5 ms·

I'm not so confident about nuclear. Nuclear power generation is a very high end technology, where humankind instead of climbing the learning curve, somehow mana

I'm not so confident about nuclear. Nuclear power generation is a very high end technology, where humankind instead of climbing the learning curve, somehow managed to go in the opposite direction. However, Elon Musk has demonstrated that you can take a technology where costs are going up, and make it do a 180, and take down costs by a factor of 10 and then 100. Can nuclear power generation get to be 100 times cheaper? I think so. Do we absolutely need nuclear energy to combat climate change? I don't think so. I do think a solution could be based on huge arrays of solar panels in countries like Australia, Mexico, Morroco or South Africa, and shipments of liquefied hydrogen to countries like Germany, China, Japan or Canada. The US can be quite self-sufficient, with solar panels in California, Nevada, Arizona, and wind turbines in Texas, Oklahoma and the rest of the wind corridor, and pipes of hydrogen from one state to another. I can even see the US exporting liquefied hydrogen as a natural succession from exporting LNG now.

ZeroGravitas 4y ago
It's essentially impossible for anything that drives a steam turbine to generate electricity to get that cheap. Just the cost of the steam parts, if everything else was free struggles to compete with solar today. Distributed solar also avoid transmission costs, at a certain point a point source of free energy gets outcompeted by distributed solar. So nuclear might have niches, or exciting new applications but is mostly a dead-end. Shipping hydrogen about seems likely to be a big thing though, easy transition for existing infrastructure.
- credit_guy 4y ago
  I read this opinion in that blog of Austin Vernon that made the HN rounds about a year ago [1]. In the end his argument is basically "if some energy generation method has moving parts, it has to be more expensive that one that doesn't". But let's look at the numbers. Here's a study by the EIA that looks at the capital costs and also at the Operating and Maintenance costs of various types of power plants [2]. The costs that are interesting for us are (all costs per GW alternative current of capacity): - solar without storage: $1.3 BN (page 175) - solar with 4 hours of storage: $1.8 BN (p. 180) - coal without CO2 sequestration: $3.7 BN (p. 46) - nuclear AP1000 (similar to Vogtle): $6.0 BN (p. 107) - natural gas power plant: $1.0 BN (p. 83) First of all, notice that the natural gas power plant comes to be the cheapest, despite the fact that it has turbines, including steam turbines. So the general assertion that things with moving parts are more expensive than those without is not quite correct. You can counter that the estimates are from 2019, and by now solar is probably already cheaper and it will continue to get so, and I suppose you are right. But the panels constitute only 18% of the cost of the solar panel plant, the rest being inverters, transformers, something called BOP (balance of plant), etc. So, if the panels go to zero, the cost of the solar power plant gets reduced by 20% only. On the other hand, the same is true for nuclear: only about 20% of the cost of a nuclear power plant is attributed to the nuclear part, the rest to the turbines, BOP, etc. So, even if by a miracle the nuclear part were to cost zero, the cost per GW would still be about $5 BN. This is higher than the coal power plant, because the steam generated by coal is much hotter than the one generated by the current generation of nuclear reactors. Does this prove