应对变压器涌流影响的线路零序保护防误动策略

Anti-Misoperation Strategy of Line Zero Sequence Protection Against Transformer Inrush Current

随着高阻抗变压器的广泛应用,其特殊的绕组结构导致励磁涌流过大,现场多次出现变压器空投时线路零序过流保护误动。目前尚缺乏对这一问题的有效应对策略,尝试通过应用人工神经网络(artificial neuralnetwork,ANN)对变压器涌流进行识别,从而达到制动线路零序过流保护的目的。从人工神经网络的选取、训练样本数据的确定、对零序保护的影响等方面进行研究,提出基于ANN的变压器涌流识别判据,将ANN输出与PSCAD/EMTDC仿真进行对比,验证了判据的有效性,并讨论其应用方案。防误动策略避免了复杂的整定过程和计算过程,为智能算法在保护中的应用提供了一种可行路径。

With the wide application of high impedance transformer,its special winding structure leads to excessive excitation inrush current,and the zero sequence overcurrent protection mis-operation occurs many times in the field when the transformer runs without any load.At present,there is no effective strategy to deal with this problem.This paper tries to identify transformer inrush current by applying artificial neural network(ANN)so as to avoid the zero-sequence overcurrent protection mis-operation.The paper first makes a study on the selection of artificial neural network,the determination of training sample data and the influence of zero order protection,then proposes the transformer inrush current identification criterion based on ANN,and compares the ANN output with PSCAD/EMTDC simulation.Finally,the paper verifies the validity of the criterion,and also discusses its application scheme.The proposed anti-misoperation strategy avoids the complex setting process and calculation process,and provides a feasible path for the application of intelligent algorithm in protection.