基于HHT的抗CT饱和变压器故障识别新方法

New method to identify transformer fault with anti-CT saturation based on HHT

针对电流互感器(CT)饱和影响变压器差动保护准确性和速动性的问题,提出一种基于希尔伯特黄变换(HHT)的抗CT饱和变压器故障识别新方法。该方法依据CT饱和情况下变压器区外故障电流的波形存在突变点,而区内故障电流的波形是连续的这一特点。首先将信号进行经验模态分解得到一组不同特征尺度的本征模态函数,其次对最先分解出的本征模态函数进行Hilbert变换得到瞬时频率,瞬时频率的突变点反映出电流信号的突变,根据实时监测瞬时频率两个相邻突变点之间的时间间隔可以实现对变压器区内故障、区外故障以及由区外故障转换到区内的转换性故障的准确识别。仿真实验表明,该方法能在12 ms内快速准确地识别出CT饱和情况下的变压器的各种故障。此外,该方法还能闭锁励磁涌流。

Aiming at the problem that CT saturation will affect transformer differential protection＇s accuracy and speed, this paper proposes to use Hilbert Huang Transform （HHT） to identify transformer fault under CT saturation condition. The waveform of transformer＇s external fault current contains some catastrophe points; in contrast, internal fault current＇s waveform is continuous. The new method is just based on this waveform characteristic. Firstly, a group of intrinsic mode functions （/MF） which are different in characteristic scale are obtained from Empirical Mode Decomposition （EMD）. Then, Hilbert transform is applied for the first decomposed IMF to get the instantaneous frequency. Instantaneous frequency＇s catastrophe points just reflect current＇s catastrophe points. According to time interval between the two adjacent catastrophe points, it can identify transformer＇s external fault, internal fault and transferring fault accurately. PSCAD simulation results verify that under CT saturation condition, this method can identify all kinds of faults of transformer in 12 ms. Besides, it can block the inrush current.