Bi-Level Control Strategy for EV Charging in Active Distribution Network Based on Wavelet Decomposition

Considering the increasing integration of distributed renewable energies (DREs), the active distribution network (ADN) is facing challenges to maintain generation-load balance. Meanwhile, since the electric vehicles (EVs) are usually charged together in time and space, the EV charging is typical impulsive load and may further aggravate power imbalance. A bi-level EV charging strategy is proposed to reduce the adverse impacts of DRE power fluctuation and EV power impulse on ADN. In the outer level, a multi-objective optimization model is provided based on the predicted DRE power. With the objectives of minimizing net power fluctuations and EV charging power changes, the optimal Pareto solution set is obtained through the multi-objective particle swarm optimization algorithm, on which a basic EV charging scheme in the large time scale is determined. Then, in the inner level, the wavelet decomposition is utilized to extract the high-frequency power prediction errors. According to the high-frequency components and the state of charge of EVs, the charging scheme is adjusted in a small-time scale.