Hi, apart from electricity generation, there is the other commitment to reduce greenhouse gas emissions (GHG) by 81% by 2035 compared to 1990, made by Keir Starmer at COP29.
From Statistical Review of World Energy 2024, in 1990, U.K. CO2 equivalent emissions from energy, process emissions, methane and flaring were 595 million tonnes. So by 2035 they need to be 113 million tonnes. In 2023 they were 327 million tonnes. So if other GHG sources ( e.g. from farming) are ignored, fossil fuel energy consumption has to be reduced by 65% by 2035 compared to 2023. Now, in 2023, fossil fuels, mostly oil and gas, provided 5.16 exajoules (EJ) of primary energy. If U.K. has to replace 65% of this, it has to find 3.35 EJ of new CO2-free energy. Even if we assume that miraculously U.K. will reduce its primary energy requirement by 40%, it still has to find 2.01 EJ of new CO2-free energy. This is equivalent to providing 65 GW of continuous CO2-free energy or around 22 new Hinckley Point C nuclear power plants or 145 GW of new wind assuming a very optimistic capacity factor, CF of 0.45. For refernce, CF for wind currently is 0.31(for solar it is 0.11) and installed wind capacity is around 31 GW. So almost everything has to be run on CO2-free electricity There also has to be a massive change in lifestyle on top of the 40% reduction in primary energy consumption assumed - enormous reductions, almost to zero, in meat and dairy consumption, aviation , cement and steel industries . …
All this will be impossible to achieve in the real world by 2035. Anyway, what would be the cost of all this if it were possible?
It is my intention to extend my analysis to 2035 in the near future (and when I have time). The assumption to make about nuclear generation is critical because the projections for wind & solar capacity will just result in a very unstable system that has to rely on large amounts of gas to cover intermittency. The problem is that I don't believe either the current UK government or the UK nuclear industry is capable of bringing any large new nuclear plants (other than Hinkley Point) on line by 2035. Even if they gave permission immediately to Sizewell C it will take more than 10 years to build.
There is a small chance that enough units of the GE-Hitachi SMR might be built in that timescale to make a difference. I don't think that there is any chance that RR units will built that quickly. But all of this is fiddling at the margins. Ultimately what we have politicians who insist on virtue-signalling and have no willingness to listen to serious engineering analysts.
Comments like yours take me back to a motto a highly-esteemed engineering boss of mine had on his office wall: "Nothing is impossible for those who don't have to do it".
The NESO projection that gas will supply 5% of demand by 2030 is slightly disingenuous, I think. Quite by how much I don't know.
While gas does provide balancing of supply and demand it also provides inertia, reactive power and short circuit current input ( wnen required) which renewables cannot and are essential for a stable grid.
These services can be provided even if the generator is not actually providing power just by being synchronised to the grid. Usually called spinning reserve.
How large this spinning reserve is something I cannot guess at but suspect it will be significant, especially as those services will have diminshed with the declining nuclear capcity byt that time.
Obviously running gas plants with no grid input generates CO2 and does not help the longevity of the gas turbines running on no load with a cost implication to the operator, who has to be paid adequately for the service provided.
Perhaps someone who is far more informed than I can comment.
It does seem to be a bit of a cover up is possible here to comply with the 5% (Maybe, a big maybe)
as I don't think spinning reserve capacity information is readily available?
You are correct about the need for inertia, etc. The NESO 5% gas figure is just silly - driven by politics rather than serious analysis. It is based on an assumption that exporting countries (to the GB market) will be willing to supply whatever the GB needs.
My estimate is about 3 times that and my model suggests that some gas generation will be required for about 50% of hours in the year. That will be enough to warrant keeping 5-8 GW of CCGTs running and they can be paid to operate on standby (at, say, 10% of output) to provide spinning reserve.
On the other hand, the renewables folk will say that this is all terribly 20th-century. Their view is that new technology can allow capacitor banks and batteries to provide system inertia and reactive power. I suspect that NESO is going to degrade the reliability standards by allowing much larger variations in frequency.
NESO have been gradually reducing operational minimum inertia levels. Currently they are supposed to operate at above 120GVAs which is already less than half of levels in the pre renewables grid. They are talking about aiming to reduce this to 96GVAs by year end. It can be useful to think of the reaction time that inertia provides for boosting other power by dividing the inertia by the demand level: inertia is then measured in seconds of demand. I have found that in practice NESO have been marginally conservative by ensuring that inertia remains above 4 seconds, at least so far.
However they are now very dependent on batteries to provide initial support, usually backed up with pumped storage before gas takes over. This often works very well - but not always. NESO remain extraordinarily tight lipped about their lack of understanding of what happened at 13:10 in 22nd December 2023. Initially there was a 1GW trip of IFA1, which was followed rapidly by losing almost 400MW at Cottam in the Midlands and a trip on the Moray offshore interconnector in Northern Scotland. Those who watched the frequency plunge to 49.273Hz estimate there were also other significant losses of embedded generation, perhaps totalling well over 2GW across all sources. There were 60 seconds where the recorded frequency was below the 49.5Hz SQSS standard, but an in depth enquiry was avoided because that did not exceed 60 seconds. There would have been enormous incentives to correct the data if necessary to avoid an extra second.
The reality seems to be that interactions between inverter based sources are not always predictable. Something similar underlayed the August 2019 blackout, when the voltage disturbance from a lightning strike was sufficient to knock out the entire output of Hornsea wind farm and helped trigger significant embedded generation losses as well as the loss of Little Barford CCGT.
I am not a power engineer so I haven't anything useful to say about the 2023 episode. However, thinking about the 2019 blackout as well as events in the US & Canada, I suspect that reliability margins need to be greater than they are. Losing Hornsea and Little Barford was one thing, but as we start to rely more heavily on large internal interconnectors there is the prospect of losing, for example, the planned Eastern Green link. In opting for economies of scale the system is becoming increasing vulnerable to major network failures. Power engineering principles of the past weren't daft and the idea that we can squeeze reserve margins without a substantial decrease in reliability is self-deception.
10% on an individual CCGT if that’s what you are implying isn’t realistic, with the age of most of the UK fleet you’ve always been looking at something nearer 30%, the likes of say Carrington commissioned in 2016 could be lower. The data on minimum stable export levels by generator are somewhere on Elexon but I’ve not looked for a while. It’s not strictly an electrical limitation, the emissions tend to go off the scale at low loading.
Fair point about older plants. I was thinking about ones built in the last decade but there is only Carrington. Some of the ones commissioned since 2010 might be lower too. As you mention, I doubt whether any of them can meet NOx emission limits at low levels of output. Yet another reason why they might choose to run Drax a lot more than is being suggested.
I particularly like that you have used 8 years of data to look at gas and interconnector needs. NESO and the CCC both used a single year analysis when doing the same work. This totally ignores longer duration weather trends.
On the gas case side, am I right in understanding that you’ve done historical pricing for the prior year comparison and current gas pricing for the last table? I’m interested as many forecasts for gas prices show a substantial fall in gas price from 2026/7 as more export terminals come on stream. This will be definition driven down gas power prices in the 2024 case but will make much less impact in the NESO case as more of the power is at a locked in price with CPI escalation. Would be interested in your thoughts on that?
If I understand the question in your second paragraph correctly, the answer is yes. All of the main calculations are based on 2024 market prices and 2024 gas prices. The exception was the comparison across years which is based on the historical prices of each year.
I know that many expect gas prices to fall later this decade. That reflects the overall pattern of gas prices over several decades - occasional sharp spikes followed an extended downward trend for 8-10 years. Investment in pipelines and LNG transport is an element of why that pattern prevails. I used 2024 prices to avoid comments that my results rely on specific assumptions about the future of gas prices.
I would be cautious about translating future gas prices to wholesale power prices. In part this is because I don't believe that gas generation is now the marginal source of supply to the market - that is imports. Another consideration is that we are approaching the point at which there is no requirement for gas generation, other than for reasons of system reliability (spinning reserve, etc), for a significant number of periods in the year. At that point, even if you accept the conventional view marginal supplies will definitely come from imports because biomass (with the next highest fuel cost) is subsidised. At that point, as I argue in the Technical Appendix, the market looks to be extremely unstable.
