Hacker News Reader

Comment by seanmcdirmid | Hacker News Reader

Not sure why you were downvoted, but this is correct. Outages are super expensive for a nuclear power plant, and plants take days to start and stop. Not only that, but capital costs are so large, you need to run it as much as possible to have any hope of just keeping up with interest payments.

People think waste is the biggest problem with nuclear, but I think it is huge capital costs and lack of flexibility are even larger problems (not to mention insurance for accidents, but only the government can provide that). Coupling nuclear with some kind of large scale battery would work to even out load requirements, but the only thing that might work is pumped storage, which would require a huge reservoir at a decent elevation, not to mention being pretty inefficient.

View on HN

Nuclear is mostly so expensive because of the huge capital costs, and lots of (political..) delays when building a new reactor.

Operating costs aren't so bad.

Capital costs for building nuclear reactors have gone up over time. That's partially to pay for necessary improvements to safety. But mostly down to double standards that require much higher standards in nuclear than other sources of power.

For example, coal plants release orders of magnitude more radiation than nuclear plants.

For those who, like me, were also shocked by the claim that coal plants release more radiation than nuclear plants: https://skeptics.stackexchange.com/questions/1018/do-coal-pl...

> The paper itself states that this result is only valid not considering nuclear accidents and nuclear waste, nor it considers non-radiological effects

Well yeah, nuclear plants don't release much radiation when there are no accidents and the waste is safely contained. But very often the waste isn't contained properly, and sometimes there are accidents. These events release a lot more radiation than coal plants.

> These events release a lot more radiation than coal plants.

Unless you have data that supports this statement, I'd state the opposite. It's easy to think nuclear accidents release a lot of radiation, but the actual accumulated dose from nuclear power (including accidents) for the average individual over time is very low. Coal, on the other hand, continuously spews out radioactive ash in large quantities.

Granted, other sources of radiation (radon, cosmic rays, medical x-rays, etc.) are a lot larger, but if comparing the two I'd guess coal is the larger culprit even when including nuclear accidents.

> the actual accumulated dose from nuclear power (including accidents) for the average individual over time is very low.

This is only because nobody's living in nuclear disaster zones. If people were carrying on their daily lives in the Chernobyl or Fukushima then their accumulated dose would be much much higher.

As you say, the levels from coal are small enough that other naturally occurring sources are more significant, and thus it has little practical effect on people's lives. On the other hand, nuclear accidents take large areas of land out of use for time scales long enough that they effect several generations.

The selected answer explains how a single nuclear accident can pretty much outdo all coal emissions.

I see China is building lots and lots of nuclear power stations - so it's economically viable for them.

I don't get why you're being downvoted. China is building a lot of power plants, a lot of them are nuclear (and coal, etc.).

Yes. Though that only shows that they can afford it, not necessarily that it's more economic than other forms of energy. Power production is highly regulated after all.

(It might very well be economically the best for them. But making that judgement requires more context.)

> delays when building a new reactor.

We create these problems for ourselves. Legal challenges are the number one cause of construction delays.

> For example, coal plants release orders of magnitude more radiation than nuclear plants.

See for example:

* https://www.scientificamerican.com/article/coal-ash-is-more-...

And if you have renewable sources that are much cheaper per kWh of output than nuclear, why charge those batteries with nuclear?

True except nuclear is basically a renewable itself given that we have enough fuel in reserve for 70k years or so. If we could get the capital and operating (including waste storage) costs down, it would basically be free energy. Barring achieving that, it is a problem that we should keep working on, it would be a pity to see nuclear plant development completely stopped (well, unless someone figures out profitable fusion).

I read there was 200years of nuclear (https://www.scientificamerican.com/article/how-long-will-glo...) but I can't find a source for your figures? Could you point me in the right direction?

For a long-term nuclear strategy you need breeder reactors which extract nearly 100 times more energy from the same fuel (and at the same time help a lot with the nuclear waste problem), or a way of getting Uranium out of sea water.

Thank you I learnt something

The 200 years number is for the most inefficient (and producing most radioactive waste) types of reactors, but those reactors are also the only ones USA wants you to have, so...

I see, which is might help explain breeder reactors are so few.

Most fast/breeder reactors get a knee-jerk reaction from so-called "nuclear haves". I think only France (a nuclear state) and Japan got close to so-called "plutonium economy" when it comes to reprocessing spent fuel, and even they don't use fast reactors on mass scale.

