开关柜局部放电信号降噪与缺陷识别研究

Noise reduction and defect recognition of partial discharge signals in switchgear

文中采用暂态地电压法(TEV)进行检测,设计了四种典型的缺陷模型并搭建试验平台,分别对局部放电缺陷模型进行了实验。由于变电站现场环境复杂,需要对采集的信号进行信号降噪。针对以往小波降噪都是按照经验采取固定的分解层数的问题,提出一种Mallat算法结合最优分解层数自适应算法对含噪信号进行分离与重构,结果显示该算法可以很好地滤除噪声。对重构后的局放信号提取八种时域特征参数,并采用BP神经网络对开关柜局部放电的类型进行识别,当误差准确率δ=0.002时,放电类型的识别正确率最高,能够达到97%。

In this paper,the transient ground voltage method(TEV) is used to test,four typical defect models are designed,and a test platform is built.The partial discharge defect model is tested separately.Due to the complex environment of the substation,it is necessary to perform signal denoising on the collected signals.Aiming at the problem that the wavelet denoising is based on the experience of taking a fixed number of decomposition layers,a Mallat algorithm combined with an optimal decomposition layer adaptive algorithm is proposed to separate and reconstruct the noisy signals.The results show that the algorithm can filter out noise well.The reconstructed partial discharge signal is extracted from eight time domain characteristic parameters,and BP neural network is used to identify the type of partial discharge of the switchgear.When the error accuracy rate is 0.002,the correct recognition rate of the discharge type is the highest,which can reach 97%.