Thank you for this article Gordon, clearly the result of a great deal of research and analysis. I am unable to offer anything scholarly, but it has prompted a few thoughts. Net Zero as a policy has been pursued for some time, and the contribution of wind and solar energy to the electricity grid is now significant. Given that the result to date has been increased energy costs, observable to all, doubtless causing economic damage and in some cases real hardship, it seems brave, or even courageous, to believe that the approach simply hasn't gone far enough. ( I use Sir Humphrey Appleby's terminology - a brave decision costs votes, a courageous one may lose an election.) A pragmatist would surely pause and question the direction of travel, so it's clear the show is not being run by such animals.
As an aside, there is currently a local issue with the route for power lines and associated pylons from a wind farm in rural Wales. Our local MP, a Lib Dem safe from the agonies of governing, has advocated that power lines should be run underground where they pass near to people's homes or businesses. It would be blasphemy for her to cast doubt on the whole endeavour, and she will be unaware that in passing, she has assigned zero value to landscape and countryside.
Re your first paragraph. That is why everything is blamed on gas prices. It wasn't us, guv, that pushed up prices, it was that horrible gas market. That is why I wrote my previous piece showing that, putting aside spikes in 2008 and 2021-23, the long term increase in electricity prices was entirely the consequence of policy decisions. But both politicians and the media have short memories and are uninterested in anything other than simple stories.
Re Lib Dems. I have come across similar behaviour from LD MPs concerning solar farms. They are the kind of politicians that give NIMBYs a bad name - pure hypocrisy - whereas the majority of objectors that I deal with have a much better understanding of the choices and policy issues. I have long felt that any MP (MSP/MS) who supports the development of renewable generation should thereby be taken as agreeing to wind turbines or solar panels being located in their local park and no objection from them should be permitted. The point is that policymakers want to impose the costs of their decisions on other people, but not themselves.
This is a tremendous piece of work. I know just how much effort is required to pull something like this together having done various analyses myself. So well done.
It underscores the idea that the markets have changed how they operate and backs it with hard analysis, as well as showing that the Miliband plan is in reality a high cost option. I think it's possible to discern many of your findings when closely looking at a chart of electricity supply by source with a price overlay like this one:
I have yet to complete my analysis of 2024 because of the need for data cleaning which can be very time consuming. However, it is easy to see that intraday price volatility is huge, particularly when there is a high share of renewables on the system. When it is windy there are huge swings in interconnector supply between net export and large imports, and pumped storage also plays a role. Back then batteries were small beer, and only reported for net discharge AFAICS under the OTHER categroy of generation. Downloading individual metered readings for batteries only shows the net charge or discharge over a settlement period, not the total input for charging/cooling and discharge output, so it is hard to get good information on battery performance, and unfortunately Bess Analytics has now disappeared behind a paywall having been acquired by Cornwall Insight, so now we depend on commentary from the like of ModoEnergy for analysis (they are very good though).
If we rearrange the data in price order we get a chart like this
which shows clearly that when wind output dominates prices are set by competition between wind sources, and tend close to zero or negative when we have a surplus export. Subsidising exports for the benefit of foreign consumers has never made sense to me. As capacity increases we will see more hours with wind surpluses and the surpluses will grow in low demand hours. It is clear that the structure of the market will concentrate curtailment on those wind farms that get no compensation when prices are negative. We can see this already with Seagreen and Moray East, which have yet to take up CFDs and become the dominant volumes of curtailment. The price they have been getting has been steadily reducing too, as more competition starts to come on stream from other wind farms in construction which also get no subsidy protection.
Incidentally, I think you are incorrect about ROCs in your Annex. For now they remain traded, although the market is very opaque and not openly reported unless you subscribe to a specialist price reporter like ICIS. However, the final recycle value of ROCs that includes the premium for redistribution of the cashout fund is published (often obscurely) by OFGEM. REF has a convenient table of the history at the bottom of the page here:
I am using an assumption of a 12% premium for the current year, giving a value of £72.50/ROC, as a guess that allows for slightly disappointing actual output from renewables and probably more stable demand. Up for consultation is a proposal to fix the ROC value without trading at a 10% uplift to the indexed cashout value for the life of the scheme between 2027 and 2037 when it expires, because it is anticipated that the balance between the obligation and supply could become much more volatile in future.
On your last point, I made a conscious decision to exclude recycle values. There are two reasons.
First, recycle values have varied greatly over time. In several recent years they have been well below your estimate of 12%. If I had included them I would have adopted a figure in the range 5-10%. I have discussed this on several occasions with Lee Moroney of REF. In many years the ultimate recycle value was much higher than the expected value during the year.
Second, recycle values are only available with a substantial lag and I wanted to ensure that I had comparable figures up to 2024. Since the whole ROC system is likely to change from 2027 I wanted to ensure that my estimates were cautious and defensible.
The procedure for calculating the ROC obligation has been Calculation B since the early years. That estimates the number of ROCs that eligible generators will earn using historic data on capacity factors and the the OFGEM data on capacity. That total is uplifted by 10%. The demand liable to fund ROCs is also estimated (I.e. excluding exempt energy intensive demand) and the obligation is then calculated by dividing one by the other to give a figure in ROC/MWh.
After the end of the year OFGEM calculate the total obligation by multiplying the ROC/MWh by actual eligible demand. Thus if demand is higher than forecast the total number of ROCs or cashout payments will be correspondingly higher, and vice versa. The actual ROCs earned will almost certainly be fewer than the obligation total by virtue of the 10% uplift in the forecast. Only in 2015 was generation significantly higher, resulting in a zero premium to cashout value and some carryover into the subsequent year (I would have to check whether there was a corresponding adjustment to the 2016 obligation).
If the forecasts are correct then the actual premium would be 10%, perhaps adjusted downwards for the time value of money, since cashout payments are only needed by October - essentially almost a year later on average. Those trading the market will make their own forecasts of the premium based on actual demand and generation and their own forecasts for the balance of the year, and use those to inform the price they would pay up front for ROCs. A correctly forecast premium leaves them indifferent between paying for ROCs and paying cashouts, since the former attract the recycle, while cashouts do not. A firm that underestimates the premium will in fact pay more because it will have to buy more than pro-rata cashouts. One that overestimates it will pay more simply because they overpaid.
There is an element of transaction cost and incentive that will mean that the actual receipts by generators will be slightly below the full recycle value. However, with that adjustment, that is what they can expect to be paid. If overall ROC generation falls short of forecast the ROC value rises in compensation, and it falls if generation is higher. If demand is higher than forecast ROC values will increase, and fall if it is lower.
This paper could have been a useful contribution to the debunking of Mad Miliband's lunatic lies about falling energy prices.
It's a great shame therefore that you have written such a confusing and incomplete story. Many of us are trying to inform the public of the true threat of Net Zero - not "Clean Power by 2030" which is a tiny (yet still wholly unachievable) part of the full story. You could choose to collaborate with others to get the complete truth out there, but you seemingly prefer to go it alone with little care for the confusion you will actually add to.
I am sorry that you feel that way, but I am not going to change how I work. As anyone familiar with my work will guess, I believe that collectively we benefit from people who are willing to investigate the nitty-gritty details of grand claims. In modern parlance I am a data scientist not a political campaigner. I want to test the claims made by politicians and lobbyists not engage in grand speculation. Following the aphorism of Maynard Keynes, I am economist as "humble dentist" rather than public panjandrum.
My previous comment should have had this preface ... sorry for my spastic thumbs!
Thank you for this. It addresses a problem that I, as an observer have been struggling with for a long time, both to formulate the right form of questions and where to look for the answers.
Some time ago I thought I’d try ChatGPT to get a concise answer and this is what it told me:
Could you possible look it over and correct any of the errors in my question and nuances for the reply?
The components of Net Zero policies and wind and solar power or all renewables have a cost. What are the out-turn expected UK costs in 2035 associated with the following basic element of Net Zero and renewables on the basis of current UK power needs?
Total requirement for wind farms cost in £
Total requirement for Solar Farms cost in £
Storage of electricity for 24hrs in any one day cost in £
Power Transmission network additional needs for wind and solar cost in £
Power Distribution network additional needs for wind and solar cost in £
Base load backup requirements for when the wind does not blow and the sun does not shine, nuclear or gas or oil cost in £
Replacement cost over a 25 year operational life of all installation.
Total cost for new installations for heat pumps cost in £
Total cost for all vehicle replacement by electric vehicles cost in £
Total land take for erection and placement of wind and solar farms and battery storage facilities to meet the 2035 target date and the associated cost, area in hectares and cost in £
The reply from Chat GPT
Estimating the costs and land requirements associated with the UK's Net Zero policies and the expansion of renewable energy infrastructure by 2035 involves several key components. Below is a breakdown based on current data and projections:
1. Total Requirement for Wind Farms Cost (£):
Offshore Wind: The UK aims to achieve 50 GW of offshore wind capacity by 2030. The cost of offshore wind has been decreasing, with recent projects achieving strike prices as low as £47.50 per MWh. Assuming an average capital expenditure of £2.5 million per MW, the total investment for 50 GW would be approximately £125 billion.
