Real-world use of electric vehicles (EVs) — including heavy traffic, long highway trips and short city trips — could allow current EV batteries to last 40% longer than originally forecast, according to new data from Stanford University.
Through a joint project between Stanford University’s Precourt Institute for Energy and SLAC National Accelerator Laboratory, the study suggests a typical EV battery may not need to be replaced for several additional years.
Through testing, the team found that the sweet spot for EV battery longevity is real-world use — counter to the findings of previous studies on EV batteries.
How batteries are tested
Typically, the testing of batteries involves using a constant rate of discharge followed by recharging. This cycle is repeated rapidly and many times to learn about new battery designs and if they are good or have bad life expectancy.
However, Stanford University researchers said this is not a good way to predict the life of an EV battery considering it does not factor everyday commuting.
“We’ve not been testing EV batteries the right way,” said Simona Onori, associate professor of energy science and engineering in the Stanford Doerr School of Sustainability. “To our surprise, real driving with frequent acceleration, braking that charges the batteries a bit, stopping to pop into a store, and letting the batteries rest for hours at a time, helps batteries last longer than we had thought based on industry standard lab tests.”
The impact of the study could have far-reaching ramifications for battery prices, which have already dropped in price by 90% in the past 15 years. Even with these drops in prices, batteries still account for almost a third of the price of a new EV.
How they did it
Researchers developed four types of EV discharge profiles and tested 92 commercial lithium-ion batteries for more than two years across the discharge profiles. Stanford University found that the profile that reflected actual driving behavior was more realistic and equated to higher EV battery life.
A machine learning algorithm was trained on all the data that was collected and found the different discharge profiles on battery degradation.
The study shows a correlation between sharp, short EV accelerations and slower degradation, something that was contrary to long-held assumptions of previous EV battery studies.
“We battery engineers have assumed that cycle aging is much more important than time-induced aging. That’s mostly true for commercial EVs like buses and delivery vans that are almost always either in use or being recharged,” said Alexis Geslin, a PhD student in materials science and engineering and in computer science in Stanford’s School of Engineering and worked on the project. “For consumers using their EVs to get to work, pick up their kids, go to the grocery store, but mostly not using them or even charging them, time becomes the predominant cause of aging over cycling.”
The study found a sweet spot for balancing time aging and cycle aging in the testing of the batteries that was the realistic consumer EV driving scenario. Stanford said automotive OEMs should update EV battery management software to take advantage of these findings and maximize battery longevity.
The next steps
Researchers said the study could go beyond just batteries. It could be applied to energy storage applications as well as other materials and devices in physical sciences where aging is crucial — like plastics, glasses, solar cells and more.
“This work highlights the power of integrating multiple areas of expertise — from materials science, control, and modeling to machine learning — to advance innovation,” Onori said.
The full research can be found in Nature Energy.
