基于最小二乘支持向量机的变压器励磁涌流识别方法研究

Research of magnetizing inrush current identification method based on LS-SVM

纵差保护是变压器的主保护,但在变压器空载合闸或者变压器外部短路故障被切除端电压突然恢复时会产生励磁涌流,可能会导致纵差保护的误动作,这将严重影响变压器的安全可靠运行。为此提出了一种基于最小二乘支持向量机(LS-SVM)的励磁涌流鉴别方法。选择二次谐波含量和间断角作为输入向量,利用具有高斯核函数的LS-SVM建立分类模型,对励磁涌流进行识别。仿真结果表明,该方法鉴别励磁涌流和故障电流有很高的识别率。该方法为减少变压器的误动和拒动提供了一种新的思路。

Transformer differential protection is the main protection, but in the condition of the transformer no-load switching-in or when the removed external transformer short-circuit fault terminal suddenly restores voltage, magnetizing inrush current will occur, which may result in differential protection malfunction, thus will seriously affect the safe and reliable operation of transformer. Therefore, a method based on least squares support vector machine （LS-SVM） to distinguish magnetizing inrush current is proposed. Selecting the secondary harmonic content and the dead angle as input vectors, in addition, using LS-SVM with a Gauss kernel function to establish the classification model could identify magnetizing inrush current. Simulation results show that the identification rate of magnetizing inrush current and fault current is high using this method. The method provides a new idea to reduce the transformer malfunction and refuse operation.