将形态学数字滤波器作为变压器保护的前置滤波单元,在保留故障信号特征的前提下,最大限度地抑制各种干扰的影响。在此基础上,提出一种基于电流采样值相关度的变压器保护新判据,利用变压器两侧电流采样值波形的相关度来区分内部故障和外部故障。在详细分析2种相关度计算公式对新判据影响的基础上,进行了大量的仿真实验。实验结果表明:该判据能够快速、可靠地判别变压器内外部故障。由于借鉴了传统比率差动的定值整定思想,使得新判据定值的整定简单易行,同时基于这种方法的判据,对频率偏移和谐波的影响也有一定的抵御能力,与其他励磁涌流识别判据相配合,可构成一种较理想的变压器主保护方案。
In this paper, the morphological filter is utilized as the filtering unit of the transformer protection. The characteristics of the fault signal can be preserved and the noise interferences are suppressed. Based on this, the paper presents a new transformer protection criterion to distinguish internal faults and external faults by correlation coefficients of sampled current values. After detail analysis on influence on the new criterion of the two formula of correlation, simulation results show that one formula of correlation criterion can distinguish internal fault from external faults very fast and reliably. Based on setting scheme of ratio-restraint differential protection, the setting scheme of the new criterion is simple and feasible. Besides, the new criterion can effectively restrain the interference of harmonics and frequency variation. Cooperating with the criteria to indentify inrush current of transformers, an ideal protection scheme is presented.