The integration of electric vehicles (EVs) with a decarbonizing electricity grid is crucial for achieving US emissions reduction targets. Industry forecasts project a dramatic increase in light-duty EVs and charging plugs by 2035, potentially reaching hundreds of millions globally. However, this rapid expansion introduces significant challenges. The coupling of transportation electrification with the grid is currently uncoordinated, despite shared emission reduction goals. While the grid impacts of EV adoption have been analyzed at low to near-term levels, assessing and mitigating the consequences of deep EV penetration remains a critical challenge, demanding models capable of capturing the diverse behaviors and conditions of future drivers. This study addresses this gap by focusing on the long-term planning implications for daily charging demand under high electrification scenarios, considering the multifaceted roles of charging infrastructure, controls, and driver behavior. These factors significantly influence grid operations, making long-term planning complex. Driver behavior is highly variable, with charging location, timing, and frequency directly affecting the load shape and grid demand. Implementing charging controls and altering the landscape of charging infrastructure (increasing or decreasing the availability of charging options) offer powerful tools to manage charging demand and mitigate negative grid impacts. Charging controls, often termed smart or managed charging, can reshape demand by delaying charging or modulating power delivery throughout a charging session. The design and geographic distribution of the charging infrastructure network influence driver choices and system-wide demand by affecting charging location and time of day. Access to convenient charging is also vital to avoid user inconvenience which can hinder both EV adoption and continued use. Early EV adopters disproportionately consist of affluent single-family homeowners (SFHs) with access to home charging. This contrasts with lower-income households, renters, and multi-unit dwelling (MUD) residents, who often lack home charging access despite targeted subsidies. Assuming future charging infrastructure use mirrors early adopter behavior would misrepresent future driver options and potentially overlook valuable opportunities for households, utilities, and regulators.

Existing approaches to modeling large-scale charging demand often rely on early adopter behaviors or modeler assumptions about driver behavior. Numerous studies have explored the use of charging controls to improve grid impact and EV costs, but these studies frequently have limitations. Many studies employ limited scenarios regarding charging infrastructure access, focus on centrally optimized controls instead of site-specific optimization, neglect rate schedule-driven optimizations, focus on current grid resources and conditions, or exclude grid storage and emission calculations. Previous research examining different charging infrastructure scenarios often concentrated on early adopters and failed to conceptualize infrastructure as a charging control tool. The importance of charging infrastructure for long-distance travel and high-energy days to support EV adoption has been addressed in recent studies, emphasizing the need for comprehensive models and policy considerations.

This study models daily EV charging demand under high electrification scenarios in 2035 for the US portion of the WECC grid, encompassing 11 states. The research employs a data-driven, probabilistic method to capture driver charging preferences based on real charging data, calibrating the model using a large dataset of charging sessions. The model links charging behavior clusters to driver income, housing, travel distance, and charging access. Controlled charging is implemented site by site, simulating realistic responses to electricity rates. Two primary strategies (control and infrastructure build-out) are analyzed, employing detailed models of driver behavior, control mechanisms, and grid dispatch. The study focuses on aggregate charging patterns of personal light-duty vehicles to assess generation-level grid impact. To calculate the grid impact, the aggregate electricity demand for an entire year is dispatched to a model of future grid generation resources, reflecting forecast retirements and additions of fossil fuel generators, along with increased wind, solar, and grid storage. Wind and solar generation variability is based on 2019 data. Baseline annual electricity consumption is projected to increase by 16% by 2035 due to electrification in sectors beyond transportation. Four scenarios for future charging infrastructure are modeled, varying home charging access from universal to low based on survey data. Four types of conventional charging control are modeled: SFH timers set for 9 p.m. and 12 a.m. start times, and site-level, unidirectional load modulation control at workplaces to minimize demand charges or average grid emissions. A "Business As Usual" scenario represents the current mix of control strategies. A detailed model of the grid in 2035, based on capacity expansion planning outputs, is used to calculate the grid impact at the generation level. This model incorporates announced generator retirements and additions, increased baseline demand, and increased wind, solar, and grid storage. The study analyzes both total and net demand to assess grid impacts across different electricity generation sources. The study assesses the effect of increased EV penetration on several metrics: Peak total and net demand, grid storage requirements, ramping of fossil fuel generators, and grid emissions.

