Two-Time Scale Reactive Power Optimization with Considering the Uncertainty of RESs and EV Charging Stations in Active Distribution Network

The high penetration of distributed energy sources such as renewable energy sources and electric vehicles has posed tremendous challenges to power networks. This paper proposes a two-time scale reactive power optimization strategy for active distribution network (ADN) with a high percentage of renewable energy sources, and also considers the supporting capability of the reactive power provided by electric vehicle charging stations (EVS) in the grid. In the first stage, with the reduce total network loss of the system buses being taken as a target, the improved particle swarm optimization is used for reactive power optimization in the 1-h time scale, and the settings schedule of on-load tap changers (OLTC) and switching capacitor(SC) is developed. In the second stage, the scenario generation and reduction are employed to deal with the uncertainty of renewable energy sources and electric vehicles at a 15-min time scale. This stage aims at minimizing the voltage deviation of the system buses and maintains the settings schedule of OLTC and SC in the first stage, further optimizing the reactive power output/input of RES and EVS. The Canopy-Kmeans based clustering method is able to reduce the influence of subjective factors in the selection of representative scenarios. The proposed two-stage reactive power optimization strategy was evaluated using the modified IEEE 33-bus distribution network, and the results of which verifies the effectiveness of the strategy.