Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×
AI Technology

AI Is Throwing Battery Development Into Overdrive (wired.com) 30

Improving batteries has always been hampered by slow experimentation and discovery processes. Machine learning is speeding it up by orders of magnitude. From a report: Inside a lab at Stanford University's Precourt Institute for Energy, there are a half dozen refrigerator-sized cabinets designed to kill batteries as fast as they can. Each holds around 100 lithium-ion cells secured in trays that can charge and discharge the batteries dozens of times per day. Ordinarily, the batteries that go into these electrochemical torture chambers would be found inside gadgets or electric vehicles, but when they're put in these hulking machines, they aren't powering anything at all. Instead, energy is dumped in and out of these cells as fast as possible to generate reams of performance data that will teach artificial intelligence how to build a better battery. In 2019, a team of researchers from Stanford, MIT, and the Toyota Research Institute used AI trained on data generated from these machines to predict the performance of lithium-ion batteries over the lifetime of the cells before their performance had started to slip. Ordinarily, AI would need data from after a battery had started to degrade in order to predict how it would perform in the future. It might take months to cycle the battery enough times to get that data. But the researchers' AI could predict lifetime performance after only hours of data collection, while the battery was still at its peak. "Prior to our work, nobody thought that was possible," says William Chueh, a materials scientist at Stanford and one of the lead authors of the 2019 paper.

And earlier this year, Chueh and his colleagues did it again. In a paper published in Nature in February, Chueh and his colleagues described an experiment in which an AI was able to discover the optimal method for 10-minute fast-charging a lithium-ion battery. Many experts think fast-charging batteries will be critical for electric vehicle adoption, but dumping enough energy to recharge a cell in the same amount of time it takes to fill up a tank of gas can quickly kill its performance. To get fast-charging batteries out of the lab and into the real world means finding the sweet spot between charge speed and battery lifetime. The problem is that there is effectively an infinite number of ways to deliver charge to a battery; Chueh compares it to searching for the best way to pour water into a bucket. Experimentally sifting through all those possibilities to find the best one is a slow and arduous task -- but that's where AI can help.

This discussion has been archived. No new comments can be posted.

AI Is Throwing Battery Development Into Overdrive

Comments Filter:
    • I hope your comfortable; you'll have to wait -18 years.

    • No. You want thinner phones. Thinner I say!
    • by knarf ( 34928 )

      Get a better phone, I've been charging mine close to weekly (usually every 5 days) for ages now. First with a Motorola Defy, now with a Xiaomi Redmi Note 5. The key to doing this is to dump everything off the phone you don't need, starting with Google and GSF. No Google account, no Facebook, no Shitter, no nothing. FDroid for apps, Aurora Store for those which live on the Play store, etc. Maybe you won't get to 5 days but you should be able to get 3 easily, 4 without too many problems.

      • When I don't use my phone the battery lasts longer too.
        The tank in my car also stays full longer if I don't drive it. So it's not just batteries that perform this magic.
  • by JoeyRox ( 2711699 ) on Tuesday October 13, 2020 @03:53PM (#60604206)
    Instead of groundbreaking new battery technology failing to arrive 5 years after it was promised, it'll now fail to arrive just 3 years after being promised.
    • Hey, let's apply some AI to fusion then.

    • No, for idiots who are credulous of media talking heads, this extends the length of bullshit times, because they'll always assume the AI speeds development up, when really it bleeds resources away from real engineering and diverts it to PHBs at the AI companies. So when the materials engineer says it take another 5 years to complete the materials engineering, and the talking head would have told you that the completed product will be on the market in 5 years, now the talking head will say 3 years to the opt

      • by gweihir ( 88907 )

        Unfortunately, that is pretty much the state of things. Materials research is hard. Sure, there will be a tiny faction of it where statistical classifiers and predictors can help, but that is about it. This is no revolution, that is just the typical "AI bullshit" that regards AI as the second coming or something.

  • Fancy Regression (Score:5, Interesting)

    by Thelasko ( 1196535 ) on Tuesday October 13, 2020 @03:56PM (#60604224) Journal
    I've played with AI for product development. You still need to run a DOE. [wikipedia.org] It just provides a better fit for unknown functions. However, it tends to fall apart for extrapolation. That makes it great for product refinement, but you aren't going to have any major breakthroughs with this technique. That's because you likely will get stuck at a local minimum, where some "out of the box" thinking would explore a different valley all together.
    • you aren't going to have any major breakthroughs with this technique

      I think that is due to the state of AI development rather than an inherent problem with it.

      Even simple game-playing AI used to work by studying human play, parsing that out into individual steps and calculating the highest "value" of those to play in any given situation. But some recent AI has learned to play just by being told the rules, which frees it from being able to do only those things we've done.

      Probably the sticking point in most development of physical products would be giving it a useful se

      • But what if we also used Block chain technology, in the cloud? Also combine that with 3D printing and a new buzz word I am still working on and we can have your batteries for even more money than you knew you had.
        • have your batteries for even more money than you knew you had.

          Thats called a 'loan' [provenir.com]. They've been around a long time [wikipedia.org]
          The buzzwords would be 'microfinance' [google.com] or 'fintech' [google.com]

          You'd need to use fintech in the cloud to microfinance a 3D printed Blockchain AI. Then it will cure covid while removing CO2 from the air.
          Sorry, fusion is still 20 years away.

      • My gut reaction is that would be a project of monumental proportions. It would require resources similar to the Human Genome Project. [wikipedia.org]
    • by rtb61 ( 674572 )

      Except the most obvious one. You pulse the charge at the frequency of the bond you want to break. Resonance will help break that bond, no free energy, just better absorption of energy into the molecular bond to help break it, pretty straight forward. Now Toyota America can claim that, lamers.

  • It will be funny when technology and basic economics gets us to zero emissions faster than our pathetic ambitions (tempered by the PR Dept of the fossil fools and their moronic public).
  • Sounds a lot more like the latter.
  • This reminds me of the old movie 'Demon Seed'. A company builds an AI to answer technological problems, but instead it takes too broad a viewpoint and starts asking why the problems existed in the first place. It would be funny if an AI answer to fast-charging being damaging batteries were 'don't fast charge the batteries then, Slow the metabolism of your society and economy instead, drive slower too, batteries will last much longer that way'

  • Every 5 years we get new, better batteries and new technology that needs even newer, better-er batteries.

  • First the CFO on our back, then the CEO and now the AI, sooner or later the abbrevionitis will get us,

  • How many industrial sized lithium based batteries can be made before we run out of lithium.
    Lithium concentration is defined in ppm. There's billions of tons of it in sea water, but low concentration.
    Nickel as a battery component is found in concentrations of 1 - 65% in ore and Australia has bucket loads.
    Nickel won't make small portable batteries, but with NiFe batteries having a life of 20-100 years, why wouldn't we make Ni storage batteries.

    • To quote wikipedia, "Due to its low specific energy, poor charge retention, and high cost of manufacture, other types of rechargeable batteries have displaced the nickel–iron battery in most applications." Also, charge-discharge cycle efficiency is only 65%.
  • Nice try Boffins.
    But it's still PR bullshit to get a grant.
    Give us a battery now.

    • "Miracle battery coming in 3 years!". BULLSHIT. BULLSHIT. BULLSHIT. Why is it these miracle battery assholes can't do the press release after they make something that actually works? It's infuriating because when one puts low quality journalism together with low quality science one gets an order of magnitude jump in the garbage-level of "New Tech" stories to the point that they are not only false, but quite aggravating to even read.

Time is the most valuable thing a man can spend. -- Theophrastus

Working...