Great post - I've been looking for a serious update to Mackay's numbers for ages.
One thing in curious about: developments in fixed wind turbines tech mean they are economically viable in more locations (e.g. deeper water) so can cover more area.
But how much more do they produce per square km of deployment than Mackay's estimates?
In his book, he has a nice section explaining that bigger turbines have to be spaced out more, so whilst theyre cheaper, they don't produce as much more energy as the headline 'output per turbine' would naively suggest.
But modern turbines are higher, so they presumably 'catch' more wind, and windspeeds are more consistent higher up. But I'm curious how big an effect this is.
I don't MacKays estimates but the cited study seems to assume a capacity factor of 50% for a 15 MW turbine.
Average turbine capacity of turbines installed in 2021 was 7.4 MW but all manufacturers seem to have 15+MW designs in the work [1]. So the projection is slightly optimistic IMO. The capacity factor seems slightly more optimistic at 50% as worldwide average seem to have fluctuated between 35 and 45% [2]. Although other numbers are closer to 60% for the UK [3]
Source [3] seems highly suspicious. There's ~24 GW of wind capacity in the UK, the vast majority offshore, and according to Elexon Portal (https://www.elexonportal.co.uk, a good public frontend at http://gridwatch.org.uk/), which gives realtime grid data, the maximum production was 10. So an average cap factor of 39% is, well sorry it's wrong.
And time and time again renewables always seems to take optimistic, if not ludicrous estimates.
Mackay's papers is indeed 15 years old, and what do we have now? Extremely expensive power, backed by fossil gas, because if wasn't, people would die. Avoiding 2 deg C and ecosystem collapse is not even vaguely possible anymore. In that time, 40 years ago, France got to 17 g / kWh. That's not a success and, for the literal trillions of ESG money spent, someone should be held accountable for that.
Wind turbine size has an incredible effect on total power generation. The generated power is subject to the square of the blade length, and the cube of the wind velocity (which generally increases as you get further from the ground). Costs don't increase exponentially - so it's better to have fewer big turbines than many small ones.
The result is that the wind industry (offshore - onshore is limited by physical size constraints) has been consistently getting bigger every year.
I thought the issue with wind in the UK was that its supply is (Scotland) where the demand isn’t (the south). So we’d (a) have to build loads of pylons or expensive underground cables and (b) lose a lot in transmission.
You would have to build lots of transmission, but the losses aren't particularly significant for high voltage lines — it's only about 1000 km from the Shetland islands to Southampton, and HVDC transmission losses are quoted at 3.5% per 1,000 km. Pricing seems to be a trade secret, but the suggested numbers on the Wikipedia page for the 8 GW cross-channel link were £110M for the converter stations and £1M/km for the undersea cable.
I know that a mere back-of-the-envelope calculation isn't worth much more than the used envelope it was written on (doubly so when it is based on guesstimates of the input numbers), but that would be only £1bn for 8 GW or £4bn for 32 GW (compared to actual average usage of 31.5 GW last year), which is the kind of thing that the British government shouldn't blink at but in practice actually faffs and fails at basically all the time.
(And the sector is theoretically privatised, so this would have to become a business investment, which in turns will have potential investors ask inconvenient questions like "What's the risk we have cheaper options in 10 years that make this power line redundant? And what about those fusion reactors I keep reading about in the Sunday Times? What if Scotland becomes independent and stops selling you the electricity?")
With proper high volatage direct current (HVDC) transmission, the transmission losses transporting electricity from Scotland to the south of England are not very relevant. It's like a couple of percent.
A bigger problem is just the UK's inability to complete infrastructure megaprojects on land, so the connectors would likely need to go in the sea and take a perhaps inefficient route.
There are a number of problems with wind in the UK. NIMBYism means it’s either in the north (nowhere near the consumer) or out in the sea which is both not terribly near the consumer and ferociously expensive to maintain. The UK Energy Catapult estimates that a single service vessel “truck roll” or “boat launch” (I guess) is something like £250K. Probably much more now as that figure is 10 years old. This means that it makes economic sense to wait until you have several broken wind turbines before sending out a service vessel. Couple this with the fact that they dont seem to have as long a lifetime as was promised (various reasons). Finally it is a meteorological reality that when it’s very cold in the UK and energy demands are high… it is also usually very still with no wind, and of course in the middle of winter when there are few hours of daylight helping us with solar generation.
> Finally it is a meteorological reality that when it’s very cold in the UK and energy demands are high… it is also usually very still with no wind, and of course in the middle of winter when there are few hours of daylight helping us with solar generation.
Your meteorological reality seems to not correlate with actual reality. In the UK the highest energy demand is actually correlated with high wind speeds [1]
>This reflects the variation in temperatures and wind speeds with season, with calmer, warmer conditions in summer and cooler, windier conditions in late autumn and early spring. However above the 75th percentile of demand, average wind power reduces, which occurs predominantly in winter and autumn. Understanding this downturn in wind power provides the motivation for this paper. Given our interest in high demand days, which predominantly occur in winter (figure 1, upper right), only winter days are considered.
>The tendency for lower wind power during higher winter demand is shown by the tilt of the density contours of the daily distribution (figure 1, lower left). It is also clearly seen when averaged across days of similar demand (figure 2, left). Average wind power reduces by a third between lower and higher winter demand, from approximately 60% to 40% of rated power.
Look at figure 2. Black is wind power, and the X axis is demand. Wind production capacity is down when demand is high.
The vast majority of the UK's cold winter weather comes from wind from the North-East, bringing in the much colder weather from Arctic/Siberian regions.
Ok my statement was largely based on the abstract, I only skimmed the paper. The abstract refers to the uptick for very high demand percentiles (>90%), which I guess is still much smaller than the downward trend. I apologise I got this wrong.
I frequently hear people bring up transmission losses as a concern, and genuinely curious where this idea comes from? Was this taught in schools or part of some disinformation campaign?
My understanding is that it is "simple" resistance heating of the transmission lines (P = I^2 R). Which is why high voltage lines are good ideas ( V = I R ) -- they lower the current.
Heating can be quite significant. I guess whether or not it’s economically significant depends on your cost of generation. There was a mega-outage which cut off Italy when overheated, sagging cables, struck the treetops in the middle of either another random outage or some scheduled maintenance. It’s been many years since I read the full report on the incident (which was excellent) but there was some great data in there about degrees of heating vs degree of sagging.
One piece of this really jumped out at me: the projection of overall energy demand to shrink from 2900 TWh to 900 TWh over the next 27 years. The article waives that away by pointing to efficiency gains from electrification and decarbonisation — but that’s just a stupendous change in consumption over a quite short period of time.
I would honestly like a deeper explanation of how electrification will produce such a wild decrease! That’s shrinking energy use by more than 2/3, and presumably after taking into account population/industry growth…? Or are the authors just wildly pessimistic (not…unmerited) about Britain’s trajectory over the next quarter-century? What am I missing here?
This is final energy demand and thus presumably includes energy used for heating and transportation that's currently provided via fossil fuels. In fact, that's the point: electrical motors and heat pumps are more efficient, and the final energy demand is reduced. In other words, demand for electricity goes up, total energy demand goes down.
The ballpark number for EVs is that they are 4x more efficient than ICE cars AIUI.
Harder to ball park heat pumps because it depends on the climate, but 4x is probably a reasonable guesstimate. At worse it's equal to burning things for heating (when it's too cold outside to use heat pumps, not sure that really happens in Britain), at best it's... some ridiculous factor better (when it's practically the same temperature outside and inside).
Correct, a lot of Mackay's estimates fail to account for the difference in delivered efficiency between electric solutions and their gas counterparts. For instance in his car section[1], he estimate you'd need to produce 40kwh to drive 50km. That may be close to the true energetic content of the gas burned, but you could drive that far on just 10kwh in a modern EV, meaning electrification dropped your gross energy needs by 75%.
> You’ve shown that electric cars are more energy-efficient than
fossil cars. But are they better if our objective is to reduce CO2
emissions, and the electricity is still generated by fossil power-
stations?
> This is quite an easy calculation to do. Assume the electric vehicle’s
energy cost is 20 kWh(e) per 100 km. (I think 15 kWh(e) per 100 km is perfectly
possible, but let’s play sceptical in this calculation.) If grid electricity
has a carbon footprint of 500 g per kWh(e) then the effective emissions of
this vehicle are 100 g CO2 per km, which is as good as the best fossil cars
(figure 20.9). So I conclude that switching to electric cars is already a good
idea, even before we green our electricity supply.
Marginally and variable costs it does though. A 100% capacity charging station has few variable costs compared to gasoline delivery; the grid interconnect and charging cables are equivalent to filler and tank safety inspection, not daily or weekly deliveries of fuel by pipeline + truck.
But we’re already maintaining the grid. Growing it to handle the new workload would require a fraction greater maintenance workload which would be tiny in comparison to maintaining the entire parallel infrastructure needed to get oil from a Saudi Arabia desert to a car filling up at a gas station in the middle of Wichita.
Refining oil is a power-intensive process, requiring 5 kWH of energy per gallon of gas. This is why talks about EVs requiring so much more electrical power generation are vastly overblown - they don't account for the power consumed to pump, refine, transport, and distribute the petro products needed for ICE vehicles.
I'm sure he was. It's easy for us to see 15 years later that the combination of EVs + solar/wind has a huge efficiency gain because you can get to avoid internal combustion altogether, but EVs were nascent enough in 2008 for that to not be as clear, so those assumptions didn't make it into his calculations.
This is a bit of an apples-to-oranges comparison. If you consider the energetic content of the gas burned in a car, then you need to measure the same thing for the EV - large power plants are a bit more efficient than car engines, but unavoidably far from 100%; in order to get 10kwh of electricity to an EV the current grid would burn gas or coal with an energetic content of probably 25kwh or so.
So electrification alone can't bring so large reductions in total energy consumption - it is simply one of the prerequisites for that, and it also needs to be accompanied with large scale build-up of non-fossil electricity, of renewables or nuclear.
But when you're using solar to generate the energy, what's the equivalent of burning? I'd argue that it's just "whatever the panel produces". So, if you have some panels producing 1 KWh, and you need that KWh for your car, the relevant efficiency here is transport efficiency.
Since we currently aren't covering the existing demand for electricity with reneweables, the "marginal kwh" of electrifying some car comes from fossil fuels, not solar.
Those are two separate things - electrifying some transportation need gets us X benefit, and converting some electric generation from fossil fuels to solar gets Y benefit, but the former doesn't automagically create the latter.
1) There is no reason to attribute all of the carbon intensity to the last use to be added. It is far more accurate to estimate the carbon intensity of something by the blended production mix at its time of use. Fossil fuel plants are often kept online for factors beyond the need for additional capacity at a certain time.
2) You are assuming there is no curtailment or rejection of wind, solar, and nuclear, which is categorically false. This is why there are time-of-day rates that vary so drastically. There are large amounts of renewable/low-carbon energy that would otherwise be wasted in pretty much every geography.
> If you consider the energetic content of the gas burned in a car, then you need to measure the same thing for the EV
True, but the resulting numerical "hit" for EVs is a lot smaller than for ICE vehicles.
This is largely because
1.) big stationary heat engines are much more efficient than small weight-optimized mobile ones, and
2.) the electricity grid also has some renewable energy mixed in, so a growing fraction of the electricity supply doesn't even need an (energetically imperfect) heat engine.
Mackay's book, and this discussion, are in the context of a hypothetical future where Britain is powered predominantly by solar and wind. It's not even that hypothetical anymore, because as you can see from this dashboard(https://www.energydashboard.co.uk/live) the UK is currently getting 75% of its electricity from renewable and low carbon sources. So even today, driving 50km there only requires 15kwh of gross energy.
That’s not true, because the point of comparison is already the _output_ of the electricity generation in TWh. Once the energy is available in the grid you can put it in your EV at almost 1:1.
Additionally, the parent overlooks that while a lower number is estimated, they compare against a hypothetical 2000+ TWh number instead of the tightest 900 TWh estimate.
If you want to include production of electricity you also need to include refining and distribution of gas. So many exclude that and make calculations as if the gas get to the gas station by magic
Presently running my EV just about 100% on rooftop solar energy, which is possible about 3/5ths of year in my situation. Melbourne, AUS, admittedly, but still.
Fuel delivery still requires a crap load of energy, some of it in the form of electricity - even just running the pumps 24x7 - but there's so many elements in that value chain it could fill volumes. There's far fewer even in the dirtiest of EV charging scenarios.
Probably more significantly: about 30-40% of UK energy use is gas central heating which is 90-95% efficient. Electric heat pumps (beginning to be installed in numbers the U.K.) are 300-400% efficient.
Industrial heating is also a big factor. That too is dependent on gas right now. A lot of that could shift to electrical in the next few decades. And the rest might transition to hydrogen.
Could you cite where you found those numbers? I can't find either of them:
The submission (blog post) doesn't mention the number 900 at all; the paper which the submission is about only mentions 900 in a footnote saying "Total European [energy] supply was 22,900 TWh (IEA, 2022)."; the summary pdf of said paper never mentions 900 or 600.
Figures in the paper ending in 600 occur in a few spots, but nowhere 1600. There is 21'600 TWh/year (total European energy supply, page 8), 16'600 TWh/year (idem), 10'600 TWh/year (prior studies' estimates of UK wind resources), and 2'600 km² (land occupied by buildings).
But maybe I shouldn't be drawing conclusions based on trying to search character sequences in a semi-picture format...
> They estimate that it could produce 2,895 TWh of electricity each year from solar and wind. That’s almost double its estimate for final energy demand in 2050. See the chart below. [...] We can see this when we look at other estimates of energy demand from the literature. The National Grid FES projects that Britain will need just 900 TWh in 2050. [...] Total final energy demand today is 1599 TWh.
It's right under the first main heading. Just above and below the first picture. Searching for 900 in the post took me right to it. I have no idea how you missed it.
Electrification doesn't cause such a decrease. Fixing your projection so it's not anymore a "we can't do anything, we must keep investing in BP" piece into a realistic one is what reduces it.
Honestly, anybody claiming in 2008 that PVs are too expensive so we should not invest on them is safe to ignore.
