Waveform dissimilarity factor-based protection for main transformers in wind farms

Current differential protection (CDP) is widely used as main protection for transformers due to its high reliability and selectivity. Owing to non-fundamental frequency current, magnetizing inrush and current transformer (CT) saturation, CDP may undergo slow operating speed when the internal fault occurs. When the external fault occurs, CDP may false operate. Furthermore, given the disadvantage of Hausdorff distance being extremely sensitive to abnormal data, a new metric that reveals the dissimilarity degree of two waveforms is presented, namely waveform dissimilarity factor (WDF). WDF has inherent immunity to abnormal data. Subsequently, WDF-based protection for main transformer is proposed herein. In order to reduce the adverse effects of magnetizing inrush and CT saturation and accurately identify internal faults, the reward and punishment mechanism is designed. The protection principle is elaborated and its operating characteristics are analyzed under various conditions. Finally, by constructing the DFIG-based wind farm, PSCAD is used to verify the advantages of the novel protection algorithm based on WDF. Its performance is widely assessed in various scenarios in detail, including magnetizing inrushes, minor and server internal faults, inter-turn fault, turn-to-ground fault, internal fault with CT saturation and external fault with CT saturation, data loss, abnormal data, and CT error. The results show that the novel protection scheme has the characteristics of fast, secure, and robust.