Research On Pipeline Leak Detection And Location Technology Based On Statistical Hypothesis Tests

Pipeline has become an important transport way, in parallel with railway, highway, aviation and water means. As the aging plumbing, geographical/ climatic conditions and man-made destructive factors often cause pipeline leaking accidents, it is an urgent need for pipeline leak detection and location. The existing system based on negative pressure wave is able to detect and locate the large and rapid leak, however, for slow leak; it lacks effective detection and location way. Therefore, it has important theoretical meaning and application value to research on the detective and location way for slow leak.In this paper, a new signal processing method based on second generation wavelet transform is introduced, which provides the low-frequency part of every single layer of reconstruction to the locator program. As the falling-edge feature of slow leak pressure signal difficult to extract, on the basis of the statistical properties of pipeline pressure signal, this paper propose a new leak detective method based on first-order local statistical properties and a leak location method based on the multistage hypothesis tests. Also the pipeline leak location system is built based on second-generation wavelet transform and multistage hypothesis tests.The experimental test shows that, the leak detection and location method based on second generation wavelet transform and multi-level hypothesis test can accurately determine the leak location to pressure signal of obvious trend falling edge. Moreover, to the pressure signal of slow trend falling edge, it can also ensure quick location with a certain precision range. Therefore, it provides an effective way to solve the slow leak location problem.