Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI for real-time fusion plasma behavior prediction and manipulation

control.princeton.edu

Post reply on HN
Print view View original

71–80 of 167 posts

Re: AI for real-time fusion plasma behavior prediction and manipulation

#71

Post by adrian_b »

vilhelm_s wrote: post #53
atomic128 wrote: post #12

Neutrons make hardware radioactive. Many on Hacker News fantasize about fusion (not fission) reactors. These fusion reactors will be an intense source of fast neutrons. All the hardware in a fusion reactor will become radioactive. Not to mention the gamma rays. If you have to deal with radioactive materials, why not just use fission? After 70 years of working with fission reactors, we know how to build and operate th…

The radioactivity generated from neutron activation is low-level, so you don't need to worry about accidents releasing lots of radioactivity, or about how to store waste for a long time. There are a lot of people worrying about those two things for fission reactors. Also, the fuel for fusion reactors is much more plentiful. If we went all in on fission we might run out of easily minable uranium ore in a century or so…

The radioactivity generated from neutron activation is not at all low-level, because the neutron flux is huge, providing most of the energy generated by the fusion reactor.

The intense neutron flux will transmute a very high number of atoms, so when taken out of the reactor all materials are very highly radioactive.

What can be hoped is that there may be choices for the materials used in a fusion reactor that will ensure a short enough lifetime for the radioactive isotopes, so that the radioactivity of the contaminated materials will become low-level soon enough.

The studies that I have seen have the target that the radioactive waste produced by a fusion reactor should become low-level radioactive waste after one hundred years.

To reach this target, many commonly used structural materials, like many types of steel, must be completely avoided, e.g. any steel containing nickel, molybdenum or niobium. Even the carbon from steel is a problem, because the radioactivity of C14 will persist for thousands of years.

A smaller fraction of the materials, particularly from highly activated plasma facing and near plasma components, may fail to meet current low-level waste criteria even after one thousand years.

See e.g. the report:

https://assets.publishing.service.gov.uk/media/61ae4caa8fa8f...

Re: AI for real-time fusion plasma behavior prediction and manipulation

#72

Post by iforgot22 »

avidphantasm wrote: post #68

I know AI is the buzzword du jour, but this is really ML, and really just advanced cybernetic control systems. Deep learning systems have a high enough degree of variety necessary to control short time step nonlinear systems like the plasma in a tokamak.

ML, a subset of AI...

Re: AI for real-time fusion plasma behavior prediction and manipulation

#74

Post by chriskanan »

There is a lot of AI research in the nuclear fusion space. For inertial confinement fusion (a competing technology to magnetic confinement fusion, e.g., tokamaks) the National Ignition Facility (NIF) used it for their experiment that resulted in "ignition."

My lab is collaborating with researchers at the Laboratory for Laser Energetics to use AI to improve inertial confinement fusion (ICF). We recently put out this paper [1] using Kolmogorov-Arnold Networks (KANs) to predict the outcome of ICF experiments. Currently, existing physics simulators are based on old Fortran code, are slow, and have a high error between their predictions and actual laser shots, so among other goals, we are trying to build better predictors using neural networks. This is needed since it is hard to rapidly iterate on real data, since they only have a dataset of around 300 ICF shots.

[1] https://arxiv.org/abs/2409.08832

Re: AI for real-time fusion plasma behavior prediction and manipulation

#75

Post by dang »

tolerance wrote: post #6
carbocation wrote: post #2

Is this just an announcement of a grant renewal?

I think this gives an overview of the actual research being funded. "AI for Real-time Fusion Plasma Behavior Prediction and Manipulation - Plasma Control Group": https://control.princeton.edu/machine-learning-for-rt-profil...

Thanks! We've changed the link above to that. Submitted URL was https://www.ri.cmu.edu/ai-meets-fusion-cmu-princeton-join-fo....

Re: AI for real-time fusion plasma behavior prediction and manipulation

#79

Post by wruza »

spenczar5 wrote: post #15
HPsquared wrote: post #7

Earlier quoted context omitted.

Neural networks have been used in industrial process control for many years. This is just another industrial control problem, perhaps a difficult one.

That’s really interesting! Do you have other interesting specific examples? I would have guessed that most industrial control problems were simpler sets of differential equations that could be directly estimated.

I’ve heard about a local packaging factory recently that uses an ML-first system for messaging their clients about what they will need to order soon, based on recent orders and global criteria. It’s not a simple problem, apparently. They signal the clients and start pre-producing, basically sort of algotrading themselves.

Not really an industrial control though, but close to it.

Re: AI for real-time fusion plasma behavior prediction and manipulation

#80

Post by empath75 »

meindnoch wrote: post #67
rob74 wrote: post #31

Earlier quoted context omitted.

The designs are complicated, well, because in practice it's not as simple as "put these rocks next to each other and they get hot". When you put the rocks next to each other, they not only get hot, but also emit some nasty radiation that has to be shielded. And if the rocks get a liiiitle bit too close together, they might explode, which leads to huge headaches for everyone involved, so you'd better make sure that do…

>And if the rocks get a liiiitle bit too close together, they might explode Impossible. Worst thing that can happen without carefully designed explosive lens is a nuclear fizzle.

There are lots of ways that nuclear plants can explode or fail in otherwise catastrophic ways, it doesn't need to be an atomic explosion.
Post reply on HN