back to article Yes, true, fusion reactors don't work quite yet, but, er, maybe AI can help us stop our experiments from imploding

Machine-learning algorithms may help scientists develop a warning system capable of preventing fusion reactor meltdowns in the future. Scientists have attempted to harness the power of nuclear fusion for decades in the hopes of providing clean energy. Clouds of plasma made up of positively and negatively charged ions are …

  1. Fredrick Smith


    Once the AI gets its hands on a fusion reactor, it will be unstoppable.

  2. Andy E


    The AI got it right 80% of the time. In real life that other 20% where it gets it wrong could end things very quickly.

    1. alain williams Silver badge

      Re: Kaboom!

      The nice thing about fusion reactors is that any Kaboom will be much smaller than you would get with a fission one and there is much smaller amount of radio active emissions which are far less harmful.

      1. Martin-R

        Re: Kaboom!

        The issue is more about damage to the reactor vessel - ion beams suddenly going in a straight line instead of in circles, and forces from collapsing magnetic fields

      2. HildyJ Silver badge

        Re: Kaboom!

        Also, fusion reactor experiments fail all (literally all) time. Otherwise we'd have a working reactor.

        The failures are generally not in the initiating elements (ions, lasers, whatever) it's in the containment of the plasma they create.

        1. Imhotep Silver badge

          Re: Kaboom!

          But there have been some interesting changes recently both in scale and how they are addressing containment. With multiple independent companies approaching the problem, we may see some interesting developments.

        2. MrXavia

          Re: Kaboom!

          Well no, because the experiments often work, but not producing more power than it takes to run the experiment, which is why we keep working on it

    2. Il'Geller

      Re: Kaboom!

      Machine learning.

      1. phuzz Silver badge

        Re: Kaboom!

        So what you're saying is, place the AI core right next to the reactor.

        That way the little silicon bugger has an incentive to get it right!

        1. Il'Geller

          Re: Kaboom!

          Yes, it’ a very bad idea. However it’s the only one.

    3. Aqua Marina

      Re: Kaboom!

      "...that automatically controls the tokamak rather than relying on human intervention."

      So MCAS for reactors, what could possibly go wrong!

  3. Chris G Silver badge

    We'll have fusion in the next 20/30/40 years

    So if they continue working on the AI at the same time it should be pretty good by the time fusion really is here.

    1. HildyJ Silver badge

      Re: We'll have fusion in the next 20/30/40 years

      Plus in the event of an emergency you'll be able to evacuate with your flying car'

    2. jake Silver badge

      Re: We'll have fusion in the next 20/30/40 years

      SpeakingTyping as a SAIL alumnus, that's what they told us about fusion and AI in the 1960s. We are no closer today, IMO.

      Throw HildyJ's flying cars into the above, too.

      1. Alan Brown Silver badge

        Re: We'll have fusion in the next 20/30/40 years

        "We are no closer today, IMO."

        However we DID have safer fission reactors and walked away from them.

  4. This post has been deleted by its author

    1. John Brown (no body) Silver badge

      Re: What's the training dataset?

      Fail fast, fail often?

  5. Il'Geller

    AI is a search technology, finds information into structured texts.

    Indeed, “Details from various fusion reactors, such as their plasma current, plasma energy, radiation power, strength of radial magnetic field, and more are taken into account and used as inputs into the algorithm.”

    Therefore it will work.

    1. Imhotep Silver badge

      That is an interesting definition of AI. I believe most of us in the technical fields are thinking along different lines when we use the term.

      1. Il'Geller

        This is the definition.

        If you have another definition one would you share it, please?

        AI came from NIST TREC QA as AI-parsing, which replaces n-gram parsing. This parsing is the only novelty in all Science (it’s linguistic part) since the first computer was switched on. If you know another novelty please be so kind and tell?

        1. jake Silver badge



          1. Il'Geller

            Re: Judgement

            Everything that is somehow connected, everything what happens with fusion reactors is described or can be described by texts. In other words, it all boils down to searching for what needs to be done at this particular moment in the texts, which is AI.

  6. AdamT

    That's an incredibly short amount of time - err, really?

    250ms = 1/4s which I'd have thought was an eternity in the realm of atomic physics.

    Perhaps this is more to do with thermal/fluid dynamics (which is presumably a bit more "physical world" speed) than the actual atomic interactions?

    Still feels like quite a long time though. I mean, don't most car engine management systems operate on a 100ms timer for example?

    1. Aerostat

      Re: That's an incredibly short amount of time - err, really?

      It's a long time in terms of physics, but a rather short time for a human to respond and react to it, which is what was meant and why a computer will do it instead.

  7. wiggers

    "Clouds of plasma made up of positively and negatively charged ions are placed in a donut-shaped chamber"

    Surely they need to be one or the other? It's only in colliders they are opposite charges, so they flow in opposite directions under the same magnetic field.

    1. Imhotep Silver badge

      Perhaps they meant the positively charged ions and negatively charged electrons?

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2022