Leak detection using instantaneous frequency analysis

Leaking pipes are a primary concern for water utilities around the globe as they compose a major portion of losses. Contemporary interest surrounding leaks is well documented and there is a proliferation of leak detection techniques. Although the reasons for these leaks are well known, some of the current methods for leak detection and location are either complicated, inaccurate and most of them are time consuming. Transient analyses offer a plausible route towards leak detection due to their robustness and simplicity. These approaches use the change of pressure response of the fluid in a pipeline to identify features. The method used in the current study employ a single pressure transducer to obtain the time domain signal of the pressure transient response caused by a sudden opening and closing of a solenoid valve. The device used is fitted onto a standard UK hydrant and both cause a pressure wave and acquire the pressure history. The work described here shows that the analysis using Hilbert transform (HT), Hilbert Huang transform (HHT) and EMD based method is a promising tool for leak detection and location in pipeline network. In the first part of the work, the analysis of instantaneous characteristics of transient pressure signal has been calculated using HT and HHT for both simulated and experimental data. These instantaneous properties of the signals are shown to be capable of detecting the reflection from the features of the pipe such as leakages and outlet. When tested with leak different locations, the processed results still show the existing of the features in the system. In the second part of the work, the study is based on newly method of analysing non-stationary data called empirical mode decomposition (EMD) for instantaneous frequency calculation for leak detection. First, the pressure signals were filtered in order to remove the noise using EMD. Then the instantaneous frequency was calculated and compared using different methods. With this method, it is possible to identify the leaks and also the features in the pipeline network. These were tested at different locations of a real water distribution system in the Yorkshire Water region.