基于小波分析和卷积神经网络的互感器二次多点接地故障智能检测

Intelligent Detection of Secondary Multipoint Grounding Fault of Transformer Based on Wavelet Analysis and Convolutional Neural Network

为了准确检测互感器二次多点接地故障,提出基于小波分析和卷积神经网络的互感器二次多点接地故障智能检测方法。采用基于小波阈值滤波的互感器信号去噪方法,去除互感器信号噪声;通过基于小波分析的互感器信号特征提取方法,提取去噪后互感器信号的有效特征向量;使用流行学习方法,实现互感器有效信号特征非线性降维;把降维后信号特征作为基于卷积神经网络的二次多点接地故障智能检测方法的输入样本,使用检测合格的卷积神经网络,作为互感器二次多点接地信号特征分类器,实现互感器二次多点接地故障智能检测。实验结果显示,此方法检测结果与实际互感器运行情况相符,检测结果准确。

In order to accurately detect the secondary multipoint grounding fault of transformer,an intelligent detection method based on wavelet analysis and convolutional neural network is proposed.The transformer signal denoising method based on wavelet threshold filtering is used to remove the transformer signal noise.Through the transformer signal feature extraction method based on wavelet analysis,the effective feature vector of transformer signal after denoising is extracted.The popular learning method is used to realize the nonlinear dimensionality reduction of effective signal characteristics of transformer.The signal features after dimensionality reduction are taken as input samples of the intelligent detection method of secondary multipoint grounding fault based on convolutional neural network,and the qualified convolutional neural network is used as the feature classifier to realize the intelligent detection of secondary multipoint grounding fault of transformer.The experimental results show that the detection results of this method are consistent with the actual operation of the transformer,and the detection results are accurate.