Customer-centric meso-level planning for electric vehicle charger distribution

Abstract

This paper presents a framework for the meso-level planning of electric vehicle (EV) charger distribution, addressing customer-level plugin uncertainties and regional variations. To bridge the gap between nationwide macro-level targets and localized micro-level plans, a meso-level approach is essential, considering regional diversity, market dynamics, and customer-centric factors. To address these limitations in meso-level planning, this paper proposes a customer-segmentation approach coupled with a probabilistic Monte Carlo Simulation to identify charger requirements in the cities across the Australian capital. This paper employs a K-means clustering algorithm to classify EV customers, leveraging prioritized features derived from a Random Forest Classifier based on available models in regional markets. Regional uncertainties in EV plug-in behaviors are collected from Australian scenarios to capture the inherent variability in charging patterns and accurately reflect the local demographics influencing EV adoption and usage. The results indicate the need for installing 1239 regular and 235 fast-charging ports, with a projected 29.53 % increase in charging ports by 2040 due to population and vehicle sales growth. Additionally, comparing with existing reports reveals that neglecting customer plugin behaviors and vehicle types can lead to an inaccurate estimation of charger requirements, resulting in significant investment losses. The findings underscore the importance of incorporating customer segmentation and regional variations into EV infrastructure planning to optimize resource allocation without costly mismatches.