That's close to the time frame when I realized solar was going to win just based on pure business accounting.
Thing to consider the ultimate price of a manufactured good tends to track production volume, device complexity, and energy required to produce including raw material.
Solar panels require complex machines to produce but are themselves simple, the volume is high at scale, and energy requirements are low. That points to something where the price is close to the energy and material costs.
They can surely become low, but currently they are not. The EROEI of PV panels is barely on the region where it stops being one of the largest factors in its cost.
But yeah, PV has space to improve a by a few orders of magnitude more.
In the 15 years since, public support for PV flipped from con to pro, and we've learned that much less land must be allocated to just PV.
As noted in the OC, Mackay's contribution was methodological. It easily accommodates different assumptions and updated numbers. O'Callaghan et al did just that. Yielding a new conclusion.
I could believe some reduction if there were massive efficiency increases (e.g. replacing resistive heating with heat pumps) but that seems impossible on that scale unless they’re also forecasting entire industries leaving the country.
Take a look at the amount of rejected energy we waste by using for example heat engines today. Electrical engines and heat pumps vastly reduce those losses.
IT looks like the 900 number is discussed in chapter 3 of this document [1].
As far as I can tell the current usage of 1200 TWh include electricity and combustible chimerical energy of gas. Electricity use is ~300 Twh, and Gas usage is ~800 Twh.
The proposal is that gas heating would be entirely replaced with heat pumps and most gas generation would be replaced with modular nuclear reactors and offshore wind.
The numbers are a little misleading because of the way gas and electricity are summed to get the top level numbers. A TWh of gas consumption is not the same as a TWh of electricity consumption. In thier model, 50 TWh of electricity can replace 400 TWh of gas. The challenge with this approach is that it is not show what is going on with user consumption. Are they getting more, less, or the same thermodynamic work done?
ICE vehicles turn less than 35% of their energy consumption into productive work. The rest is waste heat. That's the main reason transitioning to pure EVs must happen.
That's an inherent limitation of heat engines - if you burn the same thing in a power plant, then you could get 45% instead of 35%, but you can't possibly convert, say, 80% of burned fossil fuel into productive work.
> if you burn the same thing in a power plant, then you could get 45% instead of 35%, but you can't possibly convert, say, 80% of burned fossil fuel into productive work.
Not 80%, but combined cycle gas turbines regularly achieve up to 60% steady state efficiency.
There is research into ammonia fueled turbines that supposedly achieve even higher efficiency.
I really don’t see the point of that calculation. It’s like saying that the wind turbine only captures 5% (number made up) of the energy of the wind going through it.
Then you didn't read the article. The whole point is can the UK meet all of its energy needs with solar and wind. That includes transportation energy. The comment I was responding to was wondering how to reduce the energy needs. A big part of it is by drastically increasing the efficiency of transportation by moving from ICE to BEV.
But the change is not a 1:1 linear ratio. Demand increases, but only if there is pain that more could use. Once a room is bright enough you won't add more light.
> Exactly, but the improved efficiency is more than the loss from leaving it on.
The improved efficiency needs to be more than the drop in marginal cost which is what governs demand, nothing to do with how much more electricity you use today. And even then, that relationship isn't linear because a 30% reduction in cost can drive a 60% increase in demand because that reduction puts it in a new price bracket where a lot more people can afford it (since wealth is non-linear). This stuff is super non trivial and has all sorts of higher order effects.
There is no such projection, it seems you misread something.
The O’Callaghan et al. paper (in the Blog post here the related figure is fig 1) says that current (2023) demand is 1500TWh and current supply is 2885 TWh. There are different projections for total demand in 2050 (note all of them project a reduction of demand, due to efficiency gains), one of these is the national grid FES which projects 900 TWh. Importantly the O'Callaghan paper opts to be conservative and choose to use the current demand as the demand for 2050. This is conservative, because it is higher than all projections which all assume that we get demand reduction from efficiency gains.
I want to add that Britain's energy demand has already fallen by one quarter in the last 15 years even without the efficiency gains from large scale electrification, so a 2/3 reduction (which to stress again is not what the paper assumes) is not so outrageous.
Some of that is due to efficiency gains (good). But some is also due to deindustrialization and increased imports of energy intensive products (bad). We have to look at full lifecycle global CO2 emissions in order to perform a valid analysis of any changes.
Depends what you care about. MacKay's claim was specifically that the UK couldn't be powered this way (which appears to be falsified); if the manufacturing all ends up in China, which has the Gobi desert and is closer to the equator, and has the Three Gorges Dam and the will to build it, I don't even have to run the numbers to know they can be completely green.
China isn't even remotely "green". Most of the official statistics coming out of there are highly manipulated. In particular they haven't built even a fraction of the storage capacity that would be needed to keep their factories running when the sun isn't shining, nor are they likely to do so in the next few decades.
Lots of things could happen in the sense that they don't violate the laws of physics. But your claim about China isn't grounded in any sort of economic or political reality. The demographic issues alone make a total green energy transformation impossible.
Because that level of industrial transformation requires huge labor and capital resources that China simply doesn't have. The CCP cadres have been reporting exaggerated numbers up the chain for years in order to keep their jobs so any metrics coming out of the central government are basically junk.
We know they've got the capacity for the required level of transformation just by looking at the things they export to us, along with satellite pictures[0] of their growing power sources (and indeed all other infrastructure) of all types.
[0] Well, remote sensing in general. You can argue if a map at specific spectral lines to detect industrial byproducts is a picture or a graph.
This is the crux of it. Yes, we've reduced demand from electrical appliances, an made more efficient cars. But we've also closed all of our steel mills, aluminium smelters and other heavy industry. We still use those products, but the energy intensive processes are done elsewhere.
The UK "industrial" energy consumption today is 30% of what it was in 1970.
In the future post-fossil fuel, solar powered age, that was going to happen anyway.
This is a long term problem Europe faces: with solar projected to dominate world energy supply (if historical experience rates continue), Europe will be an energy ghetto. Energy intensive industry will move elsewhere. Going nuclear cannot save them, since their nuclear will be competing against world best case solar in sunnier places.
The better question is why do they think more efficiency will lead to lower use rather than higher consumption? I know, if the price stays the same per unit of energy and i stop spendjng so much because of efficiency gains, i have a ton of other stuff i would love to do that consumes energy so my decrease in usage would be tiny. What will decrease my usage is price increases, which can only go so far as taxes to capture proposed externalities before i revolt and elect someone who will axe the tax (as is about to happen in canada).
Thats not really what that chart looks like to me, or at least the effect is very moderate. The big drops correspond to economic shocks. Even if i ignore that the chart is topping at roughly 350 in 1975 and bottoming around 310 in what i assume is 2023, which makes it what? A 0.22% annual decrease?
You are making the assumption there won’t be some new technology that gobbles up significant energy. At one point an American’s electrical consumption might have been home lighting and a radio. Then it jumped by an order of magnitude with the refrigerator and air conditioner.
Most of the examples I can think of with that would not be localized.
For example, if we need massively more computing for an AI breakthrough, there's nothing requiring that to be localized. Similarly for production of any commodity that can be shipped.
Energy costs are so high in the modern age relative to how dirt cheap it was pre-OPEC crisis, that any sort of significant energy consumption would be a barrier to significant adoption. Your average Joe cannot afford to mine Bitcoin, for example.
If anything we have mostly seen massive improvements in energy efficiency in the last two decades, with appliances, the switch to LED lighting from incandescent, heat pumps vs other heating/air technology, etc. What new technologies have shown up get quickly innovated in the name of energy efficiency.
As an aside re: Canadian politics. New Governments don't get elected, old ones are voted out of office. The shelf life is generally 10-15 years and this government has gotten stale. It just so happens the party in position to replace them plans to remove the carbon tax (which hasn't been implemented in any meaningful way) as part of their platform
Thats true but a misinterpretation of the data in my opinion. Old ones are voted out because their policies mismatch the current pain points, which is definitely happening here. Libs were elected on a luxury beliefs platform that isn't compatible with the current state of the economy/interest rate regime/inflation. Enough people have finally realized this to grow disillusioned and change voting patterns.that the libs appear to be a bunch of kleptos and/or incompetent and/or corrupt is speeding up that realization.
If the libs had kept the klepto stuff to a less obvious level and adapted the binding parts of their luxury beliefs platform they probably would still be polling a majority and could have kept this going for quite a while longer.
> I know, if the price stays the same per unit of energy and i stop spendjng so much because of efficiency gains, i have a ton of other stuff i would love to do that consumes energy so my decrease in usage would be tiny.
Do you, though? What? I'm not sure how typical this is. Like, if you replace your creaky old gas central heating with proper insulation and a ground to air heat pump, you're going to be saving a _lot_ of energy if it's a big house. What would you use this extra energy 'budget' on?
That’s got nothing to do with these total energy calculations. When burning gas in a car you’re converting less than 1/5th of the energy in the crude oil to actually move your car. A wind powered EV is closer to 4/5th of the energy from wind ending up moving your car.
Not only is the EV itself more efficient, so is each stage before that more efficient. Something like 1/3 to 1/2 of all CO2 emissions from ICE cars occur before gasoline makes it to the car. Thus, driving an EV lowers the total energy needed without actually reducing how much people are driving.
PS: Total energy calculations are fairly arbitrary because it’s ultimately sunlight powering things, but that’s also true of fossil fuels just at really unbelievably low efficiency so that parts generally ignored.
Please explain. I keep hearing this meme, but the arguments don't make sense. For instance, people will say things like, "New World Order wants us all to use electric vehicles so they can limit our movement!" This is backwards, though, because electric cars can plug in anywhere, but gasoline stations are much more centralized and easy to regulate. If I were the Global Elite, I wouldn't want people producing power on their own rooftops, storing it in batteries, and driving around with it. Green technologies seem to lower costs and increase resilience for the average person. What am I missing?
In the past, customers were encouraged to be very inefficient with their electricity usage. Rather than fit efficient insulation, high consumption was tolerated for years to come.
And thanks to the "Base Load" myth of coal based generators, people were encouraged to use heat banks and storage HWS which loaded the system during the night.
But this could only be done by offering cheap "night rate" electricity.
In the future, our consumption must be focused on day-time peaks, as well as much better insulation, and the use of heat-pumps. And of course a steep increase in electricity tariffs will greatly encourage this.
In the future, companies which operate 24 Hrs without thought to electrical cost, will focus much more on efficiency, and even begin to shut down operations at night.
In many countries (eg Australia) there has already been a huge increase in domestic solar and battery installations.
This change will be driven by unavoidable price rises, and the recognition of our profligate wastage in the past..
Burning fuel to generate electricity creates waste heat that siphons off most of the energy. By the time electricity reaches your outlet, around two-thirds of the original energy has been lost in the process: https://yaleclimateconnections.org/2022/10/energy-loss-is-si...
> the projection of overall energy demand to shrink from 2900 TWh to 900 TWh over the next 27 years
To be clear, today's demand isn't 2900 TWh. It's 1599 TWh. And none of the people claiming Britain can be powered via renewable power are using the 900 TWh figure.
Just noting for people who haven't read the article because this comment wasn't clear to me:
Today's demand: 1599 TWh
UK’s Department for Industrial Strategy forecast demand for 2050: 1250 TWh
UK National Grid FES forecast demand for 2050: 900 TWh
MacKay’s forecast demand for 2050: 2900 TWh (I think this is what this comment is referring to)
MacKay’s 2008 forecast renewable production for 2050: 1500 TWh
O’Callaghan (new, 2023 report) forecast demand for 2050: 1500 TWh
O’Callaghan (new, 2023 report) forecast renewable production for 2050: 2895 TWh
The difference is mostly about pricing of renewables radically dropping since 2008, and the likelihood of floating offshore wind producing large amounts of supply.
Electricity demand has fallen by ~10% over the past decade while we've been introducing EVs and heat pumps and stuff.
These things are all more efficient than the hydrocarbon equivalents, further depending on the measure, if you count the gas input energy in to a gas plant, a renewable generator is much more efficient too.
So I think it's possible, although the figure ends up somewhat misleading without a lot of context.
The answer to the headline question ultimately doesn't matter. Either the answer is "yes", or we figure out how to make the answer "yes", or we all die.
There are other options... nuclear power is more expensive but if cost wasn't the issue (e.g. running out of land was) is an option.
If cost is the issue there are still risky moonshots like "throw tons of money at fusion" (attacking the cost of nuclear) and "throw tons of money at high temperature superconductor research" (attacking the amount of energy we need), and "geoengineering" (risking screwing it up worse). Not guaranteed to work, but you know, better than rolling over and dying. Also has the side benefit that a lot of the moonshots are worth trying anyways.
Hard reliance on nuclear power is one of the "we all die" options. Didn't we learn our lesson about building a whole civilization on consuming a limited resource dug out of the ground?
Is it not the case that Uranium and Thorium are not just more abundant than existing fuels, but also about a million times more energy dense? I find it hard to believe running out is a chance we'll face soon.
More abundant? Depends. There's a lot in the sea, which isn't currently recoverable. There's a lot we could do (but actually don't) with breeder reactors to make more fuel.
This means the answer to the question "how long could we last on just nuclear fuel alone?" varies from 5.7 years[0][1] to 4.3 billion years[1].
Not like the sun is renewable either. Everything is finite if you use enough of it. The problem is that the old "limited consumable resource dug out of the ground" changes our natural habitat as a side effect of creating energy, assuming we continue to use it as we do today.
(By that last bit, I mean: powering EVs from coal plants with carbon capture at exhausts might be different, idk, but probably not cheaper than just not polluting in the first place.)
"Renewable" isn't the best word for wind and solar, better would be "use-it-or-lose-it", either you harvest it or it dissipates into useless low grade heat. Using more of it doesn't deplete the supply of it.
Solar energy (and thus wind assuming otherwise constant environmental conditions) is only going to get more plentiful for next 5 billion years...
The Earth becomes uninhabitable in about a billion years due to solar brightening, but your point certainly applies to far future civilizations in solar orbit.
I think nuclear is a red herring in most of the world (too expensive and SMRs are too far away to matter), but the idea that we’ll run out of fuel for them is a nonsense.
