复杂电力系统中的故障定位：通过战略性传感器布置克服电磁时间反演

The Electromagnetic Time Reversal (EMTR) method has been receiving a lot of attention in the last years as an effective method to localize faults in electrical grids by analyzing transient signals induced by the fault. In this paper we first underline some of the limitations of EMTR when applied in real networks. Then, we propose a new metric based on spectrum similarity between the recorded signal and the fault-to-sensor frequency response of the network. Unlike similar approaches in the literature, this metric requires no information regarding the fault instant. Afterwards, we argue the need of multiple sensors to address the limitations of fault observability in complex network topologies. A novel algorithm for optimal sensor placement is introduced, which uses observability indices to determine the most effective sensor locations. This ensures maximum fault detection coverage, even in networks characterized by significant characteristic impedance variations between overhead lines and underground cables. Extensive simulations on two real medium-voltage distribution networks consisting of 32 and 161 buses respectively, are performed to illustrate the effectiveness and scalability of the proposed approach.