Will the grid be ready to support widespread EV charging in the western US? The roles of charging infrastructure access and operation

Published in Sustainability
Like

Share this post

Choose a social network to share with, or copy the shortened URL to share elsewhere

This is a representation of how your post may appear on social media. The actual post will vary between social networks

Electric vehicles (EVs) are central to many plans for reducing emissions. EVs couple transportation to the electricity system, and this coupling presents both challenges and opportunities for the future grid. As the transition to EVs depends on a reliable, affordable supply of electricity, grid planners are carefully preparing to support this increase in demand. At the same time, the grid is undergoing rapid changes to decarbonize and is facing worsening impacts of climate change.

In California, both sides of this story were in the news recently. On August 25th, regulators at the California Air Resources Board (CARB) passed a rule ensuring all new car sales from 2035 will be of zero-emission vehicles (link to the announcement). Then, just days later, the California state grid operator issued a series of “Flex Alerts” urgently asking consumers to help the grid get through an extreme heat event by conserving electricity between 4 and 9 p.m (link to the alert).

With the addition of millions of new EVs will these grid issues be exacerbated? or can flexibility in charging demand limit any negative grid impacts?

In our paper recently published in Nature Energy,  we model the Western US grid in the year 2035 to test its readiness to support deep EV adoption under a range of scenarios for future EV charging demand.

My PhD was all about the impact of EV charging on the electricity grid, and our analysis leading up to this paper developed new models for two main elements driving EV charging demand: (1) EV driver behaviors and interactions with charging infrastructure and (2) automated charging control (also called smart charging). Our modeling found that each had a powerful effect on the timing of electricity demand from EVs, which prompted the questions: how do different scenarios of charging behavior and control interact with long-term grid planning? and which scenarios of demand should we aim for?

In this study we developed 25 different scenarios to compare these types of flexibility: we varied future drivers’ access to charging options from universal at-home charging access to scenarios dominated by public or workplace charging and we modeled the impact on their decisions of where/when to charge; we modeled the timer-based response of at-home charging to three different residential electricity rate schedules; and we modeled the response of workplace charging to two commercial rate schedules, one encouraging peak minimization and one aiming to align charging with low average grid emissions. These scenarios were simulated using a detailed, data-driven model of driver charging behaviour that identified and scaled-up patterns observed in a large dataset of nearly 3 million charging sessions from 2019 in California.

We coupled this with an open-source dispatch model of generation for the US portion of the Western Interconnection grid (WECC) to estimate the emissions and generation-level impacts of each scenario for charging demand in 2035. The model of fossil fuel generators, developed by Professors Inês Azevedo and Thomas Deetjen (Available at: https://doi.org/10.1021/acs.est.9b02500), creates a cost-based merit order of the generating units and dispatches them in order of least cost to meet demand at each hour. We extended this model to include future levels of non-fossil fuel generation and grid storage, and to reflect announced generator retirements or additions.

Figure 1. a, An overview of the modelling approach. To study the grid impacts of EV charging scenarios, charging demand was simulated for each region using a model of driver behaviour, regional profiles were aggregated, and grid dynamics were modelled including non-fossil fuel generation, storage and the dispatch of fossil fuel generators.  b, The model for EV charging demand in each region as a function of neighbourhood characteristics, access to charging and driver behaviours. Please refer to the paper for more details: https://www.nature.com/articles/s41560-022-01105-7. 

Main Result: Charging Infrastructure

We found that scenarios with more daytime charging and less at-home charging infrastructure lead to better results by all the measures of grid impacts we modeled: generation capacity, storage requirements, ramping, use of non-fossil fuel generation, and emissions.  

Access to convenient charging stations is critical to encouraging and sustaining EV adoption and we do not suggest limiting home charger installations. Rather, we recommend a focus on building abundant workplace and public charging stations throughout the Western US to give both current and future EV adopters convenient access to daytime charging.

Using charging infrastructure policies as a way to improve future grid impact is a powerful idea. Our results showed that it can be a more impactful tool for improving grid impact than adding automated control. We recommend that policymakers and other modelers add infrastructure build-out to their planning as a tool for long time-scale charging control.

Main Result: Charging Control

Our analysis of the different automated charging control strategies yielded notably mixed results.

In the case of single-family home timers, the deciding factor was the alignment of the timer start time (the start of the lowest price period in the electricity rate) and high non-EV baseline demand. 9 p.m. start times created a big spike in peak net demand and forced the need for large amounts of grid storage, while 12 a.m. start times delayed charging past the period of high non-EV demand and reduced the need for grid storage.

The results for workplace control revealed a challenging tension between distribution- and generation-level grid impacts. Peak minimization control implemented in response to demand charges, which is best at protecting distribution system equipment, caused problems at the generation level because it pushed workplace charging toward the late-afternoon/early-evening crunch time with high non-EV demand.

Finally, workplace control designed to minimize average grid emissions did not substantially reduce annual emissions in our simulation. Average grid emissions are lowest in the middle of the day in the western US, thanks to large amounts of solar generation, and optimizing for this signal condensed workplace charging demand into peak solar hours. On days with excess solar generation this change helped reduce grid emissions, but in 2035 those days were still relatively rare. On days where all non-fossil fuel generation was already being used, this shift increased midday demand for fossil fuel generators and unfortunately, as average and marginal grid emissions were not aligned, the fossil fuel generators on the margin which were used to meet that added demand were often dirtier than average.

Conclusions and Future Directions

As most control schemes being implemented today aim to reduce site-level or driver-level costs, the design of rate schedules that can lead to real emission reductions and better grid impacts is a critical challenge going forward. We think it is also critical, however, that policymakers look beyond rate design and automated control to consider the role of charging infrastructure policies. 

In the Western US we showed the benefit of expanding daytime charging, but the best charging option could be different in other regions with different grid and driving conditions. 

As millions of new charging stations are rolled out over the next decade, their role in shifting driving behaviors and determining large-scale grid impacts should be carefully considered.

Link to the paper: https://www.nature.com/articles/s41560-022-01105-7 

Citation: Powell, S., Cezar, G.V., Min, L. et al. Charging infrastructure access and operation to reduce the grid impacts of deep electric vehicle adoption. Nat Energy (2022). https://doi.org/10.1038/s41560-022-01105-7 

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in