Yeah, this seems really stupid. I can see an argument for prioritizing renewables for new capacity (and can also see the argument for not), buy if the nuclear plants are already built, why de-prioritize dispatching their power now?
Of course this isn't 100% due to free markets as the subsidies of renewables allow them to offer in extremely cheaply, but it exacerbates the problem. Even without the subsidies, this would happen with current designs. Now of course there is "resiliency" value to having a nuke plant (example: when the sun doesn't shine) that is a value by itself, but it isn't being priced and paid for.
I get there's this utopian vision of solar, and the gains in the industry have been impressive, but those gains are backed by some really deep market subsidies and policy choices - which, to be fair, were needed to get the technology off the ground. However, it's not there yet, and it didn't get there under its own power. What we're seeing here is a repercussion of those policy choices, not the a market stabilization in favour of a superior player.
Generators bid in their short run marginal cost for energy and the cost to provide certain reserve services and the unit commitment and economic dispatch algorithms choose which resources to turn on and what level they run at based on some cost minimization function.
Before markets, the incumbent utility would run all their nukes and coal as base load. Now with regional markets you have lots of utilities and private entities competing. Once massive amounts of cheap renewables (armed with subsidies to boot) entered the supply mix, it pushed out the most expensive generation to where instead of running ~24/7, they began running far less. Go look at any nuke plants capacity factor (percentage of the year running) in the US by year (you can Google it). You'll see a rather sharp drop. It can only go so low before the plant has to be retired as the cost to run it is greater than the break even point. CAISO operators obviously don't want that, but absent some kind of side payment from the state (unpopular with economists), it's the inevitable outcome.
I just wanted to make the distinction that the entity running the market and making sure the grid doesn't blackout usually doesn't make any decisions regarding what the rate payers at the retail level pay for. Even if they know a plant is required for reliability, it's a difficult process to figure out how to accommodate the plant outside the market. It could be argued that these markets in the US need to figure out how to better quantify the risk from losing these nukes and reflect that as part of the price (as well as other technologies). So far that hasn't occurred. Some areas have long-term capacity markets (example: PJM), but California does not (I don't think). Even then, the capacity markets have had many issues.
What this really means is that their cost function is broken.
Power plants really have three types of costs.
The first is the construction cost. This is a sunk cost for anything that already exists.
The second is fixed operating costs. This is things like property tax and security and maintenance, which you could save if you shut down the plant permanently, but you don't save at all by having it run at e.g. 50% capacity.
The third is variable operating costs. The cost of generating the incremental MWh, given that the plant is operational. For fossil fuel plants this is dominated by fuel. For nuclear plants (and renewables) this is basically zero.
Only the second and third should be a factor in whether a plant that already exists gets shut down. But more importantly, only the third should be a factor in where you buy power.
If you have a nuclear power plant which you're paying to keep operational but not consuming 100% of its output, you did the math wrong. (Or your electricity prices are regularly zero or negative.)
All that to say the way things are done are not perfect, but they have had a lot of brilliant people looking into them for a long time. There is a lot of economic theory out there on the usage of short run marginal costs for offers and renewables are insanely cheap in comparison to most other technologies.
When you talk about paying for a nuke plant, but not consuming 100% of it's output, that's not exactly what I'm talking about. I would guess that many nukes simply have to run and can't be committed/decommitted like many other conventional technologies. They essentially become a price taker at that point and whatever they are making is likely not enough to cover what they need to stay in business. I hope that's more clear.
Also, zero or negative prices are indeed becoming increasingly common in certain parts of the US.
Edit: the capacity factors I saw posted publicly for nukes in 2020 weren't near as bad as I thought. Regardless, there are still a lot of planned retirements as I understand it, in the next few years. With Southern Company's Vogtle way over budget and behind schedule, I can't imagine another nuke being built in the US of that kind anytime soon either.
Do you think these cost functions properly price disruptions? I'm saying this because the US is an order of magnitude less reliable than in Europe, and most of these costs are pushed onto consumers. E.g. When the power goes out and I have to throw away food or lose a day of work, PG&E has never compensated me for that. It's externalized. Now imagine a scenario where lots of people rely on the grid to charge their cars.
I'm really skeptical of the obsessive hyper-optimization that's going on. If you have a situation where there is an extremely cheap energy provider that is highly unreliable, it could drive out of business the more expensive, reliable provider, and that needs to be taken into account. The zero or negative spot prices that were common in Texas as a result of windpower look to be an example of that pathology at work.
So for a single living alone, that's a bad deal, as suffering a power outage for 1 day and then being given $108 will not make them whole -- it wont even pay for the spoiled food, let alone the inconvenience of being without AC or internet for a day.
