Robust Identification Algorithm for Distribution Network Topology Based on Mutual-information Bayesian Network

Based on the micro phasor measurement unit (μPMU), this paper proposes a new identification method of distribution network topology. The nonlinear relationship among distribution network topology, photovoltaic (PV), load and μPMU measured voltage is fitted by Bayesian network (BN). The interval of continuous nodes is measured by introducing the grid division of maximal information coefficient (MIC). MIC solves the problem that traditional BN needs to specify the interval number manually when processing continuous data, and is difficult to be applied to cases with lots of continuous variables. The photovoltaic-load data generated by the Latin hypercube sampling (LHS) can realize the uniform distribution of the scene in the sample space. The training process of BN is simplified and the generalization ability of topology identification is improved. The effectiveness of the method is verified by a simulation example. The simulation results show that the proposed method has the same identification accuracy and higher timeliness compared with the real-time estimation matching method. The identification time does not increase linearly with the increasing of the number of feasible topologies. The proposed method has good applicability. Even in the case of partial failure of the μPMU or lack of key data, such as load and PV data, the proposed method can ensure high identification effect, and thus has strong robustness. © 2021 Automation of Electric Power Systems Press.