State Estimation Of Distribution System With Distributed Generations Based On PMU Synchronous Information

With the continuous enhancement of automation system and management system for distribution system, the importance of state estimation for distribution system becomes increasingly prominent. Distributed Generations connected to the network changes the structure and trend of the distribution system, so it is necessary to make special treatment during state estimation. Phasor measurement unit (PMU) has phase angle measurement added to the traditional measurement data in the configuration of power system and the measurement data provided by it can help improve the accuracy of estimation. The main content of this paper is as follows:Approaches to estimate the current distribution system state are summarized and are compared to analyze their merits and demerits as well as applicability. After PMU measurement is introduced, mixed measurement state estimation model is established based on the combination of limited PMU measurement and traditional SCADA measurement. Depending on different categories, the distributed power connected to the distribution network is individually equivalent to four different kinds of node types to be processed. As for the optimal configuration of PMU, when the system can be fully observed, merits and demerits of current optimal configuration algorithm are analyzed. Considering it is difficult to realize full observation of system, optimal configuration algorithm to determine the PMU configuration position is proposed based on the number of outgoing lines and injection power of the bus, and its validity is verified by simulation in IEEE-57 node case. Since there is bad data in actual measurement data, approaches to detect and identify the bad data in current state estimation are summarized, and current approaches are compared. Aiming at the measurement data including gross error, one improved method of robust least square method is proposed in this paper. Identification classification is performed based on residual error of measurement data in combination with k-means clustering principle, and weight factor in the state estimation is changed accordingly so that the influence of measurement data including gross error on state estimation accuracy is reduced. MATLAB software is used to make simulation verification in IEEE-33 node distribution system. The simulation results prove that the approaches proposed in this paper can restrict effectively the influence of gross error on the state estimation accuracy so that the state estimation accuracy is improved.