Event Debrief: EV User Behavior Plays 'Pivotal' Role in Electrifying Transportation, Says Christine Gscwendtner

Key Takeaways

• Increased variability in the power system resulting from greater renewable energy generation and the ongoing electrification of transport is creating challenges for grid operators to balance supply and demand.
• Electric vehicles (EVs) can be a source of much-needed flexibility for the power system if users can be induced to optimize charging times and locations as a form of demand response.
• Current studies on demand response do not adequately account for heterogeneity in user behavior when it comes to driving and charging EVs. A portfolio of demand response solutions that takes into account different user behavior will provide the greatest amount of flexibility to the grid.

As the share of power generation from renewable energy continues to grow, so does the variability of the overall power system, challenging the ability of grid operators to consistently balance supply and demand. At the same time, the increasing electrification of transport is creating new sources of demand that exhibit their own, dissimilar load profile. Electric vehicles (EVs) are typically charged at night when drivers return home, creating significant demand at a time when the sun is no longer shining. Increasing electric vehicle sales and greater penetration of variable renewable energy (VRE) power sources are expected to exacerbate this mismatch of electricity supply and demand.

Solving this challenge will require greater flexibility in the power system. The behavior of EV users themselves could play a role in modulating electricity demand and providing that flexibility, said Christine Gschwendtner, Postdoctoral Research Fellow at Harvard Kennedy School’s Belfer Center, during a November 20 Energy Policy Seminar. While EVs can create difficulties for grid operators (for example, by creating high demand peaks at inopportune times), they can also be a source of support by providing flexible demand response on command. Compared to other potential sources of flexibility, such as supply, transmission, and greater energy storage, demand response could help smooth load requirements at a relatively lower cost. “Charging behavior is decisive for whether EVs are a challenge or a solution for the electricity system," said Gschwendtner.

Gschwendtner observed that the energy transition is creating a paradigm shift in the electricity system that places users at the center Historically, the power system has been unidirectional, with electricity flowing from producers to consumers, but today’s increasingly decentralized system arms consumers with their own power supplies and new digital tools to engage with the grid in novel ways. 

User behavior, said Gschwendtner, cannot be fully controlled. In the case of EVs, charging behavior and grid contexts are oversimplified in current models. As a result, current projections of the impact of demand response measures may be overoptimistic if approaches do not sufficiently account for user behavior, which can be highly variable.

To account for heterogenous charging behavior by EV owners, Gschwendtner developed an “agent-based model” that incorporates different behavioral patterns to evaluate their potential for providing flexibility to the grid. Using inputs that included driving behavior, charging behavior, and different macro scenarios for levels of EV adoption, Gschwendtner modeled various load profiles within different contexts. The model was built around four main factors that Gschwendtner found contributed most to heterogeneity in user behavior: battery state of charge, the time of day, charging station locations, and planned travel. Other factors, including charging cost, parking duration, and charging power were also found to play a role, but were less important. Perhaps unsurprisingly, these different factors contribute to different driving and charging behaviors that lead to different impacts on system flexibility. Adjusting each factor results in a different reduction in total peak power demand and amount of load shifted to other parts of the day, with the “state of battery charge” factor resulting in the highest reductions and load shifting across all scenarios.

While system planners hesitate to rely on demand response because of the heterogeneity in user behavior, Gschwendtner stressed that the agent-based model shows that portfolios of different demand-side flexibility sources, including a variety of technologies, can help to offset the variability in user behavior. Heterogeneous plug-in behaviors can help reduce peak demand by distributing load throughout the day and in different geographical locations. The specific needs of different grids and contexts should determine which combinations of behaviors and automated charging processes are used to encourage flexibility. Overall, Gschwendtner concluded that “integrating technical solutions with diverse behavior aspects is crucial" to positioning EVs as a source of flexibility rather than a source of inconvenience for grid operators.