基于峭度和时域能量的局放脉冲提取算法

提出了一种针对电力电缆局部放电脉冲的提取方法，首先根据信号峭度的变化定位局部放电脉冲峰值所在位置，然后根据信号能量的变化采用滑动能量搜索的方法以局部放电脉冲峰值为中心向两边计算搜索局放脉冲的边沿，从背景噪声中准确提取局放脉冲信号。该方法可以将局部放电检测信号中纯噪声信号部分与包含局部放电脉冲信号的部分区分开来，且不受局部放电脉冲个数和脉冲所在位置的限制，并具有步骤简洁、计算量小的特点。通过与小波阈值去噪、经验模态分解（EMD）两种主流电力电缆局部放电脉冲提取方法的仿真实验对比和实际实验对比，验证了该方法的有效性与准确性。

Partial Discharge Pulse Extraction Algorithm Based on Kurtosis and Time Domain Energy

This paper proposes a power cable partial discharge (PD) pulse extraction method. Firstly, the PD pulse peak is located according to the variation of signal kurtosis, and then according to the signal energy's changing the method of sliding energy search is adopted to calculate and search the edge of PD pulse from the centered PH pulse to the both sides, so as to extract the PD pulse in the noise background accurately. This method can distinguish the full noise signal section from the section that contains partial discharge pulse signal in partial discharge detection signal, and won't be constrained by the number of partial discharge pulses and position of pulse, and also has features of simple step and small amount of calculation. By comparing the simulation experiments of two mainstream power cable PD pulse extraction methods (the wavelet threshold de-noising and the empirical mode decomposition) with the actual experiments, the effectiveness and veracity of this method have been verified.