3 ms·

To be fair, to make the average output of a PV plant comparable to the average output of a nuclear plant you need to add some storage, which drives up the cost

To be fair, to make the average output of a PV plant comparable to the average output of a nuclear plant you need to add some storage, which drives up the cost a bit. Probably not enough to account for the extra cost of nuclear though.

kragen 4y ago
You do, although it's still unclear how much storage you'll need to add. On the other hand, while you can't turn PV plants on at night, you can't turn conventional nuclear plants off, and so to go all nuclear you need some form of load following: either gas peakers, or PV that carries a higher load during the day, or storage, or giant resistor banks. Evidently naval reactors are a lot better at this, though.
- adrianN 4y ago
  As far as I know it's not terribly difficult to model how much storage you'll need. It has a couple of variables, like how well connected your grid is, or how well you can shape demand, but then it's just looking at past weather data and deciding how low you want the probability for blackouts to be. Here for example is a toy implementation: https://model.energy/ https://model.energy/ From what I've heard the amount of storage is not very large if you assume sufficiently large grids and the total system cost is expected to be lower than what we currently pay for electricity.
  - kragen 4y ago
    Demand response is the biggest unpredictable factor, I think, but demand is also unpredictable. Power plants are commonly depreciated over 30 or even 40 years. Electricity consumption in the US doubled during the 01960s, and it doubled in PRC during the 02010s. If the transport sector in the US went all-electric, that would double electrical consumption again even without increasing energy use. How can you predict what fraction of your users will be using electric vehicles 15 years from now, and whether they'll charge them in the daytime or at night? That depends on, among other things, whether they'll go back to working in offices, whether the offices will have chargers, and how much cheaper it will be for them to charge during the day than at night. And that, in turn, depends on what kind of time-of-use rate schedules you can get the public utility commission to approve. Will people insulate their houses more so they don't need to heat them at night? That depends not only on the rate schedule but their future expectations of the rate schedule, as well as what their house buyers' expectations of the rate schedule. How much will superinsulating your walls raise the house's sale price? How about home TCES — if daytime electricity is sufficiently cheap, such forms of energy storage might become popular as a cheaper alternative to superinsulation, and maybe suppress the demand for nighttime electricity further in cold areas; but we don't have any mass-market experience with them right now. Will building codes, or insurance underwriters, impede the wide adoption of TCES after the first homeowner files a massive insurance claim to replace their hardwood floor ruined by a calcium-chloride spill? Will dirt-cheap rail-shipped carnallite drop the price of TCES further than calcium chloride possibly could? Will new carnallite deposits be found, closer to large cities? How about the future of industry? Mass PV rollout will drop the price of energy dramatically, so industrial processes that are currently unprofitable because they use too much energy will become profitable. They might outcompete more energy-efficient incumbents. The first example of this might be aluminum replacing steel in more and more uses; today it's so much more expensive than steel that it's only used in places where its lighter weight is a big advantage, but with lower energy costs (and shorter shipping distances) it might replace steel for many more purposes. Minimills recycling existing steel might be able to fulfill the entire steel demand for quite a while — steel mills that smelt iron from ore would shut down entirely, and the 30-year-depreciated power plants built to supply those mills might find themselves without the demand they were built for. And this might happen by 02039. This is maybe the most predictable outcome of much lower prices for electrical energy, but probably won't be the most dramatic one. So shifts in electrical demand will depend on cultural change, diffusion of innovations, regulatory capture, consumer expectations, and cost-driven substitution in the radically changed economic landscape. The last time the price of energy dropped so dramatically might have been Watt's steam-engine in 01776. The dislocations in society that resulted would have been very hard to predict.
    - pfdietz 4y ago
      Specifically, aluminum might replace steel because steel is being moved to an electrically reduced metal, with electricity separating oxygen from the oxide (indirectly perhaps, using electrolytic hydrogen). But aluminum is already electrically reduced. The cost advantage of using cheaper coke to make iron goes away.
      - kragen 4y ago
        Steel moving to being electrically reduced could happen for regulatory reasons, as you seem to be suggesting, or it could happen because the electricity becomes cheaper than the coke, at least during the daytime.
        pfdietz 4y ago
        Even so, this reduces the cost difference between iron and aluminum, so we'd see more applications shift to aluminum.
        kragen 4y ago
        Agreed! To be more specific, the volume electrochemical equivalent of trivalent iron is 8.83 μm dm²/A/hour (and that's the form of iron in hematite), and for aluminum it's 12.4 μm dm²/A/hour, calculated in units(1) as 26.982 dalton / (2.70 g/cc) / 3 e. Usually metal prices are quoted per kg, not per liter, so it might be better to calculate these as 190 μg per coulomb for iron and 93 μg per coulomb for aluminum: iron would still be cheaper, costing 49% as much (assuming the Faradaic efficiencies are comparable and the other aspects of the processes are of insignificant cost, which are obviously wrong assumptions), but right now aluminum costs US$3.10/kg (USGS MCS 2022) and steel is more like US$1.30/kg (same source: US$110 billion ÷ 87 million tonnes raw steel). This is a much smaller gap than I had expected, so maybe the switch to electrically reduced iron (directly, as in electrolytic iron, or indirectly) won't be as big a change as I thought.

