Method for identifying pipeline leakage

Abstract

The invention discloses a method for identifying pipeline leakage, which comprises the following steps: aiming at two observation signals of adjacent sensors, using the signals as an input signal and a reference signal of a self-adapting filter respectively, and extracting a weight vector of the filter during convergence, wherein the weight vector is the optimal estimation of the correlation functions of the two observation signals; judging that a pipeline probably leaks when the weight vector of the filter during the convergence has a peak value, otherwise judging that the pipeline has no leakage; and when the judgment result is that the pipeline probably leaks, calculating the weight vector complexity value of the part beyond one neighborhood of the envelope peak value of the weight vector, and inputting the weight vector complexity value into a decision algorithm to finally determine whether the pipeline leaks really. The method has the following technical advantages: not only the noise-signal ratio of signals is improved, but also the correlated characteristics about leakage signals shared in the two signals are highlighted, and the decision algorithm is adopted to perform quantitative analysis on the complexity of the self-adapting filtering weight vector, eliminate noise interference and identify whether a signal source is a real signal source having the leakage.

Description

technical field [0001] The invention relates to a pipeline leakage detection technology, in particular to a pipeline leakage identification method. Background technique [0002] Timely and correct detection of pipeline leakage is the premise of leakage location and the first step of engineering leakage control. [0003] Early engineering leak detection personnel judged whether the water supply pipeline leaked by listening to the sound produced by the leak. This method of identifying leaks that relies on the experience and knowledge of engineers has large subjective factors, low efficiency, and easy missed detection. With the development of technology, the detection methods currently used mainly include: continuous flow detection method, minimum night flow and pipeline pressure signal wavelet analysis, etc. There are shortcomings such as time-consuming detection, limited types of detection pipelines, and the detection area is greatly affected by the ground environment, which...

AI Technical Summary

Problems solved by technology

Aiming at the influence of fixed non-leakage sound sources inside and outside the pipe on leak detection, Ai et al. combined the linear predictive coding cepstral coefficient (LPCC) and hidden Markov model (HMM) of the leak sound signal, and Toshitaka and Akira proposed to use the power spectral density of the leak signal , destruction factor based on AR model, etc. are used as the characteristics of the leakage sound signal, combined with the support vector machine (SVM) to distinguish whether the relevant components in the sensing signal come from leakage or various non

Images