Abstract: Abstract: In order to reduce the multi-point leakage positioning error of urban pipeline, a method of locating multi-point leakage in pipeline based on improved complete ensemble empirical mode decomposition (ICEEMD) sample entropy was proposed. Firstly, the ICEEMD algorithm was suggested, to denoise and decompose the multi-point leakage sound signal into real leakage signal component and redundant component; sample entropy analysis was used to remove redundant components, and obtain effective signal. Secondly, observed signal was separated by blind source separation and maximum likelihood estimation, which obtains independent effective leakage signal.Then, the cross-correlation time delay method was used to calculate the time difference between the two signals reaching the upstream and downstream sensors. Finally, the precise location of a single leakage point was calculated. The results show that the method has better signal decomposition effect and can greatly improve the accuracy of multi-point leak location of pipelines.

Key words: Key words: multiple leaks, improved complete ensemble empirical mode decomposition, sample entropy, blind source separation, maximum likelihood estimation