My understating of the issue is that one need to keep the plasma is the adequate state for
1. the reaction to occur (i reckon we can do that for a fraction of a microsecond or something)
2. make it self-sustaining (we're not here yet)
3. Obviously without producing an H-bomb. Neighbours might get unhappy.
I am not a physicist, but AFAIK the problem is to maintain the plasma in a steady state. This could mean that they will try to keep the plasma stable for 8 minutes until it collapses.
The characteristic frequencies in the plasma are much higher. Kilohertz and up. The steady state will be reached within seconds, so to the plasma 8 minutes is basically "infinity". But at that point the all the power that is leaking out of the plasma (50MW of heating times a Q of 10 implies that the plasma will output half a GW) will have heated the wall to the limit (and that is with active cooling of the walls). One could beef up the cooling to run at actual steady state (for an hour, or a day or a year), but it would not tell you much extra, so they didn't bother.
Tokamaks need a changing current in the central solenoid. So while 8 minutes is infinity for a stellarator, there are important performance benchmarks at 8 seconds, 8 minutes, 8 hours, etc. You trade pulse length for ohmic heating.
Tokamak-style reactors like ITER rely on a rapidly-changing magnetic field to induce an electric current in the plasma, for stability of the plasma. The magnet that provides this rapidly-changing field (the 'central solenoid') can't keep increasing in field strength indefinitely -- so there's a time limit to the life of the plasma. There are other ways to supplement the current in the plasma, which can help extend the life of the plasma indefinitely.
> But this is beyond ridiculous, so many years of R&D for 8 minutes of sustained acticivty activity? That sound like a joke
...You know the duration of the plasma pulse being 8 minutes doesn't mean they do it once and then the whole project is over and they need to start dismantling it, right? After a pulse, the plasma is discharged, you have a few minutes of rampdown and cooling, and then you can do it again. IIRC in the first experimental phase they expect to do ~30,000 pulses.
That’s like saying “spending $30,000 for a license plate holder sounds like a joke”. This machine is being built for the information needed to make real reactors. Empirically derived scaling laws, confinement physics, and divertor physics all greatly benefit from having their regimes pushed past where they have been. It’s a science machine.
ITER is aiming for net-positive energy in terms of heat coming off the plasma (but the energy will just be dumped); DEMO is aiming to actually generate electrical power.
