Balancing a nationwide power grid is very complex. Some energy sources can be started and stopped instantly, but are limited - water. Others are plentiful, but unpredictable - wind. Others are predictable, but take a long time to start and stop - gas, coal(several hours), nuclear(1 day to start, fast to stop, but very expensive). A balanced grid will need all of them, will need them in quantities which can cover faults in the big producers(a nuclear reactor makes 700-800 MW). They will need them built in the right place, because while more power cables can be built, you can't transfer a lot of power on very long distances, for cost and grid stability reasons.
> Others are predictable, but take a long time to start and stop - gas, coal(several hours), nuclear(1 day to start, fast to stop, but very expensive).
The start time is long but that does not say much about the overall operations.
> Modern nuclear plants with light water reactors are designed to have maneuvering capabilities in the 30-100% range with 5%/minute slope, up to 140 MW/minute
> In France, with an average of 2 reactors out of 3 available for load variations, the overall power adjustment capacity of the nuclear fleet equates to 21,000 MW (i.e. equivalent to the output of 21 reactors) in less than 30 minutes.
The variability of France’s nuclear fleet is harder on the generator (valves and structures susceptible to thermal stresses in particular), and a possible contributor to their inability to keep their fleet in good repair.
Arguably, if cost effective, nuclear is best run at full output as consistently as possible, with other systems buffering that supply with demand (hydro storage, batteries, demand response, etc).
Despite the insistence that Closed Cycle Gas Turbines can't react quickly, because they're by far the largest component that we could start and stop the UK does in fact very quickly increase and decrease output from the CCGTs. For example this morning 2.79GW at 0600 to 3.89 at 0700.
There are much faster options, batteries, import, even the pumped storage is seconds instead of minutes - if available, but CCGT is just not that slow to change compared to the weather. In that same period the wind power went from 10.9GW to 11.4GW. 500MW is a lot of power but it's not more than 1.1GW
An interesting complicating factor here is that much of the UK's installed base of CCTG stations were built during the 90s with the intention of replacing many of the smaller coal-fired stations, which would typically be doing 2-shift operations (i.e., day and evening). Now, those CCGT stations are increasingly used to counterbalance renewables, and (as you point out) are now running on much shorter cycles than they were designed for.
A report from a few years back (which I'm afraid I've utterly forgotten the source) examined the data on this, and argued that as a result of this changed pattern of use, these CCGT stations were now not achieving nearly the kind of efficiency figures they were designed for, which from a carbon point of view is not good news - we might still be emitting lots of the stuff, but just not getting as much practical benefit from it as we used to.
Now, I'm not meaning to suggest that this is a disaster, or that is somehow invalidates the entire of concept of renewables, but it does point to the need to be careful about what we take to be a useful measure of progress - and that merely the quantity of supply to the grid in GWH isn't necessarily it.
And the article under discussion here is of course picking away at another strand of this same idea - when we connect these generators together, it gives rise to system-level effects, and we need to be thinking about the outcomes, both beneficial and harmful, in system-level terms as well.
The only thing that matters for climate change is the amount of gas burned - the efficiency with which we turn it into work is irrelevant. If we reduce the amount of gas burned, at a cost of burning the remaining gas less efficiently, that's still a win.
In a certain sense one should expect lowering usage to inevitably lower efficiency, as a sort of inverse corollary to Jevon's paradox (which states that as efficiency rises, total usage does too).
The problem is that building wind turbines in Britain has opportunity costs.
For simplicity: assume a status quo of 100% gas. We are burning 100 units of gas for that per year.
Now assume by building a crazy amount of wind turbines we could satisfy 95% of the UK's power demand with renewable. However, for the remaining 5% we'd need to burn 50 units of gas.
In this scenario, efficiency of burning gas drastically plummeted, but so did overall gas use.
However now the question is: for the resources invested into building all those turbines, could we have gotten a better climate bang than 50 units of gas saved?
(All numbers made up, obviously. In practice, we can probably make the economics work. Though we might need to deregulate the grid. It's crazy to pay wind turbines for not running. At least mine bitcoin or smelt aluminum or something.)
