Going back to the scale of those plants: how well do they scale down? I'm guessing there are some lower limits where efficiency tapers off.
I'm guessing there is nothing to be gained but I'm wondering, given how one kvetches about the state of the grid, would more smaller nukes be better than large centralized ones? The grid wires have to all be connected and running everywhere, but I wonder if more smaller nukes might make the grid more robust.
I'm guessing the paperwork & red tape does not scale thus it's less work to make one big one as opposed to multiple smaller ones across the various NIMBY areas.
It's a complex topic, so I'll try and simplify it somewhat:
Large units have the following advantages:
- Higher nameplate capacity for a given staff load
- Lower cost output (tied to the above point, lower OPEX)
- Many mature existing designs in a variety of sizes (C6 is around 675MWe for example, EPR is 1,600MWe)
Large units have the following disadvantages:
- Higher CAPEX for construction
- A trip drops a tremendous amount of capacity off the grid, so your grid needs to be setup to deal with that
Small units have the following advantages:
- Faster construction
- Lower CAPEX
- Fewer staff
- A trip drops a much smaller amount of capacity off the grid, making contingency requirements less
Small units have the following disadvantages:
- Incremental capacity additions are small
- More expensive per MW for a given staff load
What Ontario Hydro did here in Ontario was build 4-packs. Multi-unit plants with common turbine halls increased the economy of scale factor and that's how we had previously built coal plants. While our first commercial nuke was a single unit (Douglas Point, 220MWe), the first large-scale setup at Pickering was 4x515MWe units, followed by Bruce A at 4x750MWe units, followed by Pickering B at 4x516MWe units, followed by Bruce B at 4x860MWe, followed by Darlington A at 4x880MWe.
One of the advantages of the above approach is that it minimized the effects of a large single unit loss, because we had a LOT of units. The maximum single unit trip is 880MWe out of a nuclear capacity that was, at the time, about 14,000MWe.
The grid is designed to connect large generators to centres; it uses large transmission corridors to move massive amounts of power from these sites to local distribution. Local distribution itself is not normally setup to have large quantities of embedded generation (this is the issue large quantities of residential solar creates, the grid isn't setup for it). Think of these transmission corridors as backbones. These are much larger and much more robust than local distribution infrastructure and therefore much more resilient.
Transitioning the traditional infrastructure, with limited large backbones, to a more dense spiderweb of "microgrids" with embedded generation will add considerable cost, both in terms of CAPEX, as well as OPEX. Maintaining even larger amounts of transmission is going to make things significantly more expensive.
I expect how we'll see SMR's deployed are in sites similar to how Ontario Hydro built our plants. You'll have a lot of units at a single site, which simplifies management and security and minimizes transmission infrastructure. That's just my take on it however, how it actually bears-out may be quite different.
The first commercial SMR to come online in North America will likely be the MMR at Chalk River, unless the BXWR 300 at Darlington beats it. OPG was well ahead of the game with an active EA and valid site license already, and they are using a cookie-cutter design based on a scaled-down traditional BWR design by GE, so it has the least risk of complication.