Austin Vernon's point? Not necessarily. First, the fact that gas power plants are very cheap is reason to hope. If we make nuclear reactors that don't use steam (which is corrosive) but some other gas, then it's possible the cost could go down significantly. Well, less than one year ago China hooked to the grid exactly such a plant [3]. It uses Helium as a coolant, which means that the turbines can be cheaper (they should be similar to the ones used in a gas-firing plant). It also runs much hotter than a regular pressurized water reactor (about 700 Celsius vs 300 Celsius), which means the efficiency is higher. Can the US build such reactors? I don't see why not, see for example Xe-100 [4]. But the first step is to get back to the nuclear technology learning curve. [1] https://austinvernon.site/blog/nuclear.html https://austinvernon.site/blog/nuclear.html [2] https://www.eia.gov/analysis/studies/powerplants/capitalcost/pdf/capital_cost_AEO2020.pdf https://www.eia.gov/analysis/studies/powerplants/capitalcost... [3] https://en.wikipedia.org/wiki/HTR-PM https://en.wikipedia.org/wiki/HTR-PM [4] https://www.energy.gov/ne/articles/x-energy-developing-pebble-bed-reactor-they-say-cant-melt-down https://www.energy.gov/ne/articles/x-energy-developing-pebbl...
  - pfdietz 4y ago
    Gas is cheap because it substantially reduces the number of heat exchangers. Transferring heat across a solid/fluid interface is kind of expensive. The more of that you have to do, the more expensive your power plant will be. A simple cycle combustion turbine involves no heat exchangers at all. Heat is generated in the compressed air by combustion; the waste heat is carried away in the air + combustion products exiting the turbine. A combined cycle plant does have a boiler and a condenser for the steam section, but that part is only producing 1/3rd the power. For nuclear to get cheap we'd need something like an open air Brayton cycle, but that would involve running air either through the reactor or through a high temperature heat exchanger made of silicon carbide. It's interesting (well, to me) to consider how much easier that sort of open cycle system would be on Titan, where one could use 90K N2 rather than 300K air as the input fluid. Titan may be the best place in the solar system for nuclear power.
    - credit_guy 4y ago
      That's the idea behind the gas-cooled fast reactor [1]: The reference reactor design is a helium-cooled system operating with an outlet temperature of 850 °C using a direct Brayton closed-cycle gas turbine for high thermal efficiency. Helium does not absorb neutrons when it passes through the core, and if the nuclear fuel is well insulated (such as the TRISO fuel that Xe-100 plans to use), then you are safe to send it to spin the turbine. If you have any concerns that the cooling gas may become contaminated somehow, then you use a heat-exchanger. At this point nobody is using direct cooling-gas-to-turbine cycle. But after people gain experience with cooling-gas-to-heat-exchanger, at some point they'll move to the direct cycle. [1] https://en.wikipedia.org/wiki/Gas-cooled_fast_reactor https://en.wikipedia.org/wiki/Gas-cooled_fast_reactor [2] https://en.wikipedia.org/wiki/Nuclear_fuel#TRISO_fuel https://en.wikipedia.org/wiki/Nuclear_fuel#TRISO_fuel
      - pfdietz 4y ago
        A problem with these reactors using TRISO fuel is it greatly increases the volume of waste. The spent fuel is now mixed with a large amount of carbon. Cask storage becomes less practical. Also, any time the temperature is above 550 C you can't use ordinary steels, as they will creep. So these reactors have materials challenges.

