Conventional methods for sizing EV charging infrastructure often rely on peak-day assumptions and simple heuristics. These approaches typically assess battery investment in isolation, with limited consideration for local grid import/export constraints, renewable integration, or EV arrival variability. Additionally, they often ignore potential revenue from market participation, such as flexibility services.

To address these limitations, this paper presents results from a novel Excel/Python-based modelling suite developed to optimise the sizing of EV charging infrastructure. The tool integrates local solar PV generation, on-site battery storage, and EV arrival and departure patterns. Using probabilistic simulation and a Mixed Integer Linear Programming (MILP) optimiser, it evaluates the economic performance of different infrastructure configurations under uncertainty, aiming to maximise profitability while operating within grid constraints.

This analysis focuses on several key questions: How do seasonal EV arrival patterns influence optimal infrastructure sizing? What are the primary drivers of profitability, including charger count, import/export constraints, and site location? What combinations of charger capacity, battery size, and PV installation provide the strongest economic performance? To what extent can curtailment strategies and participation in flexibility markets improve site viability?

The tool is applied to real-world EV charging data from Scotland, combined with synthetic PV generation profiles for multiple locations and time-series pricing data from United Kingdom (UK) flexibility markets. Scenarios model deployments of 1 to 40 chargers, each rated at 150 kW. A Monte Carlo analysis captures the sensitivity of economic outcomes to stochastic EV arrivals and solar output. Machine learning techniques are then used to identify key profitability drivers and inform robust design decisions under uncertainty.

Results show that smaller charging sites (e.g. ~10 chargers) with limited grid access are frequently uneconomic unless supplemented by additional revenue streams. Participation in UK flexibility markets can increase revenues by up to 50%. Curtailment during local peak usage periods can also enhance economic performance by reducing costly grid import peaks. Seasonality plays a critical role in determining infrastructure value - sites with concentrated demand during specific periods (e.g. holidays) may yield better returns when correctly sized.

Our findings demonstrate that integrating battery storage, PV generation, and arrival/departure profiles within a sizing framework is essential for designing economically robust EV charging sites.