How Machine Learning could help develop Fast-Charging Batteries | Top up in 10 minutes
Stanford Study reveals Fast-Charging batteries being aided by Machine Learning
Stanford University researchers are using machine learning (ML) to create improved fast-charging batteries. In addition to using machine learning to speed up scientific analysis by searching for trends in results.
The researchers used it in conjunction with information gained from experiments and equations driven by physics to discover and explain a mechanism that reduces the lifetimes of fast-charging lithium-ion batteries.
“It was the first time this approach, known as scientific machine learning, was applied to battery cycling,” said study leader Professor Will Chueh.
According to him, the findings challenge long-held theories about how lithium-ion batteries charge and discharge and provide researchers with a new set of guidelines for designing longer-lasting batteries.
The research, published in Nature Materials, is the most recent partnership between Stanford, SLAC, MIT, and Toyota. The aim is to combine fundamental science and industry expertise to create a long-lasting EV battery that can charge in 10 minutes.
The new research builds on two previous initiatives in which the group used more traditional ML types to speed up battery testing and the process of narrowing down several possible charging methods to find the ones that function best.
Although these studies aided researchers in making much quicker progress—for example, cutting the time required to assess battery lifetimes by 98 percent—they did not expose the underlying physics or chemistry that caused sure batteries to last longer than others, as the latest study did.
Combining all three methods could theoretically cut the time it takes to carry a new tech for fast-charging batteries, from the lab bench to the market by up to two-thirds, according to Chueh.