[Tade0]

There's one problem with nuclear that probably will never be overcome: it takes around 20 years from commissioning to the eventual connection of a nuclear power plant to the grid.

You can have an equivalent in solar, wind + storage in less than three years. That's 17 years of not producing electricity.

[Turing_Machine]

"You can have an equivalent in solar, wind + storage in less than three years."

A nuclear power plant can crank out 5 gigawatts.

There aren't any solar or wind plants of that capacity that were constructed in any amount of time, much less three years.

[philipkglass]

A utility scale solar farm can go from first concrete pouring to selling electricity in one year. And you can build a lot of farms in one year in parallel. For example, last year the United States installed ~10 GW of utility scale PV, mostly in farms under 200 MW each.

Based on EIA data, in 2016 the capacity factor of US nuclear power was 92.5%. Utility scale PV was 27.2% and wind was 34.7%: https://www.eia.gov/electricity/monthly/epm_table_grapher.cf...

Based on 2016 capacity factors you need to install a total of 17 gigawatts-peak of American utility scale PV or 13.3 GWp of wind to match the annual energy production of 5 gigawatts-peak of nuclear reactors. Last year the US completed about 10 GWp of utility scale PV and 8.2 GWp of wind (5.57 real annualized gigawatts, assuming same 2016 capacity factors going forward).

Storage, though? Nobody's going to install sufficient storage in the next 3 years that you could shut down the nearest reactor without burning more natural gas. Large scale storage is just starting to enter the mainstream.

[Turing_Machine]

"For example, last year the United States installed ~10 GW of utility scale PV"

So if a nuclear plant has 92% capacity factor, and a solar plant has 27%, that means a 5 GW nuclear plant would require about 17 GW of solar to replace it. In other words, there was less than 2/3 of a nuke's capacity of solar installations last year. Total. For the entire country.

[philipkglass]

"A nuke's capacity" -- not quite. There are only a handful of nuclear plants running with net capacity over 5000 MW. The largest nuclear plant in the US is Palo Verde: 3942 MW from 3 reactors. The sole reactor to enter operation in the US last year, Watts Bar 2, generates 1165 MW net: https://www.iaea.org/PRIS/CountryStatistics/ReactorDetails.a...

So multiplying capacities and capacity factors, for 2016 we get:

2016 nuclear: 0.925 * 1165 = 1078 MW

2016 utility PV: 0.272 * 10000 = 2720 MW

2016 utility wind: 0.347 * 8200 = 2845 MW

Last year's large scale wind and solar installations would be expected to produce as much electricity, annually, as 5 reactors like Watts Bar 2. Watts Bar 2 is pretty typical as far as capacity goes. The four new AP1000 reactors under construction in the US are each 1117 MW net, for example.

I personally hope that the US continues to build enough reactors to regain and maintain the institutional competence that was lost after the last wave of builds ended in the 1980s. Nuclear power has an excellent safety record, no air pollution, and is easier to manage at high penetration levels than wind/PV. But when I'm crunching the numbers I estimate that renewables will grow faster in North America for the foreseeable future; renewable projects scale down better, are cheaper per MWh generated even adjusting for capacity factor [1], and are much faster to go from first concrete pouring to first revenue.

[1] Cheaper per MWh than observed real-world prices for reactors currently under construction. Some older reactors can produce electricity at significantly lower costs.

[Tade0]

Last year China added 23GW of wind power.

Adjusted for the capacity factor that's over 5.75GW. Sure, it's not one plant, but quite possibly most of that capacity wasn't commissioned until 2013. Better still, they're just getting started with this.

The beauty of this power source is that it scales great - instead of erecting one 7MW turbine you can do that with any number of them, limited only by your budget.

Meanwhile it takes 5 years to construct a single 1GW block of nuclear and you can't just simultaneously start building another one next to it.

Nuclear won't win, because it's expensive and doesn't scale well.

[Turing_Machine]

"Nuclear won't win, because it's expensive and doesn't scale well."

https://en.wikipedia.org/wiki/List_of_largest_power_stations

I see several nuclear plants on that list, but not one solar or wind plant.

I think perhaps claims about the latter being "less expensive" and "scaling well" should wait until they've actually been economically operated at scale.