Research on Noise Reduction Method for Leakage Signal of Water Supply Pipeline

Abstract

The negative pressure wave (NPW) signal caused by leakage of water supply pipeline contains a lot of noise due to the influence of water flow state and circuit noise, which will adversely affect the accuracy of leakage detection and leak location. To solve this problem, a novel pipeline pressure signal denoising method based on variational mode decomposition (VMD) is proposed in the paper. Firstly, the key parameters which affect the performance of VMD method: IMFs K and penalty factor $\alpha $ are optimized by genetic algorithm (GA) without prior knowledge. At the same time, different parameters are combined to form different reconstructed signals based on the correlation coefficient (CC). Secondly, the optimal combination of parameters is adaptively determined by using a fitness evaluation function based on the information entropy (IE). Thirdly, based on the optimal parameter combination, reconstruct the signal to achieve noise reduction. Finally, the denoised signal will be used for leak detection and leak location calculation. In order to verify the effectiveness of the proposed method, simulation analysis and laboratory experiments are carried out. Simulation analysis results show that the noise suppression effect of our method is better than that of wavelet, empirical mode decomposition (EMD) and VMD. The highest SNR obtained by our method is 16.187. The laboratory experiment results showed that the NPW signal denoised by proposed method had less error and more stable results in location, with an average relative error of 2.97%.

Keywords

• Negative pressure wave
• leak location
• variational mode decomposition
• genetic algorithm
• correlation coefficient
• information entropy