It won't be long (perhaps in the next 2-3 years) before the UK grid will be able to operate for periods without any CCGTs running at all. We've already come quite close this winter, with record low CCGT output and record high wind turbine production.
Wind turbine output, although variable, is also fairly predictable: so good modelling and scheduling should ensure that when CCGTs do operate, they can run as efficiently as possible and not be spinning up and down too frequently.
This was blown way out of proportion by the press for about a week a few months ago. Someone got hold of the National Grid contingency plan for situations where projected power generation doesn't meet demand, and somehow extrapolated that into "the National Grid are poised over the disconnect button, and will be certain to use it any day now".
... most of the modern light water nuclear reactors are capable (by design)
to operate in a load following mode, i.e. to change their power level once
or twice per day in the range of 100% to 50% (or even lower) of the rated
power, with a ramp rate of up to 5% (or even more) of rated power per minute.
One trouble is that changing the power output does put stress on components because of thermal expansion and contraction, potentially shortening their lifespan, but it something that can be designed for.
Varying output from a nuclear plant is mostly achieved by simply releasing the generated steam into the atmosphere instead of sending it through the turbine[1].
But operating a nuclear plant in this fashion pushes up the price per MWh considerably given their very high cap-ex and op-ex. And while fuel cost is negligible for nuclear, creating more nuclear waste per useful MWh generated is a further drag on costs.
So as a solution, it "works" if the nuclear plant does not have to compete in terms of price with other sources of electricity. But nuclear fails to compete on cost even if operated continuously - it's uncompetitive with cheap, quick to deploy, low op-ex, modern tech like CC gas turbines or renewables in most western electricity markets and can only survive with government subsidy[2].
It seems obvious that nuclear can not compete against natural gas when natural gas is priced cheaply and the pollution caused by fossil fuel is put on society rather than the operator. A combined grid of renewables and fossil fuels has been the primary strategy in most European countries and was working very well in keeping prices low until Russia invaded Ukraine.
The big problem is that energy prices are set based on the most expensive unit that needs to be turned on to meet demand. Renewables do not tend to be that during periods of low supply, as low supply of energy in the eu market generally means sub-optimal weather conditions for renewables. It is going to be either fossil fuels, nuclear, or battery. If we take out fossil fuels then that leaves battery or nuclear. Neither is very economical without subsidies. Governments (and tax paying citizens) are however very keen on grid stability and thus willing to spend a lot of money to keep it running.
I get why it works this way because the alternative would be to force the fossil generators to sell at the renewable price and thereby making it uneconomical for them to operate which leads to brown-outs. I just think the societal costs the up very high because EVERYONE is paying a premium on power, and the total sum of that premium is only going to increase as we move more and more stuff over to electricity.
I therefore wonder if the market couldn't be structured in a better way which would still ensure that the fossil backup generators are adequately compensated but smoothes the extra cost over the remaining cheap GWh. Something like a meditating party which is aware of the production costs and buys up the daily power and sells it on at an averaged price. There are probably good reasons why this wouldn't work, but I am too stupid to figure them out.
That's what financial hedging and the like do - users and producers can sell power at guaranteed prices in exchange for missing out on price volatility. Ultimately, it's probably fairest that the wholsesale power market works the way it does to ensure guaranteed power - It costs money to provide a 100% service guarantee.
It's worth noting there are some demand response initiatives and the like that are approaching this from the other side - they will pay a user to not use power at particular times of high load. If you don't want to pay a premium on power, I suspect there will be providers happy to oblige, so long as you are willing to forgo the 100% service guarantee.
With the past winter it seems that people who do not want to pay a premium on power would prefer if the government then stepped in and paid it for them. With the Energy Price Guarantee program, customers don't need to bear the cost of 100% service guarantee.
At this point there isn't really any part of the energy grid that governments do not subsidize. They subsidize companies that provide grid stability. They subsidize renewables that provide capacity. They subsidize the customer who buy energy. They subsidize the grid infrastructure that transports the energy. They subsidize the interconnection between countries that enables trade between countries. They subsidize the cleaning up and associated costs from pollution.
