Research And Implementation Of The Infrasonic Signal Processing And Location Algorithm Of Pipeline Leakage

Over the past few years,oil pipeline leaks often occur due to the aging and erosion of pipelines and human-made destroy,which contributed to economic losses,environmental pollution,and people's health.Therefore,it is critical to pinpoint leakage locations promptly and accurately.With the rapid development of oil pipeline detection techniques and study of infrasound signal propagation characteristics,the detection of pipeline leakage has become possible through receiving and processing infrasound signals collected by sensors.With regard to the infrasound signal detection techniques in oil pipelines,this thesis firstly analyzes the existing methods and then points out their advantages and disadvantages.Next,we improve the wavelet signal denoising method which is usually used in leakage detection,and propose a novel leakage location algorithm based on adaptively locking original signals.The main work of this thesis is as follows:1.We first study the generation and propagation characteristics of leakage acoustic signals in fluid pipelines,thus demonstrating the possibility of applying infrasound theory into pipeline monitoring and alarming.Moreover,we investigate the noise in the leakage signals and then present an improved wavelet signal denoising algorithm.In particular,(1)Whitening test is used to determine the number of decomposition layers.After performing the wavelet transform(WT),the improved method will check that whether the wavelet coefficients have white noise characteristic.Once the white noise characteristic is disappeared,the wavelet decomposition will stop,which can solve the decomposition determination problem in wavelet signal denoising.(2)According to the statistical property of white Gaussian noise under WT,we adopt a threshold calculation method based on the 3? criterion for wavelet coefficients in each layer.After the calculation of thresholds,wavelet coefficients will proceed in terms of the soft threshold function.The wavelet coefficients in each decomposition layer will be removed if their amplitude values are smaller than thresholds;otherwise,the coefficients will be contracted and remained.Consequently,the proceed wavelet coefficients are the optimally estimated values.What's more,by performing inverse wavelet transform,a relatively smooth signal can be reconstructed without losing details of signal spikes.Simulation results show that the improved wavelet signal denoising algorithm could effectively improve the signalto-noise ratio(SNR).2.A novel leakage location algorithm is proposed.This algorithm firstly estimates the value of the original signal length,and correlation analysis will start to distinguish the original signal segment from the noised signals received by sensors at both sides of the pipeline.The leakage location can be calculated via cross-correlation analysis of original signals.This process should be repeated and get the mean value of calculated locations,which further improves the location accuracy.Simulation results demonstrate that the novel leakage location algorithm could not only improve the accuracy of time-delay estimation but also minimize the influence on cross-correlation outcomes brought by noise.