Hacker News Reader

Comment by jillesvangurp | Hacker News Reader

True, however, solar and wind are proving to be the killer technology in the sense that they are forcing coal plants to actually go extinct at a rapid pace. They are being forced out of the market based on cost. Nuclear never had that power of persuasion because it was too expensive. That's why we have so many coal plants. Nuclear is part of the problem, not the solution. To be part of the solution it will have to be vastly cheaper than it is today. Probably by at least one or two orders of magnitude.

That's an interesting research problem to work on the next decades. By the time that happens or not, there won't be any coal plants remaining and most gas will be on the way out as well (considering that is already barely competitive today).

View on HN

Nuclear is upfront expensive, but the incredible efficiency wins hugely long term, AFAIK?

Nope. Investment has dried up in nuclear exactly because the ROI is garbage compared to the trend lines in renewables + storage + gas turbine plants. A lot of posters here have a sort of smugness about being fans of nuclear power, blaming it's decline purely on "irrational" fears. They blithely ignore that even in authoritarian states where there's no meaningful political opposition to nuclear projects, we see the same declining interest. It just takes too much capital too long vs alternatives now. Absent a global carbon tariff nuclear will likely never again be competitive with gas supplementing renewables.

Gas is on the way out as well. Renewables undercut them routinely. Gas plants are expensive to operate and in a peaker plant role they have to compete with cheap battery which means they are not running most of the time and only when all else has failed. And as we saw in Texas, that's not a good plan either when gas supplies run tight.

E.g. the Australians vastly reduced their need for gas plants just by installing a few hundred MW of battery and saved a lot of cash in the process. ROI on those batteries has been quite high very early on. Not having turn on their peaker plants is a big deal.

Really? China has been building plants quite steadfastly. Fukushima seemed to have put the breaks on that but still, there are plants being built. Yet Western companies, like Areva, seemed to have completely lost their game in making the plants and nobody's willing to buy them from China or Russia. And sure, there's a high capital cost but putting the solar panels or wind turbines to match the capacity of one modern nuclear plant isn't exactly cheap either.

Buying solar or wind gets you about 3x the capacity for the same price in about 6-12 months. It's a lot cheaper and a lot less risk. And that's excluding operational cost. If you want capacity on a budget in a hurry, wind is your best bet right now. Nuclear is for the opposite scenario where you have decades time and are not in a hurry to get capacity online and don't mind paying an order of magnitude more.

The Chinese hedged their bets a few decades ago and wind + solar won there as well. China is breaking records in terms of growth of their renewable capacity and are a key technology supplier in that space and also a big part of why it got so cheap.

Yeah, it's this, combined with how cheap and effective gas turbines are, co2 be damned. It's really hard to argue for several billion dollars on a 10 year min delay until revenue vs renewables plus gas that builds out in under a year and is net positive ROI within a shockingly short time.

Even though I'm generally pro nuclear, I've come to the conclusion that absent some innovation that changes the cost structure, it's an increasingly dead option. I was hopeful NuScale's approach would do that, but it's pretty clear now they're missing their needed numbers.

Nuclear operating cost is comparable to offshore wind, coal, and combined cycle gas, and around 4X more expensive to operate than onshore wind or solar. (Of course, coal and gas would be more expensive if they had to pay a fair price for their emissions)

Nuclear is also eye-poppingly expensive to decommission.

Finally, note that operating cost doesn't include the sort of re-engineering and upgrades that are routinely required, and have doomed plants like San Onofre.

Most levelized energy costs do include cost of decommissioning and maintenance.

On the other hand, estimates of wind and solar do not include the cost of storage to actually make intermittent sources viable. Which is understandable, because there's really no plan to provision this much storage without a massive breakthrough in storage solutions. Almost all plans for a predominantly wind and solar grid assume that something like hydrogen, synthetic natural gas, or something else will provide massive amounts of cheap storage. Until then, it's fossil fuels to fulfill off-peak demand.

Modern nuclear plants (with 50+ years of safety lessons) for baseload and a mix of other sources (wind, solar, etc) sounds ideal.

Peak energy consumption happens in the evening, when solar isn't producing any energy. And nuclear costs just as much to run at partial capacity as it does to run a full capacity. So if you build enough nuclear plants to support the grid when solar and wind isn't producing, then there's no reason to build any solar and wind. Nuclear makes the intermittent sources redundant.

If nuclear isn't subsidized by military uses its energy negative when you factor in mining, enrichment, and operation.

Not just upfront. Nuclear plants need staffing, security (lots of it) throughout its operational life, waste handling and storage, constant audits (and related bureaucracy) for safety, security, etc. None of that stuff is free. Additionally, nuclear plants work with materials that actually decay when exposed to lots of neutrons which I imagine needs occasional maintenance as well You need trained staff as well and a few propeller heads that actually understand the physics. These people need educations and salaries. None of that stuff is free. That's why nuclear is currently considered the most expensive form of energy in the market.

