Research on VMD based adaptive denoising method applied to water supply pipeline leakage location

Abstract

In the study of water supply pipeline leakage location, the noise will bring significant errors to the location estimation. The traditional denoising method for non-stationary random signals mainly dependent on the estimation of noise parameters, and these methods are confined to the removal of random noise, ignoring periodic noise and components unrelated to leakage in pipeline. In this paper, an adaptive denoising method based on VMD is proposed. By defining a parameter called signal clarity, modes closely related to leakage can be adaptively selected under unknown background noise, and finally the selected signal mode is reconstructed. In the simulation analysis, the proposed method can accurately extract the mode component of the leakage signal from the mixed noise, and improve the signal-to-noise ratio. In actual experiments, this method improves the error of the leakage position compared to the conventional denoising method. Experiments show that the minimum relative error is 0.17% and the maximum is 2.18%.