Capacity Expansion Optimization of Determined Charging Stations Based on Accessibility Analysis and Improved Discrete Choice Model

Abstract

Reasonable capacity expansion optimization of determined charging stations is important for the popularization of electric vehicles (EVs). In this article, Gaussian two-step floating catchment area method and temporal clustering are adopted to study the unbalanced spatial and temporal distribution of charging station accessibility, as well as to determine candidate sites for capacity expansion optimization. Then, a two-level optimization model for capacity expansion of determined charging stations is proposed to reduce the computational complexity caused by the non-linearity and non-convexity of choice probability formulas when the discrete choice model is applied to statistically describe and analyze discrete behaviors such as users’ EV purchase preferences and charging station heterogeneity. In the proposed method, the lower-level model is described as a maximal coverage location problem by expressing the error term of the utility function as a linear combination of the random vectors from IID normal distribution and IID Gumbel distribution, which effectively simplifies the computing process. Finally, the two-level model can be transformed into an integer linear programming problem to optimize the capacity of determined charging stations. Experimental results show that the purchase rate of EVs is significantly improved, and the accessibility of charging stations during rush hours is more balanced than before.