Application of Empirical Mode Decomposition based adaptive filtering algorithm to suppress DSI in partial discharge detection

HUANG CHENG-JUN,QIAN YONG,JIANG XIU-CHEN

2005-01-01

Abstract:Discrete spectrum interference (DSI) is one of the main interferences in online partial discharge monitoring system, usually comes from communication system, high frequency relay protection system, broadcast, etc. The energy of DSI is so high that, in most cases, partial discharge signals are overwhelmed. Without eliminating the DSI properly, we can hardly analyzing the partial discharge accurately. Among the existing tools, adaptive filter receives much more attention than others, due to its ability of suppressing DSI without needing the prior knowledge of the interference frequency. In this contribution, we employ it to eliminate the DSI in partial discharge detection. For DSI with single main frequency, adaptive filter does work. However, when the number of main frequency increases from one to more, the performance of the filter degrades. In some case, it even diverges, without output. For solving this problem, we incorporate Empirical Mode Decomposition (EMD) with the common adaptive filter, and propose a new scheme for suppressing DSI. By EMD, the DSI with multi-frequency are adaptively decomposed into different Intrinsic Mode Functions (IMF), with different frequency component into different IMF. In this way, the problem of suppressing multi-frequency DSI is reduced to the problem of suppressing single-frequency DSI, and accordingly, the problem can be solved with ease.