kragen 4y ago

You do, although it's still unclear how much storage you'll need to add. On the other hand, while you can't turn PV plants on at night, you can't turn conventional nuclear plants off, and so to go all nuclear you need some form of load following: either gas peakers, or PV that carries a higher load during the day, or storage, or giant resistor banks. Evidently naval reactors are a lot better at this, though.

adrianN 4y ago
As far as I know it's not terribly difficult to model how much storage you'll need. It has a couple of variables, like how well connected your grid is, or how well you can shape demand, but then it's just looking at past weather data and deciding how low you want the probability for blackouts to be. Here for example is a toy implementation: https://model.energy/ https://model.energy/ From what I've heard the amount of storage is not very large if you assume sufficiently large grids and the total system cost is expected to be lower than what we currently pay for electricity.
- kragen 4y ago
  Demand response is the biggest unpredictable factor, I think, but demand is also unpredictable. Power plants are commonly depreciated over 30 or even 40 years. Electricity consumption in the US doubled during the 01960s, and it doubled in PRC during the 02010s. If the transport sector in the US went all-electric, that would double electrical consumption again even without increasing energy use. How can you predict what fraction of your users will be using electric vehicles 15 years from now, and whether they'll charge them in the daytime or at night? That depends on, among other things, whether they'll go back to working in offices, whether the offices will have chargers, and how much cheaper it will be for them to charge during the day than at night. And that, in turn, depends on what kind of time-of-use rate schedules you can get the public utility commission to approve. Will people insulate their houses more so they don't need to heat them at night? That depends not only on the rate schedule but their future expectations of the rate schedule, as well as what their house buyers' expectations of the rate schedule. How much will superinsulating your walls raise the house's sale price? How about home TCES — if daytime electricity is sufficiently cheap, such forms of energy storage might become popular as a cheaper alternative to superinsulation, and maybe suppress the demand for nighttime electricity further in cold areas; but we don't have any mass-market experience with them right now. Will building codes, or insurance underwriters, impede the wide adoption of TCES after the first homeowner files a massive insurance claim to replace their hardwood floor ruined by a calcium-chloride spill? Will dirt-cheap rail-shipped carnallite drop the price of TCES further than calcium chloride possibly could? Will new carnallite deposits be found, closer to large cities? How about the future of industry? Mass PV rollout will drop the price of energy dramatically, so industrial processes that are currently unprofitable because they use too much energy will become profitable. They might outcompete more energy-efficient incumbents. The first example of this might be aluminum replacing steel in more and more uses; today it's so much more expensive than steel that it's only used in places where its lighter weight is a big advantage, but with lower energy costs (and shorter shipping distances) it might replace steel for many more purposes. Minimills recycling existing steel might be able to fulfill the entire steel demand for quite a while — steel mills that smelt iron from ore would shut down entirely, and the 30-year-depreciated power plants built to supply those mills might find themselves without the demand they were built for. And this might happen by 02039. This is maybe the most predictable outcome of much lower prices for electrical energy, but probably won't be the most dramatic one. So shifts in electrical demand will depend on cultural change, diffusion of innovations, regulatory capture, consumer expectations, and cost-driven substitution in the radically changed economic landscape. The last time the price of energy dropped so dramatically might have been Watt's steam-engine in 01776. The dislocations in society that resulted would have been very hard to predict.
  - pfdietz 4y ago
    Specifically, aluminum might replace steel because steel is being moved to an electrically reduced metal, with electricity separating oxygen from the oxide (indirectly perhaps, using electrolytic hydrogen). But aluminum is already electrically reduced. The cost advantage of using cheaper coke to make iron goes away.
    - kragen 4y ago
      Steel moving to being electrically reduced could happen for regulatory reasons, as you seem to be suggesting, or it could happen because the electricity becomes cheaper than the coke, at least during the daytime.
      - pfdietz 4y ago
        Even so, this reduces the cost difference between iron and aluminum, so we'd see more applications shift to aluminum.
        kragen 4y ago
        Agreed! To be more specific, the volume electrochemical equivalent of trivalent iron is 8.83 μm dm²/A/hour (and that's the form of iron in hematite), and for aluminum it's 12.4 μm dm²/A/hour, calculated in units(1) as 26.982 dalton / (2.70 g/cc) / 3 e. Usually metal prices are quoted per kg, not per liter, so it might be better to calculate these as 190 μg per coulomb for iron and 93 μg per coulomb for aluminum: iron would still be cheaper, costing 49% as much (assuming the Faradaic efficiencies are comparable and the other aspects of the processes are of insignificant cost, which are obviously wrong assumptions), but right now aluminum costs US$3.10/kg (USGS MCS 2022) and steel is more like US$1.30/kg (same source: US$110 billion ÷ 87 million tonnes raw steel). This is a much smaller gap than I had expected, so maybe the switch to electrically reduced iron (directly, as in electrolytic iron, or indirectly) won't be as big a change as I thought.