Multiple Leaks Detection in Fluid Pipeline Based on Improved VMD and TWSVM

Abstract

When multiple leaks occurs at the same time, the influence of leak acoustic wave signals is superposed in the leak acoustic wave signal, which affects the propagation and attenuation rule of the transient acoustic signal. The traditional leak detection method is impossible to effectively detect the multiple leak signals under the condition of mixed signal. Thus, a detection method of multiple leaks is proposed based on improved variational mode decomposing (VMD) and twin support vector machine (TWSVM). Since the decomposition model number of VMD affects the feature extraction, the energy function is used to improve VMD. Therefore, The decomposition model number of VMD is determined by energy function of each component. The improved VMD is applied to acoustic wave analysis and processing. Additionally, the multiple leaks features are extracted, which are input into TWSVM, TWSVM to detect multiple leaks. By analyzing the acquired data of multiple leaks in experimental pipeline, the results show that the proposed method can accurately detect the number of multiple leaks in pipeline.