A Precise Location Method of Grounding Fault in Multi-branch Renewable Energy Distribution Network Based on Mask Mechanism

In the current power distribution system, the rapid growth of the number of loads and distributed power sources has changed the fault characteristics of the system, and the applicability of the traditional passive fault location method based on the fault characteristics of the single power source in the system has decreased. The active location method that generates controllable fault characteristics by injecting signals can effectively avoid the influence of distributed power sources, but it faces the problem of difficulty in calibrating the reflected signal of the fault point and the problem of pseudo fault points in complex topologies. In response to the above two problems, a precise location method for grounding faults in multi-branch new energy distribution networks based on the Mask mechanism is proposed. Inspired by the Mask mechanism in machine learning, the system information at the time of the fault is used to construct an auxiliary vector to operate the collected fault signal. By reducing the possible range of the fault point wave head, the difficulty of calibration is reduced and the generation of pseudo fault points is eliminated. In the PSCAD model built based on a real distribution network topology of a certain place, the effectiveness and the transition resistance capability of the proposed method under different fault conditions are verified. The simulation results show that the positioning error of the proposed method does not exceed 1%, which solves the two major problems faced by the active positioning method in the current multi-branch new energy distribution network.