Adaptive denoising method for leakage detection of liquid pipelines using automatic variational mode decomposition

Abstract

Pipeline leakage detection is an important measure to ensure the national economy and public safety. This paper aims to develop an adaptive denoising method to achieve leakage signals under a low signal-to-noise ratio. This method relies on signal processing to extract sub-modalities. It improves variational mode decomposition for signal denoising structurally, called automatic variational mode decomposition (AVMD), enabling it to select effective modes without using prior knowledge. It achieves progressive mode decomposition by increasing constraint criteria and changing the objective function of VMD and sets an automatic bandwidth adjustment rule based on the energy ratio between modes. It also sets an iterative condition considering the power concept of gradually decreasing signals to achieve the target mode number without advanced setting. Several examples, including simulating linear signals and nonlinear signals, as well as real-life applications, are demonstrated to show that AVMD is superior to VMD and other existing improved methods in reducing relative errors and improving the separation effect for cross signals, and AVMD can effectively suppress noise.