August 19, 2016
Assistant Professor of Engineering Systems, Massachusetts Institute of Technology
Electrifying transportation is one of the most promising ways to significantly cut greenhouse gas emissions from vehicles, but so-called range anxiety – concern about being stranded with an uncharged car battery – remains a barrier to electric vehicle adoption. Is range anxiety justified given current cars and charging infrastructure?
It’s a question my research group and I addressed in a paper published in Nature Energy, by taking a close look at this problem with a new model.
Specifically we asked: when looking down on the geographic area of the U.S. from a bird’s-eye view, how many personal vehicles on the road daily could be replaced with a low-cost battery electric vehicle, even if daytime charging isn’t available? Our analysis is, to our knowledge, the most expansive yet detailed study to date of how current and future-improved electric vehicle technology measures up to people’s energy-consuming behavior.
We found nearly 90 percent of vehicles on the road could be replaced by a low-cost electric vehicle available on the market today. What’s more, this number is remarkably similar across very different cities, from New York to Houston to Los Angeles. That is, there is a high potential for electrification of cars in both dense and more sprawling cities in the United States.
To realize this potential, however, the needs of prospective electric vehicle drivers have to be met on all days, even high-energy ones, such as days that require long-distance travel.
Two key innovations can enable this. The first is to predict the days on which drivers are likely to exceed the car’s range, which our model is designed to do. And the second is institutional or business-model innovation to provide alternative long-range vehicles on those high-energy days. For example, conventional cars and eventually low-carbon, long-range alternatives, might show up at a user’s door at the click of a button. This need may last for some time even as battery technology improves and charging infrastructure expands.
Vehicle range is not a single number
An electric vehicle’s range is typically thought of in terms of a fixed number, but the number of miles covered on a single charge changes with factors including driving speed and style, and outdoor temperature. To understand the range of a car we need to look beyond the car itself to how people are behaving.
Over the last four years in my research group, we’ve built a model (called ‘TripEnergy’) of the second-by-second driving behavior of people across the United States, how they are likely to use heating and cooling systems in their cars, and how various electric and conventional vehicles would consume energy if driven in this way.
This approach gives us a probabilistic view of electric vehicle range. For example, for the Nissan Leaf, we find that 74 miles is the median range – based on driving patterns, half of the cars on the road in the U.S. would be able to travel this far, and half would not. (A Ford Focus electric performs similarly.) There is a distribution in this range, which demonstrates how widely actual performance can vary. We estimate, for instance, that five percent of 58-mile trips could not be covered on one charge, and five percent of 90-mile trips could.
Evaluating electric vehicle technology against driving behavior
With the TripEnergy model in hand, we asked how many cars on the road could be replaced with a low-cost electric vehicle available today. We considered a case where drivers can only charge once daily, for example at home overnight. This allowed us to study a situation where only limited changes are needed to existing public charging infrastructure and cars can use power plants that would otherwise sit idle overnight.
We found that, given how people are driving across the U.S., 87 percent of cars on an average day could be replaced with a current generation, low-cost electric vehicle, with only once-daily charging. This is based on the driving behavior of millions of people across the U.S. across diverse cities and socioeconomic classes.
Switching from conventional to electric vehicles for those cars would cut emissions by an estimated 30 percent, even with today’s fossil fuel based supply mix. In total, the trips taken by those cars represent roughly 60 percent of gasoline consumption in the U.S.
This large daily adoption potential is remarkably similar across both dense and more sprawling U.S. cities, ranging from 84 percent to 93 percent.
While it’s true that people behave differently across cities – in how they use public transport, whether they own a car, and how often they drive the cars they own – when they do drive, we found that a similar number of cars in different cities fall within the range provided by a low-cost electric vehicle.
Returns on technology improvement
What if batteries improve, and allow for longer driving range for the same cost as current-generation lithium ion batteries?
The federal research agency ARPA-E has set a target for batteries to store roughly two times more energy by weight than today’s batteries in electric vehicles. If that technical target is reached, we estimate that the 87 percent daily adoption potential estimate would rise to 98 percent, and the gasoline substitution potential would rise from 61 percent to 88 percent. The 2017 Chevy Bolt and 2018 Tesla Model 3 are expected to achieve roughly similar increases in potentials at an increased cost compared to today’s Nissan Leaf, though these costs are still close to the average cost of new cars. The Tesla Model S travels even further but costs significantly more.
Even with substantial battery improvements, however, other types of powertrain technologies will be needed to cover those days with the highest energy consumption. This need may persist for some time, even with expanded charging infrastructure, due to a small number of very-high-energy days.
The upshot on range constraints
For people to overcome range anxiety and feel comfortable buying an electric vehicle, they need to know their needs will be met on all days, including high-energy days. Predicting when this will occur – and in advance when buying a vehicle on how many days this will occur – is something that our model is well-suited for.
Our model can, with limited input on travel distance, time and location, predict the probability of exceeding the car’s range, and point to days where drivers need to turn to other, longer-range cars, for example withing households, or even within communities and through commercial car-sharing programs. The results also shed light on the quantity of long-range cars that would be needed at the population level, a gap to be filled by private sector innovation as well as national and local policy.
Reasonable financing to help distribute the upfront costs over the car’s lifetime and increasing the opportunities for charging, even if only once-daily, would also encourage EV adoption.
In all, our analysis shows that current electric vehicles can meet most daily driving needs in the U.S. Improved access to shared, long-range transport, alongside further-advanced batteries and cars and decarbonized electricity, provide a pathway to reaching a largely decarbonized personal vehicle fleet.