Today batteries are used everywhere, from small electronics to cars. Development and production of power sources takes a lot of time and money, and most of all resources requires testing prior to sale is necessary to identify the lifetime and distribute the classes to meet the needs of different customers. Still life was determined by numerous cycles of charging and discharging, but with the increase in battery capacity, it takes more time. To help, as always, was artificial intelligence taught him to give accurate predictions based on only five cycles.
The development of artificial intelligence has been the researchers from the Massachusetts Institute of technology and research centre Toyota. Instead of the many cycles of recharge and use the battery pack, they offered to hold only five cycles, and give these data processing computer algorithm. To identify life, he uses hundreds of millions of data points, and draws attention to voltage drop and other factors, indicating the complete discharge. According to researchers, the prediction accuracy reaches 95%.
According to researchers from Toyota of Patrick herring, thus machine learning significantly accelerates the development of new batteries and significantly reduces costs for both research and production. Moreover, the researchers suggest that technology can help to optimize the charging process, so that it is replenished as quickly as possible — about 10 minutes.
It is noteworthy that the Massachusetts Institute of technology often conducts research in the field of batteries. For example, in September 2018, it developed a power source that absorbs carbon dioxide.
You must have something to say about the new work of scientists — their opinions can be shared in the comments. Don’t forget to join our Telegram chat, where there is always lively discussions on the theme of science and technology!