Onshore Wind: Plans include doubling onshore wind capacity by 2030. With current capacity around 15 GW, this implies an additional 15 GW. At an estimated £1.25 million per MW, the total cost would be approximately £18.75 billion.
2. Total Requirement for Solar Farms Cost (£):
Solar Power: The UK government aims to increase solar capacity to 70 GW by 2035, up from approximately 16.9 GW as of mid-2024. This requires an additional 53.1 GW. Assuming an average cost of £0.75 million per MW, the total investment would be around £39.8 billion.
3. Storage of Electricity for 24 Hours Cost (£):
Energy Storage: National Grid estimates that the UK will need 20-25 GW of energy storage capacity by 2035. Assuming a cost of £500,000 per MW, the investment required would range from £10 billion to £12.5 billion.
4. Power Transmission Network Additional Needs for Wind and Solar Cost (£):
Transmission Upgrades: To accommodate increased renewable generation, significant investment in the transmission network is necessary. Estimates suggest that £170-£210 billion will need to be invested in the grid by 2050, with a substantial portion required by 2035. This includes projects like the Eastern Green Links, with combined costs of approximately £3.4 billion.
5. Power Distribution Network Additional Needs for Wind and Solar Cost (£):
Distribution Upgrades: Upgrading local distribution networks to handle decentralized generation and increased demand (e.g., from electric vehicles and heat pumps) is essential. While specific figures are less clear, a significant portion of the £170-£210 billion grid investment will be allocated to distribution enhancements.
6. Baseload Backup Requirements Cost (£):
Nuclear Power: The UK plans to have up to 24 GW of nuclear capacity by 2050. Projects like Sizewell C are expected to contribute 3.2 GW, with costs estimated between £20 billion and £35 billion. Assuming similar costs for additional capacity, the total investment could range from £150 billion to £280 billion.
Gas with Carbon Capture and Storage (CCS): Maintaining gas-fired plants equipped with CCS as backup is also considered. The costs for these technologies are evolving, but significant investment will be required to ensure reliability during periods of low renewable output.
7. Replacement Cost Over a 25-Year Operational Life of All Installations:
Renewable Energy Installations: Assuming an average operational life of 25 years, the replacement cost by 2050 would be comparable to the initial investments:
Offshore Wind: ~£125 billion
Onshore Wind: ~£18.75 billion
Solar Power: ~£39.8 billion
Total: ~£184 billion
8. Total Cost for New Installations for Heat Pumps (£):
Heat Pumps: The UK aims to install 600,000 heat pumps annually by 2028. Assuming this rate continues to 2035, approximately 4.2 million units would be installed. At an average cost of £10,000 per unit, the total investment would be around £42 billion.
9. Total Cost for All Vehicle Replacement by Electric Vehicles (£):
Electric Vehicles (EVs): With approximately 32 million cars on UK roads, transitioning entirely to EVs by 2035 is ambitious. Assuming an average EV cost of £30,000, the total investment would be £960 billion. However, this cost is typically borne by consumers rather than the government.
10. Total Land Take for Erection and Placement of Wind and Solar Farms and Battery Storage Facilities to Meet the 2035 Target Date and the Associated Cost:
Land Requirements:
Onshore Wind: 45 GW of onshore wind would require around 5,000 square kilometers, or 2.1% of the UK's land. However, 99% of this area remains usable for other purposes like agriculture, meaning only 50 square kilometers (5,000 hectares) are exclusively occupied by turbines.
Solar Farms: Achieving 70 GW of solar capacity by 2035 would require significant land area. Assuming an average of 5 acres per MW, this equates to approximately 141,640 hectares. Efforts are being made to utilize rooftops and brownfield sites to minimize the use of greenfield land.
Associated Costs: Land acquisition costs vary widely based on location and land type. Utilizing existing structures (e.g., rooftops) can reduce costs, while greenfield developments may incur higher expenses due to land value and potential planning challenges.
Conclusion:
The transition to a Net Zero economy by 2035 involves substantial investments across various sectors, including energy generation, infrastructure, and transportation. While these figures provide a broad overview, actual costs will depend on technological advancements, policy decisions, market dynamics, and societal adoption rates.
The investment costs for different types of generation are 50-75% of the actual figures. That is because all of the official figures that ChatGPT uses have been grossly skewed by renewables optimism and do not reflect what you learn if you analyse company accounts.
As an example, the solar capital cost is given as £0.75 million per MW. The actual current figure is about £1.25 million per MW. The ChatGPT figure is what BEIS/DESNZ used to claim in 2020 was the cost before reality reasserted itself. All of the other figures are equally wrong.
I had always assumed that, for the majority of times, GB imported as it was cheaper than generating (by any means). Obviously there would be times when importing was the only way to “keep the lights on” but that these times would be infrequent. Is this assumption correct?
In a sense, imports are always driven by price, i.e. that it is cheap to import than to run alternative generators. That is true when there is extreme system stress because with such events the cost of either demand reductions or running backup generators is so high.
The critical questions for keeping the lights on is (a) whether import capacity is fully committed, and (b) if not, whether neighbouring countries are willing to increase their exports (our imports), which they may not be willing to do if they are operating under similar stress.
For the original 2GW IFA interconnector it was true that its supply cost was very low, taking advantage of the 5.6GW nuclear complex at Gravelines. It took the pressure off expanding the Dungeness site, including sidelining anti-nuclear protests.
BritNed was built after Greenpeace protests led by Zac Goldsmith halted the development of Kingsnorth D which would have been a modern coal station. Instead the MPP3 power station was built at the other end of the line at Maasvlakte next to the HVDC terminal. That had a marginal supply cost advantage in that the very biggest bulk carriers could deliver coal, whereas the Kingsnorth berths were slightly more limited. That very marginal freight saving is totally inadequate to pay for the cost of the link, so it has increased costs, and now we also see reduced supply security because the effect of Germany becoming a net importer has been to divert output inland. Indeed, you often see Kent being used as a transmission line for power from France and proxy Belgium to the Netherlands. During the French Grand Carenage it was used to route Belgian supply to France.
It is somewhat harder to disentangle the effects of NEMO from Belgium, where reactor shutdowns like those in France and closure too have complicated the supply picture. In addition, Belgium is small and linked in to the Netherlands, France and Germany via Luxembourg. It's worth looking at the analyses of European power flows by Paul-Frederik Bach
Norway is in principle another low cost supplier based on its hydro production. However, increased interconnection has driven up prices in Southern Norway, which is mulling sharp cutbacks in exports to its customers: Denmark, Germany, Netherlands and the UK. It also acts as a through route for some Swedish exports where prices have been similarly affected in the South. The problem for the UK is
Norway-> UK 1.4GW NSL
Norway-> Denmark 1.6GW -> UK 1.4GW Viking Link
Norway -> Netherlands 0.7GW-> UK 1GW BritNed
Norway-> Germany 1.4GW -> UK 1.4GW NeuConnect from 2028
Much of our interconnector supply is directly at stake.
In general the Single Electricity Market is very efficient at transmitting shortage pricing across much of Europe. The latest countries to suffer are the Baltics which recently cut their links to Russia and Belarus. With cables between Finland and Estonia and Sweden and Latvia severed by anchors (another vulnerability that caused lengthy interruption on IFA in the past) and a lack of local dispatchable capacity prices have soared.
Inside the UK interconnector use is significantly driven by lack of regional dispatchable generating capacity and transmission constraint. This is particularly true for London and the South East. So we pay nearly £10,000/MWh to persuade Belgium to export to keep the lights on in London (or perhaps in France which we were supplying at the time).
Thank you for providing the background to the development and operation of various interconnectors. I was not trying to justify the reliance upon imports. What I wanted to do was to challenge the assumption that the GB market is an isolated system with the marginal price set consistently by gas generation, which is, in my view, a hangover from 10 years or more ago.
Equally I think it is ridiculous that NESO uses derating factors of 60% to 75% for imports over various inteconnectors for the capacity market. In essence that ignores exactly the kind of event that you refer to in your last sentence. We should not rely on any interconnectors in reliability calculations unless there are contracted supplies at reasonable prices. I doubt whether any exporter would agree to such contracts but at least we should test the market.
