利用电流波形特征识别变压器励磁涌流和故障电流

提出一种利用波形特征快速区分变压器励磁涌流和内部故障电流的新方法。该方法基于涌流波形畸变严重，会呈现出尖顶波的凹弧特征，而故障电流基本保持正弦特征的思想，首先找出10ms数据窗内采样值的绝对值取得最大值的点，然后计算该点两侧相邻的2个采样步长组成的大梯形面积与2个单采样步长组成的小梯形面积和的差的总和来区分变压器励磁涌流和内部故障。动模实验表明：该方法能够快速切除变压器内部故障，动作时间约为10ms，所需时间较短，识别灵敏度高，并且实现方便，不受电流非周期分量的影响。

New Scheme To Discriminate Magnetic Inrush Current from Internal Fault of Power Transformer Based on the Characteristic of Current Waveform

A new algorithm based on the waveform characteristics is presented for the discrimination of the magnetic inrush current from the fault current of a power transformer. The method is implemented based on the principle that the waveform of magnetic inrush current seriously distorts while the fault current nearly keeps sinusoid. Taking the sample data of within 10 ms after protection starts, the proposed method finds the point whose absolute value reaches maximum, and calculates the difference around the peak between area of big trapezium which is made of two sample intervals and that of two small trapezium which is made of a sample interval. According to the sum of difference, the discrimination between magnetic inrush current and fault current of the power transformers can be made. Experimental results indicate that the method is not only effective under different fault conditions and fault clearance time being limited to 10 ms, but also can be realized easily and hardly affected by decaying DC components.