The study reveals significant increases in electricity consumption with deep EV adoption. At 50% adoption, annual consumption increases by 5% over the 2035 baseline, reaching a total increase of 22% over 2019 levels when considering all electrification sectors. Full electrification (100% adoption) results in an 11% increase from EVs and a 28% overall increase over 2019 levels. The timing of this increased electricity use is crucial, as grid impacts vary significantly with different demand profiles. The addition of EV charging to the baseline 2035 demand has the effect of shifting the peak demand on a typical weekday from 5 p.m. to later in the evening in several home charging scenarios, or to mid-morning in daytime charging scenarios. High home charging pushes the peak to 7 p.m., while daytime charging creates mid-morning peaks. With 50% EV adoption, the increase in peak total demand ranges from 3% to 9%, while 100% adoption leads to a 9–26% increase. Daytime charging scenarios generally increase peak demand more than high home charging scenarios, except for scenarios with 9 p.m. timers. Analysis of net demand (total demand minus non-fossil fuel generation) shows a different impact: home charging scenarios exert greater stress on the remaining fossil fuel generators than daytime charging scenarios. Due to high solar generation during the day, peak net demand consistently occurs in the evening. The Business As Usual scenario increases peak net demand 1.6x more than the Low Home, High Work scenario with 50% EVs, and 1.8x more with 100% EVs. The Universal Home access scenario with 9 p.m. timers increases peak net demand by 3.3x or 3.4x. The study finds that 10 GW of storage suffices to support at least 50% EV adoption. The grid supports more EVs with greater daytime charging and fewer EVs with higher home charging. In the best-case scenarios (Low Home access, Business As Usual, or High Home access with midnight or random timers), the grid can support 100% EV adoption. The worst-case scenario (Universal Home access and 9 p.m. timers) supports only 59% EV adoption. Charging controls, particularly 12 a.m. SFH timers and randomized timers, substantially increase the level of EV adoption the grid can support. The minimum 4h grid storage needed varies widely between scenarios, with daytime charging scenarios requiring considerably less storage than high home charging scenarios. Switching from Business As Usual to Low Home, High Work access could reduce storage costs by US$0.7 billion to US$1.5 billion with 50% adoption and US$1.6 billion to US$3.4 billion with 100% adoption, depending on storage cost forecasts. Even with 10 GW of grid storage, substantial 1h ramps in demand for fossil fuel generators occur, impacting grid reliability. Home charging increases ramping, whereas daytime charging decreases it. Random and 12 a.m. SFH timers can reduce ramping in some scenarios, but changing charging access has a larger effect than adding control. In scenarios with 50% EV adoption, total annual excess non-fossil fuel generation decreases, most rapidly in scenarios with more daytime charging. The added grid CO2 emissions per mile of EV charging are substantially lower than those from internal combustion engine vehicles (ICEVs), demonstrating a 4x or 3x improvement depending on the comparison (ICEV average or sedan, respectively). Daytime charging scenarios yield lower CO2 emissions per mile than home charging scenarios. Sensitivity analysis shows that the main conclusions remain robust across variations in grid capacity (solar, wind, gas, and coal generation).

This study's findings highlight the pivotal role of charging infrastructure in improving EV grid integration. In future grids with increased renewable generation, the timing of charging becomes paramount. Shifting charging from home to daytime significantly improves grid impact metrics, including ramping, non-fossil fuel generation utilization, storage requirements, and emissions. This is robust across varying EV adoption levels. The results strongly support expanding daytime charging access; restricting home charging could negatively impact adoption and equity. Policymakers must prioritize convenient, affordable, and widely available public daytime charging options. The study reveals a conflict between system-level and site-level benefits. Peak minimization control, while common at commercial sites, increases late-afternoon demand, straining the grid at the generation level. Balancing these objectives requires further research. Workplace control designed to align charging with low average grid emissions doesn't yield meaningful reductions due to high variability in generator dispatch order and marginal emission profiles. Future electricity rates should better align with wholesale prices and account for daily grid generation conditions. Different assumptions about baseline demand and generation resources could influence the relative benefits of daytime versus nighttime charging. Seasonal effects could also impact regional results. Future research should explore the coupling of EV charging with other electrification sectors and different grid decarbonization pathways. However, the time of day of charging remains a critical factor.

This research demonstrates the potential of charging infrastructure to enhance EV grid integration. Shifting charging to daytime significantly improves various grid impact metrics. Policymakers should prioritize convenient and widespread daytime charging options. While some charging controls offer benefits, others have substantial negative consequences. Future research should address the trade-offs between distribution and generation-level impacts and design better-harmonized electricity rates. The strategic deployment of charging infrastructure is a powerful tool for mitigating the grid impacts of deep EV adoption.

This study models only personal, light-duty vehicles, excluding commercial vehicles. Seasonal effects due to temperature changes are neglected. The model does not represent transmission, interconnection, or congestion, so curtailment or export of excess generation is not explicitly modeled. The study relies on projections of future grid resources and baseline demand, which are subject to uncertainty.