Accurate battery lifespan predictions allow for the faster development of novel battery materials and the selection of appropriate use practices for deployment longevity. Traditional battery testing, on the other hand, can take years to reach thousands of cycles. Machine learning algorithms were employed by scientists to forecast how long a lithium-ion battery would survive. The technique could lower the cost of battery development.
Consider a clairvoyant informing your parents how long you would live on the day you were born. Battery chemists who use new computational models to calculate battery lifetimes based on as little as a single cycle of experimental data may have a similar experience.
In a new study, researchers at the U.S. Department of Energy’s (DOE) Argonne National Laboratory used machine learning to forecast the lifespan of a variety of different battery chemistries. The scientists can properly calculate how long different batteries will continue to cycle by analyzing experimental data acquired at Argonne from a batch of 300 batteries representing six distinct battery chemistries.
In a machine learning algorithm, scientists train a computer program to make inferences on an initial collection of data, and then use what it has learnt from that training to make conclusions on another set of data.
Battery lifespan is of vital importance for every user for every different type of battery application, from cell phones to electric vehicles to grid storage. Having to cycle a battery thousands of times until it breaks can take years; our method offers a kind of computational test kitchen where we can quickly establish how different batteries will function.
Noah Paulson
“Battery lifespan is of vital importance for every user for every different type of battery application, from cell phones to electric vehicles to grid storage,” said Argonne computational scientist Noah Paulson, one of the study’s authors. “Having to cycle a battery thousands of times until it breaks can take years; our method offers a kind of computational test kitchen where we can quickly establish how different batteries will function.”
“Right now, the only method to evaluate how the capacity in a battery declines is to actually cycle the battery,” said Argonne electrochemist Susan “Sue” Babinec, another study author. “It’s highly pricey and takes a long time.”
According to Paulson, determining a battery’s lifetime might be difficult. “The reality is that batteries do not live forever, and how long they last is determined on how we use them, as well as their design and chemistry,” he said. “Until recently, there hasn’t been a good way to predict how long a battery will survive. People will want to know how long it will be before they have to spend money on a new battery.”
The study was notable for relying on substantial experimental work done at Argonne on a range of battery cathode materials, including Argonne’s proprietary nickel-manganese-cobalt (NMC)-based cathode.
“We had batteries that represented different chemistries, that have different ways that they would degrade and fail,” Paulson said. “The value of this study is that it gave us signals that are characteristic of how different batteries perform.”
Further research in this area, according to Paulson, has the potential to shape the future of lithium-ion batteries. “One of the things we’re able to do is train the algorithm on a known chemistry and have it make predictions on an unknown chemistry,” he explained. “Essentially, the algorithm may steer us in the path of new and improved chemistries with longer lives.”
In this approach, Paulson believes that the machine learning method could speed up the creation and testing of battery materials. “Assume you have a new material and have cycled it a few times. You may use our method to anticipate its lifespan and then decide whether to continue cycling it experimentally or not.”
“If you’re a researcher in a lab, you can discover and test many more materials in a shorter period of time because you can analyze them faster,” Babinec noted. The study’s findings were published in the online edition of the Journal of Power Sources as “Feature engineering for machine learning enabled early prediction of battery lifetime.”
The University of Oxford’s Department of Engineering Science saw the commercial necessity for industries like these to be able to forecast how long their batteries will survive. Members of the research group created machine learning models to imitate battery life behavior, but they required a means to visualize their inputs and outputs in a way that would appeal to individuals outside of academia.