On gas prices - while gas generation accounts for a small share of electricity produced, it frequently sets prices as the marginal generator, so I believe gas plays a role.
On costs - it’s quite interesting to hear across European countries that renewables reduce bills, especially when, for example, wind generation costs without firming are same or higher than market prices today.
It is not so simple. Your first point is the standard view. My argument is that in the current system the marginal generators are imports not gas plants- i.e. gas generators run up to the point where their marginal cost matches the price of supply from marginal suppliers of imports. In terms of conventional economics, the classic model of system pricing was/is based on a model in which imports are small and importers are price-takers. That is not the case today. There are also issues around the pricing of import capacity that I haven't discussed but which add an extra layer of complexity.
The claim in various European countries that renewables reduce bills rest on the usual confusion of what I will call "pure power costs" and meeting a system demand curve. In many European countries wind & solar generators rely on socialising the cost of transmission - e.g. it is the transmission system operator that covers all of the cost of connections to the grid and transmission, even for offshore wind farms.
Imports might just be gas generation pricing in disguise - just happening elsewhere. But as you said, that’s the standard view, which is not entirely right. And capacity constraints do add a thick layer of complexity. Thanks for taking the time to reply!
I would have gone along with your first sentence until I looked carefully at the composition of generation in EU countries that export directly or directly to the GB market. France is the major contributor and gas accounts for less than 5% of French generation. In addition, the French system is highly regional. Gas is the marginal fuel only in the south - ie exports to Italy and Spain. It is nuclear output that is crucial for GB imports.
The other major sources of exports to GB are Norway/Denmark (NSL/Viking links). None of that is gas. Gas does matter in the Netherlands & Belgium (Britned/Nemo links) but these are small in the overall picture. There are other things going on, like French exports to Germany and Belgium but there are constraints on interconnector capacity and even Germany only produces 15% of its electricity from gas.
My point is that what may have been true in 2017-18 is no longer true today.
Most people still think (including those who should know better) that electricity prices are determined by a merit order auction, as operated under the pool price system that was abolished in 2001, with all generators being paid the clearing price. That is a very long way from the truth.
Particularly since OFGEM bill methodology evolved into the OFGEM cap, retailers have been effectively required to make advance hedge purchases. Until the energy crisis that assumed purchase 12 -18 months in advance, though in practice that was only partly achieved.
The energy crisis made hedging enormously risky for generators and retailers: the consequence of a generator breakdown could be
In having to buy alternative supply at bankrupting prices even if the mark to market collateral requirement hadn't exhausted their borrowing capacity; retailers ran the risk that prices were liable to fall leaving them with expensive purchases and departing customers, and likewise a huge stretch on balance sheets to provide hedge collateral. That is why the cap is now reset quarterly, but still most electricity is effectively bought ahead of time.
The sellers of that are ultimately generators who know their costs and can be reasonably certain of being able to supply if called on. That is nuclear (including from France by interconnector), biomass like Drax, and gas CCGT (in the past coal when it was not constrained), and some other elements of interconnector based supply. Gas based hedges price off futures gas prices - not the short term day to day prices that can be much more volatile. Nuclear breakdowns reduced the willingness to make advance hedge sales, or only at very high prices.
The reality of electricity prices for renewables generation is that CFD (average cost about £150/MWh) and FiT based generators (average cost over £200/MWh) get fixed prices that have to be paid by retailers and are passed through to bills. ROC based generators are guaranteed their subsidies when they generate, but most do not sell significant volumes of forward hedges because they are always at risk of having to buy in costly alternative supply if the wind doesn't blow. That means they will sell in short term markets when weather forecasts allow them to have a better idea of production. It also means that on windier and sunnier days renewables have an important role in setting prices with their pricing determined by their subsidies and the alternativevif curtailment revenue.
Forward hedges tend to be at a constant level of supply 24x7, or at least for a block of hours per day for peak load. These positions have to be "shaped" to match real world demand patterns with lower demand overnight and at weekends and higher demand in rush hour. In terms of cost it is the difference in price between the high and low demand hours that matters. That is increasingly being set by the margin required by batteries to charge up and discharge, and by demand flexibility/price rationing. There is still an element determined by the costs of ramping up and down for CCGT flexibility.
Of course there is also the rebalancing required to accommodate the level of renewables generation. When it is high, renewables generators need to offer at a low enough price for CCGT to prefer cutting output and reselling the gas they bought to supply that, subject to the curtailment alternative. CFD generators do not care what these prices are, since they will simply make up the difference in CFD payments. When renewables output is low prices will be set by demand, mostly industrial demand that backs off or uses its own generators. Interconnector supply may fall as supply is likely to be tight on the Continent too. But if a retailer that has bought a hedge from a CCGT generator ends up selling the CCGT contract back and replacing it by wind at the same price, (or equivalently if the generator buys the supply from the wind farm and doesn't generate to meet his contract, using the wind instead), then the retailer's cost is unaltered whatever the transaction price for the substitution.
There is still a market that operates an auction on a pay as cleared basis - Nord Pool. It acts to provide the marker for CFD payments, which simply determines which pocket CFD generators get paid from, although its pricing is used also for other renewables base prices, but ROC subsidies now averaging close to £100/MWh are not included, neither are renewables greenwash certificates called REGOs that sell for £10-15/MWh.
Pricing is a lot more complex than it used to be, and Gordon is right that international prices now have a very important role. Gas is not really setting prices: the margins between daily gas price and CCGT sales prices can vary hugely over a day. CCGT is pricing as best it can in the market, which is determined by increasingly by other competition..
I agree fully that policymakers and many commentators operate on a model of price determination that is very out-of-date - really for a system with a single spot market and minimal trade. A couple of other points:
A. I have over-simplified the role of imports in the market. Import capacity is limited and auctioned. When there is excess supply of imports the price of capacity adjusts to ration import supply, but modelling that is nearly impossible when there are multiple interconnectors with transit costs between them. Further, this introduces statistical problems of simultaneity.
B. Your points about the complexity of PPAs and hedging markets - both spot and futures - are important. However, I think more renewable output is being sold ahead via PPAs with traders taking the weather risk. The trading arms of operators like Orsted, SSE, SPR, etc have large contracts with their wind farms. Octopus does the same for solar facilities. Often that is the basis on which the operators sell majority shares in the solar and wind farms to passive investors. In effect, the operator reaps a capital gain on the shares that it sells but takes on future hedging risks via its trading arm - Orsted & SSE do this on a large scale. It is a quasi-Ponzi game in which shareholders in the operating companies don't really understand what is going on. Orsted relies on the Danish government and SSE on its networks to underwrite the risks.
C. You might read my REF paper on price variability. I think that Ofgem have gone down completely the wrong route for setting the price cap. My view is that they should permit/encourage the use of lagged average market price which is the standard flexible pricing model in many western EU countries. What they have done, even with 3-month price caps, is to impose a large hedging cost on customers, which has increased over time as the bankruptcy of many energy suppliers revealed the consequences of not hedging.
On your point C I agree entirely that OFGEM should not impose a large hedging cost on retailers. I wrote to OFGEM a long time ago (long before the energy crisis blew up) suggesting that they should base prices around monthly averages, and I also wrote to the FCA pointing out that OFGEM really lacked any competence in understanding the financial exposures involved in hedging, and that the provision of fixed forward deals and caps were sales of sophisticated derivatives that they would regulate closely if they were treated properly as financial products. I tried to get the FCA simply to provide some advice to OFGEM, but they refused to countenance it at the highest levels.
The effect of the cap hedging requirement was to bring forward a lot of demand into the market, which meant that prices were higher than they otherwise would have been, and that effect rolled forward as the cap persisted.
On PPAs: From what I can see, there are PPAs that take all the dirty work away of ensuring that actual output is sold, curtailment payments applied for, nominations made to Elexon, ROCs applied for and given value etc. But the actual pricing offered on those tends to be hand to mouth and based on actual volumes and day ahead markets. Very few offer guaranteed pricing or volume guarantees. This is visible in wind farm accounts, which I know you have studied also. Of course, CFDs completely eliminate price risk anyway.
Hi, apart from electricity generation, there is the other commitment to reduce greenhouse gas emissions (GHG) by 81% by 2035 compared to 1990, made by Keir Starmer at COP29.