ZeroGravitas 4y ago

It's essentially impossible for anything that drives a steam turbine to generate electricity to get that cheap. Just the cost of the steam parts, if everything else was free struggles to compete with solar today. Distributed solar also avoid transmission costs, at a certain point a point source of free energy gets outcompeted by distributed solar. So nuclear might have niches, or exciting new applications but is mostly a dead-end. Shipping hydrogen about seems likely to be a big thing though, easy transition for existing infrastructure.

credit_guy 4y ago
I read this opinion in that blog of Austin Vernon that made the HN rounds about a year ago [1]. In the end his argument is basically "if some energy generation method has moving parts, it has to be more expensive that one that doesn't". But let's look at the numbers. Here's a study by the EIA that looks at the capital costs and also at the Operating and Maintenance costs of various types of power plants [2]. The costs that are interesting for us are (all costs per GW alternative current of capacity): - solar without storage: $1.3 BN (page 175) - solar with 4 hours of storage: $1.8 BN (p. 180) - coal without CO2 sequestration: $3.7 BN (p. 46) - nuclear AP1000 (similar to Vogtle): $6.0 BN (p. 107) - natural gas power plant: $1.0 BN (p. 83) First of all, notice that the natural gas power plant comes to be the cheapest, despite the fact that it has turbines, including steam turbines. So the general assertion that things with moving parts are more expensive than those without is not quite correct. You can counter that the estimates are from 2019, and by now solar is probably already cheaper and it will continue to get so, and I suppose you are right. But the panels constitute only 18% of the cost of the solar panel plant, the rest being inverters, transformers, something called BOP (balance of plant), etc. So, if the panels go to zero, the cost of the solar power plant gets reduced by 20% only. On the other hand, the same is true for nuclear: only about 20% of the cost of a nuclear power plant is attributed to the nuclear part, the rest to the turbines, BOP, etc. So, even if by a miracle the nuclear part were to cost zero, the cost per GW would still be about $5 BN. This is higher than the coal power plant, because the steam generated by coal is much hotter than the one generated by the current generation of nuclear reactors. Does this prove Austin Vernon's point? Not necessarily. First, the fact that gas power plants are very cheap is reason to hope. If we make nuclear reactors that don't use steam (which is corrosive) but some other gas, then it's possible the cost could go down significantly. Well, less than one year ago China hooked to the grid exactly such a plant [3]. It uses Helium as a coolant, which means that the turbines can be cheaper (they should be similar to the ones used in a gas-firing plant). It also runs much hotter than a regular pressurized water reactor (about 700 Celsius vs 300 Celsius), which means the efficiency is higher. Can the US build such reactors? I don't see why not, see for example Xe-100 [4]. But the first step is to get back to the nuclear technology learning curve. [1] https://austinvernon.site/blog/nuclear.html https://austinvernon.site/blog/nuclear.html [2] https://www.eia.gov/analysis/studies/powerplants/capitalcost/pdf/capital_cost_AEO2020.pdf https://www.eia.gov/analysis/studies/powerplants/capitalcost... [3] https://en.wikipedia.org/wiki/HTR-PM https://en.wikipedia.org/wiki/HTR-PM [4] https://www.energy.gov/ne/articles/x-energy-developing-pebble-bed-reactor-they-say-cant-melt-down https://www.energy.gov/ne/articles/x-energy-developing-pebbl...
- pfdietz 4y ago
  Gas is cheap because it substantially reduces the number of heat exchangers. Transferring heat across a solid/fluid interface is kind of expensive. The more of that you have to do, the more expensive your power plant will be. A simple cycle combustion turbine involves no heat exchangers at all. Heat is generated in the compressed air by combustion; the waste heat is carried away in the air + combustion products exiting the turbine. A combined cycle plant does have a boiler and a condenser for the steam section, but that part is only producing 1/3rd the power. For nuclear to get cheap we'd need something like an open air Brayton cycle, but that would involve running air either through the reactor or through a high temperature heat exchanger made of silicon carbide. It's interesting (well, to me) to consider how much easier that sort of open cycle system would be on Titan, where one could use 90K N2 rather than 300K air as the input fluid. Titan may be the best place in the solar system for nuclear power.
  - credit_guy 4y ago
    That's the idea behind the gas-cooled fast reactor [1]: The reference reactor design is a helium-cooled system operating with an outlet temperature of 850 °C using a direct Brayton closed-cycle gas turbine for high thermal efficiency. Helium does not absorb neutrons when it passes through the core, and if the nuclear fuel is well insulated (such as the TRISO fuel that Xe-100 plans to use), then you are safe to send it to spin the turbine. If you have any concerns that the cooling gas may become contaminated somehow, then you use a heat-exchanger. At this point nobody is using direct cooling-gas-to-turbine cycle. But after people gain experience with cooling-gas-to-heat-exchanger, at some point they'll move to the direct cycle. [1] https://en.wikipedia.org/wiki/Gas-cooled_fast_reactor https://en.wikipedia.org/wiki/Gas-cooled_fast_reactor [2] https://en.wikipedia.org/wiki/Nuclear_fuel#TRISO_fuel https://en.wikipedia.org/wiki/Nuclear_fuel#TRISO_fuel
    - pfdietz 4y ago
      A problem with these reactors using TRISO fuel is it greatly increases the volume of waste. The spent fuel is now mixed with a large amount of carbon. Cask storage becomes less practical. Also, any time the temperature is above 550 C you can't use ordinary steels, as they will creep. So these reactors have materials challenges.