Flexible DC distribution network fault protection method based on transient characteristics differences of fault current

Abstract To improve the rapidity and accuracy of fault detection in multi-terminal flexible DC distribution networks with modular multilevel converters. A protection method based on the transient characteristic difference of fault current is proposed. First, the parameters of the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) are optimized using the Aquila Optimizer algorithm. So that, the feature extraction effect is optimal. Secondly, the distinction in the magnitude of internal and external fault currents is deduced through detailed fault characterization. Thus, the energy moment eigenvalue of the fault current is used to construct the internal and external fault criterion. Then, depending on the distinction between the faulty and sound pole currents during internal fault. The ratio of the singular value of the positive and negative pole currents is taken to create the differentiation criterion of faulty pole. This protection scheme can reliably and sensitively recognize various types of short-circuit faults by only using DC line single-end current signals. Finally, the validity and robustness of this method are confirmed by simulating and analyzing under different working conditions.