From Statistical Review of World Energy 2024, in 1990, U.K. CO2 equivalent emissions from energy, process emissions, methane and flaring were 595 million tonnes. So by 2035 they need to be 113 million tonnes. In 2023 they were 327 million tonnes. So if other GHG sources ( e.g. from farming) are ignored, fossil fuel energy consumption has to be reduced by 65% by 2035 compared to 2023. Now, in 2023, fossil fuels, mostly oil and gas, provided 5.16 exajoules (EJ) of primary energy. If U.K. has to replace 65% of this, it has to find 3.35 EJ of new CO2-free energy. Even if we assume that miraculously U.K. will reduce its primary energy requirement by 40%, it still has to find 2.01 EJ of new CO2-free energy. This is equivalent to providing 65 GW of continuous CO2-free energy or around 22 new Hinckley Point C nuclear power plants or 145 GW of new wind assuming a very optimistic capacity factor, CF of 0.45. For refernce, CF for wind currently is 0.31(for solar it is 0.11) and installed wind capacity is around 31 GW. So almost everything has to be run on CO2-free electricity There also has to be a massive change in lifestyle on top of the 40% reduction in primary energy consumption assumed - enormous reductions, almost to zero, in meat and dairy consumption, aviation , cement and steel industries . …
All this will be impossible to achieve in the real world by 2035. Anyway, what would be the cost of all this if it were possible?
Gautam - thank you for the comment.
It is my intention to extend my analysis to 2035 in the near future (and when I have time). The assumption to make about nuclear generation is critical because the projections for wind & solar capacity will just result in a very unstable system that has to rely on large amounts of gas to cover intermittency. The problem is that I don't believe either the current UK government or the UK nuclear industry is capable of bringing any large new nuclear plants (other than Hinkley Point) on line by 2035. Even if they gave permission immediately to Sizewell C it will take more than 10 years to build.
There is a small chance that enough units of the GE-Hitachi SMR might be built in that timescale to make a difference. I don't think that there is any chance that RR units will built that quickly. But all of this is fiddling at the margins. Ultimately what we have politicians who insist on virtue-signalling and have no willingness to listen to serious engineering analysts.
Comments like yours take me back to a motto a highly-esteemed engineering boss of mine had on his office wall: "Nothing is impossible for those who don't have to do it".
Professor Hughes,
The NESO projection that gas will supply 5% of demand by 2030 is slightly disingenuous, I think. Quite by how much I don't know.
While gas does provide balancing of supply and demand it also provides inertia, reactive power and short circuit current input ( wnen required) which renewables cannot and are essential for a stable grid.
These services can be provided even if the generator is not actually providing power just by being synchronised to the grid. Usually called spinning reserve.
How large this spinning reserve is something I cannot guess at but suspect it will be significant, especially as those services will have diminshed with the declining nuclear capcity byt that time.
Obviously running gas plants with no grid input generates CO2 and does not help the longevity of the gas turbines running on no load with a cost implication to the operator, who has to be paid adequately for the service provided.
Perhaps someone who is far more informed than I can comment.
It does seem to be a bit of a cover up is possible here to comply with the 5% (Maybe, a big maybe)
as I don't think spinning reserve capacity information is readily available?
You are correct about the need for inertia, etc. The NESO 5% gas figure is just silly - driven by politics rather than serious analysis. It is based on an assumption that exporting countries (to the GB market) will be willing to supply whatever the GB needs.
My estimate is about 3 times that and my model suggests that some gas generation will be required for about 50% of hours in the year. That will be enough to warrant keeping 5-8 GW of CCGTs running and they can be paid to operate on standby (at, say, 10% of output) to provide spinning reserve.
On the other hand, the renewables folk will say that this is all terribly 20th-century. Their view is that new technology can allow capacitor banks and batteries to provide system inertia and reactive power. I suspect that NESO is going to degrade the reliability standards by allowing much larger variations in frequency.
NESO have been gradually reducing operational minimum inertia levels. Currently they are supposed to operate at above 120GVAs which is already less than half of levels in the pre renewables grid. They are talking about aiming to reduce this to 96GVAs by year end. It can be useful to think of the reaction time that inertia provides for boosting other power by dividing the inertia by the demand level: inertia is then measured in seconds of demand. I have found that in practice NESO have been marginally conservative by ensuring that inertia remains above 4 seconds, at least so far.
However they are now very dependent on batteries to provide initial support, usually backed up with pumped storage before gas takes over. This often works very well - but not always. NESO remain extraordinarily tight lipped about their lack of understanding of what happened at 13:10 in 22nd December 2023. Initially there was a 1GW trip of IFA1, which was followed rapidly by losing almost 400MW at Cottam in the Midlands and a trip on the Moray offshore interconnector in Northern Scotland. Those who watched the frequency plunge to 49.273Hz estimate there were also other significant losses of embedded generation, perhaps totalling well over 2GW across all sources. There were 60 seconds where the recorded frequency was below the 49.5Hz SQSS standard, but an in depth enquiry was avoided because that did not exceed 60 seconds. There would have been enormous incentives to correct the data if necessary to avoid an extra second.
The reality seems to be that interactions between inverter based sources are not always predictable. Something similar underlayed the August 2019 blackout, when the voltage disturbance from a lightning strike was sufficient to knock out the entire output of Hornsea wind farm and helped trigger significant embedded generation losses as well as the loss of Little Barford CCGT.
I am not a power engineer so I haven't anything useful to say about the 2023 episode. However, thinking about the 2019 blackout as well as events in the US & Canada, I suspect that reliability margins need to be greater than they are. Losing Hornsea and Little Barford was one thing, but as we start to rely more heavily on large internal interconnectors there is the prospect of losing, for example, the planned Eastern Green link. In opting for economies of scale the system is becoming increasing vulnerable to major network failures. Power engineering principles of the past weren't daft and the idea that we can squeeze reserve margins without a substantial decrease in reliability is self-deception.
10% on an individual CCGT if that’s what you are implying isn’t realistic, with the age of most of the UK fleet you’ve always been looking at something nearer 30%, the likes of say Carrington commissioned in 2016 could be lower. The data on minimum stable export levels by generator are somewhere on Elexon but I’ve not looked for a while. It’s not strictly an electrical limitation, the emissions tend to go off the scale at low loading.
Fair point about older plants. I was thinking about ones built in the last decade but there is only Carrington. Some of the ones commissioned since 2010 might be lower too. As you mention, I doubt whether any of them can meet NOx emission limits at low levels of output. Yet another reason why they might choose to run Drax a lot more than is being suggested.
Great article.
I particularly like that you have used 8 years of data to look at gas and interconnector needs. NESO and the CCC both used a single year analysis when doing the same work. This totally ignores longer duration weather trends.
On the gas case side, am I right in understanding that you’ve done historical pricing for the prior year comparison and current gas pricing for the last table? I’m interested as many forecasts for gas prices show a substantial fall in gas price from 2026/7 as more export terminals come on stream. This will be definition driven down gas power prices in the 2024 case but will make much less impact in the NESO case as more of the power is at a locked in price with CPI escalation. Would be interested in your thoughts on that?
If I understand the question in your second paragraph correctly, the answer is yes. All of the main calculations are based on 2024 market prices and 2024 gas prices. The exception was the comparison across years which is based on the historical prices of each year.
I know that many expect gas prices to fall later this decade. That reflects the overall pattern of gas prices over several decades - occasional sharp spikes followed an extended downward trend for 8-10 years. Investment in pipelines and LNG transport is an element of why that pattern prevails. I used 2024 prices to avoid comments that my results rely on specific assumptions about the future of gas prices.
I would be cautious about translating future gas prices to wholesale power prices. In part this is because I don't believe that gas generation is now the marginal source of supply to the market - that is imports. Another consideration is that we are approaching the point at which there is no requirement for gas generation, other than for reasons of system reliability (spinning reserve, etc), for a significant number of periods in the year. At that point, even if you accept the conventional view marginal supplies will definitely come from imports because biomass (with the next highest fuel cost) is subsidised. At that point, as I argue in the Technical Appendix, the market looks to be extremely unstable.
Thank you for this article Gordon, clearly the result of a great deal of research and analysis. I am unable to offer anything scholarly, but it has prompted a few thoughts. Net Zero as a policy has been pursued for some time, and the contribution of wind and solar energy to the electricity grid is now significant. Given that the result to date has been increased energy costs, observable to all, doubtless causing economic damage and in some cases real hardship, it seems brave, or even courageous, to believe that the approach simply hasn't gone far enough. ( I use Sir Humphrey Appleby's terminology - a brave decision costs votes, a courageous one may lose an election.) A pragmatist would surely pause and question the direction of travel, so it's clear the show is not being run by such animals.
