One day, the demand for electricity to charge electric vehicles could overwhelm the power grid, unless the electricity sector is ready for the challenge.
With a growing fleet of electric vehicles on the road, grid planners depend on accurate estimates of load patterns to calculate electricity demand. A team of Stanford University researchers have assembled a scalable probabilistic model for load demand that can be applied to a flexible range of populations and account for a wide range of factors. In California, the model found that by 2030 – in a scenario where most EV owners choose to charge their vehicles when they drive home every night – peak charging demand would more than double higher than if drivers charged all day at home, work and public stations. By 2030, electric vehicle charging will account for a significant share of electricity demand in advanced economies. Therefore, keeping peak demand as low as possible would reduce the need for new generators and transmission lines.
“We wanted to create a model framework for long-term planning that captures real-life driver pricing patterns and accounts for uncertainty,” said Ram Rajagopal, lead author of the study, which was published in the March 1 edition of Energy appliedRajagopal is an associate professor in the Department of Civil and Environmental Engineering at Stanford.
Today’s 7 million electric vehicles worldwide are expected to grow to 400 million by 2040. In order to sustain this nearly 60-fold increase, the world must make substantial updates to the infrastructure supporting electric vehicles, including generation capacity, transmission and distribution, smart grid technologies and around 300 million easily accessible charging stations.
The shift from gasoline-powered vehicles to electric vehicles is key to decarbonizing human activity. Opportunities for drivers to charge their electric vehicles need to be convenient and plentiful for this change to happen. Future policy decisions may facilitate simple, reliable and affordable pricing.
“We anticipate this will be used by people in the utility industry, as well as government, who want a data-driven approach to studying future scenarios,” said Siobhan Powell, co-lead researcher of the study. and doctoral student in the Stanford department. of Mechanical Engineering.
The researchers’ model aims to give network planners and policy makers this information: where, when, how, how much and how often drivers will charge. On a regular laptop computer, the model could simulate charging data for 100 million electric vehicles in about 10 minutes, the researchers said.
More diverse population
Today, the majority of electric vehicle charging in California takes place in single-family homes. However, the researchers predict that people with less access to home charging – such as apartment renters and condominium owners – will switch to electric vehicles and the fraction of home charging will decline. At the same time, workplaces could install more charging stations and electric utilities could encourage charging at different times of the day. The model can take into account several changes in factors.
“In the model, we capture the behavior of the driver, which includes when they charge, when they charge, and how much energy they use. We also capture the location, such as the house or place of work,” said Gustavo Cezar, the other co-PI. Cezar is a civil and environmental engineering PhD student and engineer at Stanford’s SLAC National Accelerator Laboratory.
Some of today’s drivers also use automatic timers to dictate when their cars are charging. Timers track electricity prices and charge vehicles at the cheapest time. However, if a large number of timers were used, they could cause an increase in demand for electricity on the grid. As of 11:00 p.m., electric vehicle charging demand in California could peak in the range of 8.725 GW in 2030. Typical daily peak demand for all electricity in California is currently in the range of 25 GW , so electric vehicle charging would account for a large portion of all electricity demand. .
This could create a number of problems, especially if the network is not equipped to handle this level of demand. In comparison, the scenarios without timer control and where the load was distributed over the day led to a peak demand of less than 4 GW.
“Grid operators have to worry about reliability and cost. Without planning, in the worst case, they could end up with an outage, or even if they can meet all the demand, end up with high electricity costs,” Powell said.
Policy decisions made in this spirit could incentivize charging at a more distributed rate, which could curb investment in electricity infrastructure while maintaining grid reliability. The framework created by the researchers can also be augmented to include other factors such as day of the week, season, holiday, or region.
But, Cezar said, simply switching to electric vehicles from combustion-powered cars, motorcycles and small trucks might not be enough on its own to decarbonize light-duty vehicles. He said that of all the talk about switching to electric vehicles, pollution from the generation of electricity that charges them gets too little attention.
“Here in California, we primarily use solar power during the day, so we’re going to have to source fossil fuel generation to power overnight charging, especially if the focus is on residential charging” , did he declare. “The next step is to figure out how to sustainably increase the number of electric vehicles.
Cezar said the team is working to disaggregate its model and apply it to narrower places, like neighborhoods and campuses, to help planners avoid overloading local power distribution systems.
The title of the article
Scalable probabilistic estimates of electric vehicle charging given observed driver behavior
Publication date of articles
Conflict of Interest Statement
The authors declare that they have no known competing financial interests or personal relationships which might appear to influence the work reported in this article.
Warning: AAAS and EurekAlert! are not responsible for the accuracy of press releases posted on EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.