And every step further removed the customer gets the less incentive there is to actually reduce consumption, particularly at the most expensive times of day. Subsidising that period when it’s only going to get monumentally more expensive to supply relative to the rest of the day seems kind of wrong.
Your reference for [1] just states that bypassing the turbine is a thing, not that it's normally used.
First, reactors are in a stable equilibrium when operating, so one will actually increase their power by increasing the rate at which heat is removed (and v.v.). Alas, that's workable only within some small range.
A reason[1] load-following with PWRs was originally difficult is that traditionally PWRs use boron concentration in primary loop to regulate power and that can be decreased only slowly. The reason it's done that way is that it's the easiest way to ensure that power is adjusted uniformly throughout the core; if instead some control rods were partially inserted, the top part of the core would operate at lower power (and thus lower fuel burn-up) than the bottom part, which would cause compounding control issues later on.
France is using their PWRs in load-following mode by (a) having additional less absorptive control rods ("gray rods") that can be inserted fully to adjust power by smaller increments (b) more complicated schemes to decide which combination of available actuations to use to change power. See https://hal.science/hal-01496376/document for a paper that tries to optimize control designs so that power changes are more possible (and describes how the control schemes work).
Note that the total heat capacity of even just the primary loop in usual reactors is quite large: in PWRs it usually requires ~0.5s of full power output of the reactor to warm it by 1degC, so this can easily cover, say, ~5% variations for something like a minute.
[1] Another is that reactors are not stateless due to xenon poisoning.
I am skeptical that renewables are cheaper than nuclear when one factors in the impossible amounts of energy storage required to make them meet the same reliability guarantees that nuclear (and fossil) can meet - indeed, as far as I know, there exists no proven, cheap, scalable technology to store power at grid scale at all.
Alas, in the real world because of public opinion and political pressure, it's almost impossible to build new nuclear power plants. And those that get build are crazy expensive and overengineered, and invariable overrun their schedule and budget.
What impossible amounts? You can play around with a small model here: https://model.energy/ It's simplified of course, but the estimates should be in the right ballpark.
The anticorrelation is mostly on longer timescales. Winter vs summer for example. On a daily or hourly basis the anti correlation is pretty much non existent and it's on these timescales that you would time shift demand. A quite common occurence in Europe is that large parts of Europe during winter have almost non existent air pressure differences for days or weeks on end. During these times neither wind nor solar is very helpful and other solutions are needed, not all European countries have the pumped storage capacity for that. LNG to the rescue I suppose, now that Russia is limiting supply.
I was wondering if someone was going mention that. Pumped hydro is great, but it's not scalable. You need favorable geography to make it economical at all, and in the end it doesn't store enough energy to do more than smooth over transient grid fluctuations lasting a few hours. The UK is, relatively speaking, quite well provisioned with pumped hydro - its largest storage facility is Dinorwig in North Wales, which is built into a mountain with very favorable geometry - it has nearly 6 times as much capacity as the next biggest station. It can store enough energy to run the entire UK for... about 16 minutes. That's not going to do the trick if the grid runs entirely off wind and solar and you have a dark, calm day, let alone the weeks at a time that weather can be unfavorable. And there isn't anywhere to put another hundred Dinorwigs, never mind the budget.
It's because of this that there's a lot of talk about wild ideas like pressurizing abandoned mines and so on - there are a lot of mines around. But then we're back to the "proven technology" sticking point.
The suitable geography is considerably less rare than the nuclear and carbon industries jointly like to pretend. This has been confirmed by multiple studies (I have posted them at least 3 times before because this talking point is sadly rather common).
Nowhere is currently "well" provisioned for pumped hydro given a solar and wind grid coz while they existed for over a hundred years they have never had to store that much energy. Newer, larger ones are being built around the world. Australia will be well provisioned soon.
Go back in time 10 years when solar and wind first became economic and people made similar comments about how little of it there was (1% of total power!), ignoring the unit economics completely. We are at that exact same inflexion point with pumped hydro.
