Optimal allocation of distributed generation and electric vehicle charging stations based on intelligent algorithm and bi-level programming

To facilitate the development of active distribution networks with high penetration of large-scale distributed generation (DG) and electric vehicles (EVs), active management strategies should be considered at the planning stage to implement the coordinated optimal allocations of DG and electric vehicle charging stations (EVCSs). In this article, EV charging load curves are obtained by the Monte Carlo simulation method. This article reduces the number of photovoltaic outputs and load scenarios by the K-means++ clustering algorithm to obtain a typical scenario set. Additionally, we propose a bi-level programming model for the coordinated DG and EVCSs planning problem. The maximisation of annual overall profit for the power supply company is taken as the objective function for the upper planning level. Then, each scenario is optimised at the lower level by using active management strategies. The improved harmonic particle swarm optimisation algorithm is used to solve the bi-level model. The validation results for the IEEE-33 node, PG&E-69 node test system and an actual regional 30-node distribution network show that the bi-level programming model proposed in this article can improve the planning capacity of DG and EVCSs, and effectively increase the annual overall profit of the power supply company, while improving environmental and social welfare, and reducing system power losses and voltage shifts. The study provides a new perspective on the distribution network planning problem.