基于波形相似度的同杆双回线路故障识别

Fault Identification of Double-Circuit Transmission Lines on the Same Tower Based on Waveform Similarity

为提升同杆双回线路保护算法的可靠性,利用同杆双回线路故障后近故障端前行波电流和远故障端反行波电流在波形上的关系,提出了一种基于波形相似度的同杆双回线路故障识别方法。首先利用原始数据计算线路两端的电压、电流暂态量;经相模变换解耦之后,选取同一模量计算两端对应的前行波电流和反行波电流;采用滑动窗的方法提取余弦相似度系数作为特征向量;经SMOTE方法平衡区内外故障样本数据后,引入概率神经网络(ProbabilisticNeuralNetwork,PNN)进行区内外故障识别。基于PSCAD/EMTDC的实验结果表明,该方法在多种故障条件下都能够准确识别出区内外故障,并且在高阻接地故障、噪声干扰和CT饱和等情况下也具有较为优异的表现。

In order to improve the reliability of the protection algorithm for double-circuit lines on the same tower,using the relationship between the forward wave current at the near-fault end and the back-travel wave current at the far-fault end after the failure of the double-circuit line on the same tower,a new method based on waveform similarity is proposed.Fault identification method for double circuit lines on the same tower.First use the original data to calculate the voltage and current transients at both ends of the line;after phase mode transformation and decoupling,select the same modulus to calculate the corresponding forward wave current and reverse wave current at both ends;use the sliding window method to extract the cosine similarity.The coefficient is used as a feature vector;after the SMOTE method balances the internal and external fault sample data,Probabilistic Neural Network(PNN)is introduced to identify the internal and external faults.The experimental results based on PSCAD/EMTDC show that the method can accurately identify internal and external faults under various fault conditions,and it also has excellent performance under high-resistance ground faults,noise interference and CT saturation.