基于经验模式分解处理局部放电数据的自适应直接阈值算法

A New Adaptive Direct-threshold Algorithm to Partial Discharge Data Processing Based on Empirical Mode Decomposition

根据局部放电信号的特征,将经验模式分解(EMD)应用于局部放电信号分析中,提出了处理局部放电数据的自适应直接阈值(ADT)算法。首先,将基于EMD的时空滤波方法应用于局部放电数据的预处理中。与传统滤波方法相比,该方法无需预定义滤波器系数,而且能够充分保留原始信号本身所固有的非平稳特征。其次,为了最大限度的抑制噪声干扰,进而提出了ADT算法。该方法不存在小波方法中的小波基选取问题,以多分辨率的EMD为基础,结合3σ准则自适应地确定分解尺度和阈值,是一种完全的数据驱动型方法,具有较好的自适应能力和综合处理性能。仿真数据和试验数据的处理结果表明了该方法的有效性。

This paper proposes a novel method using empirical mode decomposition（EMD） for the analysis of partial discharge（PD） signals buffed in excessive noise. First of all, a time-space filtering based on EMD is introduced to the pretreatment of PD data. Compared with traditional filtering, the time-space filtering does not need to define the coefficient of filter and is able to preserve the features of signal. Further more, a new way called adaptive direct-threshold（ADT） algorithm on the basis of 3σ rule and EMD is stated in this paper in order to restrain noise maximally. This technique overcomes the difficulty in choosing wavelet. It is fully data-driven and is able to determine the decomposition order and threshold adaptively, so it has strong adaptive capacity and preferable processing performance. The simulation and experimental results demonstrate effectiveness of the proposed algorithms.