Battery development is crucial for clean energy growth and the uptake of electric cars. They need to become safer, cheaper, and able to hold more energy for longer. However, improving batteries relies on a trial-and-error process, which is very slow.
Fortunately, machine learning may be able to help, speeding up the discovery process substantially. Artificial intelligence is already being used by researchers and battery companies worldwide for this purpose. It will likely claim a key role in leading to batteries that can push the limits.
At Stanford University’s Precourt Institute for Energy, there’s a lab with six refrigerator-sized cabinets designed to drain batteries of their life as fast as possible. The cabinets provide all the performance data to train AI to build a better battery.
Materials scientist William Chueh from Stanford said:
“At the end of the day, we see our job as accelerating the pace of battery R&D. Whether it’s discovering new chemistry or finding a way to make a safer battery, it’s all very time-consuming. We’re trying to save time.”
There’s an infinite number of ways to deliver charge to a battery, and AI can sift through them in but a fraction of the time compared to hands-on experimentation. AI was made for this! It can find optimal solutions in a vast search space much more efficiently than humans.
The only problem when it comes to AI’s helping with battery-building is the lack of data. Data Scientist Bruis van Vlijmen, working on battery analytics at Stanford, said:
“Historically, battery data has been challenging to acquire because it’s not shared between researchers and companies. There’s a high level of secrecy or proprietary information.”
The University of Chicago came up with a solution to help with this data-shortage problem: a platform called Data Stations, where different groups can contribute information without ever giving outsiders direct access to their data. The pool of information allows researchers to upload machine-learning models to the platform for training.
The researchers can’t see the details of the data, but they can see whether exposure to it improves their AI’s ability to make predictions about batteries. The hope is that the Data Station will ease people’s fears about losing proprietary data to competitors while allowing researchers to use that information to develop better batteries.
But even without a gigantic database to work with, researchers have managed to improve recharging rates and predict battery health, among other things, with AI. Stanford University and MIT researchers collaborated with the Toyota Research Institute this year on a study that led to cutting EV charging times down to ten minutes. Meanwhile, Cambridge and Newcastle researchers designed an AI that can predict battery health with ten times more accuracy than current industry standards.
Batteries could be the cornerstone for lowering carbon emissions related to transportation and electricity. Therefore, it’s only fitting that we tap into big data and computer learning to speed the process up as much as possible.
