Comment by pasabagi | Hacker News Reader

Comment by pasabagi | Hacker News Reader

Parent Full thread

Just thinking the problem through, I think that's very unlikely, unless you're in a pathological situation.

So, for a nuclear plant to be profitable in a competitive market, it needs to produce energy for less than its competitors. If its competitors don't have to pay to capture carbon (through money or output) then they are always going to be cheaper per watt, and the plant is going to become unprofitable.

The only way that would work is if you had chronic electricity undersupply (v. likely but also pathological) or if nuclear plants were way more efficient than alternatives (the opposite is true).

Generally the argument for nuclear is you need a diverse range of energy sources to have a stable power supply, not that they are a great 'bang-for-buck' option. (An argument I agree with, fwiw).

EDIT, PS: It's worth keeping in mind that carbon capture is largely an unsolved technical problem, and I think it's perfectly feasible that it will never be solved. The basic idea of burning fossil fuels for energy, then using energy to recoup the carbon from the atmosphere, just makes no sense at all. There's no world where this would be cheaper than just not releasing the CO2 in the first place.

If you made all the alternatives also pay to capture their carbon cost, I think nuclear would be very competitive.

Obviously it blows fossil fuels out of the water, but hydroelectric also uses huge amounts of concrete, and IIRC the energy return on investment (MWh you get out per MWh it takes to construct) for solar and wind is quite poor and so I'd expect that to take a while to pay off too.

Nuclear has a vastly worse EROI than grid PV solar. Nuclear lasts longer which somewhat offsets construction costs, but needs a vast workforce continually and a maintenance and power simply to operate.

As a sanity check, sub 2c/KWh solar can’t require that much energy simply because energy costs money. Nuclear on the other hand ends up having ~1,000 workers drive somewhere for 50 years which also takes energy, but that’s often ignored.

That said, people love to come up with all kinds of irrelevant comparisons using vastly outdated data to make things look worse for solar while ignoring Nuclear’s downsides. For example France’s capacity factor fell below 70% when they tried to ramp up production and things would look even worse on a 100% nuclear grid.

It seems highly unlikely that grid PV solar has a higher EROI than nuclear and I can't find anything that supports this either (I find plenty of graphs supporting the opposite).

You can easily come up with some wildly different numbers but …

2013 Meta analysis puts nuclear at 14:1 EROI. https://www.sciencedirect.com/science/article/pii/S030142151...

You can find numbers all over the place for solar here’s something back in 2015 that goes as high as 34.2 : 1 depending on design and location: https://www.sciencedirect.com/science/article/abs/pii/S13640... and of course the best numbers have improved dramatically over the last 7 years. At the other end, Solar in Alaska obviously has much worse numbers.

It’s really interesting how much you can twist these numbers. For example the electricity to run equipment like pumps, computers etc at a nuclear power plant. It’s easy to call them as an input when the power plant is offline, but during operation are they actual cost or should the be treated as an inefficiency where only the external inputs count.

Yeah while EROI is supposed to be some sort of objective value, it is of course extremely sensitive to the data you put in. It is hard to even agree on what to include and where to draw the system boundaries. LCA suffers from the same issue but I do feel like EROI gets abused more for ideological purposes than LCA, I think LCA as a concept is a bit more robust by now. I mean this in a general sense, I'm not pointing fingers here.

Most of those are from fossil fuel or nuclear shills. You can get wildly varying numbers based on what you include and exclude.

Meta analyses say current/upcoming techs are about as good or maybe a bit better, but energy payback time is significantly lower due to reduced lifetime. Additionally the GHG impact of nuclearand wind isn't quite proportional to energy payback time because of concrete.

https://ieeexplore.ieee.org/document/8118077

https://www.sciencedirect.com/science/article/pii/S030142151...

Really though the answer is both are fine and nuclear probably wins once you include a mix of wind, hydro and storage. And embodied emissions will go down as low carbon energy becomes a bigger share.

The main arguments against new nuclear are political (making new oil barons but with nuclear is stupid among other problems) and opportunity cost (using a nuclear reactor to replace 1GW of fossil fuels 80% of the time uses the same resources as using renewables to replace 10GW of fossil fuels ~40% of the time -- some of which is uncorrelated -- but at much longer delay).

I largely agree with what you said, but that 80% capacity factor is rather nuanced.

Nuclear’s capacity factor varies quite a bit depending on the local grid. France’s averaged below 70% while US nuclear reactor’s where generally over 90% largely due to the local energy mix. Solar is running into similar issues as it scales, but they both can greatly benefit from energy storage. The issue is nuclear power is so expensive it makes energy storage less appealing.

So, if we are talking about a potential grid with 30% nuclear 10% hydro and the rest solar and wind then nuclear could have a 90+% capacity factor and mostly be offline in the spring and fall when seasonal demand is low. Unfortunately that means Nuclear isn’t going to scale and help with global warming. On the other hand if you want a 90% nuclear 10% hydro grid then nuclear is going to have something like a 40% capacity factor without a great deal of energy storage.

In the context of "turn the most gas and coal plants off per dollar ASAP", I think 80% is pretty reasonable. Especially when you include things like plants that were paid for but never opened and various safety shutdowns.

Similarly 25-30% for solar and 60% for wind with them being anticorrelated enough that you can call an equal mix ~40% over all.

The other benefit is solar+wind incentivises storage development in a way that nuclear does not. Cost-wise, for many applications it's optimal to overbuild by a factor of 2 or 3 just to get energy in the sides of the energy curve. This results in energy that has been paid for, and can be turned on or off in milliseconds but has no value at 11am.

Many storage technologies become viable at $10-20/MWh that need further development to be viable at higher price.

