> Renewables are about an order of magnitude cheaper than nuclear.
I think there are a lot of complications that are encountered when doing this sort of calculation. You might get a max output of "$15/MWh" or "$112/MWh" but that's only a portion of the picture.
Firstly Nuclear is X amount of power essentially on tap 24/7 where X is between some Min and Max value. From what I understand this makes grid management much simpler and reduces storage, transmission, and conversion losses. Even in the best case situation of "renewable" energy you'll see, in the absolute best case claimed by pumped hydro storage, you're seeing about an 87% efficiency.
Suppose you had a nuclear power plant that could output 100W consistently. Suppose you wanted to build a similarly efficient "renewable" energy source. How much more peak power would it need to output to be competitive? If we assume the power source is 0% efficient at night and 100% efficient during the day, off the bat, you'd need to generate 200W continuously during the day. Then you'd need to generate an extra 13% on those extra 100Ws to make up for storage losses (1.13*100 ~= 113). That means 100W from Nuclear is ~= 213W "renewable". Obviously this is a terrible calculation and not at all accurate but it's just a sample of some of the math that needs to be done.
Another factor that I've been skirting around is that "renewable" sources of energy are often not so renewable. Specifically things like like solar require a huge amount of rare metals & complex fabrication. Right now none of those externalizations are represented in the price of solar because there are no carbon taxes in countries that manufacture them.
Money in this case is a proxy of what needs to be measured and should not be the only factor we optimize for.
Nuclear is really good at producing consistent power. Unfortunately, that's not what we want. Peak demand for power in the evening is several times what it is during the middle of the night.
Except that it is exactly what we want for the level of demand that we never fall below, which is still about half of the peak demand. And if we use nuclear for that, it significantly reduces the amount of storage we need for renewables.
Also, nuclear combines well with thermal storage because you don't incur double conversion losses. If you store heat from the reactor during the day when solar is generating, use it during the peak demand in the evening when it isn't, and then run directly from the reactor overnight, you don't really need any other storage.
> Except that it is exactly what we want for the level of demand that we never fall below, which is still about half of the peak demand. And if we use nuclear for that, it significantly reduces the amount of storage we need for renewables.
The bottom half of the demand is the easy half, why would you use your most expensive power for the easy stuff?
> Also, nuclear combines well with thermal storage because you don't incur double conversion losses. If you store heat from the reactor during the day when solar is generating, use it during the peak demand in the evening when it isn't, and then run directly from the reactor overnight, you don't really need any other storage.
In other words, you're running your nuclear at a 50% duty cycle, doubling it's price. I don't think there is much appetite for power that costs over $200 per MWh.
I think this is, again, too simplistic a model for understanding the costs of different power systems. Industrial use cases (think data centers, factories, etc) likely reduce usage during night time hours but they surely don't drop by 50% usage.
> The bottom half of the demand is the easy half, why would you use your most expensive power for the easy stuff?
Factoring in carbon footprints nuclear would likely beat our battery based storage options if carbon output was taxed. Nickel–metal hydride battery and LiPo are not carbon neutral and are consumable components.
> In other words, you're running your nuclear at a 50% duty cycle, doubling it's price. I don't think there is much appetite for power that costs over $200 per MWh.
50% duty cycle means 50% fuel usage (not really but not not) meaning you're not really "wasting" energy here.
Another question to ask ourselves: we make the Total Cost of Ownership of a plant far better than it even currently is? We can do this by extending the lifespan of a plant to amortize construction. If you could design a power plant to last 100 to 200 years, that could be amortized further factoring into lower costs. We can also do this by reducing construction costs or improving efficiency.
I think those are accomplish-able goals (at least judging by China's adoption of Nuclear). Current plants in the USA operate for ~40 years. That means current plants were built with ~$today-50 (or more) years of physics and material science knowledge.
For context, that's the 70s. In the 70s:
- C was invented
- TTL logic was just starting to be used
- Stainless steel did not exist <- this statement was wrong
Edit: philipkglass pointed out that stainless steel is quite a bit older. Disregard that last item.