For a household of 2.5, they would get $270, which would pay for the spoiled food and inconvenience of buying it and maybe get $30 left over for the inconvenience of no A.C., cell phone running out, not being able to use your computer or internet, do your laundry, perhaps not being able to use your mobile phone, etc. And we aren't even assuming PEV or people working from home. Then they would also lose 1 day of wages.
So it would seem to me that just to make someone whole you should probably double or triple that, say to 20K/MWh, but then this is supposed to be a penalty rate, so it should cost more than what it would take to make a household whole, because we don't want the utility to be indifferent between making the household whole and providing the electricity, we want the utility to be heavily incentivized to avoid the disruption.
$9000/MWh isn't crazy for a few intervals once a year, but if a generator clears 200 MW for each hour of the day for a day in the day ahead market and then has issues and can't provide that power in the real time market, it has to buy back that power at the spot price. That can get insanely expensive fast. Now imagine the owner of that generator has other generators in the same boat and it just stacks. These aren't trivial prices. Similarly, the load serving entities have to buy power at that spot price if they don't have it locked in at a lower price in the forward market (hard to do when events occur over multiple days). A little googling will turn up some chatter about many of those companies nearly defaulting from the ERCOT scarcity pricing during the February event.
Exactly! And it should be 30K, not 9K, due to my arguments above. You are pointing out that renewables, while unreliable, are so cheap they are driving the reliable electricity out of business. That means something isn't being correctly priced, because we as a nation require reliability. So I propose to put a price on that reliability, and you point out it can drive the unreliable provider out of business, which is kinda the point -- that unreliable provider will need to purchase insurance from the reliable provider to cover them, or "side payments" as you call it, in case something goes wrong. This puts a price on having the nuclear provider available in case the wind doesn't blow, and creates payments to that provider, but via market mechanisms. If the wind provider is confident in their battery storage they don't need to purchase insurance from the nuclear provider. But the point is to stop externalizing the costs of unreliable power. This is to end up in a situation where we as a nation are able to fund reliable sources of electricity rather than everyone switching to wind and the nation going dark when the wind stops blowing, throwing up it's hands and saying "oops", or "wow, I can't afford this. Gonna declare bankruptcy now, see you on the other side with more cheap spot prices". That's how you do it with the market mechanisms - by putting much bigger penalties on unreliability than we have now and requiring proof of insurance for being able to handle those payments.
> These are wholesale prices, not retail
Of course, because we are talking about how much to penalize providers for disrupting the lives of households - remember we are trying to price reliability. The providers pay wholesale, so that is the relevant cost function to target, not retail.
One question to ask yourself is who is going to pay for this block of nuke capacity? It would be very expensive to pay for that on a 24/7 basis "just in case". The market already clears reserves for a variety of contingency events and it's certainly a balancing act to ensure you have enough without over paying for something you'll only need in ultra rare occasions.
Part of the market philosophy is also not to signal out any one technology. You put in some requirements and let the economics find the cheapest solution. So you could ask for more reserves, but they probably wouldn't just come from nuke.
There is an alternative reality where anti nuclear activist [1] never succeeded and we continued to make great progress on reactors. Ultimately, reducing fossil fuel use and corresponding GHG levels.
One of the very things the environmental groups fought against was one of the things that could have saved us.
"The road to hell is paved with good intentions."
But sure, nuclear power is not something that would work for every nation on earth. But it would certainly work well for North America and the EU, as well as China, India, Australia, and Japan, which together account for the majority of the world's energy needs.
This also ignores the rest of the intended refinement chain: thorium and other reactor fuel sources.
Aka, those 50m t of ore globally work out to around 0.5-1m t of fuel though we’re extracting the cheapest stuff first so expect prices to rise well before you extract all of it. There are also options to extract vastly more energy from that same ore, but again they all make nuclear more expensive.
Light-water reactors, which are the more/most prevalent design. Heavy-water reactors, like CANDUs, can use non-processed uranium in 'natural' concentrations:
* https://energyeducation.ca/encyclopedia/CANDU_reactor
* https://en.wikipedia.org/wiki/CANDU_reactor#Purpose_of_using...
The heavy-water is deuterium oxide, where there "hydrogen" atom has one proton and one neutron, whereas the far more common hydrogen has no neutrons in the nucleus:
They do use up to 30% less mined uranium as your link points out, but also produce a lot more high level nuclear waste.
On the other hand, breeder reactors don't necessarily make weapons-grade material. The Integral Fast Reactor in the US produced a mix of four plutonium isotopes, which was totally unsuitable for weapons, and harder to enrich to weapons-grade than natural uranium.
Using breeders multiplies the energy from uranium ore by a factor of a hundred. With such low fuel consumption, it also makes seawater extraction pretty easy, which extends our fuel supply to many millions of years.