That’s a reasonable point, but I think it’s still overestimating the value of adding new nuclear power plants when you already have significant nuclear capacity.

It’s clear the average capacity factor falls when you add more nuclear power. In other words rather than looking at the average capacity factor as being X%, you need to consider how much the average capacity factor falls with each additional reactor and then calculate what the effective capacity factor of that reactor must have been to hit the new average.

To illiterate if 10GW averages to 80% and 11GW averages to 79% then the 11th GW must have effectively been 69%, because (80 * 10 + 69) / 11 = 79.

I would say the main arguments against nuclear are political; the waste storage problem and a fear for accidents. There can also be laws against building new nuclear plants or taxes that disproportionally disadvantage nuclear. The huge capital costs don't speak in favor of nuclear either, but I do wonder how the other aspects affect even considering to build new plants.

I don't really follow your opportunity cost argument. Do you mean the price of 1 plant equals the price of 10GW renewables? Are you taking into account the fact that you have to replace the renewables after about 20 years while you can run a new plant for about 60 years?

At $0.8/gross watt and 25% capacity, solar breaks even at around 20 years at 2c/kWh. At 3c/kWh you can also cover interest and upkeep and make a net profit. Prices are falling double digit percentages every year. The story is similar for wind, although prices are no longer falling rapidly.

2-3c/kWh for nuclear takes care of your variable costs of operation. You will never pay back the building costs at 3-4c/kWh.

The financial payback time for renewables is shorter than the build time for the nuclear plant, whereas the nuclear plant will need subsidy to stay on after 20 years or so.

Ie. if I have $10 billion right now, I can spend it on nuclear. In 15-20 years I will have ~0.8-1GW of net capacity that will begin to produce energy at a marginal cost of 3 to 4 cents per kWh and nothing in the bank.

Or I can spend it on ~4GW net of solar. In 20 years I will have my original money back, a modest profit, and a mostly worn out system that will conservatively produce ~1-2GW net at a marginal cost of around 1c/kWh for another ten years. Additionally industry trends imply that half of my money will buy me 6-8GW net of capacity in the same space if I simply replace the panels with 30-40% efficient ones (as over half of the current costs are land, permiting, and frame).

The outlook is not quite so compelling for wind, or high latitudes, but it still blows nuclear out of the water.

The story also looks a little better for nuclear if you're anticipating green hydrogen or ammonia costs to be prohibitively high or you decide to bulldoze the fossil fuel plants rather than using them as part of a chemical storage system. Additionally if you are anticipating energy growth nuclear looks good in the short term (but after 60 years we're basically back at square one arguing about thermal pollution rather than CO2)

That's the LCOE perspective, but the costs for the end consumer include ancillary functions that the TSOs have to procure or build out themselves. So far I have not seen an electricity grid with a high penetration of renewables that has resulted in lower total costs, which isn't that weird considering the price for a highly distributed grid is complexity.

Thermal pollution doesn't have to exist for newer plants, you can use the waste heat for district heating. If we assume there are no political forces that try to stop this (big if), then it becomes a purely economical calculation to see if it makes sense.

In another comment you mentioned 40% capacity factor in a system with equal parts wind and solar. This seems a little high right now. The best performing offshore parks are around 55-60%, but it's usually lower. For onshore it's 25-35% and offshore is 35-50%.

Have you got any solid info on current and projected total system costs? I can only find repetitions of a single UK somewhat optimistic study citing 10-14gbp (and limited in scope to 30% renewable mix) or vague fear mongering from fossil fuel interests.

As you probably know, it's messy and not entirely straightforward to get any decent numbers on the whole picture. Which is my roundabout way of saying that my sources could be better, but I do believe it helps paint a picture. Total system costs will increase because of stabilization measures that need to be taken now that we're moving towards a grid with a lot of variability, this includes but is not limited to various types of storage and frequency stabilization. The increased transmission capacity between countries and an increased number of transmission cables to connect all the remote wind turbine parks to the grid will also add to the costs.

[0] looks at Germany specifically, the report was put out by VDE which is a large technical branch organization [1].

Check out the slide on page 17 [2], it has projected wholesale costs for 2030. The median price isn't going down and the span is getting much bigger than during previous time periods. This doesn't show the total system costs, but we would clearly be in a grid with a high variability and this would result in cost increases as mentioned. This can be seen in [4], on page 16 they specifically mention "The rapid increase in investments in electricity distribution networks could moreover be correlated to the increase of the share of distributed electricity generation (especially from renewable energy sources) in the period and electrification in general." And yes, I know that correlation is not causation.

There is also this report [3] that does look at the total picture. The conclusion there is that it is better to have a mix than 100% renewables basically.

[0] https://www.euractiv.com/section/electricity/news/german-ele...

[1] https://en.wikipedia.org/wiki/VDE_e.V.

[2] https://www.ceps.eu/wp-content/uploads/2019/02/03082019%20CE...

[3] https://www.sciencedirect.com/science/article/pii/S030626191...

[4] https://trinomics.eu/wp-content/uploads/2020/11/Final-Report...

After selling electricity from a solar power plant you should have enough money to build it’s replacement in less than 20 years.

Both Nuclear and Solar get subsidies, but grid Solar generally doesn’t need them.

I was wondering if that was part of the calculations or not. It's capital that needs to be accounted for if you want to compare energy sources.

The existence of subsidies shouldn't really matter, the end consumer pays for them some way or another and you factor this into the total capital cost.

Obviously it's impossible to burn fossil fuels to capture carbon and remain carbon-negative. But that doesn't necessarily hold true for nuclear, solar, and wind. It might still be true, but it's plausible.

Reply on news.ycombinator.com