Predictive Control of EV Overnight Charging with Multi-Session Flexibility

TL;DR

The paper tackles reducing CO$_2$ emissions from domestic overnight EV charging by enabling multi-session planning with model predictive control (MPC) over horizons up to seven days. It formulates a predictive net-carbon objective $oxed{ ilde{J}_0}$ and discretizes the forecast window, enforcing SOC bounds and daily energy needs while minimizing emissions. Long-range carbon forecasts are generated by adapting NESO data with CarbonCast-based error scaling, including a linear growth term $ ilde{oldsymbol{ u}$ for forecast error and probabilistic session timings around typical commuting hours. Across year-long simulations and regional UK data, multi-session MPC achieves substantial emission reductions (up to ≈46% relative to uncontrolled charging), with effectiveness modulated by driving/charging patterns and local grid carbon intensity; the work also discusses emission-trading market implications and the potential need for aggregators when scaling to fleets.

Abstract

The majority of electric vehicles (EVs) are charged domestically overnight, where the precise timing of power allocation is not important to the user, thus representing a source of flexibility that can be leveraged by charging control algorithms. In this paper, we relax the common assumption, that EVs require full charge every morning, enabling additional flexibility to defer charging of surplus energy to subsequent nights, which can enhance the performance of controlled charging. In particular, we consider a simple domestic smart plug, scheduling power delivery with the objective to minimize CO$_2$ emissions over prediction horizons of multiple sessions -- up to seven days ahead -- utilising model predictive control (MPC). Based on carbon intensity data from the UK National Grid, we demonstrate significant potential for emission reductions with multi-session planning of 40 to 46\% compared to uncontrolled charging and 19 to 26\% compared to single-session planning. Furthermore, we assess, how the driving and charging behaviour of EV users affects the available flexibility and consequentially the potential for emission reductions. Finally, using grid carbon intensity data from 14 different UK regions, we report significant variations in absolute emission reductions based on the local energy mix.

Figures (5)

• Figure 1: $C_{gen}[l]$, $l\in\{1,\ldots, 48 \times 7\}$ (black line): actual half-hourly carbon intensity measurements for UK National Grid electricity generation during the first week of January 2022, published by the National Energy System Operator (NESO) nationalenergysystemoperator; $\hat{C}_{gen}^{(1)}[l]$ (black dotted line): NESO's published sequence of one-step-ahead forecasts; $\hat{C}_{gen}[l]$ (blue line): our $l$-step-ahead forecast sequence (\ref{['eq:forecastCgen']}), estimated by transferring CarbonCast side-information maji2022.
• Figure 2: State of Charge (SOC) profiles over six simulation days for three charging strategies: uncontrolled (solid blue), MPC with a 1-day horizon (dashed blue), and MPC with a 4-day horizon (dotted blue). The black line shows the carbon intensity signal (C$_{\text{gen}}$). Gray background indicates night charging windows, during which the vehicle is plugged in; The figure highlights how predictive strategies shift charging to periods of lower carbon intensity.
• Figure 3: Cumulative CO$_2$ emissions over six simulation days for the same three charging strategies. Emissions increase stepwise with each charging event. MPC strategies result in significantly lower cumulative emissions.
• Figure 4: Impact of user flexibility on average carbon intensity ($C_{EV}$) during one month of simulated EV charging (January 2023, UK national carbon intensity data). Each gray bar represents a different daily plug-in time window, ranging from 20 to 4 hours. Four lines correspond to daily energy demands of 5, 10, 20, and 30 kWh. Shorter plug-in durations and higher energy demands both increase average carbon intensity. Daytime charging scenarios (right side) typically result in significantly higher emissions than overnight charging at comparable plug-in durations.
• Figure 5: Regional variation in average carbon intensity ($C_{EV}$) of EV charging across 14 UK regions during January 2023. Results are shown for uncontrolled charging and three predictive MPC strategies with forecast horizons $N = 1, 2, 4$ days. Smart charging reduces emissions in all regions, with relative reductions ranging from 16% to 83.9%. Longer prediction horizons yield larger reductions across regions.