I feel like this has been absent in many pro-nuclear arguments (I always assumed we had much more)
The other fuel used would be plutonium and thorium, I believe. My understanding (not a nuclear scientist) is that there's a cycle of enrichment and decay that produces more fissile materials than it consumes.
Interestingly, from my perusal of wikipedia, it seems like you effectively can recycle the uranium isotopes indefinitely in such a reactor, as well as generating fissile plutonium and thorium.
Again, the problem is that generating fissile plutonium is how you make a bomb
China is leading the way, both in terms of adding coal capacity right now to address the power outages while also building out nuclear capacity for the future.
The ability to act rationally like this when popular movements turn irrational is one advantage of an authoritarian system. No doubt many in the west are looking somewhat enviously at China.
So, the Chinese are spending way more money than they should on building super duper expensive coal plants that can’t effectively run on the kind of coal that they can get. And they have a coal embargo against Australia, so they can’t import the stuff that they could actually burn effectively.
Meanwhile, solar panels have continued to get progressively less expensive, so that even in China, it is now cheaper to run solar panels than it is to burn coal.
Increased uranium prices will not significantly affect energy cost, as fuel price is very small fraction of nuclear plant operation.
These days there are ways of getting uranium from sea water [1], basically making the supply practically limitless. This way is currently un-economical, but as current reserves are depleting, it will become viable option as well. (Along with researching for more reserves)
So yes, it was totally possible to decarbonise electricity production long ago.
[1] https://www.forbes.com/sites/jamesconca/2016/07/01/uranium-s...
Breeder reactors can make fuel during use.
We only need enough fuel for one reactor.
That seems like plenty of time to
a) build renewables
b) figure out fusion
c) discover/make accessible more uranium just like we did with oil
[1] https://cen.acs.org/materials/Fishing-uranium-ocean-spider-s...
Nuclear fuel can increase during use. Running out isn't required, the way it is with consumable fuel
It's not a technical problem (solution is well known), it's a policy problem.
Existing power infrastructure has already paid its capital costs, meaning they have a massive incumbent advantage. Coal and gas power also haven't historically had to pay for their externalities, which is another distortion in their favor. This means that without extra incentives, you're stuck with coal until the plants need to be replaced, which essentially means until something truly magical happens or forever, whichever comes first.
So, you can look at the subsidies for solar as offsets for the incumbent advantages and externality subsidies enjoyed by fossil fuels. Meanwhile, the LCOE of solar has improved exponentially with scale and semiconductor improvements, which gets close to the 'something magical.'
My local solar company advertises itself as "alternative to unreliable PG&E", it is expensive and not many can afford it but I can see why some people might spend their healthcare or education budget on buying those expensive solar panels.
https://en.m.wikipedia.org/wiki/Diablo_Canyon_earthquake_vul...
Also note, the linked page states there are four "nearby" faults -- it doesn't state the distance to them, but included is the San Andreas fault, and I'm relatively confident we know it's capable of producing >= 7.5 magnitude earthquakes.
I don't know, nor claim to know, much of anything about how susceptible [or not] this plant is to seismic events -- and I'm generally quite pro-nuclear -- but it's worth noting that there are other seismic concerns here, and potentially some questionable incentives for PG&E who is the one attesting to the capabilities of the area & this specific plant.
According to USGS seismologist, Jeanne L. Hardebeck, the Shoreline Fault has potential to trigger an earthquake of 6.4–6.8 magnitude
Earthquake sizes are usually expressed via Richter scale or Moment Magnitude Scale values, which both attempt to quantify the total energy released by the quake. These scales do NOT quantify the force experienced at any particular location, which depends heavily on distance from the epicenter and local geology. For that kind of quantity, you would use a seismic intensity scale (https://en.m.wikipedia.org/wiki/Seismic_intensity_scales).
So it's wrong to say that a quake "...would not be 8 when it reaches Diablo Canyon" because the value '8' is an MMS or Richter value. Instead, you could say that "...the damage potential would be reduced by the distance to Diablo Canyon."
If they're based on historical data, on the worst known past earthquake/s in the region, then they may be effectively black-swan blind. Eg, they may be failing to take into account that the worst past events always exceeded the prior worst event before it, and therefore prior worst events were not, and are not, a reliable measure for future ones.
Building earthquake-proof nuclear reactors requires, at minimum, taking the worst past earthquake nearby, and assuming the next will be worse yet, and building for that one (or the one after, etc).
Or better yet, just don't build nuclear plants on top of fault lines, regardless of what historical data shows about it.
It doesn't speak well to their competence, but "inspect and maintain 1,000s of miles of line" is different than "inspect and maintain 1,000 items in one place."
Moreover, it sounds suitably silly to be used for such silly nonsense.