common.At present,most of the main research work of leak detection methods based on acoustic signal processing is based on the direct collection of leakage signals from the pipeline,and this method will be more limited in the actual leak detection work.Based on the actual leak detection situation in the work,this paper mainly studies the method of leak detection using the ground leakage acoustic signal.Firstly,based on Fourier transform theory,the paper explores the characteristics and changing rules of leakage signals under different leakage conditions(such as different pressures,propagation distances,pipes,etc.),and finds that the increase in pressure leads to an increase in signal power without affecting the frequency range.,and the two presented a good linear relationship;within a certain range,the signal power will exponentially decay with the increase of the propagation distance,but the frequency range is basically constant;the difference in the pipe will not have a significant impact on the leakage signal;combined with the correlation coefficient Principles of coherence function,comparative analysis of external leakage signals and integrated leakage signals under the same conditions,found that when the signal to noise ratio is high,the two have good similarities,indicating that the leakage signal collected at the ground,the main The source is an outside source.Secondly,according to the problem that the leakage signal is disturbed by noise,wavelet analysis theory is used to study the denoising processing of noisy leakage signal.Using different wavelet functions and threshold denoising methods,the signal-to-noise ratio and the correlation coefficient of the signal before and after denoising are used as evaluation indexes to process the periodic signal and the leakage signal respectively,and it is found that the wavelet transform can effectively suppress the noise in the noisy signal.In part,improve the signal to noise ratio of the signal.Finally,from the point of view of stochastic features,the autocorrelation function theory and Approximate Entropy Theory(Ap En)are used to describe the "repeatable" feature of the leakage signal,and the leakage signal is distinguished from the noise signal;and the Lempel-Ziv complexity is selected.(LZC)calculates signal complexity as a second random feature that distinguishes between leakage and noise signals.The BP neural network was constructed using two characteristic parameters: Ap En and LZC.The BP network was trained with abundant data samples,and the identified leakage signal and noise signal were used for identification test.The recognition accuracy rate was 90%,and this method was proved.The presence of leaks can be identified from common interference noise.Applying this leak identification method to actual leak detection work can significantly increase work efficiency and reduce the cost of leak detection.