Power System Topology Identification Using Neural Networks Part I - Line Processing

摘要

Power system state estimation is a basic function in power system control and analysis. The topological structure of the power system changes during power system operation and must be checked before state estimation to avoid gross error during state estimation. This paper present a model for identifying anomalies in power system topology based on neural networks. The procedure is divided into two parts: nodes and branch processing. This paper focuses on branch processing. Several types of neural networks were tried out on different power system network configurations. The results obtained using multilayer percepton with backpropagation learning algorithm are presented and analyzed.