Can you expand on what relevance this has to the pragmatic thermal process based point raised in the comment above yours?
The aluminium smelters I've had contact with have all avoided any shutdowns (back up generators, etc) because any significant breakdown in continuous processing involves several days of tear down in order to restart.
I am literally within 20 feet of someone who worked 5 years as a general foreman in charge of the smelter portion of the Portland aluminium smelter* who is not embracing the idea that this a process you can just turn off and turn on again later w/out a large chunk of effort.
The relevance is that you would have a base load to keep the system hot which is separate from the electricity that does the work.
I think at a steel smelter, since the heat does the work, they aren't as easily separable. You would still use less energy at the times when you processed less metal.
Aluminium smelters were traditionally some of the biggest loads on the grid. They were built near hydroelectric dams and had contracts with electricity providers to absorb excess energy and to pause operations on short notice when other demand is high. In exchange for this they get it very cheap, which is good because they need a lot of it. Variable energy supply is not new to them.
Now I've never worked at either and I'm just spitballing ideas on the internet, but you haven't either and you are too.
The aluminium smelters I've had contact with have all avoided any shutdowns (back up generators, etc) because any significant breakdown in continuous processing involves several days of tear down in order to restart.
I am literally within 20 feet of someone who worked 5 years as a general foreman in charge of the smelter portion of the Portland aluminium smelter* who is not embracing the idea that this a process you can just turn off and turn on again later w/out a large chunk of effort.
* https://en.wikipedia.org/wiki/Portland_aluminium_smelter