The largest pumped hydro facility in the world is Fengning in China, at 40GWh, and the second largest is Bath County in Virginia, at 24 GWh. Dinorwig's 9GWh is really not too shabby. Even Fengning would only power the UK for just over an hour. This is simply not the same order of magnitude for the storage you'd need to make it through a gray UK winter on renewable energy alone.
What's the longest period without wind and sun you're willing to provision for before you give up and tell the population they'll have to do without electricity for a bit? A day, a week, a month? Numerically, how much storage would that actually need? How many stations, how big? You'd need over a hundred Fengnings to power the UK for a week. Where would they go? I'm all for renewables + storage but you can't handwave these questions as FUD, it's a serious problem.
I suspect that if we committed to categorically eliminating fossil fuels, including peaker plants, the first time the lights went out because the weather was bad you'd have people clamoring to build nuclear power plants. Statistically, it'll happen at some point no matter how much storage you provision.
>6.5 Fengnings or equivalent should be enough for a 94% renewable grid in the UK.
That doesn't follow at all from your article, which is about the US. You can't just extrapolate from a different country at a lower latitude with different weather patterns and vastly more space to put things like onshore wind/solar farms without running into NIMBYIsm, not to mention more hours of sunlight just from spanning 4 timezones. 6 hours of storage is not even close to enough for reliable renewable power in the UK. It wouldn't even cover a single windless winter night.
And even if we take it at face value, the scenario you linked involves masses of overbuild, over the course of nearly 30 years ("by 2050"), and still leaves 6% of energy coming from carbon combustion. If we start building nuclear plants now, even if we accept your premise that they take 20 years to build (they needn't, especially with scale), then we can get to zero carbon almost a decade earlier - and with minimal land use.
It's not like it's impossible - France went all in on a nuclear grid.
>That doesn't follow at all from your article, which is about the US. You can't just extrapolate from a different country
This is FUD.
You absolutely can if you are discussing orders of magnitude which we were.
Our fundamental disagreement wasnt about whether it was 8x fengnings or 6.5x but rather whether it was of the order of 65 or 6.5.
>It's not like it's impossible - France went all in on a nuclear grid.
Not impossible, just at great expense and it wasnt worth it. In 5 years less of France's electricity will be nuclear than it is now while still spending vast sums on new plants. They're officially hoping renewables will make up the difference.
You get a lot of storage "for free". Some examples are house heating and hydro-electric that you would have the dams for anyway for flood control. Also, fuel based power-plants are very low capex so it's totally reasonable to have 50% demand able to be met by fossil fuels and then just keep them off whenever you have enough renewables.
Nit: the steam is not released to atmosphere, it is sent directly to the condensers. Treated water to use in boilers and turbines without leaving deposits and damaging them is precious, so it is a closed cycle and not vented to atmosphere whenever possible
But using a nuclear power plant as a backup when there is no wind doesn't make any sense. If you build a nuclear power plant you might as well use it, costs the same either way. And if you use it, why building wind?
The problem is taking the most expensive power source with a large portion of the costs being the initial investment and then not running it 100% is economical suicide.
Not if the market provides incentives to curtail, which it does with negative pricing. If other energy sources can curtail enough at the same price, they’ll do it. Otherwise nuclear will. The costs you’re alluding to can’t be avoided, but they’ll be spread across the system.
This is all predicated on the market operator actually having the systems in place to signal the need for curtailment effectively, of course. That’s a whole different question.
Most reactors in service operate at a constant load, and don't vary output according to demand. Certainly in the UK they do not. Sometimes reactors are operated for extended periods at reduced load for various reasons (eg: to conserve fuel and extend the time before a refuelling shutdown is required), but they don't vary output day-to-day.
A nuclear power plant can't just "keep spinning without a load" - all that energy has to go somewhere! If a nuclear plant is disconnected from the grid (tripped), the nuclear reaction must be stopped (eg: by inserting control rods into the core).
Of course it can, just short the generator coils and you have a free brake. The turbine should then still have resistance and shouldn't overspeed. Or just idk, use it to pump some water in a loop or discharge through some resistors. Getting rid of power isn't that hard if you want to do it. Simplest solution would I suppose be to just have an outside radiator that brings the steam to cooling tower levels of manageability so you can throttle the turbine with just a valve.
