Transient-Based Leak Detection and Monitoring of Water Pipes Using Complementary Ensemble Empirical Mode Decomposition (CEEMD) Method

Abstract

Transient signal occurs when there is sudden change of pressure inside the pipeline especially due to pressure surge or opening and closing valve. Decomposition method applied in this study to analyze the transient signal and remove noise that contaminate the signal. In this paper, an adaptive decomposition algorithm that is Complementary Ensemble Empirical Mode Decomposition (CEEMD) proposed in order to overcome the problem occur in Hilbert-Huang Transform (HHT) method. This method takes over the Empirical Mode Decomposition (EMD) method that purpose as pre-processing method in HHT. This improvement made to overcome mode mixing and reconstruction error that lies in EMD method. CEEMD method used to extract the IMF’s component of a contaminated signal to remove undesirable noise signal. Instantaneous analysis using Hilbert Transform (HT) apply for selected IMF to locate the feature’s position along the pipeline. The result prove that proposed decomposition method show success with the percentage of error between measured and analysed distance only below that 4% for all features extracted.