We haven’t looked very hard at all for uranium. Nor have we tried very hard to use the uranium we already have efficiently. Nor have we tried very hard to use thorium as an alternative. Some combination of these would stretch out the supply of uranium we have essentially indefinitely.
We are all going to die! But nuclear power is to expensive, so let's sit in the roads and block traffic and subsidize Teslas for rich people.
Either we are under an existential threat, in which case nuclear power is an amazingly cheap way to save 8+ billion people, or there is some other agenda at play.
The primary constraint limiting the speed at which we move away from fossil fuels is cost. Investing in more expensive alternatives like nuclear increases the risk of the threat becoming existential, and the damage it does if it doesn't become existential, compared to investing in cheaper alternatives like solar and wind.
It is not a case of "we should do everything" because we can't afford to do everything. If we could afford to do everything we could easily do a small subset of everything and solve the problem.
The article is literally about showing that there is in fact enough solar and wind energy available to fulfill expected (not artificially constrained or "deindustrialized") demand.
The article doesn't have a plan for storage, and relies on a plurality of energy from floating offshore wind, which has never been relied on at any scale. There are 4 windfarms in the world generating just 193MW. You are going to bet humanity on that?
Superconductors are probably a red herring. It's technically possible to make a 1 Ω global power grid for only a few hundred billion in raw materials, a superconductor isn't going to help much with anything except the material cost, including but not limited to the cost of actually installing that cabling, the geopolitics of where to put it, conflicts involving it.
A superconducting cable probably also makes attempts to damage the thing easier — to get 40,000 km of aluminium down to 1 Ω, it needs to have a cross section of 1 m^2, which is kinda hard to damage, though also you don't really want a single cable because that, with current global electrical demand and reasonable (i.e. currently in use) choices for the voltage, would be in the order of 1.5 mega-amperes and match Earth's geomagnetic field at a distance of about 11 km.
My understanding is that that if they have nice properties they'll enable substantially more efficient motors/generators thanks to strong magnetic fields. I can't say I'm that confident in that knowledge though.
No, it is not very true - even if you look at the worst case "do nothing" scenarios in the IPCC report, they are very bad, but very, very far from "we all die".
Using such hyperbole is counterproductive and makes people discount the actual risks. If we're talking about, for example, a billion people being displaced due to climate change disrupting food production or flooding areas which are currently densely populated, and many millions dying in that process, that is bad enough by itself to justify taking action, and there is no need to resort to ridiculous unjustified exaggerations implying that we're all going to die.
Even if the answer were "we'll still need 50% from fossil fuels" it still cuts CO2 in half. If the answer were "we still need 10% for unusual cases", it still means that fossil fuels stop being the problem, at least for that one country.
This isn't insurmountable. It doesn't have to be perfect. Even a real but incomplete effort makes a genuine difference.
The problem, unfortunately, remains the US, who has a large and powerful minority who is convinced it's all a hoax. The solution doesn't have to be complete but it does have to be something. With many millions of people actively making it worse, even 100% in the UK doesn't come anywhere near close.
That's what I thought was so great about this article. Even if you quadrupled the estimate of total rooftop solar, it is still just a drop in the bucket compared to the total energy generation needed. It's almost not worth talking about because it's such a minor amount of capacity.
I didn't read the paper, but from personal experience:
1. Rooftop solar depends on lots of individuals to make an investment to put solar on their houses. If you say "let the government subsidize it", that doesn't really make much sense because the government can get a much better return on each dollar spent by investing in more efficient technologies.
2. Many roofs are not suitable for solar. They are either facing the wrong way, at a bad angle, shaded, or, in my case, too "origami-like". Solar panels need to have minimum clearance fro ridge lines on a roof, which can drastically reduce the total coverable area.
1. Most residences not being suitable for significant generation.
2. The return on investment is decades long (yes, so is double glazing, but that is useful for far more households), not that most households can afford it at all.
3. The total generation capacity is piddle, especially for government investment.
4. The materials that go into making PV panels are horrible to extract.
I'm not saying PV panels are useless, but they are not much of anything. Not something my lecturers at uni liked hearing/reading, but lo and behold pretty much nothing substantial has changed since. They're too busy blowing Sustainable Development smoke up their own arses though.
rooftop power is basically the most costly and least efficient mode of generating power. It many places, it easily costs 10x for the same nameplate capacity, and because of suboptimal locations and angles, it only produces a fraction of that nameplate.
Consumer rooftop is about the same price range as nuclear.
Commercial rooftop (warehouses, big box stores etc) is a bit cheaper than nuclear, with the most expensive being the same cost as an average nuclear plant, and the low end for new commercial rooftop solar being the same as the running costs of already built nuclear:
It still shows rooftop clearly to clearly have the worst LCOE. I bet the assumptions for Rooftop dont include suboptimal builds, like builds in San Francisco, on a north facing roof, under a tree, with storage.
Rooftop can be good in niches, but it is hardly a panacea.
Your point is false. Rooftop PV saves (if your roof is suitable) significant amounts on a domestic energy bill even in the UK. I know this from having lived in two houses in the UK that had them, and the calculations that went into deciding to install the first of those, and that panels/inverters etc. have become much cheaper since then. Also, last time I was living in the UK, the grid cost looked like it was billed as a separate line item to the per-unit (kWh) cost, so this really is about the domestic price itself and not just funky billing hiding the cost of the grid like you wrote in another comment: https://news.ycombinator.com/item?id=38158442
The grid scale alternatives — solar farms, wind farms, hydroelectric — can be better or worse in different ways and different times of year, but normal people can't put full size ones on their own property and such wind and hydro scale non-linearly anyway so small ones aren't as cost effective[0].
Furthermore, the reason for the link under which this is being discussed, is that the UK has land area constraints. Suitable areas in the UK for grid scale PV are basically all farms and/or designed National Parks or AONBs and/or designated green belt, and the exceptions (like disused military bases and airfields) have people clamouring over them to build new towns. Sometimes people can get grid scale PV past planning permission, but it's hard work and upsets people with power who want the countryside to look like farmland. Rooftop PV circumvents that.
[0] Except possibly geothermal/ground source heat pump. Those still provide cost effective benefit when you can do them, but the wide price range means I'm not sure which is better.
Rooftop power (or other local generation) has a big economic incentive behind it, which is that the energy it generated essentially goes for residential rates as opposed to wholesale rates (which is at least in part actually born out by a reduction in the utilization of the grid). So I expect it to continue to happen, despite it being less efficient in theory than grid-scale. Local battery storage has a similar incentive, but even more accessible (at this point, a home battery system is likely a better house upgrade investment than solar).
Selling rooftop energy for residential rates isn't economically sustainable, and has already been eliminated in California unless you are grandfathered in.
This is because only a small fraction of the residential rates goes to production costs, and the rest go to distribution infrastructure and operations. Distributions and infrastructure costs/ kWh go with more residential production, not down.
Think of it this way. With commercial power you might pay 0.10/kWh production and and 0.30wh distribution. You can make your own rooftop for 0.35/kw, but the grid still costs the same or more, so that gets added to your bill.
Residential rates for rooftop solar only ever made sense as a huge subsidy for early adopters.
I don't mean net metering, I mean if you use 1kWh from your solar panels that's 1kWh you aren't buying from the grid. That's the big advantage which doesn't depend on subsidies.
The key point is that the difference between "wholesale electricity" and "retail electricity" is based on total costs which don't change if you buy 1kWh less from the grid - while it's currently priced "per kWh", it's not really a marginal cost but a fixed one, so people in aggregated would have to pay the same total amount to the grid no matter how much home solar panels are built.
There is a cost advantage currently, but that is a side-effect of how we currently do accounting of grid costs, not reflecting any actual advantage in reducing total costs across the country.
Exactly. Lets say you and your neighbors each pay $100. You put in solar for $90, and save $10. Because most of the power grid costs are fixed, your neighbor is now has to pay $200 to pick up the slack.
This is why California is now considering grid connection taxes, which would charge homeowners for electricity they generate at home.
I thought this was a really excellent post. The thing I liked best about it was how it presented the different set of numbers and didn't try to say "one is wrong or one is right", but instead tried to explain where the numbers came from, and what the outcome would be if some assumptions were wrong so the reader can do their own analysis. What I found particularly helpful:
1. Explaining the difference between MacKay's original "technically possible" vs. "practically possible" supply numbers. I agree with the article, the world has changed a ton since 2008 and I do think much more of that technical possibility is now practical due to changes in tech and attitude.
2. One thing I was cautious about is that the lion's share of final 2050 supply in the updated numbers comes from floating offshore wind, which in my understanding is the least technologically "ready" solution. Can someone with more knowledge comment on this? Is floating wind really as "production ready" as would be needed to match these numbers?
Floating wind is production ready and in production in the UK. The flagship example was Hywind in Scotland which has been in production since 2017.
There are several other offshore projects using floating wind, more being converted and a lot more planned for conversion or development. There is quite a lot of investment and encouragement from government. See https://www.great.gov.uk/international/content/investment/se... for a reasonable overview.
Floating wind is very low in novelty, there is just a cost trade-off depending on water depth. If your water is too deep then structures become too expensive and floating is cheaper in deep water. There is some room for improvement in floating which will get costs down, but it is totally fine to just deploy straight away.
But if you can install offshore wind on small structures like <30 metres then get that done first, it's cheaper.
We already have plenty of words and phrases that mean 'large' or 'most', so it's a shame this phrase is losing its specific meaning and joining with myriad interchangeable synonyms that mean something more pedestrian.
> To be clear: this does not mean that this is the ‘optimal’ electricity mix in 2050. Not least because energy storage costs would be very high. We would probably want to diversify a bit, not least to help with grid balancing. Before all of the nuclear fans get mad: I think there’s room for nuclear in there too.
> But the point still stands: it seems we have a lot of untapped solar and wind resources and they could make up a large chunk of our grid, even if they’re not 100% of it.
There's one component that could not have been a part of the 2008 analysis: batteries.
Global annual manufacturing capacity is currently enough to produce 50min worth of storage for the whole world (as a fraction of annual electricity production).
That's not a lot and it's not utilized fully, but still well within the capabilities required to shave off the evening and morning peaks - assuming batteries last more than 5 years, which is a conservative estimate.
Nuclear would have been a great component here, but IIRC Hinkley Point C is still under construction and will remain so until 2027.
For stuff like this I assume batteries will have 50% capacity after 1000 cycles even though this always gets surprised responses insisting they're good for far more cycles, as I think lowballing is a safer assumption. You can always just multiply as appropriate for different assumptions about cycles.
However, that last 50% is still useful in this context, so that's half an hour every day going down to 15 minutes/day after 3 years; assuming it's exponential decay, I think that's a steady state equivalent to 4.3-ish times the production rate? Not sure though[0].
Same graph also forecasts 6.79 TWh/y manufacture in 2030, which is ~2 hours each day instantaneous or 8.6 hours steady-state if my previous assumptions were correct. At that scale, I suspect you also need to start accounting for the impact from variable day lengths meaning there won't be demand to fully discharge the batteries 365 days of the year, and also that demand is lower at night while all the demand numbers I've given assume constant use.
A realistic estimate also needs to account for the impact of nighttime renewables getting used preferentially, further lowering the overall frequency with which the batteries need to go through a complete cycle, and conversely increasing battery demand to account for how well the grid they're connected to can compensate for renewable-unfriendly weather, but that's not the topic of this sub-thread.
Also also, I expect these batteries will probably mostly get used by cars.
Finally, I think recycling is a red herring because I can't believe that getting metals out of batteries is harder than getting the same metals out of literal rocks in the ground.
[0] integral from 0 to infinity of 2^(-x/3) for a 3 year capacity halving time?
A common issue with all predictions from a few years ago is that they failed to predict the 90% fall in cost of solar panels. We see a similar thing today when battery storage is dismissed as being too expensive.
Unfortunately, current policies are often based on predictions from a few years ago.
Battery storage has excellent long term potential but costs are falling much more slowly than with solar panels. There are some significant constraints on raw materials supplies. Those can eventually be worked through but it will take longer.
You may be right of course, but I believe the reason for the slow drop is mostly because most of the battery factories currently under production are not online yet and EV makers are still grabbing all the batteries they can get. VW just started building one of their planned 6 last year.
Once the transition to EV's is mostly done and car sales fall to normal levels there will be an immense surplus of battery production capacity.
My understanding is also that LFP batteries have basically eliminated the raw materials bottleneck, which was another development that few analysts (or anyone else) were able to predict just a few years ago.
>> My understanding is also that LFP batteries have basically eliminated the raw materials bottleneck, <<
the casuality between the raw material supply bottleneck and LFP would be difficult to demonstrate. LFP has lower energy density and would have required more lithium to get equivalent KWh than what is normally required by high-energy, high-nickel batteries (eg, NCM/A); therefore higher lithium demand, but ...
The price of lithium, and other battery raw materials, however, plunged by 75% since last Novemember peak (still falling) -- it's widely understood that the drop is largely caused by slower than expected EV demand (and oversupply).
>> but I believe the reason for the slow drop is mostly because most of the battery factories currently under production are not online yet... <<
GM forecasted earlier this year that their (Ultium) cell cost would drop to $87/KWh by 2025; just last week, GM reported a 40% drop in their cell cost YoY. Sure, there is 400+GWh new US battery production coming up in 2025-2026 and the commodification/mass-production of made batteries would accelerate further price drop.
>> There are some significant constraints on raw materials supplies. <<
which was true for all kinds of stuff from construction materials to EVs last year -- the prices of battery raw materials, eg, lithium, nickel and cobalt, plunged by as much as 75% since last November peak.
Battery economics have really changed in the last 5 years without a lot of notice. California has installed 5GW worth in the last 4 years. Max demand in California is about 50GW. What's driving batteries is the spread between price per MHW at 12am and 6PM.
When it comes to storage there are also a lot of other technologies that may be viable if solar/wind are built out to the point where they are particularly over-subscribed (meaning that on a particularly windy/sunny day they provide much, much more than 100% of demand).
For example, using things like clean hydrogen or Power-to-methane processes that can create gas to be used in existing peaker plants.
Curtailment already happens routinely today for solar PV in some countries (eg Spain) - though it's due to running out of grid capacity in the right places, not oversupply on the grid as a whole. Economically for PV farms which are being curtailed building colocated hydrogen storage makes some sense.
