Underdetermined blind source separation method of pipeline leakage signals based on empirical mode decomposition
SUN Jie-di1, HAO Ya-Li1, WEN Jiang-tao2,LIXiang-wen3
1. College of Information Science and Engineering, Yanshan University, Qinhuangdao 066004, China2. Key Laboratory of Measurement Technology and Instrumentation of HeBei Province,Yanshan University, Qinhuangdao 066004, China
Abstract:It is usually mixed with a variety of interference vibrations and noise in the high-pressure gas pipeline leakage stress wave signals. In order to extract the effective leakage signal from the mixed signals, it proposes an underdetermined blind source separation method based on EMD. Firstly, the mixed signals collected from several sensors are decomposed by EMD. The decomposed intrinsic mode functions (IMFs) are analyzed and acquired the normalized kurtosis features. Then select the principal IMFs which containing most of the leakage information to analyze and reconstruct. The reconstructed signal and the original observation signals compose of a new observation signal matrix, from which separate the effective leakage signal. This proposed method solves the underdetermined blind sources separation. The experimental results show that it can extract leakage signal effectively.