ASSA-VMD-SI and Frechet method of pipe vibration for noise reduction and leakage identification

Abstract

Pipe is an essential component of city. The working condition of underground pipe is complex. There is a lot of noise in the acquired vibration signals, which would interfere with the feature analysis and leakage identification. Pipe leaking could lead to the accidents such as ground subsidence, waterlogging, etc. Therefore, it is of great importance to identify the pipe leakage. In this paper, an ASSA-VMD-SI (Adaptive Sparrow Search Algorithm-Variational Mode Decomposition-Self Information) and Frechet method of pipe vibration for the noise reduction and leakage identification was proposed. First, the SSA was combined with adaptive sine–cosine and Cauchy-Gaussian variational methods. The penalty factor and modal decomposition number of the VMD were optimized. Then, the vibration signal was reconstructed based on the self-information distance between the intrinsic mode function and the original signal. Finally, the multi-features of vibration were extracted. The Frechet similarity between the baseline and test parameters was calculated to identify the leakage state. Experiments showed that this method could filter the noise of the vibration signal. Multiple leakage states could be identified in real time as well.