As an aside, there is currently a local issue with the route for power lines and associated pylons from a wind farm in rural Wales. Our local MP, a Lib Dem safe from the agonies of governing, has advocated that power lines should be run underground where they pass near to people's homes or businesses. It would be blasphemy for her to cast doubt on the whole endeavour, and she will be unaware that in passing, she has assigned zero value to landscape and countryside.
Re your first paragraph. That is why everything is blamed on gas prices. It wasn't us, guv, that pushed up prices, it was that horrible gas market. That is why I wrote my previous piece showing that, putting aside spikes in 2008 and 2021-23, the long term increase in electricity prices was entirely the consequence of policy decisions. But both politicians and the media have short memories and are uninterested in anything other than simple stories.
Re Lib Dems. I have come across similar behaviour from LD MPs concerning solar farms. They are the kind of politicians that give NIMBYs a bad name - pure hypocrisy - whereas the majority of objectors that I deal with have a much better understanding of the choices and policy issues. I have long felt that any MP (MSP/MS) who supports the development of renewable generation should thereby be taken as agreeing to wind turbines or solar panels being located in their local park and no objection from them should be permitted. The point is that policymakers want to impose the costs of their decisions on other people, but not themselves.
This is a tremendous piece of work. I know just how much effort is required to pull something like this together having done various analyses myself. So well done.
It underscores the idea that the markets have changed how they operate and backs it with hard analysis, as well as showing that the Miliband plan is in reality a high cost option. I think it's possible to discern many of your findings when closely looking at a chart of electricity supply by source with a price overlay like this one:
https://i0.wp.com/wattsupwiththat.com/wp-content/uploads/2024/08/Generation-jan-2023-1722888879.3157.png
I have yet to complete my analysis of 2024 because of the need for data cleaning which can be very time consuming. However, it is easy to see that intraday price volatility is huge, particularly when there is a high share of renewables on the system. When it is windy there are huge swings in interconnector supply between net export and large imports, and pumped storage also plays a role. Back then batteries were small beer, and only reported for net discharge AFAICS under the OTHER categroy of generation. Downloading individual metered readings for batteries only shows the net charge or discharge over a settlement period, not the total input for charging/cooling and discharge output, so it is hard to get good information on battery performance, and unfortunately Bess Analytics has now disappeared behind a paywall having been acquired by Cornwall Insight, so now we depend on commentary from the like of ModoEnergy for analysis (they are very good though).
If we rearrange the data in price order we get a chart like this
https://i0.wp.com/wattsupwiththat.com/wp-content/uploads/2024/05/Gen-by-Price-Jan-2023-1716328267.549.png
which shows clearly that when wind output dominates prices are set by competition between wind sources, and tend close to zero or negative when we have a surplus export. Subsidising exports for the benefit of foreign consumers has never made sense to me. As capacity increases we will see more hours with wind surpluses and the surpluses will grow in low demand hours. It is clear that the structure of the market will concentrate curtailment on those wind farms that get no compensation when prices are negative. We can see this already with Seagreen and Moray East, which have yet to take up CFDs and become the dominant volumes of curtailment. The price they have been getting has been steadily reducing too, as more competition starts to come on stream from other wind farms in construction which also get no subsidy protection.
Incidentally, I think you are incorrect about ROCs in your Annex. For now they remain traded, although the market is very opaque and not openly reported unless you subscribe to a specialist price reporter like ICIS. However, the final recycle value of ROCs that includes the premium for redistribution of the cashout fund is published (often obscurely) by OFGEM. REF has a convenient table of the history at the bottom of the page here:
https://www.ref.org.uk/energy-data/notes-on-the-renewable-obligation
I am using an assumption of a 12% premium for the current year, giving a value of £72.50/ROC, as a guess that allows for slightly disappointing actual output from renewables and probably more stable demand. Up for consultation is a proposal to fix the ROC value without trading at a 10% uplift to the indexed cashout value for the life of the scheme between 2027 and 2037 when it expires, because it is anticipated that the balance between the obligation and supply could become much more volatile in future.
On your last point, I made a conscious decision to exclude recycle values. There are two reasons.
First, recycle values have varied greatly over time. In several recent years they have been well below your estimate of 12%. If I had included them I would have adopted a figure in the range 5-10%. I have discussed this on several occasions with Lee Moroney of REF. In many years the ultimate recycle value was much higher than the expected value during the year.
Second, recycle values are only available with a substantial lag and I wanted to ensure that I had comparable figures up to 2024. Since the whole ROC system is likely to change from 2027 I wanted to ensure that my estimates were cautious and defensible.
The procedure for calculating the ROC obligation has been Calculation B since the early years. That estimates the number of ROCs that eligible generators will earn using historic data on capacity factors and the the OFGEM data on capacity. That total is uplifted by 10%. The demand liable to fund ROCs is also estimated (I.e. excluding exempt energy intensive demand) and the obligation is then calculated by dividing one by the other to give a figure in ROC/MWh.
After the end of the year OFGEM calculate the total obligation by multiplying the ROC/MWh by actual eligible demand. Thus if demand is higher than forecast the total number of ROCs or cashout payments will be correspondingly higher, and vice versa. The actual ROCs earned will almost certainly be fewer than the obligation total by virtue of the 10% uplift in the forecast. Only in 2015 was generation significantly higher, resulting in a zero premium to cashout value and some carryover into the subsequent year (I would have to check whether there was a corresponding adjustment to the 2016 obligation).
If the forecasts are correct then the actual premium would be 10%, perhaps adjusted downwards for the time value of money, since cashout payments are only needed by October - essentially almost a year later on average. Those trading the market will make their own forecasts of the premium based on actual demand and generation and their own forecasts for the balance of the year, and use those to inform the price they would pay up front for ROCs. A correctly forecast premium leaves them indifferent between paying for ROCs and paying cashouts, since the former attract the recycle, while cashouts do not. A firm that underestimates the premium will in fact pay more because it will have to buy more than pro-rata cashouts. One that overestimates it will pay more simply because they overpaid.
There is an element of transaction cost and incentive that will mean that the actual receipts by generators will be slightly below the full recycle value. However, with that adjustment, that is what they can expect to be paid. If overall ROC generation falls short of forecast the ROC value rises in compensation, and it falls if generation is higher. If demand is higher than forecast ROC values will increase, and fall if it is lower.
This paper could have been a useful contribution to the debunking of Mad Miliband's lunatic lies about falling energy prices.
It's a great shame therefore that you have written such a confusing and incomplete story. Many of us are trying to inform the public of the true threat of Net Zero - not "Clean Power by 2030" which is a tiny (yet still wholly unachievable) part of the full story. You could choose to collaborate with others to get the complete truth out there, but you seemingly prefer to go it alone with little care for the confusion you will actually add to.
https://x.com/EyesOnThePriz12/status/1896562028195815886
https://drive.google.com/file/d/1bgDmcUapJCgwrplegj_0MM_IbGNE_c9v/view
I am sorry that you feel that way, but I am not going to change how I work. As anyone familiar with my work will guess, I believe that collectively we benefit from people who are willing to investigate the nitty-gritty details of grand claims. In modern parlance I am a data scientist not a political campaigner. I want to test the claims made by politicians and lobbyists not engage in grand speculation. Following the aphorism of Maynard Keynes, I am economist as "humble dentist" rather than public panjandrum.
My previous comment should have had this preface ... sorry for my spastic thumbs!
Thank you for this. It addresses a problem that I, as an observer have been struggling with for a long time, both to formulate the right form of questions and where to look for the answers.
Some time ago I thought I’d try ChatGPT to get a concise answer and this is what it told me:
Could you possible look it over and correct any of the errors in my question and nuances for the reply?
The components of Net Zero policies and wind and solar power or all renewables have a cost. What are the out-turn expected UK costs in 2035 associated with the following basic element of Net Zero and renewables on the basis of current UK power needs?
Total requirement for wind farms cost in £
Total requirement for Solar Farms cost in £
Storage of electricity for 24hrs in any one day cost in £
Power Transmission network additional needs for wind and solar cost in £
Power Distribution network additional needs for wind and solar cost in £
Base load backup requirements for when the wind does not blow and the sun does not shine, nuclear or gas or oil cost in £
Replacement cost over a 25 year operational life of all installation.