The thing is, they don't really want to do it if they can save fuel by shutting down.
Not a nuclear engineer, but Im curious to know how they do it - throttling nuke is hard. I only know the stream "dumping" method.
There's good reason why they are hard to throttle. For starters thermal contraction shortened lifespan; but also because the nuclear cycle itself doesn't lend itself to throttling safely - nuclear products create "retarded (?) neutrons" which are the cornerstone of a stable control system (as opposed to prompt neutrons) and also significant amounts of neutrons poisons which are normally "burned" at equilibrium steady state power levels but which accumulate if you throttle down (therefore be needing even more prompt neutrons).
My understanding is that the more you need to rely on prompt neutrons for your neutron balance the more unstable your reactor (starting them up, therefore, is delicate). Throttling the power upsets this balance by at least two different mechanism.
There's this incredible project to build a 10GW solar farm in Morocco (1/3 of UK peak consumption) and bring the power to the UK via HVDC cable. Amazingly they estimate only 10% losses despite being over 3800km long:
In the worst case, couldn't the excess power simply be used in electrolyzers to generate hydrogen? They may not be very efficient but it's better than throwing free energy away.
> Surely HVDC links between Scotland and England could be built?
The article covers this and explains why it's not enough. Provisioning time for the links exceeds projected generation capacity increases in the Scotland.
It seems like the planning isn't so great, though. They expect to have another 4 GW of links by 2029, which is enough for the wind overshoot today, but there will be a lot more wind by 2029... so, double the investment and build 8 GW of links by 2029.
The dumb thing is that electricity transmission and distribution are usually fixed. This already doesn't make sense because it's peak demand that drives the capex. Opex is peanuts.
But the retail buyer doesn't usually see the negative/low electricity prices of high-supply+low-demand time periods for their "inefficient" uses that should still be economic.
> "But the retail buyer doesn't usually see the negative/low electricity prices"
There are electricity suppliers in the UK who offer prices linked to the wholesale price, including actually paying you to use electricity if the price goes negative. Quite useful for flexible loads such as EV charging!
Strangely there already is one between Scotland and Wales and two more are proposed (see the article).
I suspect NIMBYism is a big part of the explanation. Airborne AC links are efficient but ugly. Underwater AC links are tolerated by Nimbies, but inefficient. So you end up with underwater HVDC links.
You are absolutely right. Building pylons through much of England means dealing with highly organized NIMBYism - from people who care a lot about their local 'environment' but very little about catastrophic floods or fires in other countries, and highly opportunistic landowners who can name their price for use of their land.
The equipment costs (eg: converter stations) for an HVDC link are more expensive, but by running under the sea you save on a lot of other costs (ie: civil engineering, land acquisition, etc).
Building any new transmission line through densely-populated England is extremely expensive. Even if you can secure the necessary land and wayleaves, nobody wants them running near their house and spoiling the views, so significant segments have to run underground in tunnels, greatly increasing costs.
Besides, the UK is not that small when linking England and Scotland. The proposed Eastern Green Link 2 (EGL2) is 440 km long: there are many existing HVDC connections much shorter than that around the world!
> "why would this be necessary when the entirety of Great Britain is one synchronous grid?"
Because there are bottlenecks in capacity on the synchronous grid that restrict the amount of power that can be moved from north-to-south (or vice-versa).
It works out better/cheaper/easier to bypass those bottlenecks with efficient undersea HVDC links than to try and build more terrestrial AC transmission lines.
That's a shame, I wasn't entirely onboard with the logistics of crossing the massive fault lines along the route .. but I admired the ambition and scope of the project.
Be interesting to see if this is the end or just a pause waiting for fresh capital.
Ok sure apparently I'm a week out of date, but several other projects are in the works and many of these have already been built. There isn't a problem with the technology.
Thank you. People laughed when I suggested an HVDC link between North America and Europe.
Nordstream 1 was 1222km, and Britpipe now, is 60km shorter.
