基于随机矩阵理论的局部放电脉冲快速检测方法

Fast Detection Method of Partial Discharge Pulse Based on Random Matrix Theory

局部放电(partial discharge,PD)的检测是开展电力设备状态评估的重要手段之一。由于现场背景噪声及干扰信号影响,采集到的局部放电信号往往淹没在噪声中。噪声抑制算法是提高局部放电脉冲检测的有效方法,但这些噪声抑制算法往往需要针对所有采集的数据展开,当数据量特别大的时候将严重影响其实时性。因此,提出了一种基于随机矩阵谱分布理论的局部放电脉冲快速检测新方法。利用特高频传感器接收到的PD时间序列构造高维随机矩阵,再根据随机矩阵理论下时间序列模型的经验谱分布理论,提出了基于滑动时间窗的局部放电脉冲快速检测方法。仿真及实验室测试结果表明,所提方法能快速地识别出时间窗数据中的PD脉冲。

Partial discharge(PD)detection is one of the important means to carry out the condition assessment of power equipment.Due to the influence of noise and interference signal,the PD signals are often submerged in noise.The noise suppression algorithm is an effective method to improve PD pulse detection,however,the algorithms are often needed to be performed for all the collected data,which can affect the real-time of the proposed algorithms especially when the amount of data is very large.Therefore,a new method for rapid detection of PD pulse based on random matrix spectral distribution theory was proposed.The PD time-domain signals received by ultra high frequency sensors were used to construct a high-dimensional random matrix.According to the empirical spectrum distribution theory of time series model under the random matrix theory,the partial discharge pulse was detected in each time window.The simulation results show that the proposed method can quickly identify PD pulse of the data window.