Distribution System State Estimation based on AP Clustering and Least-square Algorithm

A state estimation method for distribution network with bad data is proposed. The horizontal and vertical similarity of the load are used to identify the bad data based on the affinity propagation (AP) clustering algorithm. Then, the bad data is changed reasonably so that it can be used as normal data. Based on the weighted least-square method, a mathematical model for solving the state estimation of the distribution network is established, which is insensitive to the initial value selection and can better meet the state estimation requirements of the distribution network. Finally, the effectiveness of the method proposed in this paper is verified by using actual grid load data and IEEE14-bus examples.