3 ms·

Traditionally, nuclear is not dispatchable either, it's baseload, and in particular these SMRs are not dispatchable. There are a few variable output reactors in

Traditionally, nuclear is not dispatchable either, it's baseload, and in particular these SMRs are not dispatchable. There are a few variable output reactors in France, but they are very expensive. Molten salt reactors could perhaps have a thermal storage component, but... Nuclear is "firm" in that its always running. This is a good characteristic. I'm more optimistic about advanced geothermal that has been developed in recent years, using the advances in drilling developed for fracking. If we can use the waste heat from nuclear for something productive, like desalination, then I think it has a much better chance. But coastal communities will also have access to offshore wind, which has really high capacity factors and will be far far cheaper than nuclear could be.

politician 4y ago
An array of SMRs is dispatchable even though a single SMR is not. It’s a scheduling problem.
- epistasis 4y ago
  Oh really, what's the ramp rate of this design? I haven't heard anybody claim that ability for SMRs. What makes an array more dispatchable than a single SMR? I don't follow how that could be possible.
  - njarboe 4y ago
    Elsewhere in the thread, credit_guy quotes 20% to 100% in 96 minutes[1]. [1]https://news.ycombinator.com/item?id=32370392 https://news.ycombinator.com/item?id=32370392
    - natmaka 4y ago
      Not sure 96 minutes will be sufficient in all load-following cases (bar a huge and therefore very expensive fleet of reactors). A gas turbine is way ahead here. Moreover in existing reactors the frequency of such adjustments is limited, for example "no more than twice a day, with a minimal 5-hours pause between modulations, and no more than 150 modulations per year". Approaching limits adds maintenance time (at shutdown state). I don't know if SMRs alleviate this, but given the causes I doubt so.
      - labawi 4y ago
        We will likely have sufficient battery storage where 1.5h would be acceptable though not quite ideal. It still requires quite a bit of storage, but its entirely doable for seasonal variations (the harder problem) and perhaps even daily cycles (the lesser problem).
throw0101a 4y ago
> Traditionally, nuclear is not dispatchable either, it's baseload, and in particular these SMRs are not dispatchable. Thermally, generally not. But one can design a system so that the steam is not sent to the turbines and so electrical power is not generated.
lamontcg 4y ago
> Molten salt reactors could perhaps have a thermal storage component If you can build one of those economically and deal with the caustic molten salt then you can build a concentrated solar thermal plant with molten salt for storage even more economically with no worries about a nuclear disaster.
- jonhohle 4y ago
  While nuclear disasters are acutely catastrophic, they are also exceptionally rare. In the US, nuclear power has 1/100th the deaths per TWh compared to hydro electric, 1/500th oil, and 1/1000th coal. I really hope the US can adopt a more aggressive approach to using more fuel, even if that means it may be closer to weapons grade at some point in its cycle. This would eliminate a significant amount of waste leaving catastrophic failure as the only real downside. I would be very happy with a nuclear+solar future (which is currently my present as well).
  - lamontcg 4y ago
    The last nuclear meltdown was literally a trillion dollar disaster. Wind and Solar don't have failure modes like that so you don't even have to think about it. This is why Nuclear tends to get more expensive over time because nobody can accurately calculate the risk of any design decision, so engineers and managers tend to make them more complex over time.
credit_guy 4y ago
The design we are talking about here has the capability to load follow: The unique features of a NuScale plant allow its modules to respond to meet the power generation demand in the evenings by increasing from 20% to 100% power in 96 minutes [1] https://www.nuscalepower.com/newsletter/nucleus-winter-2019/nuscales-diverse-energy-platform https://www.nuscalepower.com/newsletter/nucleus-winter-2019/...
- epistasis 4y ago
  Thank you! I'm surprised that I've not heard this before, and I never noticed it the many times I've been to NuScale's site. I'm clearly in left field when it comes to concerns about nuclear, in that I mostly am worried about wasted effort and capital on a tech that I do not foresee being able to keep up, rather than concerns about "safety" or belief that it is in anyway necessary for decarbonization. So I appreciate seeing some new information that make me more hopeful that SMRs will succeed.
- dmurray 4y ago
  Nuclear reactors aren't fundamentally incompatible with load following: you can stick some control rods in or out and get more or less heat energy distributed to your turbines. Nuclear power plants are generally used for base load for economic reasons: they are really expensive to build and have close to fixed hourly costs to operate, compared to say natural gas plants which are cheap to build and have operating costs proportional to the amount of gas they burn. Consequently we only build nuclear where we expect it can be utilized pretty much all the time, and we don't optimize nuclear plants for quick scaling up and down.