Boston to Lisbon is 5100km. Churchill Falls (home of a giant hydro dam project in Labrador Canada which got screwed by Hydro-Quebec because the only via transit was through Quebec), would be just under 4000km subsea.
It wouldn't make much sense: eastern US/Canada and western Europe have about the same profile (same kind of wind/hydro/solar/... sources); it would make more sense to connect regions with different profiles, like the Scotland/England example of the article (high-wind/low-population to a high-population zone) or high-sunlight to a low-sunlight (like southern europe/northern africa to northern europe)
The cable and power transmission parts of that scheme were sound - the routing across one of the more volcanic and faulted geological regions in the world was sketchy.
OK sure, that particular project may have been halted but it's not because of technical problems. And plenty of HVDC transmission lines have already been built. The key thing about HVDC is that it follows a dropping price curve similar to semiconductor manufacturing so prices will continue to come down.
That is true for almost every technology related to energy. We can't satisfy global demand for solar panels or wind turbines or batteries or other forms of storage either. We also can't satisfy demand for heat pumps or building insulation.
1.6 GW per reactor for the latest ones under construction (Hinkley Point C) and in development (Sizewell C). Each site has 2 reactors for a total of 2 x 2 x 1.6 GW = 6.4 GW.
Although this is largely just replacing the UK's existing fleet of reactors, almost all of which will have shut down by the time Hinkley Point C comes online. Of the current 5 operating UK nuclear power stations, only Sizewell B is scheduled to operate beyond 2028.
> "They will need them built in the right place, because while more power cables can be built, you can't transfer a lot of power on very long distances"
One of the reasons offshore wind has been so economic & successful in the UK is they can usually plug in to existing, redundant transmission lines left behind by decommissioned coal and nuclear power stations, which are often on the coast. It's relatively cheap to connect to the grid when the infrastructure is already there waiting: you just need to build the cables from the turbines to the shore.
You can transmit a lot of power long distances with HVDC systems. 2GW systems are in development (TenneT 2GW platform & 525kV DC cables) & HVDC interconnectors can be several hindered km long…
No power infrastructure is either quick or cheap. I would think a subsea interconnector would be quicker to build than an offshore windfarm (as there are fewer components - no turbines or structures, and even fewer cables!)
The UK is leading the world in grid interconnection and offshore wind build out (though all owned by none UK entities), so if they aren’t building quickly enough I don’t know anyone who is….
The North Sea Link between Norway and the UK is already up and running and has a capacity of 1.4 GW. It is currently (10:53 UTC) supplying 3% of the UK load.
The quickest gas generation (gas engines) can go from cold start to fully ramped up in 4-5 minutes. A typical OCGT/CCGT is a bit slower and has a higher start cost (and a CCGT won't reach peak efficiency for hours). Pumped storage hydro takes 20 seconds or so.
However, turning generation on or off isn't the only way the grid is balanced in the short term - turning up/down tends to be a big part of it too and most conventional generation can do that faster (sometimes a lot faster) than startup/shutdown.
It depends on the type of natural gas plant. Some of them are designed for efficiency which takes longer to spin up and down while some are peaker plants which can spin up in a matter of seconds/minutes.
The statement that we need all of them is not correct.
Grid forming inverters and large battery storage will replace gas peak plants in the future.
First to go are however the old coal and nuclear plants as they become unprofitable.
There is a hierarchy of time availability of power supply:
1. power available when you want it, and you can choose on the fly
2. power available when you want it as long as you know in advance
3. power available at a time that you don't choose, but you can predict
4. power available at a time that you can neither choose not predict
Examples are (roughly) 1: gas or hydro, 2: nuclear or coal, 3: sun or tidal, 4: wind. You can also think of demand types that require each of these levels or better. Of course each of these categories contains its own sliding scale of how far in advance you have to decide or can predict. Wind is not completely unpredictable, but it is further down this hierarchy than almost any other source of generation.
Moving generation up this hierarchy, or demand down it, is always going to give some benefit. Well designed power markets should make sure that there is some fair incentive for any such step.
Usually the problem with this kind of thing, is that the initial capex requirements are high, and it's just not profitable given how rarely it happens.
