利用双重特征量鉴别变压器励磁涌流的高阶统计原理

励磁涌流是制约变压器差动保护成功率的主要原因之一，为此提出一种基于Jarque-Bera系数来识别励磁涌流的新策略。对变压器区内故障电流和励磁涌流波形的整体形态进行特征分析，利用两者波形正弦度不同所导致的正态性差异，结合Jarque-Bera算法在正态性检验领域中的优势，对变压器差动电流进行绝对值处理后求取杰 氏系数，根据该系数大小选取合适的阈值来辨识励磁涌流和故障差流。仿真测试了变压器各种工况下的差流波形。与二次谐波制动方案进行对比分析，表明所提方法原理清晰，且具备良好的抗干扰性能，即使在电流互感器饱和的情况下，也能迅速、准确地识别出励磁涌流。最后，通过现场试验录波数据分析，进一步证明了该方法的正确性。

High order statistical theory for identifying transformer inrush current by using dual characteristics

Inrush current is one of the main factors restricting the success rate of transformer differential protection. This paper proposes a new strategy for identifying the inrush current based on the Jarque-Bera coefficient. The overall morphology of the fault current and inrush current waveforms is analyzed. The difference in normality caused by the different sine degrees of the two waveforms is used. Combined with the advantages of the Jarque-Bera algorithm in the field of a normality test, the absolute value of the transformer differential current is processed to obtain the Jarque-Bera coefficient. Depending on the size of the coefficient, the appropriate threshold is selected to identify the inrush and fault differential currents. The differential current waveform of the transformer under various working conditions is simulated and tested. The comparison with the second harmonic braking scheme shows that the proposed method offers a clear working principle and good anti-interference performance, and can quickly and accurately identify the inrush current even when the Current Transformer (CT) is saturated. Finally, the correctness of this method is further proved by the analysis of field test wave data. This work is supported by the National Nature Science Foundation of China (No. 52077120).