Hacker News Reader

Comment by pdonis | Hacker News Reader

> It takes decades to build a new plant.

No, it doesn't. It only takes a few years. What takes decades is resolving all the NIMBY lawsuits.

Vogtle, Summer, Flamanville 3, and Olkiluoto 3 weren't delayed by NIMBY lawsuits. They were delayed by construction screw ups. Nuclear power plants are complex, and if they are required to be constructed to high standards, delays and overruns are almost inevitable.

A PV field, on the other hand, is much simpler, with much more inherent redundancy, and can tolerate construction mistakes without either safety implications or substantial loss of capacity.

Three of the plants you mentioned were 40+ years ago, right? Seems a bit outdated to point to in 2020.

None of the plants I mentioned were 40+ years ago. They are all contemporary. Vogtle is still under construction, Summer was cancelled a couple of years ago ($9B down the drain), Flamanville 3 is still under construction, and the last is near completion. You may be confused with earlier reactors at some of those sites.

Olkiluoto 3 is a new design though. Making a second exact same one again will be much faster and cheaper as you have now worked out most of the issues.

Olkiluoto 1 and 2 were both built quickly and on schedule. Mainly because they were already proven designs with tens of other units already built around the world (and the second reactor is practically identical with the first one)

> Olkiluoto 3 is a new design though. Making a second exact same one again will be much faster and cheaper as you have now worked out most of the issues.

The second exact same one is actually Flamanville 3, the other one with the massive cost overruns.

Yes. Though they did start building it before Olkiluoto 3 was up and running so can't really use all the learnings from it.

Vogtle 1 and 2 were completed on 1987 and 1989. They will be online until 2047 and 2049. Vogtle 3 and 4 are currently under construction and will be completed in 2021.

I was refering to 3 and 4 (and Summer 2 and 3)

Virginia-class nuclear submarine(GE "S9G" PWR using HEU cores, 210 MWth) takes as little as 1-2 year from laydown to launching. 19 such hulls is in service, 11 in construction.

Thats still pretty small and usage specific - modern 1000 MW electricity producing nuclear power plants produce about 3000 MW of heat, so about 15x the naval nuclear reactor in your example.

Like I said, US had commissioned around 20 of those reactors in the past couple decades.

Fair enough - still what I'm trying to say is that those 20 cover the thermal output of a single modern reactor & many are needed even just to get rid of seriously polluting power sources, like the brown coal/lignite fired power plants used in many places in Europe.

How is the cost per MWh in construction and ops?

The Chinese have started building two reactors of similar conception as Flamanville 3 and Olkiluoto 3 after these, and they both came online in 2018 after a little less than 6 years of construction. Proof enough that it's totally political, if you ask me.

Looking at this page [1] it seems China takes about 5 years to go from start of construction to commercial operation of a reactor.

I'm pretty sure they have no NIMBY lawsuits to content with, so this seems like a realistic timeframe.

[1]: https://en.wikipedia.org/wiki/List_of_nuclear_reactors#China

Sure it does. It takes one decade for America to build 1.7 miles of rail. That's for a thing that is done daily across the globe. Doing a thing that's done rarely? Yeah, I have no hope that there is an American company today that can do it.

By the time we build expertise it will be 2080. And all that capital will be tied up until then. Who even knows if fission will be the right tool then.

So in practice is does take decades.

Reply on news.ycombinator.com