Nuclear requires huge amounts of cement and steel to build. Both of these resources are enormous sources of CO2. Cement clinker manufacturing, in particular, is a horrible emitter, responsible for 8% of global CO2 emissions all by itself.
That’s one type of building material among many that are needed in huge quantities.
That's a pretty weak argument. Those are once-off building requirements, not something that is repeated. If you compare the cement used by the one nuclear powerplant to the closest city it would be like a drop in the ocean.
There's been large improvement in cements specifically in the past few years. It can be recycled, enforced, and supplanted by other materials.
Concrete sold today can be bought with an advertised lifespan of 100 years. If you replace a coal plant with a nuclear plant how many years would it need to be net carbon negative? What is the grams of CO2 / kWh of nuclear vs coal vs solar?
Concrete resorbs CO2 from the air after you pour it. This carbonation of concrete can cause corrosion to rebar or other re-inforcments. A large portion of that CO2 is eventually reabsorbed back by the concrete. Depending on the mix this process can adsorb 1/3 to 2/3rds of the CO2 produced during production. However, this is a slow process and we keep making new concrete.
It's not so easy when you have to meet it even when there is no wind or sunlight for twelve hours straight.
> In other words, you're running your nuclear at a 50% duty cycle, doubling it's price.
You're still running it at a 100% duty cycle, you're just shifting the generation times. You have a reactor that can produce heat enough to generate 1GW continuous power, so it stores that heat for six hours a day during peak solar generation, then generates 2GW for six hours to cover peak demand in the evening, then generates 1GW directly for twelve hours overnight. It's generating 1GW worth of heat at all times, you're just converting the heat to electricity at different times of the day.
The problem isn't 12 hours of no wind or solar (and, come on, that's not even realistic).
The problem are week long wind droughts with heavy cloud cover. That's where there's a clear argument for geothermal, hydro, and hydrogen. Maybe natural gas with carbon sequestration. If nuclear wants to join these options, it needs to prove itself as something that can be built. And as somebody who doesn't believe that we are close at all to natural gas with carbon sequestration, I think we are a lot closer to making that work than we are to having the managerial and technical skill capacity to build new nuclear.
Nuclear is not the only option, we have so many more.
The key point about hydrogen is that the cost of storage capacity (to be distinguished from costs related to charging and discharging rates) is extremely low. So if you need a system to cover very rare extended outages, or to store energy from summer for use in winter, hydrogen is much better than batteries.
Also, hydrogen can be mixed into natural gas, without changing much infrastructure. This is of course only a transitional measure, not an end goal, but doable now and a very cheap one.
When we get to high levels of renewable penetration on the grid, we are going to have tons of excess energy, and capacity far past demand.
We have that with current generation systems too; we don't run natural gas at full bore all the time, usually. Most generators run at a capacity factor far below 100%.
The difference is that renewables generate power at zero marginal cost, while a natural gas turbine has fuel costs and wear and tear.
So we are going to have massive amounts of super cheap energy, for those applications that are not time sensitive.
Hydrogen production through electrolysis would be a great way to store this extra, nearly free energy. However this is unfortunately not a perfect match the intermittent surplus of renewables, since the capital costs of electrolyzers makes it so that constant production would be best. But at least some surplus energy will go this route.
Storage has several options, for example we could store hydrogen only for short periods before converting to methane or ammonia or other more stable chemical forms. We could convert pipelines perhaps.
Hydrogen has a lot of difficulties, but it is also, even now, an essential part of our economy and we work with it at scale. It will be a while before "green" hydrogen from electrolysis is cost competitive with fossil fuel derived hydrogen, but it will almost certainly happen.
There's a great, but long, two part series on hydrogen in Europe from what most would call a very skeptical stance, yet as a fellow hydrogen skeptic it convinced me that hydrogen will play a much larger role in the future:
Nope. Load following is something different from a planned shutdown. You can order your reactor to support load following, most reactors in Europe are built with that capability. This means that for a boiling water reactor, the reactor will regulate itself due to feedback from the cooling system: Slower coolant flow due to less load on the generators will increase the temperature in the reactor core, creating more steam bubbles that slow down the reaction and therefore reduce power output. Pressurized water reactors (PWRs) can also be configured for load following, however, they are slower and need to employ control rods.