If there's no curtailment (and if you have sites available in the country) then pumped storage makes a lot more sense.
But either way, building sufficient storage to do anything other than peak-shave is very, very, very expensive, so we'd be better off with nuclear even at a 2-3x cost overrun.
Pumped storage and hydrogen are, paired with wind or solar, much, much cheaper than nuclear power.
We'd even be better off if all solar and wind power were converted into hydrogen first and only then turned into electricity. Not because thats cheap but because nuclear power is that expensive.
Nuclear power only becomes cost competitive when you use it to share costs with the nuclear-military industrial complex - which is why it is built.
For intraday variation, yes I agree - intermittent renewables+storage works well. But for longer term variability (eg wind droughts) or even worse, if you wanted to cover seasonal variation in production (solar PV can have enormous seasonal variability in northerly latitudes), the amount of storage you need in MWh is absolutely enormous. This is why nuclear still winds up being much more sensible.
Nuclear power doesnt store energy. It will pump out 2GW all the time whether you need 6GW or you need 0GW. It doesn't really help in that respect at all because it can't be scaled up and down. It's not a peaker.
Seasonal variation is best dealt with by creating and storing hydrogen underground. It's easy to store lots of MWhs this way. This isnt as efficient as a battery or pumped storage when roundtripped (~50% not 90%) but if seasonal variation is 4-7% of your power then that just means you'll have to overproduce in aggregate by 10-20% to accommodate this. That's not really a problem given how dirt cheap solar and wind are - about that much is being lost to curtailment already.
The funny thing is, even if all electricity generated by solar and wind were stored as gas before being used it would still work out a bit cheaper than nuclear power. Nuclear is that expensive.
Seasonal variation in PV production is much much more than 4-7% in northern Europe. In the UK daylight hours vary between 8hrs in winter and 16.5hrs in summer (more in Scotland), and on top of that winter has many more overcast days meaning you get a lot less irradiation during those hours. As a result, total irradiation in December is many multiples lower than that in June.
Given that, it's not (currently) economically viable to build storage+PV on anything approaching the scale we would need to replace baseload generation completely.
>Seasonal variation in PV production is much much more than 4-7% in northern Europe.
4-7% isn't seasonal variation in PV production. It's the amount of grid energy that needs to be supplied under a hypothetical 100% solar/wind/pumped storage grid when:
* Solar output is low
* and wind output is low
* and ~8 hours of pumped storage is depleted.
Obviously the wind blows at night and pumped storage works at night and electricity demand is lower at night. Obviously seasonal storage would not have to be tapped every night.
>Given that, it's not (currently) economically viable to build storage+PV
Why would that be given? It's categorically false. The opposite is true. Maybe what you meant to say was that "GIVEN 2GW of solar energy was being produced today" AND "coire glas is under construction" because it's highly economically viable, maybe we should lend a hand in the form of subsidies in scaling that up?
Or did you want to push those subsidies to a form of energy with a 5x higher LCOE?
I'm not saying that Solar PV isn't economically viable at all (I've invested in Solar PV projects for my day job in fact, and obviously wouldn't have done so if I didn't think this).
My concern is the "end game", where all current baseload generation gets disconnected from the grid (including current nuclear when it reaches end of life). Solar PV's current low LCOE doesn't help you a lot in that case because in order to actually deliver the energy needed at the right time you will need to do some combination of (1) build a lot more PV than "needed" and routinely curtail it significantly in summer (ie significantly derate nameplate capacity) and (2) build a lot more storage to deal with seasonal and other variations.
This is when nuclear starts to look like it makes sense as part of the energy mix in my view.
>Nuclear power doesnt store energy. It will pump out 2GW all the time whether you need 6GW or you need 0GW. It doesn't really help in that respect at all because it can't be scaled up and down. It's not a peaker.
i.e. nuclear power needs peakers or storage too. not as much as solar and wind but it still needs to be paired with peakers or storage.
>build a lot more storage to deal with seasonal and other variations.
This is what makes the most economic sense currently because pumped storage + hydrogen storage + solar + wind <<< nuclear power costs.
Even IF you assumed nuclear power plants could be peakers (they can't) or that Fukushima/Chernobyl events had a probability of 0% (they dont), it would still not make any economic sense to build thrm when those 4 technologies can handle everything together at a much cheaper cost and can be built in 1/2 - 1/20th the amount of time.
> But either way, building sufficient storage to do anything other than peak-shave is very, very, very expensive, so we'd be better off with nuclear even at a 2-3x cost overrun.
This is only true if you imagine batteries as the only storage technology. But complementary storage technologies, like hydrogen, allow 100% RE to be achieved at a cost that will likely be lower than what nuclear would allow.
A much more common issue is that many people forget that neither wind nor solar provide baseload and even with the cheapest batteries on the planet, we won’t even close to buffering a national grid’s demand for days.
Having said that, any future electricity mix must include nuclear power. France proves that only nuclear allows for deep decarbonization unless you have lots of hydropower and a small population like Norway.
For people who still think that wind and solar alone can do it, I just recommend looking at Germany which had to resort to restart even communist East Germany era coal plants to stabilize the grid after the nuclear phase out.
I think people are letting perfect be the enemy of good, sure they output carbon, but they seem like a much more affordable option for buffering dips in renewables output, compared to batteries or nuclear.
That's a good problem to have, though. Inefficiencies from following "the wrong path" just mean the end solution isn't quite as good as it might have been, but if it was good enough to solve the problem in the first place then anything you can recover by switching tack is almost pure upside.
I love this kind of hypothetical analysis, but I'd also like to point out there are already people making real strides towards this on an individual level in the UK. Youtube channels like ElectricVehicleMan catalog what a conscientious person can practically accomplish by themselves with readily available solutions today. I've particularly been impressed by the synergies between rooftop solar, battery storage, and combined heat-pump hydronic heat/hot water solutions.
It seems plausible that even a person in a typical row house could offset most of their household consumption with solutions that will end up with a reasonable return over time.
The book provides an excellent overview of how different forms of energy production and consumption add up and which energy solutions could make a real impact. I strongly recommend reading it as context for these updated numbers.
I haven't read the book, but my impression from reading this Blog post is very poor.
He seems to have written a book dismissing renewables by assuming fixed technology/costs from 2008 for projection into the future (if everybody did this companies would not invest in anything). On top of that he even hand waved the rest away by saying their installations would not be accepted by the public. That seems to me that he was set out to dismiss solar and wind and just looked for numbers to confirm this.
Generally I believe if you want to show the feasibility of a technology you should be conservative in your estimates, and if you want to dismiss it you should be optimistic in your estimates. Ideally you show both conservative and optimistic projections.
> He seems to have written a book dismissing renewables by assuming fixed technology/costs from 2008 for projection into the future
Your impression is incorrect.
As the blog post notes, the problem is that two changes in particular - the reduction in demand, and the absolutely crazy drop in the cost of PV - were far larger than anything mckay was able to reasonable foresee in 2008.
> That seems to me that he was set out to dismiss solar and wind and just looked for numbers to confirm this.
Ok thanks for the correction, as I said the impression was really only based on my interpretation of the numbers in the Blog post. I will have a look at the book then.
Some discussion here about storage, but I am interested in modeling around dynamic load shedding / smart grid / “virtual batteries”.
Seems to me that as the energy mix moves more towards renewable, to the extent that the renewable-skeptics’ prediction that variability is an issue comes true, then we would have to build gas peaker plants and start charging more for electricity at peak times. In response to this increased market rate delta it would become more viable to invest in dynamic pricing and load shedding/deferring tech.
So there is a modeling exercise which looks at the peak time price premium for various levels of increase in peaker plants required as the input variable, and compares that to the viability of virtual batteries at those price deltas as the output. I haven’t seen anything along these lines.
I wonder why oversized hot water tanks don't get more play here. You can heat up the water with excess electricity at basically any time during the day and it should stay hot for about 24 hours if it is well insulated.
In New Zealand they were remotely controlling residential customer hot water tanks and heaters with energy stored in hot oil using a “ripple” signal on the power lines … probably in the eighties by the age of the equipment I saw.
Because people want their hot water to be consistent. Even just a simple timer that kicks on the hot water in the morning would work, but most people want to have their sink still be hot even when they aren't showering.
> it should stay hot for about 24 hours if it is well insulated.
It takes 50 kWh to heat a 46 gallon tank up to 140 F. That's a ton of energy. Hotter, larger tanks lose even more energy.
Instead, get a tankless heater, backing a small heat pump water tank. You get water as hot as you can possibly want, heated whenever you want, and it never goes cold.
The whole point of over-sized water tanks isn't to be the most efficient, only to supply a lot of demand when at times of the day when there is surplus energy. Right now, solar and wind are being curtailed more and more.
Also, water consistency isn't that big a problem. Today's 120v heat pump water heaters store water at higher temperatures and using a mixing valve to deliver the desired temperature of water. I am just saying, surely electric water tanks, mixing valves, and temperature sensors are orders of magnitude cheaper than the amount of batteries needed to heat an equivilant amount of water.
Tanks can be insulated and larger tanks have less surface area per volume so keep heat longer.
My parents 1988 switched their hot water to a plan where it was all heated at night. 6 people making no effort to conserve water ran out twice in all those years. Yes the.water was hotter in the morning. But the last shower of the day still needed to be mixed with some cold or it was too hot.
Tankless is worse than a large tank. Tankless needs a lot more energy now, less over the full day, but when you turn the water on it needs a lot now. A tank easily adjusts to use power when the power company needs it. Sun shining or wind blowing, then heat water to use up whatever is extra. Clouds and no wind, just use the stored energy.
Maybe a battery is more efficient, but tanks are cheap
I think you missed that the proposal was a tankless heater backed by an air-source electric heat pump heated tank. The idea is that the heat pump may, or may not, get the water to the desired temperature; the on-demand heater just makes up the difference.
You don't need that backup. Fix the problem, which is either you don't have enough storage or you don't have enough generation. As I tried to make clear people don't need much storage.
The variability of wind/solar can also be handled without time of use pricing by the grid operator including terms in the energy purchase or interconnection agreement requiring the generating facility to maintain some degree of consistency to the output. Then it is up to the generator to figure out how to do that, perhaps with batteries.
The problem we saw in Ontario when the market was introduced in 2001 or so was that it was politically unfavorable to have the extremely high prices, even for an hour let alone long term, that would encourage an investor to build a storage facility.
I think ideally you set up the market so that supply or demand can move to meet these fluctuations. Agree it’s politically sensitive, but if you artificially flatten the price curve then you remove consumer incentives to participate in solving the problem.
I suspect industrial/commercial power usage is the big area for innovation here, but would love to see a breakdown of where the low hanging fruit is.
> Seems to me that as the energy mix moves more towards renewable, to the extent that the renewable-skeptics’ prediction that variability is an issue comes true, then we would have to build gas peaker plants and start charging more for electricity at peak times.
There are a lot of different types of variability. You're talking about sudden short-term variance in supply, which is easy to deal with- just build more renewables. Bad weather doesn't cover up entire states except for extreme events, and you can just turn it off when it isn't needed. You can build a lot of overcapacity if the alternative is to pay peaker plant rates.
Increased variance like I interpret most people talking about it is those extreme weather events not necessarily hurricanes, but things like a week of no wind or heavy clouds happening to cover every panel in a distribution area. The fear being that you would still need gas plants or huge batteries to run for that one week a year, at extreme cost. The variance averages out most of the time, but not all the time.
Virtual batteries work with the former, but not the latter.
> In response to this increased market rate delta it would become more viable to invest in dynamic pricing and load shedding/deferring tech.
It's definitely not a tech problem. It's an incentives problem. The tech was always incredibly simple, and it does literally exist already- you can buy internet-connected thermostats. All you need to do is connect Nest to your local electricity distribution company and tell it how many degrees colder/warmer you will tolerate per $ saved. 45% of a house's energy use is in controlling the temperature of air and water (and that's not counting the fridge, which is another 7%).
It's as much on the suppliers as it is on consumers, IMO. Electrical distributors are some of the laziest, worst-run companies in the country. Half of them can't even do billing right; I know dozens of people who have been double charged or never charged or charged for their neighbor- nobody wants to read their electrical bill, so nobody cares. The average US household spends ~$2450 on electricity annually, and the amount you can save for how complicated it is is just below that mental threshold.
I don't see it getting better without legislation. Most obviously, a push for subsidized smart meters that don't use 1930s tech to measure electricity. Then a standardized (extensible) API and/or reporting requirement, so that devices can know the current price of electricity. A standard for transmitting that info over the house circuits themselves, if you're feeling fancy. Direct-to-consumer subsidies from grid authorities for things like ancillary services, power factor correction, and frequency stabilization.
> You're talking about sudden short-term variance in supply, which is easy to deal with- just build more renewables.
Easy as in simple, but I think this dramatically skews the price viability. If I need 2x overcapacity then the price to the consumer is 2x per MW of base capacity, and it’s no longer viable to use solar over gas.
> It's definitely not a tech problem
I disagree. I’m aware of some existing options, my claim is that with a bigger delta, more options become viable to research and implement. For example there was a thread recently where we discussed modulating energy usage in aluminum smelting, which requires a new design for the furnace to keep the temperatures stable. (This tech already exists, but AFAICT it’s not cost-effective to deploy widely.)
There are lots of industrial processes which could conceivably modulate their power consumption, but it’s not currently cost-effective to even design these improvements at current levels of peak premium.
Tech is downstream of incentives, is what I am saying, and price signals can be a good incentive; many claim that 100% renewable is not viable because of the cost of closing that last 1% of daily variability, I am hypothesizing that the system as a whole could, with appropriate price signals, build the tech to make the demand curve much more mutable.
This gets at seasonal variability; if we have a week with lower energy production, then the peak-premium goes up, and maybe we turn off the marginal industrial, residential, and commercial consumers.
You need most of that over capacity anyway. They still keep generators from the 1920s operational just in case a storm cuts off one town from the grid, turn on those generators and let the linemen fix it a week later
I’d like to see numbers backing up this assertion, because everything I have seen suggests the opposite at grid scale. Sure, remote towns will have backups, but the major metro areas do not run at substantial overcapacity ratios. (Else, peakers would not be a thing.)
