Scientists at MIT, Stanford and the Toyota Research Institute (TRI) have developed an AI algorithm that they say can accurately predict the useful life of Li-ion batteries. The algorithm analyzed factors such as voltage decline to make predictions as to how many more cycles a battery would last. The predictions turned out to be within 9 percent of the battery’s actual cycle life.
The algorithm also categorized batteries as having either a long or short life expectancy based on the first five charge/discharge cycles. These predictions proved accurate 95 percent of the time.
One aim of the project is to identify an efficient way to charge batteries in ten minutes. The researchers think the prediction model can be used to decrease the time it takes to validate batteries with new chemistries and to grade batteries with longer lifetimes. The algorithm may also help recyclers identify cells in used EV batteries that have sufficient remaining life for secondary uses.
The researcher have made the dataset used to train the algorithm publicly available.
“The standard way to test new battery designs is to charge and discharge the cells until they die. Since batteries have a long lifetime, this process can take many months and even years,” said Peter Attia, a Stanford engineering doctoral candidate who co-led the research. “It’s an expensive bottleneck in battery research.”
“For all of the time and money that gets spent on battery development, progress is still measured in decades,” said Patrick Herring, a TRI research scientist who contributed to the study. “In this work, we are reducing one of the most time-consuming steps – battery testing – by an order of magnitude.”
The work is described in the journal Nature Energy.