The problem with renewables (for nuclear) is that it reduces the potential revenue for a nuclear plant over its 6-7 decade lifespan to well below any level where it makes sense to build them. If you budget for 1960s era grid pricing it all looks beautiful. Then fast forward 6 decades and you have solar bids trending towards 0.01$/kwh when most nuclear plants in operation are basically closer to 0.15$/kwh.

That's the whole problem. It's the most expensive option in the market by far. a 10X improvement would make it competitive with current solutions. Except that will take a few decades. So, realistically, you'd need a second order of magnitude improvement at that point. Yes, I believe renewables can do another 10x cost reduction over the next decades; possibly more. Ballpark 0.005$/kwh is probably worth investing in for a plant that comes online around 2045 or so. That's a modest 3x cheaper than the lowest solar/wind bids right now. 25 years is a long time. In any case, last time I checked, such plants are not being planned anywhere and most nuclear bids are 10x or more expensive than that. Out of the gate they are 2-3x more expensive than solutions in the market right now. That's for a plant that has not been built yet and will take decades to get built.

What about decomissioning and caring for the nuclear waste? That seems to bei very expensive as the decomissioning of the Greifswald plant in former east Germany is going to cost more than 6 billion €.

This is an extremely complicated matter, to be honest. It isn't difficult to store waste on site. There are also much cheaper technologies out there that we've invented over the last 50 years but one of the major reasons we don't have them in practice is because they are "unproven", as in physically untested at large scale (though they have been through simulations and small scale testings at DoE labs).

Long term storage isn't too bad of an issue either, but there's two camps. One is to do smaller sites where we can just bury local material in the ground. The other is having a large mass site (think Yucca Mountain), which means higher scrutiny because there's more chances of something going wrong (standard failure analysis) and you're packing more material together which increases total radiation levels. Either one will work, but both are bureaucratic nightmares. The result of which has been constantly changing plans, which drives up costs very quickly as you change gears (e.g. spend tons to survey the US for good sites, more to verify Yucca Mountain is good, start digging, cancel because NIMBY, start again, cancel, survey other areas, repeat).

But one factor I want to mention is that by the nature of the physical processes materials can't be both extremely dangerous and long living. The danger literally comes from mass being expelled from the atoms. High level radioactive materials have and always will be stored on site, as this is the safest place for them (where they can be monitored).

There are still technical challenges, don't get me wrong, but the whole process is also a bureaucratic nightmare on top of that and we know what that does to costs.

For some better facts I'll refer you to an article written by a HN user that is a reactor scientist: https://whatisnuclear.com/waste.html

There's actually extremely small amounts of nuclear waste. The entirety of the waste produced by USA's nuclear electricity generation fits in a volume the footprint of a football field and less than 10 yards high [1].

The cost of storing waste is minuscule in comparison to the amount of electricity generated. Europe's waste repository in Finland costs ~800 million Euros [2]. Which is an order of magnitude less than a nuclear plant. And the repository can accommodate the fuel produced by several plants.

In short, waste disposal accounts for a single-digit percentage of nuclear power operating costs.

1. https://www.energy.gov/ne/articles/5-fast-facts-about-spent-...

2. https://en.wikipedia.org/wiki/Onkalo_spent_nuclear_fuel_repo...

There is far far more to nuclear decommissioning cost than just spent fuel storage. The costs are truly staggering and cannot be breezed past so simply.

Average decommissioning costs are $460-$730 million dollars per GW of capacity [1]. This is about 10% of the construction cost for most power plants in the GW range.

Decommissioning adds $0.80 per MWh of electricity generated, assuming a capacity factor of 90% and a lifetime of 80 years. By comparison, the cost of natural gas power is $40 to $80 per MWh.

You and I seem to have very different ideas of what is a "staggering" cost.

1.https://www.world-nuclear.org/information-library/nuclear-fu....

As far as I know building a long-term repository for the spent nuclear fuel is so simple... not a single nation solved it. The most advanced project (Onkalo, in Finland) may open in 2023.

Decommission costs are low... when planned. When a project at scale starts, however, things are usually less fun. Take a look at the UK. The UK discounted provision for decommission costs: £100+ billion in 2013 https://www.theguardian.com/environment/2013/jun/23/britain-...

£161 billion in 2017 https://web.archive.org/web/20170516093449/https://www.gov.u...

£234 billion in 2018. Costs were so quickly exploding there is a new way to forecast, which dissimulates... err... is more... well... https://www.gov.uk/government/publications/nuclear-provision...

https://www.theguardian.com/environment/2020/nov/27/uks-nucl...

Decommissioning small and old reactors costs more, and entombing may, at least apparently (short-term), reduce the cost. In theory. Let's check a real and ongoing case: Oyster Creek. According to the EIA its construction costs were $488 million (2007 USD) ( https://www.eia.gov/nuclear/state/archive/2010/newjersey/ ). As soon as the decommission project started the Nuclear Regulatory Commission announced that it will cost "about $1.4 billion to shut down the plant". Not for an immediate and complete decommission, because the plant will stay in a “safe store” condition until 2075, with dismantling ((...)) set for a period between 2075 and 2078 ( https://www.powermag.com/oldest-u-s-nuclear-plant-shuts-down... ). Then new problems (costs!) may arise. Let's bet that, as usual, the taxpayer will pay and the real cost will be hidden.

