Optimization of orderly charging strategy of electric vehicle based on improved alternating direction method of multipliers

When it comes to solving the optimization problem of wind power and large-scale electric vehicle (EV) syner-gistic access to the power grid, issues like the difficulty of accurately predicting wind power, the massive amount of information communication, the difficulty of handling EV constraints, and the privacy leakage of EV owners are critical. A distributed framework based on power grids, EV load aggregates, and single EV is established in this paper. Based on this framework, an EV rolling real-time optimization model considering wind power forecast errors and EV forecast information has been established. An improved alternating direction multiplier approach is utilized to solve the model, and a small amount of information is transmitted between the two agents in each iteration to achieve the parallel solution of each agent's optimization goals. Without requiring global coordi-nation, the proposed architecture can achieve the goal of synchronous iteration. Finally, an example of a regional power grid is used, with EVs divided into three regions. The improved ADMM algorithm now has seven mul-tipliers, and the problem is solved, with the load variation, wind power consumption, and EV before and after optimization compared. To evaluate the effectiveness and convergence of the improved ADMM algorithm, as well as the impact of prediction information, electricity power sales price, and vehicle owner preference coef-ficient on dispatching outcomes, the operator's operational cost and the EV owner's charging satisfaction are examined. The example's simulation results show that the developed model achieves the goal of cutting peaks and filling valleys to absorb wind power, reducing operating costs and improving EV owner satisfaction.

Automated summary

This paper proposes a method for solving the optimization problem of wind power and large-scale EVs synergistic access to the power grid based on a distributed framework. By establishing a rolling real-time optimization model and employing an improved ADMM approach, the parallel optimization of each agent is achieved. The effectiveness and feasibility of the method are verified through example simulation.