2 ms·
What's your view on the Deepmind experiments to use NNs to stabilize the plasma fields, won't better control of the plasma ameliorate some of the material issue
by whizzter 5y ago
What's your view on the Deepmind experiments to use NNs to stabilize the plasma fields, won't better control of the plasma ameliorate some of the material issues? (I think a link to that was posted here some weeks ago)
- omgJustTest 5y agoThe Deepmind control of plasmas is interesting, and I think that some of the more recent successes at NIF have been related to more run-of-the-mill deep-learning based control. The problems above are not just a plasma instability problem, which has impeded progress. The interior walls of the fusion reactors are just pretty extreme environments, and the extremity is made worse by the fact that if you control plasma flow with magnetic fields, you trade the ability to control magnetic fields with dose -to-the-magnets-. Achieving high magnetic fields also require high currents -> super conductors -> low temperatures (0.1-20 Kelvin).