US: https://cleantechnica.com/2020/12/01/us-nuclear-site-cleanup...

Moreover if there is a serious glitch during decommission or at a waste repository site, are bets are off. You can obtain an insurance policy for anything, AFAIK even for a space trip, but no-one covers major nuke risk (reimbursements have a hard limit).

> As far as I know building a long-term repository for the spent nuclear fuel is so simple... not a single nation solved it. The most advanced project (Onkalo, in Finland) may open in 2023.

Back here in reality, there are several operating nuclear waste facilities: https://en.wikipedia.org/wiki/Deep_geological_repository#Nuc...

> £161 billion in 2017 https://web.archive.org/web/20170516093449/https://www.gov.u...

> £234 billion in 2018. Costs were so quickly exploding there is a new way to forecast, which dissimulates... err... is more... well... https://www.gov.uk/government/publications/nuclear-provision...

Do you realize that these figures include cleanup from nuclear weapons development?

> Decommissioning small and old reactors costs more, and entombing may, at least apparently (short-term), reduce the cost. In theory. Let's check a real and ongoing case: Oyster Creek. According to the EIA its construction costs were $488 million (2007 USD) ( https://www.eia.gov/nuclear/state/archive/2010/newjersey/ ). As soon as the decommission project started the Nuclear Regulatory Commission announced that it will cost "about $1.4 billion to shut down the plant". Not for an immediate and complete decommission, because the plant will stay in a “safe store” condition until 2075, with dismantling ((...)) set for a period between 2075 and 2078 ( https://www.powermag.com/oldest-u-s-nuclear-plant-shuts-down... ). Then new problems (costs!) may arise. Let's bet that, as usual, the taxpayer will pay and the real cost will be hidden. US: https://cleantechnica.com/2020/12/01/us-nuclear-site-cleanup...

Even if the pessimistic figures pan out and the existing nuclear fleet costs $70 billion to clean up, that's still within the $.4B to $.7B per GW range I provided.

> there are several operating nuclear waste facilities

22 sites.

Theoretically "actives" ones are:

- Olkiluoto which is in fact Onkalo (see https://en.wikipedia.org/wiki/Olkiluoto_Nuclear_Power_Plant#... ) which is "expected to be operational in 2023." (see https://en.wikipedia.org/wiki/Onkalo_spent_nuclear_fuel_repo... )

- Loviisa which is AFAIK a site dedicated to the local reactors(?), and only for "low- and intermediate-level operational wastes". Onkalo is the planned site, and is NOT active ( see https://www.world-nuclear.org/information-library/country-pr... )

- Wolseon (Korea) is for Low and Intermediate Level waste only (see https://en.wikipedia.org/wiki/Wolseong_Low_and_Intermediate_... ). Its total capacity (800k drums) is far away, as for now it only contains 100k drums (see https://www.korad.or.kr/korad-eng/html.do?menu_idx=37 )

- the "Waste Isolation Pilot Plant" (New Mexico) is, well, a "Pilot Plant" (doesn't come as a surprise). Moreover "The waste is from the research and production of United States nuclear weapons only", and only for 10k years. (See https://en.wikipedia.org/wiki/Waste_Isolation_Pilot_Plant ). In some dream one may deposit powerplant waste there, however back in reality...

"Back here in reality" (uh, uh) there is no active site in any huge nation which nuke-produces power for decades (US, France, Germany, Japan...) It will become more and more difficult for the public to believe that all this is perfectly managed, back here in reality.

> Do you realize that these figures include cleanup from nuclear weapons development?

Do you realize that the civil and military were so intimately tied during those phases that no one (even, in France, the supreme public finance audit authority, namely la 'Cour des comptes') can untangle the mess?

> Even if the pessimistic figures pan out

The UK had less than 20 nuclear powerplants and upon starting serious decommission they think that decommission costs will amount to at least £234b Even if (at best) only ~1/3 (?) will be necessary for purely civil reactors and sites that's ~£78b (~110bUSD)

The US have 58 active nuclear power plants (96 reactors) (see https://www.eia.gov/energyexplained/nuclear/nuclear-power-pl... ), approximately 3 times more.

>The cost of storing waste is minuscule in comparison to the amount of electricity generated. Europe's waste repository in Finland costs ~800 million Euros [2]

If this repository is planed for 100 000 years (in germany it's 1 mio years) that are 8000€ a year. That doesn't sound like a realistic number to me. And it's by no way europes waste repository, it's for the 2 finnish nuclear plants.

At this point it's gas plants that are forcing coal out of business. Solar/wind don't compete well with coal without extensive storage infrastructure that doesn't exist yet.

Reply on news.ycombinator.com