Research On The Consistency Of Pipeline Leak Detection Model And Application

Acoustic sensor has become one of the most important means of pipeline leak monitoring because of its high sensitivity and accurate positioning.The SVDD fault diagnosis modeling method has been widely used in pipeline leak diagnosis modeling because it only need the characteristics of normal signal samples.When signal transport to the both ends of the pipe,signals'waveform,amplitude and leakage signal frequency domain energy distribution will change a lot.The method of fixed-scale signal filtering will be difficult to ensure accurate filtering effect.Subjective judgments and personal experience will have a deep impact to the effect of artificial selection.It is difficult to ensure that the established leakage diagnosis model has high accuracy and consistency,which will increase the detection system false positives and the probability of occurrence of false negatives.In view of the above problems,this paper presents an optimization filtering method for pipeline leakage acoustic signal and sample automatic selection method for SVDD leak diagnosis modeling.By analyzing the energy distribution in the frequency domain of the normal signal and the leakage signal,the upper and lower limit frequency range of the signal filtering is determined.By means of wavelet packet decomposition and reconstruction method,this paper analyzed the influence of different reconstructed signal frequency ranges on the Gaussian property of reconstructed signal,the correlation of reconstructed signal with the original signal,and the signal-to-noise ratio in the reconstructed signal.The optimal reconstruction of the optimization filtering method based on the three characteristic variables is proposed and the optimal reconstruction of the acoustic signal is realized.Based on the calculation of the acoustic signals' positive and negative interval and the signal-to-noise ratio of the interval sub-signal,the distribution of the signal-to-noise ratio of the sub-signal in the acoustic interval is analyzed.A method of signal-to-noise ratio threshold for interval sub-signal based on"3?" criterion is proposed,which realizes the automatic acquisition of training sample signal for leakage diagnosis and modeling.Based on the above method,the validity and accuracy of the method are tested by establishing the SVDD leak diagnosis model.This paper designs a leak monitoring server program and the following functions are completed:TCP/IP data collection,acoustic signal optimization filtering,SVDD diagnosis algorithm based on time domain statistical characteristics,and leakage GIS map display and other functions.Simulated leak experiments and nearly seven months of continuous operation of the system proved that the system and method are effective and reliable...