[ben_w]

> The issue with renewables is that it can be night or cloudy or still across thousands of square miles at once

If it was only thousands of square miles, it wouldn't be a problem at all.

Whole of the UK is about 100,000 square miles, not sure how much more if you also include the offshore areas suitable for wind.

Texas is about 270,000 square miles with the same caveat, and (I think) less interconnect capacity to other networks than the UK.

> and happen randomly based on weather

Wasn't that literally the cause of the French reactors having problems? The national weather causing a correlated output reduction in many reactors at the same time?

[AnthonyMouse]

> If it was only thousands of square miles, it wouldn't be a problem at all.

> Whole of the UK is about 100,000 square miles, not sure how much more if you also include the offshore areas suitable for wind.

> Texas is about 270,000 square miles with the same caveat

It isn't "only" thousands of square miles, weather events commonly span areas the size of what you're talking about. It's obviously going to be night across the whole region at once. It's occasionally calm across the whole continental United States. Not often, but it happens.

> Wasn't that literally the cause of the French reactors having problems? The national weather causing a correlated output reduction in many reactors at the same time?

It was a confluence of factors, one of which was French law that prohibited power plants from putting coolant water higher than a certain temperature back into the river. When the river flow is low and the intake temperature is high, meeting the regulatory requirement necessitated reducing heat generation, i.e. power output. The same is true for any thermal power plant (e.g. coal or natural gas). This happened following a period of anti-nuclear sentiment that resulted in labor shortages in the industry, causing other reactors to concurrently be offline for maintenance for an unusually long period of time. Obviously you can make any generation system unreliable through mismanagement/government opposition.

The coolant temperature limit is a design issue. There are known designs that avoid it, e.g. situate the plant on a larger river or body of water that would provide enough cooling water even on the hottest of days, or use cooling towers instead of river water.

Conversely, it's not obvious how you design a wind turbine that can provide power when there's no wind.

[ben_w]

> The coolant temperature limit is a design issue. There are known designs that avoid it, e.g. situate the plant on a larger river or body of water that would provide enough cooling water even on the hottest of days, or use cooling towers instead of river water.

> Conversely, it's not obvious how you design a wind turbine that can provide power when there's no wind.

Sure it is, and you've pretty much said the solution yourself on a previous comment with a different context:

> You're talking about an individual nuclear plant rather than the whole grid.

Just as all the reactors can have a correlated failure for whatever reason, so too can all the wind. But wind and nuclear aren't the only systems.

I'm not going to investigate if the wind really does go to zero on both sides of the Rockies — though that does seem unlikely, I'm willing to just agree it does for the argument.

Dunkelflaute — lack of both wind and solar — normally only last 24 hours even in the much smaller Germany. Of course you have to size your systems to cope with the once-every-generation events: how long this is is a matter of observation, and I can't be bothered to download and process the satellite data to find that out for the USA for this comment (but it's not hard, that was part of my first proper job about 20 years ago for unrelated research about squid). Also, while high-altitude winds are not normally affected in the same way as surface winds, and flying turbines have been demonstrated to take advantage of this, e.g. https://en.wikipedia.org/wiki/KiteGen I'm just going to ignore them here.

The USA's current average (annual) electrical draw is 455.3 gigawatts, while the installed hydro capacity has a nameplate capacity of 102.8 GW with another 12 GW known. How long could they provide this, I do not know, but these things do already exist, and that's already a quarter of your instantaneous needs all by themselves.

Batteries exist, and are getting very cheap: while production is currently limited, the supply needed for the electrification of cars is by itself enough to get over any Dunkelflaute — the total need per year only requires an average household to cycle a car-sized battery 1-3 times. And that's without any support from hydro.

And the maximum size for a grid isn't a country: the USA, despite Texas being awkward, already has some interconnects to Canada and Mexico. While I'm not suggesting it's likely or coming any time soon, China has been massively expanding in many areas including the production of aluminium, and they now produce enough by themselves to build every three years a complete global power grid with one ohm resistance. They've already suggested connecting themselves to South America — if they do or don't is more a political decision about soft power projection rather than anything else, same as the Belt and Road initiative.

And that would work without any support from hydro or batteries, and would still work just fine even if you only went one of wind and PV rather than diversifying.