Inrush Current Identification Based on Mathematical Morphology and LSSVM

Differential protection has long been used as the main protection of transformers because of its fast action and high sensitivity. Inrush current may be identified as fault current and lead to maloperation of the differential protection. Moreover, facing the more and more complex power system, the existing methods still have some shortcomings, and the discrimination rate is low in the case of current transformer (CT) saturation and large remanent magnetization. In this paper, the waveform features are obtained by the mathematical morphology gradient (MMG) operator. Then the morphological pattern spectrum (MPS) is used to extract the differences between inrush current and fault current. Finally, the normalized spectrum value and gradient criterion are both used as the input of the least-squares-support vector machine (LSSVM) to classify the inrush current and fault current. The accuracy of the proposed method is verified on the data generated by a simulation model established in Matlab/Simulink. Simulation results indicate that the proposed method has a good classification accuracy under large data set and will not be affected by residual magnetization of the transformer and CT saturation.