Total cost for new installations for heat pumps cost in £
Total cost for all vehicle replacement by electric vehicles cost in £
Total land take for erection and placement of wind and solar farms and battery storage facilities to meet the 2035 target date and the associated cost, area in hectares and cost in £
The reply from Chat GPT
Estimating the costs and land requirements associated with the UK's Net Zero policies and the expansion of renewable energy infrastructure by 2035 involves several key components. Below is a breakdown based on current data and projections:
1. Total Requirement for Wind Farms Cost (£):
Offshore Wind: The UK aims to achieve 50 GW of offshore wind capacity by 2030. The cost of offshore wind has been decreasing, with recent projects achieving strike prices as low as £47.50 per MWh. Assuming an average capital expenditure of £2.5 million per MW, the total investment for 50 GW would be approximately £125 billion.
Onshore Wind: Plans include doubling onshore wind capacity by 2030. With current capacity around 15 GW, this implies an additional 15 GW. At an estimated £1.25 million per MW, the total cost would be approximately £18.75 billion.
2. Total Requirement for Solar Farms Cost (£):
Solar Power: The UK government aims to increase solar capacity to 70 GW by 2035, up from approximately 16.9 GW as of mid-2024. This requires an additional 53.1 GW. Assuming an average cost of £0.75 million per MW, the total investment would be around £39.8 billion.
3. Storage of Electricity for 24 Hours Cost (£):
Energy Storage: National Grid estimates that the UK will need 20-25 GW of energy storage capacity by 2035. Assuming a cost of £500,000 per MW, the investment required would range from £10 billion to £12.5 billion.
4. Power Transmission Network Additional Needs for Wind and Solar Cost (£):
Transmission Upgrades: To accommodate increased renewable generation, significant investment in the transmission network is necessary. Estimates suggest that £170-£210 billion will need to be invested in the grid by 2050, with a substantial portion required by 2035. This includes projects like the Eastern Green Links, with combined costs of approximately £3.4 billion.
5. Power Distribution Network Additional Needs for Wind and Solar Cost (£):
Distribution Upgrades: Upgrading local distribution networks to handle decentralized generation and increased demand (e.g., from electric vehicles and heat pumps) is essential. While specific figures are less clear, a significant portion of the £170-£210 billion grid investment will be allocated to distribution enhancements.
6. Baseload Backup Requirements Cost (£):
Nuclear Power: The UK plans to have up to 24 GW of nuclear capacity by 2050. Projects like Sizewell C are expected to contribute 3.2 GW, with costs estimated between £20 billion and £35 billion. Assuming similar costs for additional capacity, the total investment could range from £150 billion to £280 billion.
Gas with Carbon Capture and Storage (CCS): Maintaining gas-fired plants equipped with CCS as backup is also considered. The costs for these technologies are evolving, but significant investment will be required to ensure reliability during periods of low renewable output.
7. Replacement Cost Over a 25-Year Operational Life of All Installations:
Renewable Energy Installations: Assuming an average operational life of 25 years, the replacement cost by 2050 would be comparable to the initial investments:
Offshore Wind: ~£125 billion
Onshore Wind: ~£18.75 billion
Solar Power: ~£39.8 billion
Total: ~£184 billion
8. Total Cost for New Installations for Heat Pumps (£):
Heat Pumps: The UK aims to install 600,000 heat pumps annually by 2028. Assuming this rate continues to 2035, approximately 4.2 million units would be installed. At an average cost of £10,000 per unit, the total investment would be around £42 billion.
9. Total Cost for All Vehicle Replacement by Electric Vehicles (£):
Electric Vehicles (EVs): With approximately 32 million cars on UK roads, transitioning entirely to EVs by 2035 is ambitious. Assuming an average EV cost of £30,000, the total investment would be £960 billion. However, this cost is typically borne by consumers rather than the government.
10. Total Land Take for Erection and Placement of Wind and Solar Farms and Battery Storage Facilities to Meet the 2035 Target Date and the Associated Cost:
Land Requirements:
Onshore Wind: 45 GW of onshore wind would require around 5,000 square kilometers, or 2.1% of the UK's land. However, 99% of this area remains usable for other purposes like agriculture, meaning only 50 square kilometers (5,000 hectares) are exclusively occupied by turbines.
Solar Farms: Achieving 70 GW of solar capacity by 2035 would require significant land area. Assuming an average of 5 acres per MW, this equates to approximately 141,640 hectares. Efforts are being made to utilize rooftops and brownfield sites to minimize the use of greenfield land.
Associated Costs: Land acquisition costs vary widely based on location and land type. Utilizing existing structures (e.g., rooftops) can reduce costs, while greenfield developments may incur higher expenses due to land value and potential planning challenges.
Conclusion:
The transition to a Net Zero economy by 2035 involves substantial investments across various sectors, including energy generation, infrastructure, and transportation. While these figures provide a broad overview, actual costs will depend on technological advancements, policy decisions, market dynamics, and societal adoption rates.
Sources:
lordslibrary.parliament.uk
This shows the dangers of using ChatGTP!
The investment costs for different types of generation are 50-75% of the actual figures. That is because all of the official figures that ChatGPT uses have been grossly skewed by renewables optimism and do not reflect what you learn if you analyse company accounts.
As an example, the solar capital cost is given as £0.75 million per MW. The actual current figure is about £1.25 million per MW. The ChatGPT figure is what BEIS/DESNZ used to claim in 2020 was the cost before reality reasserted itself. All of the other figures are equally wrong.
I had always assumed that, for the majority of times, GB imported as it was cheaper than generating (by any means). Obviously there would be times when importing was the only way to “keep the lights on” but that these times would be infrequent. Is this assumption correct?
In a sense, imports are always driven by price, i.e. that it is cheap to import than to run alternative generators. That is true when there is extreme system stress because with such events the cost of either demand reductions or running backup generators is so high.
The critical questions for keeping the lights on is (a) whether import capacity is fully committed, and (b) if not, whether neighbouring countries are willing to increase their exports (our imports), which they may not be willing to do if they are operating under similar stress.
For the original 2GW IFA interconnector it was true that its supply cost was very low, taking advantage of the 5.6GW nuclear complex at Gravelines. It took the pressure off expanding the Dungeness site, including sidelining anti-nuclear protests.
BritNed was built after Greenpeace protests led by Zac Goldsmith halted the development of Kingsnorth D which would have been a modern coal station. Instead the MPP3 power station was built at the other end of the line at Maasvlakte next to the HVDC terminal. That had a marginal supply cost advantage in that the very biggest bulk carriers could deliver coal, whereas the Kingsnorth berths were slightly more limited. That very marginal freight saving is totally inadequate to pay for the cost of the link, so it has increased costs, and now we also see reduced supply security because the effect of Germany becoming a net importer has been to divert output inland. Indeed, you often see Kent being used as a transmission line for power from France and proxy Belgium to the Netherlands. During the French Grand Carenage it was used to route Belgian supply to France.
It is somewhat harder to disentangle the effects of NEMO from Belgium, where reactor shutdowns like those in France and closure too have complicated the supply picture. In addition, Belgium is small and linked in to the Netherlands, France and Germany via Luxembourg. It's worth looking at the analyses of European power flows by Paul-Frederik Bach
http://pfbach.dk/
Norway is in principle another low cost supplier based on its hydro production. However, increased interconnection has driven up prices in Southern Norway, which is mulling sharp cutbacks in exports to its customers: Denmark, Germany, Netherlands and the UK. It also acts as a through route for some Swedish exports where prices have been similarly affected in the South. The problem for the UK is
Norway-> UK 1.4GW NSL
Norway-> Denmark 1.6GW -> UK 1.4GW Viking Link
Norway -> Netherlands 0.7GW-> UK 1GW BritNed
Norway-> Germany 1.4GW -> UK 1.4GW NeuConnect from 2028
Much of our interconnector supply is directly at stake.
In general the Single Electricity Market is very efficient at transmitting shortage pricing across much of Europe. The latest countries to suffer are the Baltics which recently cut their links to Russia and Belarus. With cables between Finland and Estonia and Sweden and Latvia severed by anchors (another vulnerability that caused lengthy interruption on IFA in the past) and a lack of local dispatchable capacity prices have soared.
Inside the UK interconnector use is significantly driven by lack of regional dispatchable generating capacity and transmission constraint. This is particularly true for London and the South East. So we pay nearly £10,000/MWh to persuade Belgium to export to keep the lights on in London (or perhaps in France which we were supplying at the time).
Thank you for providing the background to the development and operation of various interconnectors. I was not trying to justify the reliance upon imports. What I wanted to do was to challenge the assumption that the GB market is an isolated system with the marginal price set consistently by gas generation, which is, in my view, a hangover from 10 years or more ago.
