Data-centric Simulation Model for Electric Vehicle Charging Station Load Profiling and V2G Controls

The Electric Vehicle (EV) ecosystem significantly benefits from simulations that provide precise operational insights grounded in real-world data. This research proposes a data-centric simulation paradigm to generate 24-hour load profiles of Electric Vehicle Charging Stations (EVCS). Specifically designed to serve as high-fidelity inputs for real-world simulators and digital twins, these profiles model charging execution, queuing delays, user responses to varying conditions, EVCS operational metrics, and Vehicle-to-Grid (V2G) interactions. The paradigm utilizes an engineered real-world dataset derived from public charging operations within the Western Electricity Coordinating Council (WECC) grid regions. The framework integrates five interdependent stochastic modules: (i) an extended Non-Homogeneous Poisson Process (NHPP) utilizing cubic spline-interpolated thinning to model continuous diurnal arrival variations; (ii) a capacity-constrained M / G / c queuing system incorporating balking and reneging mechanisms to model user decisions better; (iii) an hourly method-of-moments bounded log-normal formulation preserving charging dynamics, time-sensitive variations, and skewness with greater fidelity than static fits; (iv) an empirical power rate formulation constrained by EVCS capacities for Grid to Vehicle (G2V) charging ; and (v) introducing scenario based V2G control models enforcing V2G discharging and State of Charge (SoC) safeguards. The synthetic outputs are validated directly against the empirical dataset baseline ( R 2 = 0.8838, \(MAE = 0.83% \) , and two-sample Kolmogorov-Smirnov p = 0.6216 > 0.05), exhibiting strong distributional consistency, and benchmarked against state-of-the-art literature models. Furthermore, scaling to a 50% EV adoption scenario on the IEEE 315 MW WSCC 9-bus benchmark system reveals a peak demand of 144.99 MW, maintaining a 53.97% transmission headroom margin while embedded V2G controls effectively mitigate ramps and peak demands. The work establishes a robust data-centric CPS-based framework for EV-grid integration and V2G application research.