Impacts of EV Charging on the Grid
Project team: Siobhan Powell, Gustavo Cezar, Liang Min, Ines Azevedo, and Ram Rajagopal

The electricity grid and transportation sector are undergoing simultaneous, rapid, and unprecedented transformations to reduce emissions. Coupled through electric vehicle charging, the two transformations can both hinder and support each other: the grid must provide electric vehicle drivers with reliable, affordable electricity and convenient access to charging stations; electric vehicle charging can impact the grid’s transformation in turn by increasing demand, accelerating equipment aging or forcing upgrades, aligning or misaligning with renewable generation, or even providing grid services. In this work, we focus on that coupling, by understanding what shapes electric vehicle charging demand and how it should be reshaped to improve the impacts on the electricity grid.
Studying drivers’ charging behaviour is the first step toward understanding charging demand. Electric vehicle charging behaviour is highly heterogeneous, shaped by individuals’ travel patterns, access to charging infrastructure, and personal preferences. We propose a novel methodology to include driver behaviour in a model of large-scale electric vehicle charging demand for applications in long-term planning. The methodology builds in knobs for future scenario design based on data-driven modeling of driver behaviour, clustering drivers and charging sessions. We calibrate the methodology using a large data set of nearly four million charging sessions from Northern California in 2019.
Charging control is a powerful tool widely used to modify charging profiles. Studying the connections between charging control, electricity rate design, and drivers’ charging behaviour is the second step toward understanding charging demand. We first investigate controlled charging at a small scale, studying the impact of workplace charging control for different electricity rate schedules on the aging of a distribution transformer. Then, we propose a novel methodology for representing such control in large-scale models of charging demand. The proposed methodology uses machine learning to directly model the mapping from uncontrolled to controlled aggregate demand.
Finally, we apply this understanding of how drivers’ charging behaviour, charging control, and access to charging infrastructure shape and reshape demand to study the future large-scale impacts on the electricity grid. We focus on the Western US, and model grid dispatch in 2035 under a range of charging scenarios to evaluate the effects of increasing or decreasing the deployment of home or workplace charging infrastructure and of the widespread deployment of charging control in response to different electricity rate schedules.
Publications
- Electric vehicle charging : understanding driver behaviour and charging controls to improve impacts on the electricity grid
S Powell
PhD Dissertation, Stanford University - Controlled workplace charging of electric vehicles: The impact of rate schedules on transformer aging
S Powell, EC Kara, R Sevlian, GV Cezar, S Kiliccote, R Rajagopal
Applied Energy 276, 115352 - Large-scale scenarios of electric vehicle charging with a data-driven model of control
S Powell, G Cezar, E Apostolaki-Iosifidou, R Rajagopal
Energy 248 - Charging infrastructure access and operation to reduce the grid impacts of deep electric vehicle adoption
S Powell, G Cezar, L Min, I M L Azevedo, R Rajagopal
Nature Energy 7 - Scalable probabilistic estimates of electric vehicle charging given observed driver behavior
S Powell, GV Cezar, R Rajagopal
Applied Energy 309, 118382