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 tau 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.