Community-Based Electric Vehicle Charging Station Allocation Using Regional Customer Diversities

Failing to account for the diverse charging behaviors of electric vehicle (EV) users across different regions often leads to inaccuracies in planning community charging facilities, resulting in either overestimating or underestimating charger demand. This paper addresses this gap by proposing a customer segmentation-based approach for optimal charger distribution to assist policymakers and distribution network operators in the early stages of EV adoption. This work comprises two main parts: representing plugin behaviors through customer segmentation and optimal charger allocation in community distribution networks. Customer segmentation is conducted using an unsupervised clustering algorithm, with EV adoption features selected based on their importance scores using regional EV availability. Primary plugin behaviors, based on Australian contexts, consider arrival time, charger preferences, driving patterns, and initial state of charge, while secondary behaviors, such as plugin intervals and energy needs, are modeled using probabilistic simulations. Utilizing secondary plugin variables, this research employs a two-stage multi-objective optimization approach to find optimal locations and sizes of community chargers in the IEEE-33 bus system coupled with a road network. The results reveal that customer plugin diversities impact charger distribution plans, highlighting the critical need for customer segmentation accounting for regional variations.