采用LMD和互相关的管道泄漏检测信号噪声抑制

In order to overcome the limitations of the commonly used empirical mode decomposition in the undershoots or overshoots phenomenon and the drastic end effect, which will introduce false decomposition and the limitations of the correlation coefficient shifting method that it relies on the prior knowledge about signal and noise and it can't be applied in the complex pipeline environment, a noise suppression method for the pipeline leak detection with the local mean decomposition (LMD) and the cross-correlation method is proposed. To realize the suppression of the noise signal, firstly the detection signals are decomposed adaptively into several PFs (product functions) by LMD, and then the useful decomposition components are selected based on the cross-correlation method to construct the leakage signal. The simulation and experimental results show that the method proposed in this paper with LMD and cross-correlation shifting is suitable for different pipeline environment and the capability of noise suppression is strong.