Leak acoustic detection in water distribution pipelines

The leak acoustic signals collected on pipelines play an important role in detecting a leak or leaks in buried pipelines. The traditional detection methods have shown some promise in detecting leak in the absence of a fixed non-leak acoustic source occurring in or outside the detected pipeline. However, in practice, the leak signals are inevitably corrupted with these non-leak sounds as usual. In this case, the leak cannot be easily detected by the traditional methods. In this paper, a new feature extraction and leak detection system using approximate entropy is proposed to discriminate the leak signal from the non-leak acoustic sources. According to the generation mechanism of leak acoustic signals, the self-similarity characteristics of leak signal are investigated. And the autocorrelation function is adopted to describe the self-similarity of leak signal. The autocorrelation function values for the delay τ larger than the signal correlation length, not the signal itself or its entire autocorrelation function, is used to extract or evaluate the self-similarity degree of the leak signal by the approximate entropy algorithm. A neural-network approach has been developed as a classifier, which uses the identified self-similarity features as the network inputs. The proposed leak detection method has been employed to identify the leak in the buried water pipelines, and achieved a 92.5% correct detection rate.