It would be really hard. Many businesses in a metro area will have their own private backup generators which contributes to over capacity but doesn't get tracked by utility numbers. Sometimes utilities have programs where a business will on demand switch to their backup generator while the grid is on, thus getting that business off the grid. (private residences also have generators at times, it is common for businesses but both would have to be tracked.)
> I am hypothesizing that the system as a whole could, with appropriate price signals, build the tech to make the demand curve much more mutable.
This is an interesting concept. We used to live in a way that was much more directly connected to the amount of energy we could capture each year; in a good year, everyone would have a lot of food, and in a bad year, there would be famines.
Making ourselves dependent on an energy harvest seems like a way to intentionally reintroduce the problem of mass deaths when the harvest doesn't turn out as hoped. Is that something we want?
Heavy clouds often do cover multiple states (or European countries) simultaneously. But the bigger problem for places like the UK and the USA is that the major grid interconnects run East so when peak daily demand hits it's already getting dark in the places from which they can easily import electricity.
Dynamic pricing and load shedding can reduce the need for peaker plants and storage systems. But as a side effect it will also drive energy intensive heavy industries offshore. Certain types of industrial facilities can't just start and stop, and every minute they're offline they're losing money. Major countries have to keep certain strategic domestic industries operating regardless of the cost or environmental impacts; it's just too risky to depend on imports that could be interrupted at any time due to a war or other geopolitical crisis.
The missing piece is renewable fuels. Having some kind of hydrogen or ammonia, biodiesel etc. from solar and wind and available to supplement real-time generation when needed would make it much more feasible to drop fossil fuels.
Those would require a massive surplus of Solar, Wind and Hydro which we currently do not have. It may come to that but right now the surplus happens at best during a few minutes to 10's of minutes at peak solar in the summer, the rest of the day (and the rest of the year) we are running at a substantial deficit that is still made up from fossil fuels or nuclear.
The efficiency is pretty awful, though. It's quite hard to make the economics of it work (even if the electricity is free) with current fuel prices, which is why it's not really happening. Either the tech needs to get a lot better or the fuel needs to get a lot more expensive (which I don't imagine would be a popular option).
It's much easier to make the economics work than it would be to make the economics of batteries work for seasonal storage or rare event backup.
If you look at cost optimized solutions to providing steady electrical output from solar + wind in high latitudes, hydrogen is strongly featured. It becomes much more expensive to use just short term storage like batteries.
Hydrogen is often useless, but it's sometimes essential. Hydrogen critics bleat about efficiency without understanding the limits of that argument.
One thing I didn't appreciate until recently is how big Britain's EEZ is. If floating wind turbines can be made practical, there is huge opportunity not just for energy self-sufficiency, but for export too.
P.S. even if the numbers have been superseded, MacKay's original book is still worth reading because it's so fantastically clear in how it lays out the basis for estimation.
If you haven’t read the book “Sustainability without the hot air” by David MacKay then I strongly recommend it. He makes some assumptions that may not stand the test of time but overall it’s well reasoned and he explains the maths and logic well.
His passing is a real loss to UK science.
I got to see him speak once, very engaging and his passion was clear.
One thing i don't like about this analysis is when final energy is expressed in TWh, which is a unit normally reserved for electricity. That creates confusion.
MacKay's entire analysis is posed using KWh (and multiples) as a standard unit for energy. Why? From memory he stated it was because everything would otherwise need to be converted to Joules (or multiples) and KWh were at least a common enough energy unit that people "think" in them for resource estimates.
I just finished, “How Big Things Get Done”, I think on a recommendation some time back from someone on HN.
He says that among the projects that tend to be on time and on budget, roads, solar and wind are three of them. While he doesn’t say it, I read this as “all large successful projects start as small successful projects”. Once you’ve built 10 miles of road the next 10miles is mostly more of the same, subsurface conditions notwithstanding. Once you’ve installed three wind turbines in a field installing the rest looks much the same (getting the first one in required solving a bunch of transportation problems of course). The teams just get a little faster with each one, because they are iterating on a pattern they already know.
You try to build a nuclear power plant and it might not show up until after the politicians who pushed for it to be built have retired. Which means it might not show up at all because all of the skin has left the game. But if I try to cap my career as governor with a new wind or solar farm? I may actually get to cut the ribbon.
It makes me feel a bit better about our prospects that solar and wind are easier logistical problems than repeating the old patterns.
> You try to build a nuclear power plant and it might not show up until after the politicians who pushed for it to be built have retired.
You can try to reactivate nuclear power plants that were shut down.
I state I live in (New York) closed 2 nuclear reactors in 2020 and 2021, each providing more than 1 GW of clean electricity. Both reactor were closed because of political pressure. If we were to apply the reverse political pressure, I think we could have them up and running in 5 years, if not sooner.
Yes - this is (part of) the motivation for the current idea of "small modular reactors". The idea with SMRs is that if you build the units in a factory, not on site, then things like unexpected ground conditions necessitating costly design changes can be mitigated. Making them smaller might also reduce things like extremely tight tolerances on certain parts leading to potential delays due to having to "throw it out and start again".
It's closer to the model used for nuclear submarines which have a better track record in terms of delivery to time and cost than civilian nuclear power.
I just looked up a typical gas turbine plant. The US Energy Information Administration (what is that, PR for the DOE?) thinks that a 'typical' gas turbine plant is 600MW. The DOE says nuclear plants clock in around 1GW, but is that single reactor or a four reactor core system?
I feel like most of this is barking up the wrong tree though. If you're trying to get people to use a new power plant type, you may be looking at small cities and large towns. At which point you might be comparing your situation to how much power you can get from the nearest metropolis via high tension power lines, and how much more secure your winter power supply would be if you didn't rely on those lines working.
And a cursory search suggests those can vary from 40MW to just under 600.
So can you sell Geneseo a 50 MW reactor/solar/wind farm?
What is the running cost of doing something like this? How often do they need to be replaced and can we afford to do that for the long term?
Also - what about the geopolitics. The reason solar is cheap is because china. Do we want all our energy needs to be dependent on China? Although we are still building Hinkley Point C with Chinese investment anyway so ¯\_(ツ)_/¯
CGN do still have a stake in Hinkley Point C but they were (probably forcibly) bought out of Sizewell C by the UK govt last year and it's pretty clear future investment from them is no longer welcome.
Wants to run the full grid on intermittent energy and only one sentence on what happens if there is no wind (on top of typically no sun in the UK).
Using nuclear to compensate doesn’t make any sense. If you need to replicate your wind production capacity with nuclear (keep in mind wind, even offshore, can go to zero) and given that nuclear costs the same thing whether you switch it on or not, you might as well spare the construction of wind farms and go full nuclear.
As the author points, energy storage is uneconomical.
I am also surprised with the idea that electricity consumption will go down. With heating going electric, cars going electric, and mass immigration (hence population growth), that seems counter intuitive to me.
I completely agree: his introductory mention of the importance of good numbers to inform policy got me hopeful, and then he ignored storage which is possibly the most critical constraint for high renewable penetration.
I’m hoping grid scale batteries follow a similar cost reduction to solar but the latest projects which have been built do not inspire confidence. But for some reason this issue is always hand-waved away by the most ardent proponents of renewables.
People treat this issue like a football match in that they cheer for their favourite team and ignore their failings.
Grid scale batteries wouldnt be the most cost effective short term storage - pumped storage like coire glas in scotland is.
If the UK had another 10 or 15 of those (perfectly doable and economic) that would provide ~8 hours of storage. That would be enough by itself to provide a 94-96% solar/wind powered grid.
For that last 3-4% of seasonal/peaking storage and backup nothing beats hydrogen yet. It's only 50% efficient roundtripped but it stores very cheaply for very long periods.
If you turned all solar and wind energy into hydrogen and only burned that for electricity it would still be cheaper than power from a nuclear plant.
Batteries are probably best used in cars or places which can't connect to pumped storage (which are pretty rare).
Do you have any figures for the hydrogen storage vs nuclear point? I’m not saying you’re wrong but it goes against intuition.
Both hydro storage and hydrogen have the same problem: not all locales have elevated valleys or underground caverns. I’ve heard some talk about using natural gas pipelines for hydrogen storage, but that has its own issues.
The LCOE for nuclear power is roughly ~5x that of solar and wind.
If you assumed that the capex and running costs to electrolyze and store 1GWh of power in an underground cavern is roughly the same as it takes to produce it, and that it's about 50% efficient when roundtripped then that would make the stored wind and solar roughly 3/5ths of the cost of nuclear power.
This still makes it a lot more expensive than peaking with natural gas. Just because hydrogen storage is competitive with nuclear power doesn't mean it's economically competitive in general. Nuclear power is just way more expensive than people assume, and generally won't get built without a lot of really lavish subsidies (e.g. like Hinkley Point C getting paid 3-5x as much for the same kwh of electricity as everybody else).
LCOE doesn't account for storage, grid extension necessary and for the fact that nuclear can work for 100 years but solar panels and wind turbines have to be replaced after 20-30. Also, wind and solar where? Solar and wind vary greatly based on where they're installed, yet everyone throws these values around like a solar panel in Greenland would be as efficient as one in Sahara.
It doesn't account for storage (obviously, I think I was pretty clear that this would increase the price from 1/5th to < 3/5ths) or for grid upgrades but LCOE absolutely does account for the projected lifespan of the generating assets.
100 years is, by the way, ludicrously optimistic for a nuclear power plant lifespan. If we ran nuclear power plants that long past their projected lifespans then quite a few would go kaboom.
What did you base your storage cost calculation on? LCOE doesn't account for nuclear running much longer than 20-40 years. And 100 years is not ludicrous, 2 units in the US have been approved to work for 80 years, 100 is not far off if properly maintained.
>What did you base your storage cost calculation on?
Check the link above.
>LCOE doesn't account for nuclear running much longer than 20-40
I can assure you that LCOE does not assume a nuclear power station lasts 20 years. It assumes a normal lifetime - neither too short nor a lifetime that would induce catastrophic risks (e.g. 100 years).
I've seen no evidence that standard LCOE models under or overestimate solar/wind/nuclear power lifetimes at different rates and it would be presumptuous to assume that they do without evidence.
Your link doesn't say anything about what you proposed regarding storing energy using hydrogen and how much it's supposed to cost. LCOE assumes that nuclear plants will last for 40 years which is wildly off so yes, it favors renewables. And no, running plants for 100 years would not induce catastrophic risks if you engineer for it.
No, I asked you for figures and you quoted irrelevant ones. My belief is based on following costs in the energy industry. I’m happy to change my beliefs, but not without good reason.
Several hours? Wind is down to a small fraction for several days regularly [1]. Electricity outage should never happen, so you need to protect against a very high percentile of adverse events. 8h of storage won't do.
>Graham says that the CSIRO modelling showed that at very high levels of wind and solar, a maximum of half a day’s average demand was needed for storage. In some areas of the grid, only around three hours might be needed.
>This is an important point, because some renewable critics say that about a week’s worth of storage is needed
Something a lot of people forget or don't notice is that solar and wind anticorrelate. The model takes this into account and doesn't assume a 100% wind grid. That would require a lot more storage, but isn't realistic.
There will be some differences between Australia and the UK but it's not going to be hours vs. weeks.
It defies common sense that you can make up for wind energy being down to almost nothing for days with 8h worth of energy. Solar will never produce much in the uk, and will produce zero during the cold UK winter nights when consumption is maximal.
No need for a sophisticated mathematical model. Look at wind production last year per grid watch [1], the two peaks in term of energy consumption (around Feb 2022 and Nov 2022 in the "Last Year (Day Averages)" chart correspond to times when wind was low (and therefore gas was high), some days going to near zero. And last winter was an unusualy warm winter.
If I can find examples just looking at one year's worth of data, how is this going to survive a once in 25 or 50 years weather event with no wind for something like 3 weeks?
There is definitely a need for a model. Even more so if you start making wild claims about the weather.
>how is this going to survive a once in 25 or 50 years
You think the sun going out and the wind stops blowing for an entire month once every 25 years? That's quite a brave claim.
>And last winter was an unusualy warm winter.
I remember. Solar power was unusually productive last winter, heating requirements were a lot lower than usual. The wind didn't stop blowing for a month though.
> I am also surprised with the idea that electricity consumption will go down. With heating going electric, cars going electric, and mass immigration (hence population growth), that seems counter intuitive to me.
The article says energy consumption (which includes electricity and also fossil fuels burned for heating/transportation) will go down, but electricity consumption will go up.
But even the rise in electricity consumption will be tempered by the fact that electrical end uses are so much more efficient than fossil fuel burning equivalents, in particular for heating and transportation.
Heat pumps can move 3-4 times more energy than they consume, and EVs are likewise far more efficient at turning energy into motion.
Modern building techniques also play a big part in it. Houses build to modern building standards (see Scandinavia and Central Europe - the UK and Ireland is still lacking here) uses a hell of a lot less energy for heating than houses built even two decades ago.
Last year I was renting a house built in the early 2000s. It was poorly insulated and inside we had the thermostat set to 18c, yet our gas bill for November was €400. Over Christmas I moved into a newer house built in 2018 that was the same size, the thermostat was set to 22c and during February (which was much colder than November) the gas bill was under €200.
For grid scale, novel batteries, such as thermal and gravity, are just about to "cross the chasm" and jump onto the cost learning curve. In 15-20 years time, they'll be cheap and plentiful.
For residential, batteries are being added to appliances. Much cheaper (not panel upgrades) and much larger potential market size (apartments, condos) than "powerwall" style storage. These will follow the standard replacement rate of appliances, accelerated by incentives.
Sodium batteries are just starting to hit scale. Of course they'll replace lithium for stationary storage. I don't know enough to predict how big of a role they'll have in grid scale systems. Thermal batteries will be super competitive.
And of course, there's many more aspirants. Like flow batteries.
I know you're probably being tongue in cheek about there being no sun in the UK, but my domestic solar panels produce about 4 MegaWatt Hours each year.
The problem isn't "there's no sun in the UK". The problem is: What do you do on an overcast December day with no wind? Or worse - what do you do if the whole country has heavy snow and no wind in December?
