MSGID: 1:317/3 6274a491   
   PID: hpt/lnx 1.9.0-cur 2019-01-08   
   TID: hpt/lnx 1.9.0-cur 2019-01-08   
    Researchers now able to predict battery lifetimes with machine learning   
      
      
     Date:   
         May 5, 2022   
     Source:   
         DOE/Argonne National Laboratory   
     Summary:   
         Scientists have used machine learning algorithms to predict how   
         long a lithium-ion battery will last.   
      
      
      
   FULL STORY   
   ==========================================================================   
   Technique could reduce costs of battery development.   
      
      
   ==========================================================================   
   Imagine a psychic telling your parents, on the day you were born,   
   how long you would live. A similar experience is possible for battery   
   chemists who are using new computational models to calculate battery   
   lifetimes based on as little as a single cycle of experimental data.   
      
   In a new study, researchers at the U.S. Department of Energy's   
   (DOE) Argonne National Laboratory have turned to the power of machine   
   learning to predict the lifetimes of a wide range of different battery   
   chemistries. By using experimental data gathered at Argonne from a set   
   of 300 batteries representing six different battery chemistries, the   
   scientists can accurately determine just how long different batteries   
   will continue to cycle.   
      
   In a machine learning algorithm, scientists train a computer program   
   to make inferences on an initial set of data, and then take what it has   
   learned from that training to make decisions on another set of data.   
      
   "For every different kind of battery application, from cell phones to   
   electric vehicles to grid storage, battery lifetime is of fundamental   
   importance for every consumer," said Argonne computational scientist   
   Noah Paulson, an author of the study. "Having to cycle a battery   
   thousands of times until it fails can take years; our method creates a   
   kind of computational test kitchen where we can quickly establish how   
   different batteries are going to perform."  "Right now, the only way to   
   evaluate how the capacity in a battery fades is to actually cycle the   
   battery," added Argonne electrochemist Susan "Sue" Babinec, another   
   author of the study. "It's very expensive and it takes a long time."   
   According to Paulson, the process of establishing a battery lifetime can   
   be tricky. "The reality is that batteries don't last forever, and how long   
   they last depends on the way that we use them, as well as their design and   
   their chemistry," he said. "Until now, there's really not been a great way   
   to know how long a battery is going to last. People are going to want to   
   know how long they have until they have to spend money on a new battery."   
      
      
   ==========================================================================   
   One unique aspect of the study is that it relied on extensive experimental   
   work done at Argonne on a variety of battery cathode materials, especially   
   Argonne's patented 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 study in this area has the potential to guide   
   the future of lithium-ion batteries, Paulson said. "One of the things   
   we're able to do is to train the algorithm on a known chemistry and have   
   it make predictions on an unknown chemistry," he said. "Essentially,   
   the algorithm may help point us in the direction of new and improved   
   chemistries that offer longer lifetimes."  In this way, Paulson believes   
   that the machine learning algorithm could accelerate the development   
   and testing of battery materials. "Say you have a new material, and   
   you cycle it a few times. You could use our algorithm to predict its   
   longevity, and then make decisions as to whether you want to continue   
   to cycle it experimentally or not."  "If you're a researcher in a lab,   
   you can discover and test many more materials in a shorter time because   
   you have a faster way to evaluate them," Babinec added.   
      
   A paper based on the study, "Feature engineering for machine learning   
   enabled early prediction of battery lifetime," appeared in the Feb. 25   
   online edition of the Journal of Power Sources.   
      
   In addition to Paulson and Babinec, other authors of the paper include   
   Argonne's Joseph Kubal, Logan Ward, Saurabh Saxena and Wenquan Lu.   
      
   The study was funded by an Argonne Laboratory-Directed Research and   
   Development (LDRD) grant.   
      
      
   ==========================================================================   
   Story Source: Materials provided by   
   DOE/Argonne_National_Laboratory. Original written by Jared Sagoff. Note:   
   Content may be edited for style and length.   
      
      
   ==========================================================================   
   Journal Reference:   
      1. Noah H. Paulson, Joseph Kubal, Logan Ward, Saurabh Saxena,   
      Wenquan Lu,   
         Susan J. Babinec. Feature engineering for machine learning enabled   
         early prediction of battery lifetime. Journal of Power Sources,   
         2022; 527: 231127 DOI: 10.1016/j.jpowsour.2022.231127   
   ==========================================================================   
      
   Link to news story:   
   https://www.sciencedaily.com/releases/2022/05/220505114658.htm   
      
   --- up 9 weeks, 3 days, 10 hours, 50 minutes   
    * Origin: -=> Castle Rock BBS <=- Now Husky HPT Powered! (1:317/3)   
   SEEN-BY: 15/0 106/201 114/705 123/120 129/330 331 153/7715 218/700   
   SEEN-BY: 229/110 111 317 400 426 428 470 664 700 292/854 298/25 305/3   
   SEEN-BY: 317/3 320/219 396/45   
   PATH: 317/3 229/426