Most high fuel efficiency designs are fast reactors, or they employ extra radiation source to "burn down" the fissile material.

There are two reasons.

The first is that it gets you out of needing ordinary batteries, because instead of converting heat to electricity to batteries and back, you can take the heat generated by the reactor during the day, store it in an insulated vat of molten salt and then use that to drive turbines at night. This should be much cheaper than ordinary batteries because expanding storage capacity only requires having more salt in more vats, and salt is much less expensive than lithium batteries.

The second is that if you have a nuclear reactor for generating at night when solar doesn't, the incremental cost of also operating it during the day is lower than the solar generating cost. It can't recover its full fixed costs at the price of daytime solar, but as long as it can recover its incremental cost you still operate the reactor 24/7.

The problem with that approach is you need to scale the surprisingly expensive non-nuclear side of a nuclear plant for the peak load, not the average load. It's not just the salt, it's the turbines and generators and cooling towers. It also increases the number of heat exchangers, and also means you are not building a PWR, but rather a higher temperature reactor. These higher temperature reactors are either unproven (molten salt) or have disappointing track records (helium gas cooled).

Moltex claims that interposing a salt store would allow these components to be made to non-nuclear standards, but as far as I know no regulator has agreed with that.

I feel like you're just arguing in favor of building them.

"Unproven" is a synonym for we haven't done it yet, so the only way to prove it is to build them, so building them is the solution to that rather than a problem that should prevent building them.

And if the only problem is regulatory but there isn't actually any reason not to allow it, this too has an obvious solution.

The justification for treating the cooling system as part of the reactor presumably comes from older designs where the active cooling is necessary for safe operation. For newer passively-safe designs, what happens to the heat once it's removed from the reactor wouldn't matter any more than what happens to the electricity once it's removed from the reactor.

Unproven means exactly that. It means customers will not plop down billions to buy them. They will wait until it's demonstrated that the reactors will almost certainly work for their projected lifespans, that all the new parts aren't going to fail early, from all the unknown unknowns that new technology is subject to.

And all that means that new reactor technologies are not going to be really available for decades, and will be competing not with renewables and storage now, but the renewables and storage that will be available for purchase decades hence.

Because the price per kWh for renewable sources is artificially low. To give you some examples: Renewables still have various subsidies (e.g. in construction) while non-renewables are taxed, wind power plants in particular have shorter life spans that predicted, and the wings cannot be recycled but the landfill costs are generally not included.

Landfill costs are close to negligible.

Tax subsidies for renewables, at least in the US, are small compared to the levelized cost difference with nuclear.

If you look at Lazard's levelized cost of energy report, where wind is far cheaper than nuclear, they assume a 20 year lifespan for wind turbines. This is not unreasonable.

I countered this in another reply to you, but the levelized cost of energy is not a good way of comparing intermittent and baseload power sources since it completely ignores the costs associated with intermittency. Nuclear in particular is penalized heavily here. Since wind and solar are very intermittent, they look very cheap when you look at LCOE, but are very expensive in reality.

Of course LCoE is a simplification, but it cannot be ignored if the differences are large enough, as nuclear advocates wish to do. More detailed analysis has to be specific to particular situations. However, as I have also pointed out elsewhere, when optimizations are performed of the best way to produce output for "real" grids with real weather data, using realistic cost projections for 2030, new nuclear is left at 0%. The LCoE advantages of renewables are so great that the cost of dealing with the intermittency, even to 100% renewables, is less than the cost of using nuclear instead.

Ability to load follow depends on the type of the reactor.

We have a vicious cycle involved here, where old (if made safer) designs that can't load-follow are the only option, due to lack of funding for newer tech (like, late 1970s tech). At the same time the limits we put on nuclear mean that not only regulation-related costs are huge, reactors are made pretty much on one-off or very limited series basis, meaning we don't benefit from any kind of benefits of serial production, including better QA.

There are load-following small, modular reactors (adapted from nuclear submarines) that will even self-seal in case of rupture and that promise the option of just transporting them on trailer car. They languish in design bureau due to issues in getting funding to start even a demo plant in present political climate. China is helping move the tech a bit forward, but that's not enough.

Reply on news.ycombinator.com