Renewable capacity factor isn't predictable on a daily basis. You need a backup - which basically means having an entire grid's worth of gas turbine/coal/hydro plants in storage for a few hours a year, or having grid-scale storage. That's not really a solved problem (even the biggest batteries are a couple of orders of magnitude too small to provide an entire country's electricity supply).
Electric vehicles (with vehicle to grid) will help. But even then, we need way, way more long term storage than we currently have.
> The problem is: What do you do on an overcast December day with no wind?
You burn green hydrogen in turbines. A simple cycle turbine power plant is maybe 5% of the capital cost of a nuclear power plant of the same rated power output. And because its used so infrequently, the fuel costs are quite tolerable.
The hydrogen is stored in salt formations, which the UK has in sufficient quantity.
There are other more complicated e-fuel schemes with the same effect.
The UK should be exporting green energy. Which means installing two or three times the capacity projected, plus as many nuclear power stations as we can throw into the mix.
For a nation that relies upon imported food, it is vital that our exports are also required by the rest of the world. Being a major green energy exported - particularly if we can bottle it as synthetic liquified methane - is the way the UK needs to go.
You don't need storage if you build big. Those who import from you need the storage.
Wind goes pretty much to zero on certain days (like 3%). You don't solve that by overprovisioning.
There is also quite a high correlation with the weather across Europe, and when it is windy in the UK, chances are it is also windy in most of the major countries the UK could export energy to.
As for nuclear, again, it doesn't make sense to use it to compensate for wind. Just make it run whether the wind blows or not, and save the cost of a wind farm.
TLDR. The article claims Britain’s energy needs can be met by renewables. Almost half of estimated renewable output, 1557 of 2896 TWH, will come from floating offshore wind farms.
305 comments
[ 2.7 ms ] story [ 259 ms ] threadOne thing in curious about: developments in fixed wind turbines tech mean they are economically viable in more locations (e.g. deeper water) so can cover more area.
But how much more do they produce per square km of deployment than Mackay's estimates?
In his book, he has a nice section explaining that bigger turbines have to be spaced out more, so whilst theyre cheaper, they don't produce as much more energy as the headline 'output per turbine' would naively suggest.
But modern turbines are higher, so they presumably 'catch' more wind, and windspeeds are more consistent higher up. But I'm curious how big an effect this is.
Average turbine capacity of turbines installed in 2021 was 7.4 MW but all manufacturers seem to have 15+MW designs in the work [1]. So the projection is slightly optimistic IMO. The capacity factor seems slightly more optimistic at 50% as worldwide average seem to have fluctuated between 35 and 45% [2]. Although other numbers are closer to 60% for the UK [3]
[1] https://www.energy.gov/eere/wind/articles/offshore-wind-mark... [2] https://www.statista.com/statistics/1368679/global-offshore-... [3] https://windeurope.org/about-wind/daily-wind/capacity-factor...
And time and time again renewables always seems to take optimistic, if not ludicrous estimates.
Mackay's papers is indeed 15 years old, and what do we have now? Extremely expensive power, backed by fossil gas, because if wasn't, people would die. Avoiding 2 deg C and ecosystem collapse is not even vaguely possible anymore. In that time, 40 years ago, France got to 17 g / kWh. That's not a success and, for the literal trillions of ESG money spent, someone should be held accountable for that.
The result is that the wind industry (offshore - onshore is limited by physical size constraints) has been consistently getting bigger every year.
https://www.energy.gov/eere/articles/wind-turbines-bigger-be...
I know that a mere back-of-the-envelope calculation isn't worth much more than the used envelope it was written on (doubly so when it is based on guesstimates of the input numbers), but that would be only £1bn for 8 GW or £4bn for 32 GW (compared to actual average usage of 31.5 GW last year), which is the kind of thing that the British government shouldn't blink at but in practice actually faffs and fails at basically all the time.
(And the sector is theoretically privatised, so this would have to become a business investment, which in turns will have potential investors ask inconvenient questions like "What's the risk we have cheaper options in 10 years that make this power line redundant? And what about those fusion reactors I keep reading about in the Sunday Times? What if Scotland becomes independent and stops selling you the electricity?")
It's not a physical or geographic limitation.
And doesn't apply to offshore wind.
A bigger problem is just the UK's inability to complete infrastructure megaprojects on land, so the connectors would likely need to go in the sea and take a perhaps inefficient route.
Your meteorological reality seems to not correlate with actual reality. In the UK the highest energy demand is actually correlated with high wind speeds [1]
[1] https://iopscience.iop.org/article/10.1088/1748-9326/aa69c6
>This reflects the variation in temperatures and wind speeds with season, with calmer, warmer conditions in summer and cooler, windier conditions in late autumn and early spring. However above the 75th percentile of demand, average wind power reduces, which occurs predominantly in winter and autumn. Understanding this downturn in wind power provides the motivation for this paper. Given our interest in high demand days, which predominantly occur in winter (figure 1, upper right), only winter days are considered.
>The tendency for lower wind power during higher winter demand is shown by the tilt of the density contours of the daily distribution (figure 1, lower left). It is also clearly seen when averaged across days of similar demand (figure 2, left). Average wind power reduces by a third between lower and higher winter demand, from approximately 60% to 40% of rated power.
Look at figure 2. Black is wind power, and the X axis is demand. Wind production capacity is down when demand is high.
The vast majority of the UK's cold winter weather comes from wind from the North-East, bringing in the much colder weather from Arctic/Siberian regions.
I frequently hear people bring up transmission losses as a concern, and genuinely curious where this idea comes from? Was this taught in schools or part of some disinformation campaign?
I would honestly like a deeper explanation of how electrification will produce such a wild decrease! That’s shrinking energy use by more than 2/3, and presumably after taking into account population/industry growth…? Or are the authors just wildly pessimistic (not…unmerited) about Britain’s trajectory over the next quarter-century? What am I missing here?
Harder to ball park heat pumps because it depends on the climate, but 4x is probably a reasonable guesstimate. At worse it's equal to burning things for heating (when it's too cold outside to use heat pumps, not sure that really happens in Britain), at best it's... some ridiculous factor better (when it's practically the same temperature outside and inside).
If you replace a ICE by an BEV powered by solar cells, you actually reduce the total TWhs because of efficiecy:
- ICE: 6L of gasoline per 100km, that equals to about 60 KWh
- BEV: 17KWh for the same distance.
The same applies to heat-pumps and some industrial processes.
[1] https://www.withouthotair.com/c3/page_29.shtml
> OK, the race is over, and I’ve announced two winners – public transport, and electric vehicles.
He was also very positive about heat pumps.
https://www.withouthotair.com/c20/page_131.shtml
> You’ve shown that electric cars are more energy-efficient than fossil cars. But are they better if our objective is to reduce CO2 emissions, and the electricity is still generated by fossil power- stations?
> This is quite an easy calculation to do. Assume the electric vehicle’s energy cost is 20 kWh(e) per 100 km. (I think 15 kWh(e) per 100 km is perfectly possible, but let’s play sceptical in this calculation.) If grid electricity has a carbon footprint of 500 g per kWh(e) then the effective emissions of this vehicle are 100 g CO2 per km, which is as good as the best fossil cars (figure 20.9). So I conclude that switching to electric cars is already a good idea, even before we green our electricity supply.
So electrification alone can't bring so large reductions in total energy consumption - it is simply one of the prerequisites for that, and it also needs to be accompanied with large scale build-up of non-fossil electricity, of renewables or nuclear.
Those are two separate things - electrifying some transportation need gets us X benefit, and converting some electric generation from fossil fuels to solar gets Y benefit, but the former doesn't automagically create the latter.
2) You are assuming there is no curtailment or rejection of wind, solar, and nuclear, which is categorically false. This is why there are time-of-day rates that vary so drastically. There are large amounts of renewable/low-carbon energy that would otherwise be wasted in pretty much every geography.
If I turn on (or keep on) my light when usage isn't completely covered by renewables, should that be calculated as using all fossil fuels?
The logical extension is that everyone should calculate the carbon intensity as a binary choice which is obviously wrong.
The only reasonable way to go is to look at the actual carbon intensity average.
Probably storing.
Give it 50 years and we're going to be facing some other unexpected global disaster from unchecked battery proliferation.
Hopefully it's less dire than climate change though.
This is largely because
1.) big stationary heat engines are much more efficient than small weight-optimized mobile ones, and
2.) the electricity grid also has some renewable energy mixed in, so a growing fraction of the electricity supply doesn't even need an (energetically imperfect) heat engine.
Additionally, the parent overlooks that while a lower number is estimated, they compare against a hypothetical 2000+ TWh number instead of the tightest 900 TWh estimate.
Fuel delivery still requires a crap load of energy, some of it in the form of electricity - even just running the pumps 24x7 - but there's so many elements in that value chain it could fill volumes. There's far fewer even in the dirtiest of EV charging scenarios.
That’s nearly half the saving
The submission (blog post) doesn't mention the number 900 at all; the paper which the submission is about only mentions 900 in a footnote saying "Total European [energy] supply was 22,900 TWh (IEA, 2022)."; the summary pdf of said paper never mentions 900 or 600.
Figures in the paper ending in 600 occur in a few spots, but nowhere 1600. There is 21'600 TWh/year (total European energy supply, page 8), 16'600 TWh/year (idem), 10'600 TWh/year (prior studies' estimates of UK wind resources), and 2'600 km² (land occupied by buildings).
But maybe I shouldn't be drawing conclusions based on trying to search character sequences in a semi-picture format...
It's right under the first main heading. Just above and below the first picture. Searching for 900 in the post took me right to it. I have no idea how you missed it.
Honestly, anybody claiming in 2008 that PVs are too expensive so we should not invest on them is safe to ignore.
Thing to consider the ultimate price of a manufactured good tends to track production volume, device complexity, and energy required to produce including raw material.
Solar panels require complex machines to produce but are themselves simple, the volume is high at scale, and energy requirements are low. That points to something where the price is close to the energy and material costs.
They can surely become low, but currently they are not. The EROEI of PV panels is barely on the region where it stops being one of the largest factors in its cost.
But yeah, PV has space to improve a by a few orders of magnitude more.
You're misrepresenting the analysis.
In the 15 years since, public support for PV flipped from con to pro, and we've learned that much less land must be allocated to just PV.
As noted in the OC, Mackay's contribution was methodological. It easily accommodates different assumptions and updated numbers. O'Callaghan et al did just that. Yielding a new conclusion.
https://flowcharts.llnl.gov/
As far as I can tell the current usage of 1200 TWh include electricity and combustible chimerical energy of gas. Electricity use is ~300 Twh, and Gas usage is ~800 Twh.
The proposal is that gas heating would be entirely replaced with heat pumps and most gas generation would be replaced with modular nuclear reactors and offshore wind.
The numbers are a little misleading because of the way gas and electricity are summed to get the top level numbers. A TWh of gas consumption is not the same as a TWh of electricity consumption. In thier model, 50 TWh of electricity can replace 400 TWh of gas. The challenge with this approach is that it is not show what is going on with user consumption. Are they getting more, less, or the same thermodynamic work done?
https://www.nationalgrid.com/document/138976/download#:~:tex....
Not 80%, but combined cycle gas turbines regularly achieve up to 60% steady state efficiency.
There is research into ammonia fueled turbines that supposedly achieve even higher efficiency.
2. Demand is up in this scenario, not supply, because of all the things being electrified.
The improved efficiency needs to be more than the drop in marginal cost which is what governs demand, nothing to do with how much more electricity you use today. And even then, that relationship isn't linear because a 30% reduction in cost can drive a 60% increase in demand because that reduction puts it in a new price bracket where a lot more people can afford it (since wealth is non-linear). This stuff is super non trivial and has all sorts of higher order effects.
But there isn't any real reason to think this is implied under a "high renewable" scenario.
In-fact, some of efficiency gains are probably driven by increased costs.
The O’Callaghan et al. paper (in the Blog post here the related figure is fig 1) says that current (2023) demand is 1500TWh and current supply is 2885 TWh. There are different projections for total demand in 2050 (note all of them project a reduction of demand, due to efficiency gains), one of these is the national grid FES which projects 900 TWh. Importantly the O'Callaghan paper opts to be conservative and choose to use the current demand as the demand for 2050. This is conservative, because it is higher than all projections which all assume that we get demand reduction from efficiency gains.
That removes the need for storage (and is what every other country should be doing as well).
MacKay's work (and the update under discussion) about the UK, and also my claim about China, is about the latter.
https://youtu.be/kBMSZ7v3KxQ?si=ZxlQlOHc6r6tj7n7
[0] Well, remote sensing in general. You can argue if a map at specific spectral lines to detect industrial byproducts is a picture or a graph.
The UK "industrial" energy consumption today is 30% of what it was in 1970.
https://assets.publishing.service.gov.uk/media/62334e14d3bf7...
This is a long term problem Europe faces: with solar projected to dominate world energy supply (if historical experience rates continue), Europe will be an energy ghetto. Energy intensive industry will move elsewhere. Going nuclear cannot save them, since their nuclear will be competing against world best case solar in sunnier places.
For example, if we need massively more computing for an AI breakthrough, there's nothing requiring that to be localized. Similarly for production of any commodity that can be shipped.
If anything we have mostly seen massive improvements in energy efficiency in the last two decades, with appliances, the switch to LED lighting from incandescent, heat pumps vs other heating/air technology, etc. What new technologies have shown up get quickly innovated in the name of energy efficiency.
If the libs had kept the klepto stuff to a less obvious level and adapted the binding parts of their luxury beliefs platform they probably would still be polling a majority and could have kept this going for quite a while longer.
Do you, though? What? I'm not sure how typical this is. Like, if you replace your creaky old gas central heating with proper insulation and a ground to air heat pump, you're going to be saving a _lot_ of energy if it's a big house. What would you use this extra energy 'budget' on?
@dang
US life expectancy is declining unfortunately, but for unrelated reasons.
Not only is the EV itself more efficient, so is each stage before that more efficient. Something like 1/3 to 1/2 of all CO2 emissions from ICE cars occur before gasoline makes it to the car. Thus, driving an EV lowers the total energy needed without actually reducing how much people are driving.
