基于频谱特性的HVDC区内外故障快速识别

高压直流输电（HVDC）系统区内外故障快速判别是实现故障保护的基础。由于直流滤波器、平波电抗器以及线路分布电容等作用下，区内外故障时，直流电流信号传播至保护安装处所经历的物理边界存在较大区别，导致直流电流频带能量分布具有明显差异。该文采用总体经验模态分解法（EEMD）对HVDC系统故障时的直流电流信号进行分解，将信号分解为几个固有模态分量（IMF）之和。然后，计算各模态分量的能量，通过对比分析，找出特征区别，进而定义能量比判据用于区分正常、区内、本侧区外以及对侧区外故障。该方法仅采用直流电流信号，简单易实现，能够满足快速性和选择性要求。最后，通过对HVDC多种工况下的大量故障进行仿真分析，验证了该方法的有效性和鲁棒性。

Fast Identification of Inside and Outside Faults of HVDC Based on Frequency Spectrum Characteristics

Fast identification of internal and external faults of high voltage direct current (HVDC) is the basis of protection. Due to the influence of direct current (DC) filter, smoothing reactor and line distribution capacitance, there exists a big difference of physical boundary which the DC current transmits through to the relay location when internal and external faults occur. Therefore, the energy distribution of DC current frequency band is obviously different. In this paper, the DC current signals are decomposed by the ensemble empirical mode decomposition (EEMD) method, and the signals are decomposed into the sum of several intrinsic mode components (IMF). Then, the energy of each modal component is calculated. By comparing and analyzing, the distinguishing features are identified, and then the energy ratio index is defined to distinguish the normal, internal fault, external fault on both sides. This method only uses the DC current signal, which is easy to implement, and can meet the requirements of rapidity and selectivity. Finally, the simulation and analysis of a large number of faults in HVDC under various operating conditions verify the effectiveness and robustness of the proposed method.