The usual figures given for PWRs are something like "These reactors have the capability to regularly vary their output between 30–100% of rated power, to maneuver power up or down by 2–5%/minute" (https://en.wikipedia.org/wiki/Load_following_power_plant#Nuc...). This means that you can regulate your PWR over its power range within slightly more than half an hour at worst.
There is a lower limit to how much you can regulate your reactor down due to various limitations like Xenon poisoning. Because of that, a controlled complete shutdown is slower, because you need to get through the "icky" region between e.g. 30% and 0%. Also, you usually prepare for the longer phase the reactor will stay shut down. But all in all, shutdown is different from load following, and a conclusion that nuclear plants can't load-follow from a question about shutdown is just asking the wrong questions and drawing questionable conclusions from it.
Most solar doesn't require rare metals. CIGs does, but that's not very common. I can see complex fabrication, but pretty much everything requires that.
Besides cost, the other issue with nuclear, or any large single generator, is that they can go down unexpectedly and their output drops from whatever to 0. This is less common when generation is from many individual generators. If a wine turbine goes out, it only drops the output of the wind farm by some small fraction.
A quick search shows that the company referenced in that article, Solyndra, made CIGS panels, which, like I said, are relatively rare, in large part because they were a failed startup that went out of business.
> In a new report, the International Energy Agency (IEA) says solar is now the cheapest form of electricity for utility companies to build. That’s thanks to risk-reducing financial policies around the world, the agency says, and it applies to locations with both the most favorable policies and the easiest access to financing. The report underlines how important these policies are to encouraging development of renewables and other environmentally forward technologies.
Don't forget the massive initial buying subsidies, the ones that only last as long as the government wants them to, the ones which don't cover any of the maintainance costs.
> From what I understand this makes grid management much simpler and reduces storage, transmission, and conversion losses.
That's not so simple. Demand for energy is fluctuating hugely, and nuclear power cannot be freely switched on and off. This is also the case for large coal plants. So, with using renewable energy, grids need to become more agile, and a somewhat larger amount of freely controllable power sources is needed. Another strategy is to average out fluctuations over large areas.
This point was driven home for me when I was at Google. Googlers being flush with money, we had a local solar installer stop by a few times. I attended one of the presentations. This was in WA, by the way, which turns into Mordor for 6 months in any given year. So the guy keeps going on about how it'll pay for itself and so on, and presenting all kinds of stats like the ones you responded to. Then someone asked him how much our typical lack of sunlight affects this thing. He was like, "not much, let me show you", and he brought up a management UI for a solar installation in some school that they've recently completed. That was a mistake on his part. It was mid day, although a cloudy day IIRC, and the page showed like 10% of the possible output was being produced. Audience chuckled and some people started heading for the exit.
I think there are a lot of complications that are encountered when doing this sort of calculation. You might get a max output of "$15/MWh" or "$112/MWh" but that's only a portion of the picture.
Firstly Nuclear is X amount of power essentially on tap 24/7 where X is between some Min and Max value. From what I understand this makes grid management much simpler and reduces storage, transmission, and conversion losses. Even in the best case situation of "renewable" energy you'll see, in the absolute best case claimed by pumped hydro storage, you're seeing about an 87% efficiency.
Suppose you had a nuclear power plant that could output 100W consistently. Suppose you wanted to build a similarly efficient "renewable" energy source. How much more peak power would it need to output to be competitive? If we assume the power source is 0% efficient at night and 100% efficient during the day, off the bat, you'd need to generate 200W continuously during the day. Then you'd need to generate an extra 13% on those extra 100Ws to make up for storage losses (1.13*100 ~= 113). That means 100W from Nuclear is ~= 213W "renewable". Obviously this is a terrible calculation and not at all accurate but it's just a sample of some of the math that needs to be done.
Another factor that I've been skirting around is that "renewable" sources of energy are often not so renewable. Specifically things like like solar require a huge amount of rare metals & complex fabrication. Right now none of those externalizations are represented in the price of solar because there are no carbon taxes in countries that manufacture them.
Money in this case is a proxy of what needs to be measured and should not be the only factor we optimize for.