PS: Total energy calculations are fairly arbitrary because it’s ultimately sunlight powering things, but that’s also true of fossil fuels just at really unbelievably low efficiency so that parts generally ignored.
And thanks to the "Base Load" myth of coal based generators, people were encouraged to use heat banks and storage HWS which loaded the system during the night.
But this could only be done by offering cheap "night rate" electricity.
In the future, our consumption must be focused on day-time peaks, as well as much better insulation, and the use of heat-pumps. And of course a steep increase in electricity tariffs will greatly encourage this.
In the future, companies which operate 24 Hrs without thought to electrical cost, will focus much more on efficiency, and even begin to shut down operations at night.
In many countries (eg Australia) there has already been a huge increase in domestic solar and battery installations.
This change will be driven by unavoidable price rises, and the recognition of our profligate wastage in the past..
With Solar and Wind, there isn't any wasted heat.
To be clear, today's demand isn't 2900 TWh. It's 1599 TWh. And none of the people claiming Britain can be powered via renewable power are using the 900 TWh figure.
Just noting for people who haven't read the article because this comment wasn't clear to me:
The difference is mostly about pricing of renewables radically dropping since 2008, and the likelihood of floating offshore wind producing large amounts of supply.[1] https://www.statista.com/statistics/322874/electricity-consu...
Many comments try to allude to it. It is important what we are talking about- Primary energy, Secondary energy, Final energy or the Useful energy.
The embedded links go in further details.
These things are all more efficient than the hydrocarbon equivalents, further depending on the measure, if you count the gas input energy in to a gas plant, a renewable generator is much more efficient too.
So I think it's possible, although the figure ends up somewhat misleading without a lot of context.
If cost is the issue there are still risky moonshots like "throw tons of money at fusion" (attacking the cost of nuclear) and "throw tons of money at high temperature superconductor research" (attacking the amount of energy we need), and "geoengineering" (risking screwing it up worse). Not guaranteed to work, but you know, better than rolling over and dying. Also has the side benefit that a lot of the moonshots are worth trying anyways.
More abundant? Depends. There's a lot in the sea, which isn't currently recoverable. There's a lot we could do (but actually don't) with breeder reactors to make more fuel.
This means the answer to the question "how long could we last on just nuclear fuel alone?" varies from 5.7 years[0][1] to 4.3 billion years[1].
[0] https://globalwarming-sowhat.com/renewables
[1] https://whatisnuclear.com/nuclear-sustainability.html
(By that last bit, I mean: powering EVs from coal plants with carbon capture at exhausts might be different, idk, but probably not cheaper than just not polluting in the first place.)
Solar energy (and thus wind assuming otherwise constant environmental conditions) is only going to get more plentiful for next 5 billion years...
We haven’t looked very hard at all for uranium. Nor have we tried very hard to use the uranium we already have efficiently. Nor have we tried very hard to use thorium as an alternative. Some combination of these would stretch out the supply of uranium we have essentially indefinitely.
Mainly due to new regulations meaning projects have to re-engineer themselves before they are even complete. Leading to delays and cost overruns.
Either we are under an existential threat, in which case nuclear power is an amazingly cheap way to save 8+ billion people, or there is some other agenda at play.
It is not a case of "we should do everything" because we can't afford to do everything. If we could afford to do everything we could easily do a small subset of everything and solve the problem.
Nuclear would stop being expensive if it was committed to. Building 1 bespoke plant avery few decades is not a good approach.
https://en.wikipedia.org/wiki/Floating_wind_turbine
A superconducting cable probably also makes attempts to damage the thing easier — to get 40,000 km of aluminium down to 1 Ω, it needs to have a cross section of 1 m^2, which is kinda hard to damage, though also you don't really want a single cable because that, with current global electrical demand and reasonable (i.e. currently in use) choices for the voltage, would be in the order of 1.5 mega-amperes and match Earth's geomagnetic field at a distance of about 11 km.
Using such hyperbole is counterproductive and makes people discount the actual risks. If we're talking about, for example, a billion people being displaced due to climate change disrupting food production or flooding areas which are currently densely populated, and many millions dying in that process, that is bad enough by itself to justify taking action, and there is no need to resort to ridiculous unjustified exaggerations implying that we're all going to die.
This isn't insurmountable. It doesn't have to be perfect. Even a real but incomplete effort makes a genuine difference.
The problem, unfortunately, remains the US, who has a large and powerful minority who is convinced it's all a hoax. The solution doesn't have to be complete but it does have to be something. With many millions of people actively making it worse, even 100% in the UK doesn't come anywhere near close.
I didn't read the paper, but from personal experience:
1. Rooftop solar depends on lots of individuals to make an investment to put solar on their houses. If you say "let the government subsidize it", that doesn't really make much sense because the government can get a much better return on each dollar spent by investing in more efficient technologies.
2. Many roofs are not suitable for solar. They are either facing the wrong way, at a bad angle, shaded, or, in my case, too "origami-like". Solar panels need to have minimum clearance fro ridge lines on a roof, which can drastically reduce the total coverable area.
1. Most residences not being suitable for significant generation.
2. The return on investment is decades long (yes, so is double glazing, but that is useful for far more households), not that most households can afford it at all.
3. The total generation capacity is piddle, especially for government investment.
4. The materials that go into making PV panels are horrible to extract.
I'm not saying PV panels are useless, but they are not much of anything. Not something my lecturers at uni liked hearing/reading, but lo and behold pretty much nothing substantial has changed since. They're too busy blowing Sustainable Development smoke up their own arses though.
Commercial rooftop (warehouses, big box stores etc) is a bit cheaper than nuclear, with the most expensive being the same cost as an average nuclear plant, and the low end for new commercial rooftop solar being the same as the running costs of already built nuclear:
https://www.lazard.com/media/2ozoovyg/lazards-lcoeplus-april...
https://www.lazard.com/media/2ozoovyg/lazards-lcoeplus-april...
It still shows rooftop clearly to clearly have the worst LCOE. I bet the assumptions for Rooftop dont include suboptimal builds, like builds in San Francisco, on a north facing roof, under a tree, with storage.
Rooftop can be good in niches, but it is hardly a panacea.
Ain't no sunshine.
The point is that rooftop solar is grossly inefficient and terribly expensive compared to just about anything else someone could do.
The grid scale alternatives — solar farms, wind farms, hydroelectric — can be better or worse in different ways and different times of year, but normal people can't put full size ones on their own property and such wind and hydro scale non-linearly anyway so small ones aren't as cost effective[0].
Furthermore, the reason for the link under which this is being discussed, is that the UK has land area constraints. Suitable areas in the UK for grid scale PV are basically all farms and/or designed National Parks or AONBs and/or designated green belt, and the exceptions (like disused military bases and airfields) have people clamouring over them to build new towns. Sometimes people can get grid scale PV past planning permission, but it's hard work and upsets people with power who want the countryside to look like farmland. Rooftop PV circumvents that.
[0] Except possibly geothermal/ground source heat pump. Those still provide cost effective benefit when you can do them, but the wide price range means I'm not sure which is better.
This is because only a small fraction of the residential rates goes to production costs, and the rest go to distribution infrastructure and operations. Distributions and infrastructure costs/ kWh go with more residential production, not down.
Think of it this way. With commercial power you might pay 0.10/kWh production and and 0.30wh distribution. You can make your own rooftop for 0.35/kw, but the grid still costs the same or more, so that gets added to your bill.
Residential rates for rooftop solar only ever made sense as a huge subsidy for early adopters.
There is a cost advantage currently, but that is a side-effect of how we currently do accounting of grid costs, not reflecting any actual advantage in reducing total costs across the country.
This is why California is now considering grid connection taxes, which would charge homeowners for electricity they generate at home.
1. Explaining the difference between MacKay's original "technically possible" vs. "practically possible" supply numbers. I agree with the article, the world has changed a ton since 2008 and I do think much more of that technical possibility is now practical due to changes in tech and attitude.
2. One thing I was cautious about is that the lion's share of final 2050 supply in the updated numbers comes from floating offshore wind, which in my understanding is the least technologically "ready" solution. Can someone with more knowledge comment on this? Is floating wind really as "production ready" as would be needed to match these numbers?
There are several other offshore projects using floating wind, more being converted and a lot more planned for conversion or development. There is quite a lot of investment and encouragement from government. See https://www.great.gov.uk/international/content/investment/se... for a reasonable overview.
There is nothing in principle to prevent scale-up of floating wind, but there is an awful lot of scaling to do. Who really knows?
This report probably could use an update ten years hence, but they have done a good job with what is known and can reasonably be predicted.
But if you can install offshore wind on small structures like <30 metres then get that done first, it's cheaper.
About 1500 out of 2800 - or close to half - in the projections I'm seeing on that page.
Not the lion's share, unless we're looking at something different?
In this case, there are hydro, nuclear, PV, on-shore wind, thermal bioenergy plants, and fossil gas plants as well.
We already have plenty of words and phrases that mean 'large' or 'most', so it's a shame this phrase is losing its specific meaning and joining with myriad interchangeable synonyms that mean something more pedestrian.
> But the point still stands: it seems we have a lot of untapped solar and wind resources and they could make up a large chunk of our grid, even if they’re not 100% of it.
Global annual manufacturing capacity is currently enough to produce 50min worth of storage for the whole world (as a fraction of annual electricity production).
That's not a lot and it's not utilized fully, but still well within the capabilities required to shave off the evening and morning peaks - assuming batteries last more than 5 years, which is a conservative estimate.
Nuclear would have been a great component here, but IIRC Hinkley Point C is still under construction and will remain so until 2027.
Any sources for this?
I wonder how long these batteries last. If we can get our battery capacity to grow to 12-24 hours, it removes a lot of the concerns around solar.
The other remaining issue would be recycling that amount of batteries.
It charges from solar and cheap overnight electricity.
2022 production: 1.57 TWh of batteries - https://www.iea.org/data-and-statistics/charts/lithium-ion-b...
2022 total global electricity use: 28,527.8 TWh - https://ourworldindata.org/grapher/electricity-demand?tab=ch...
1.57 TWh / (28527.8 TWh/y) = 28.93 minutes.
-
For stuff like this I assume batteries will have 50% capacity after 1000 cycles even though this always gets surprised responses insisting they're good for far more cycles, as I think lowballing is a safer assumption. You can always just multiply as appropriate for different assumptions about cycles.
However, that last 50% is still useful in this context, so that's half an hour every day going down to 15 minutes/day after 3 years; assuming it's exponential decay, I think that's a steady state equivalent to 4.3-ish times the production rate? Not sure though[0].
Same graph also forecasts 6.79 TWh/y manufacture in 2030, which is ~2 hours each day instantaneous or 8.6 hours steady-state if my previous assumptions were correct. At that scale, I suspect you also need to start accounting for the impact from variable day lengths meaning there won't be demand to fully discharge the batteries 365 days of the year, and also that demand is lower at night while all the demand numbers I've given assume constant use.
A realistic estimate also needs to account for the impact of nighttime renewables getting used preferentially, further lowering the overall frequency with which the batteries need to go through a complete cycle, and conversely increasing battery demand to account for how well the grid they're connected to can compensate for renewable-unfriendly weather, but that's not the topic of this sub-thread.
Also also, I expect these batteries will probably mostly get used by cars.
Finally, I think recycling is a red herring because I can't believe that getting metals out of batteries is harder than getting the same metals out of literal rocks in the ground.
[0] integral from 0 to infinity of 2^(-x/3) for a 3 year capacity halving time?
Unfortunately, current policies are often based on predictions from a few years ago.
Once the transition to EV's is mostly done and car sales fall to normal levels there will be an immense surplus of battery production capacity.
My understanding is also that LFP batteries have basically eliminated the raw materials bottleneck, which was another development that few analysts (or anyone else) were able to predict just a few years ago.
the casuality between the raw material supply bottleneck and LFP would be difficult to demonstrate. LFP has lower energy density and would have required more lithium to get equivalent KWh than what is normally required by high-energy, high-nickel batteries (eg, NCM/A); therefore higher lithium demand, but ...
The price of lithium, and other battery raw materials, however, plunged by 75% since last Novemember peak (still falling) -- it's widely understood that the drop is largely caused by slower than expected EV demand (and oversupply).
>> but I believe the reason for the slow drop is mostly because most of the battery factories currently under production are not online yet... <<
GM forecasted earlier this year that their (Ultium) cell cost would drop to $87/KWh by 2025; just last week, GM reported a 40% drop in their cell cost YoY. Sure, there is 400+GWh new US battery production coming up in 2025-2026 and the commodification/mass-production of made batteries would accelerate further price drop.
which was true for all kinds of stuff from construction materials to EVs last year -- the prices of battery raw materials, eg, lithium, nickel and cobalt, plunged by as much as 75% since last November peak.
For example, using things like clean hydrogen or Power-to-methane processes that can create gas to be used in existing peaker plants.
If there's no curtailment (and if you have sites available in the country) then pumped storage makes a lot more sense.
But either way, building sufficient storage to do anything other than peak-shave is very, very, very expensive, so we'd be better off with nuclear even at a 2-3x cost overrun.
We'd even be better off if all solar and wind power were converted into hydrogen first and only then turned into electricity. Not because thats cheap but because nuclear power is that expensive.
Nuclear power only becomes cost competitive when you use it to share costs with the nuclear-military industrial complex - which is why it is built.
Seasonal variation is best dealt with by creating and storing hydrogen underground. It's easy to store lots of MWhs this way. This isnt as efficient as a battery or pumped storage when roundtripped (~50% not 90%) but if seasonal variation is 4-7% of your power then that just means you'll have to overproduce in aggregate by 10-20% to accommodate this. That's not really a problem given how dirt cheap solar and wind are - about that much is being lost to curtailment already.
The funny thing is, even if all electricity generated by solar and wind were stored as gas before being used it would still work out a bit cheaper than nuclear power. Nuclear is that expensive.
Given that, it's not (currently) economically viable to build storage+PV on anything approaching the scale we would need to replace baseload generation completely.
4-7% isn't seasonal variation in PV production. It's the amount of grid energy that needs to be supplied under a hypothetical 100% solar/wind/pumped storage grid when:
* Solar output is low
* and wind output is low
* and ~8 hours of pumped storage is depleted.
