变电站局部放电噪声的识别方法

Recognition Method of Partial Discharge Noise in Substation

局部放电能有效反映电气设备绝缘劣化的严重程度,是变电站电气设备在线监测的重要组成部分。从局放噪声的特点出发,提出一种基于频谱分析滤波、时域峭度和匹配追踪相结合的局部放电噪声分离方法。该方法首先根据能量的高低将谐频噪声从检测信号中滤除,然后通过计算残余信号的峭度寻找局放发生的信号时段,最后应用局放的先验模型结合匹配追踪的方法快速分离局放噪声信号。结果表明,该方法不仅大大提高了局放信号识别率和计算速度,而且还能很好地保留局放信号的相位信息,有助于实现局放源的测向和精确定位。

Partial discharge(PD)can effectively reflect the insulation degradation of electrical equipment.PD noise monitoring is an important part of online monitoring for electrical equipment of substation.A PD noise separation method based on spectrum analysis and filtering,time domain kurtosis and matching pursuit is proposed arising from the characteristics of PD noise.Firstly,the harmonic noise was filtered out from the detection signal according to the energy level,and then the moment of PD occurring was evaluated by calculating the kurtosis of the residual signal.Finally,the prior model of discharge signal and matching pursuit were combind to quickly isolate PD noise.The results show that the proposed method not only greatly improves the recognition rate of the PD signal and the calculation speed,but also preserves the phase and amplitude information of the PD signal,which can helps to realize the accurate direction-finding and source-localizing.