Equally I think it is ridiculous that NESO uses derating factors of 60% to 75% for imports over various inteconnectors for the capacity market. In essence that ignores exactly the kind of event that you refer to in your last sentence. We should not rely on any interconnectors in reliability calculations unless there are contracted supplies at reasonable prices. I doubt whether any exporter would agree to such contracts but at least we should test the market.
Excellent analysis!
On gas prices - while gas generation accounts for a small share of electricity produced, it frequently sets prices as the marginal generator, so I believe gas plays a role.
On costs - it’s quite interesting to hear across European countries that renewables reduce bills, especially when, for example, wind generation costs without firming are same or higher than market prices today.
It is not so simple. Your first point is the standard view. My argument is that in the current system the marginal generators are imports not gas plants- i.e. gas generators run up to the point where their marginal cost matches the price of supply from marginal suppliers of imports. In terms of conventional economics, the classic model of system pricing was/is based on a model in which imports are small and importers are price-takers. That is not the case today. There are also issues around the pricing of import capacity that I haven't discussed but which add an extra layer of complexity.
The claim in various European countries that renewables reduce bills rest on the usual confusion of what I will call "pure power costs" and meeting a system demand curve. In many European countries wind & solar generators rely on socialising the cost of transmission - e.g. it is the transmission system operator that covers all of the cost of connections to the grid and transmission, even for offshore wind farms.
Imports might just be gas generation pricing in disguise - just happening elsewhere. But as you said, that’s the standard view, which is not entirely right. And capacity constraints do add a thick layer of complexity. Thanks for taking the time to reply!
I would have gone along with your first sentence until I looked carefully at the composition of generation in EU countries that export directly or directly to the GB market. France is the major contributor and gas accounts for less than 5% of French generation. In addition, the French system is highly regional. Gas is the marginal fuel only in the south - ie exports to Italy and Spain. It is nuclear output that is crucial for GB imports.
The other major sources of exports to GB are Norway/Denmark (NSL/Viking links). None of that is gas. Gas does matter in the Netherlands & Belgium (Britned/Nemo links) but these are small in the overall picture. There are other things going on, like French exports to Germany and Belgium but there are constraints on interconnector capacity and even Germany only produces 15% of its electricity from gas.
My point is that what may have been true in 2017-18 is no longer true today.
Fair enough
Most people still think (including those who should know better) that electricity prices are determined by a merit order auction, as operated under the pool price system that was abolished in 2001, with all generators being paid the clearing price. That is a very long way from the truth.
Particularly since OFGEM bill methodology evolved into the OFGEM cap, retailers have been effectively required to make advance hedge purchases. Until the energy crisis that assumed purchase 12 -18 months in advance, though in practice that was only partly achieved.
The energy crisis made hedging enormously risky for generators and retailers: the consequence of a generator breakdown could be
In having to buy alternative supply at bankrupting prices even if the mark to market collateral requirement hadn't exhausted their borrowing capacity; retailers ran the risk that prices were liable to fall leaving them with expensive purchases and departing customers, and likewise a huge stretch on balance sheets to provide hedge collateral. That is why the cap is now reset quarterly, but still most electricity is effectively bought ahead of time.
The sellers of that are ultimately generators who know their costs and can be reasonably certain of being able to supply if called on. That is nuclear (including from France by interconnector), biomass like Drax, and gas CCGT (in the past coal when it was not constrained), and some other elements of interconnector based supply. Gas based hedges price off futures gas prices - not the short term day to day prices that can be much more volatile. Nuclear breakdowns reduced the willingness to make advance hedge sales, or only at very high prices.
The reality of electricity prices for renewables generation is that CFD (average cost about £150/MWh) and FiT based generators (average cost over £200/MWh) get fixed prices that have to be paid by retailers and are passed through to bills. ROC based generators are guaranteed their subsidies when they generate, but most do not sell significant volumes of forward hedges because they are always at risk of having to buy in costly alternative supply if the wind doesn't blow. That means they will sell in short term markets when weather forecasts allow them to have a better idea of production. It also means that on windier and sunnier days renewables have an important role in setting prices with their pricing determined by their subsidies and the alternativevif curtailment revenue.
Forward hedges tend to be at a constant level of supply 24x7, or at least for a block of hours per day for peak load. These positions have to be "shaped" to match real world demand patterns with lower demand overnight and at weekends and higher demand in rush hour. In terms of cost it is the difference in price between the high and low demand hours that matters. That is increasingly being set by the margin required by batteries to charge up and discharge, and by demand flexibility/price rationing. There is still an element determined by the costs of ramping up and down for CCGT flexibility.
Of course there is also the rebalancing required to accommodate the level of renewables generation. When it is high, renewables generators need to offer at a low enough price for CCGT to prefer cutting output and reselling the gas they bought to supply that, subject to the curtailment alternative. CFD generators do not care what these prices are, since they will simply make up the difference in CFD payments. When renewables output is low prices will be set by demand, mostly industrial demand that backs off or uses its own generators. Interconnector supply may fall as supply is likely to be tight on the Continent too. But if a retailer that has bought a hedge from a CCGT generator ends up selling the CCGT contract back and replacing it by wind at the same price, (or equivalently if the generator buys the supply from the wind farm and doesn't generate to meet his contract, using the wind instead), then the retailer's cost is unaltered whatever the transaction price for the substitution.
There is still a market that operates an auction on a pay as cleared basis - Nord Pool. It acts to provide the marker for CFD payments, which simply determines which pocket CFD generators get paid from, although its pricing is used also for other renewables base prices, but ROC subsidies now averaging close to £100/MWh are not included, neither are renewables greenwash certificates called REGOs that sell for £10-15/MWh.
Pricing is a lot more complex than it used to be, and Gordon is right that international prices now have a very important role. Gas is not really setting prices: the margins between daily gas price and CCGT sales prices can vary hugely over a day. CCGT is pricing as best it can in the market, which is determined by increasingly by other competition..
I agree fully that policymakers and many commentators operate on a model of price determination that is very out-of-date - really for a system with a single spot market and minimal trade. A couple of other points:
A. I have over-simplified the role of imports in the market. Import capacity is limited and auctioned. When there is excess supply of imports the price of capacity adjusts to ration import supply, but modelling that is nearly impossible when there are multiple interconnectors with transit costs between them. Further, this introduces statistical problems of simultaneity.
B. Your points about the complexity of PPAs and hedging markets - both spot and futures - are important. However, I think more renewable output is being sold ahead via PPAs with traders taking the weather risk. The trading arms of operators like Orsted, SSE, SPR, etc have large contracts with their wind farms. Octopus does the same for solar facilities. Often that is the basis on which the operators sell majority shares in the solar and wind farms to passive investors. In effect, the operator reaps a capital gain on the shares that it sells but takes on future hedging risks via its trading arm - Orsted & SSE do this on a large scale. It is a quasi-Ponzi game in which shareholders in the operating companies don't really understand what is going on. Orsted relies on the Danish government and SSE on its networks to underwrite the risks.
C. You might read my REF paper on price variability. I think that Ofgem have gone down completely the wrong route for setting the price cap. My view is that they should permit/encourage the use of lagged average market price which is the standard flexible pricing model in many western EU countries. What they have done, even with 3-month price caps, is to impose a large hedging cost on customers, which has increased over time as the bankruptcy of many energy suppliers revealed the consequences of not hedging.
On your point C I agree entirely that OFGEM should not impose a large hedging cost on retailers. I wrote to OFGEM a long time ago (long before the energy crisis blew up) suggesting that they should base prices around monthly averages, and I also wrote to the FCA pointing out that OFGEM really lacked any competence in understanding the financial exposures involved in hedging, and that the provision of fixed forward deals and caps were sales of sophisticated derivatives that they would regulate closely if they were treated properly as financial products. I tried to get the FCA simply to provide some advice to OFGEM, but they refused to countenance it at the highest levels.
The effect of the cap hedging requirement was to bring forward a lot of demand into the market, which meant that prices were higher than they otherwise would have been, and that effect rolled forward as the cap persisted.
On PPAs: From what I can see, there are PPAs that take all the dirty work away of ensuring that actual output is sold, curtailment payments applied for, nominations made to Elexon, ROCs applied for and given value etc. But the actual pricing offered on those tends to be hand to mouth and based on actual volumes and day ahead markets. Very few offer guaranteed pricing or volume guarantees. This is visible in wind farm accounts, which I know you have studied also. Of course, CFDs completely eliminate price risk anyway.