politician 4y ago

An array of SMRs is dispatchable even though a single SMR is not. It’s a scheduling problem.

epistasis 4y ago
Oh really, what's the ramp rate of this design? I haven't heard anybody claim that ability for SMRs. What makes an array more dispatchable than a single SMR? I don't follow how that could be possible.
- njarboe 4y ago
  Elsewhere in the thread, credit_guy quotes 20% to 100% in 96 minutes[1]. [1]https://news.ycombinator.com/item?id=32370392 https://news.ycombinator.com/item?id=32370392
  - natmaka 4y ago
    Not sure 96 minutes will be sufficient in all load-following cases (bar a huge and therefore very expensive fleet of reactors). A gas turbine is way ahead here. Moreover in existing reactors the frequency of such adjustments is limited, for example "no more than twice a day, with a minimal 5-hours pause between modulations, and no more than 150 modulations per year". Approaching limits adds maintenance time (at shutdown state). I don't know if SMRs alleviate this, but given the causes I doubt so.
    - labawi 4y ago
      We will likely have sufficient battery storage where 1.5h would be acceptable though not quite ideal. It still requires quite a bit of storage, but its entirely doable for seasonal variations (the harder problem) and perhaps even daily cycles (the lesser problem).

throw0101a 4y ago

> Traditionally, nuclear is not dispatchable either, it's baseload, and in particular these SMRs are not dispatchable. Thermally, generally not. But one can design a system so that the steam is not sent to the turbines and so electrical power is not generated.

lamontcg 4y ago

> Molten salt reactors could perhaps have a thermal storage component If you can build one of those economically and deal with the caustic molten salt then you can build a concentrated solar thermal plant with molten salt for storage even more economically with no worries about a nuclear disaster.

jonhohle 4y ago
While nuclear disasters are acutely catastrophic, they are also exceptionally rare. In the US, nuclear power has 1/100th the deaths per TWh compared to hydro electric, 1/500th oil, and 1/1000th coal. I really hope the US can adopt a more aggressive approach to using more fuel, even if that means it may be closer to weapons grade at some point in its cycle. This would eliminate a significant amount of waste leaving catastrophic failure as the only real downside. I would be very happy with a nuclear+solar future (which is currently my present as well).
- lamontcg 4y ago
  The last nuclear meltdown was literally a trillion dollar disaster. Wind and Solar don't have failure modes like that so you don't even have to think about it. This is why Nuclear tends to get more expensive over time because nobody can accurately calculate the risk of any design decision, so engineers and managers tend to make them more complex over time.

credit_guy 4y ago

The design we are talking about here has the capability to load follow: The unique features of a NuScale plant allow its modules to respond to meet the power generation demand in the evenings by increasing from 20% to 100% power in 96 minutes [1] https://www.nuscalepower.com/newsletter/nucleus-winter-2019/nuscales-diverse-energy-platform https://www.nuscalepower.com/newsletter/nucleus-winter-2019/...

epistasis 4y ago
Thank you! I'm surprised that I've not heard this before, and I never noticed it the many times I've been to NuScale's site. I'm clearly in left field when it comes to concerns about nuclear, in that I mostly am worried about wasted effort and capital on a tech that I do not foresee being able to keep up, rather than concerns about "safety" or belief that it is in anyway necessary for decarbonization. So I appreciate seeing some new information that make me more hopeful that SMRs will succeed.
dmurray 4y ago
Nuclear reactors aren't fundamentally incompatible with load following: you can stick some control rods in or out and get more or less heat energy distributed to your turbines. Nuclear power plants are generally used for base load for economic reasons: they are really expensive to build and have close to fixed hourly costs to operate, compared to say natural gas plants which are cheap to build and have operating costs proportional to the amount of gas they burn. Consequently we only build nuclear where we expect it can be utilized pretty much all the time, and we don't optimize nuclear plants for quick scaling up and down.