Obviously the wind blows at night and pumped storage works at night and electricity demand is lower at night. Obviously seasonal storage would not have to be tapped every night.
>Given that, it's not (currently) economically viable to build storage+PV
Why would that be given? It's categorically false. The opposite is true. Maybe what you meant to say was that "GIVEN 2GW of solar energy was being produced today" AND "coire glas is under construction" because it's highly economically viable, maybe we should lend a hand in the form of subsidies in scaling that up?
Or did you want to push those subsidies to a form of energy with a 5x higher LCOE?
My concern is the "end game", where all current baseload generation gets disconnected from the grid (including current nuclear when it reaches end of life). Solar PV's current low LCOE doesn't help you a lot in that case because in order to actually deliver the energy needed at the right time you will need to do some combination of (1) build a lot more PV than "needed" and routinely curtail it significantly in summer (ie significantly derate nameplate capacity) and (2) build a lot more storage to deal with seasonal and other variations.
This is when nuclear starts to look like it makes sense as part of the energy mix in my view.
>Nuclear power doesnt store energy. It will pump out 2GW all the time whether you need 6GW or you need 0GW. It doesn't really help in that respect at all because it can't be scaled up and down. It's not a peaker.
i.e. nuclear power needs peakers or storage too. not as much as solar and wind but it still needs to be paired with peakers or storage.
>build a lot more storage to deal with seasonal and other variations.
This is what makes the most economic sense currently because pumped storage + hydrogen storage + solar + wind <<< nuclear power costs.
Even IF you assumed nuclear power plants could be peakers (they can't) or that Fukushima/Chernobyl events had a probability of 0% (they dont), it would still not make any economic sense to build thrm when those 4 technologies can handle everything together at a much cheaper cost and can be built in 1/2 - 1/20th the amount of time.
This is only true if you imagine batteries as the only storage technology. But complementary storage technologies, like hydrogen, allow 100% RE to be achieved at a cost that will likely be lower than what nuclear would allow.
Having said that, any future electricity mix must include nuclear power. France proves that only nuclear allows for deep decarbonization unless you have lots of hydropower and a small population like Norway.
For people who still think that wind and solar alone can do it, I just recommend looking at Germany which had to resort to restart even communist East Germany era coal plants to stabilize the grid after the nuclear phase out.
I think people are letting perfect be the enemy of good, sure they output carbon, but they seem like a much more affordable option for buffering dips in renewables output, compared to batteries or nuclear.
Unless by "gas" you meant green hydrogen.
It seems plausible that even a person in a typical row house could offset most of their household consumption with solutions that will end up with a reasonable return over time.
https://www.youtube.com/@ElectricVehicleMan/videos
https://www.withouthotair.com/
The book provides an excellent overview of how different forms of energy production and consumption add up and which energy solutions could make a real impact. I strongly recommend reading it as context for these updated numbers.
He seems to have written a book dismissing renewables by assuming fixed technology/costs from 2008 for projection into the future (if everybody did this companies would not invest in anything). On top of that he even hand waved the rest away by saying their installations would not be accepted by the public. That seems to me that he was set out to dismiss solar and wind and just looked for numbers to confirm this.
Generally I believe if you want to show the feasibility of a technology you should be conservative in your estimates, and if you want to dismiss it you should be optimistic in your estimates. Ideally you show both conservative and optimistic projections.
Your impression is incorrect.
As the blog post notes, the problem is that two changes in particular - the reduction in demand, and the absolutely crazy drop in the cost of PV - were far larger than anything mckay was able to reasonable foresee in 2008.
> That seems to me that he was set out to dismiss solar and wind and just looked for numbers to confirm this.
Almost the opposite, in fact.
Seems to me that as the energy mix moves more towards renewable, to the extent that the renewable-skeptics’ prediction that variability is an issue comes true, then we would have to build gas peaker plants and start charging more for electricity at peak times. In response to this increased market rate delta it would become more viable to invest in dynamic pricing and load shedding/deferring tech.
So there is a modeling exercise which looks at the peak time price premium for various levels of increase in peaker plants required as the input variable, and compares that to the viability of virtual batteries at those price deltas as the output. I haven’t seen anything along these lines.
> it should stay hot for about 24 hours if it is well insulated.
It takes 50 kWh to heat a 46 gallon tank up to 140 F. That's a ton of energy. Hotter, larger tanks lose even more energy.
Instead, get a tankless heater, backing a small heat pump water tank. You get water as hot as you can possibly want, heated whenever you want, and it never goes cold.
Also, water consistency isn't that big a problem. Today's 120v heat pump water heaters store water at higher temperatures and using a mixing valve to deliver the desired temperature of water. I am just saying, surely electric water tanks, mixing valves, and temperature sensors are orders of magnitude cheaper than the amount of batteries needed to heat an equivilant amount of water.
My parents 1988 switched their hot water to a plan where it was all heated at night. 6 people making no effort to conserve water ran out twice in all those years. Yes the.water was hotter in the morning. But the last shower of the day still needed to be mixed with some cold or it was too hot.
Tankless is worse than a large tank. Tankless needs a lot more energy now, less over the full day, but when you turn the water on it needs a lot now. A tank easily adjusts to use power when the power company needs it. Sun shining or wind blowing, then heat water to use up whatever is extra. Clouds and no wind, just use the stored energy.
Maybe a battery is more efficient, but tanks are cheap
I think you missed that the proposal was a tankless heater backed by an air-source electric heat pump heated tank. The idea is that the heat pump may, or may not, get the water to the desired temperature; the on-demand heater just makes up the difference.
The problem we saw in Ontario when the market was introduced in 2001 or so was that it was politically unfavorable to have the extremely high prices, even for an hour let alone long term, that would encourage an investor to build a storage facility.
I suspect industrial/commercial power usage is the big area for innovation here, but would love to see a breakdown of where the low hanging fruit is.
There are a lot of different types of variability. You're talking about sudden short-term variance in supply, which is easy to deal with- just build more renewables. Bad weather doesn't cover up entire states except for extreme events, and you can just turn it off when it isn't needed. You can build a lot of overcapacity if the alternative is to pay peaker plant rates.
Increased variance like I interpret most people talking about it is those extreme weather events not necessarily hurricanes, but things like a week of no wind or heavy clouds happening to cover every panel in a distribution area. The fear being that you would still need gas plants or huge batteries to run for that one week a year, at extreme cost. The variance averages out most of the time, but not all the time.
Virtual batteries work with the former, but not the latter.
> In response to this increased market rate delta it would become more viable to invest in dynamic pricing and load shedding/deferring tech.
It's definitely not a tech problem. It's an incentives problem. The tech was always incredibly simple, and it does literally exist already- you can buy internet-connected thermostats. All you need to do is connect Nest to your local electricity distribution company and tell it how many degrees colder/warmer you will tolerate per $ saved. 45% of a house's energy use is in controlling the temperature of air and water (and that's not counting the fridge, which is another 7%).
It's as much on the suppliers as it is on consumers, IMO. Electrical distributors are some of the laziest, worst-run companies in the country. Half of them can't even do billing right; I know dozens of people who have been double charged or never charged or charged for their neighbor- nobody wants to read their electrical bill, so nobody cares. The average US household spends ~$2450 on electricity annually, and the amount you can save for how complicated it is is just below that mental threshold.
I don't see it getting better without legislation. Most obviously, a push for subsidized smart meters that don't use 1930s tech to measure electricity. Then a standardized (extensible) API and/or reporting requirement, so that devices can know the current price of electricity. A standard for transmitting that info over the house circuits themselves, if you're feeling fancy. Direct-to-consumer subsidies from grid authorities for things like ancillary services, power factor correction, and frequency stabilization.
Easy as in simple, but I think this dramatically skews the price viability. If I need 2x overcapacity then the price to the consumer is 2x per MW of base capacity, and it’s no longer viable to use solar over gas.
> It's definitely not a tech problem
I disagree. I’m aware of some existing options, my claim is that with a bigger delta, more options become viable to research and implement. For example there was a thread recently where we discussed modulating energy usage in aluminum smelting, which requires a new design for the furnace to keep the temperatures stable. (This tech already exists, but AFAICT it’s not cost-effective to deploy widely.)
There are lots of industrial processes which could conceivably modulate their power consumption, but it’s not currently cost-effective to even design these improvements at current levels of peak premium.
Tech is downstream of incentives, is what I am saying, and price signals can be a good incentive; many claim that 100% renewable is not viable because of the cost of closing that last 1% of daily variability, I am hypothesizing that the system as a whole could, with appropriate price signals, build the tech to make the demand curve much more mutable.
This gets at seasonal variability; if we have a week with lower energy production, then the peak-premium goes up, and maybe we turn off the marginal industrial, residential, and commercial consumers.
This is an interesting concept. We used to live in a way that was much more directly connected to the amount of energy we could capture each year; in a good year, everyone would have a lot of food, and in a bad year, there would be famines.
Making ourselves dependent on an energy harvest seems like a way to intentionally reintroduce the problem of mass deaths when the harvest doesn't turn out as hoped. Is that something we want?
Merely mention 'hydrogen' and they go into a tirade.
Though really, the biggest issue is apathy.
If you look at cost optimized solutions to providing steady electrical output from solar + wind in high latitudes, hydrogen is strongly featured. It becomes much more expensive to use just short term storage like batteries.
Hydrogen is often useless, but it's sometimes essential. Hydrogen critics bleat about efficiency without understanding the limits of that argument.
P.S. even if the numbers have been superseded, MacKay's original book is still worth reading because it's so fantastically clear in how it lays out the basis for estimation.
His passing is a real loss to UK science.
I got to see him speak once, very engaging and his passion was clear.
He says that among the projects that tend to be on time and on budget, roads, solar and wind are three of them. While he doesn’t say it, I read this as “all large successful projects start as small successful projects”. Once you’ve built 10 miles of road the next 10miles is mostly more of the same, subsurface conditions notwithstanding. Once you’ve installed three wind turbines in a field installing the rest looks much the same (getting the first one in required solving a bunch of transportation problems of course). The teams just get a little faster with each one, because they are iterating on a pattern they already know.
You try to build a nuclear power plant and it might not show up until after the politicians who pushed for it to be built have retired. Which means it might not show up at all because all of the skin has left the game. But if I try to cap my career as governor with a new wind or solar farm? I may actually get to cut the ribbon.
It makes me feel a bit better about our prospects that solar and wind are easier logistical problems than repeating the old patterns.
You can try to reactivate nuclear power plants that were shut down.
I state I live in (New York) closed 2 nuclear reactors in 2020 and 2021, each providing more than 1 GW of clean electricity. Both reactor were closed because of political pressure. If we were to apply the reverse political pressure, I think we could have them up and running in 5 years, if not sooner.
It's closer to the model used for nuclear submarines which have a better track record in terms of delivery to time and cost than civilian nuclear power.
I feel like most of this is barking up the wrong tree though. If you're trying to get people to use a new power plant type, you may be looking at small cities and large towns. At which point you might be comparing your situation to how much power you can get from the nearest metropolis via high tension power lines, and how much more secure your winter power supply would be if you didn't rely on those lines working.
And a cursory search suggests those can vary from 40MW to just under 600.
So can you sell Geneseo a 50 MW reactor/solar/wind farm?
Also - what about the geopolitics. The reason solar is cheap is because china. Do we want all our energy needs to be dependent on China? Although we are still building Hinkley Point C with Chinese investment anyway so ¯\_(ツ)_/¯
Using nuclear to compensate doesn’t make any sense. If you need to replicate your wind production capacity with nuclear (keep in mind wind, even offshore, can go to zero) and given that nuclear costs the same thing whether you switch it on or not, you might as well spare the construction of wind farms and go full nuclear.
As the author points, energy storage is uneconomical.
I am also surprised with the idea that electricity consumption will go down. With heating going electric, cars going electric, and mass immigration (hence population growth), that seems counter intuitive to me.
I’m hoping grid scale batteries follow a similar cost reduction to solar but the latest projects which have been built do not inspire confidence. But for some reason this issue is always hand-waved away by the most ardent proponents of renewables.
People treat this issue like a football match in that they cheer for their favourite team and ignore their failings.
If the UK had another 10 or 15 of those (perfectly doable and economic) that would provide ~8 hours of storage. That would be enough by itself to provide a 94-96% solar/wind powered grid.
For that last 3-4% of seasonal/peaking storage and backup nothing beats hydrogen yet. It's only 50% efficient roundtripped but it stores very cheaply for very long periods.
If you turned all solar and wind energy into hydrogen and only burned that for electricity it would still be cheaper than power from a nuclear plant.
Batteries are probably best used in cars or places which can't connect to pumped storage (which are pretty rare).
Both hydro storage and hydrogen have the same problem: not all locales have elevated valleys or underground caverns. I’ve heard some talk about using natural gas pipelines for hydrogen storage, but that has its own issues.
The LCOE for nuclear power is roughly ~5x that of solar and wind.
If you assumed that the capex and running costs to electrolyze and store 1GWh of power in an underground cavern is roughly the same as it takes to produce it, and that it's about 50% efficient when roundtripped then that would make the stored wind and solar roughly 3/5ths of the cost of nuclear power.
This still makes it a lot more expensive than peaking with natural gas. Just because hydrogen storage is competitive with nuclear power doesn't mean it's economically competitive in general. Nuclear power is just way more expensive than people assume, and generally won't get built without a lot of really lavish subsidies (e.g. like Hinkley Point C getting paid 3-5x as much for the same kwh of electricity as everybody else).
100 years is, by the way, ludicrously optimistic for a nuclear power plant lifespan. If we ran nuclear power plants that long past their projected lifespans then quite a few would go kaboom.
Check the link above.
>LCOE doesn't account for nuclear running much longer than 20-40
I can assure you that LCOE does not assume a nuclear power station lasts 20 years. It assumes a normal lifetime - neither too short nor a lifetime that would induce catastrophic risks (e.g. 100 years).
I've seen no evidence that standard LCOE models under or overestimate solar/wind/nuclear power lifetimes at different rates and it would be presumptuous to assume that they do without evidence.
Coz there is a very good reason why it gets built even when it is not economic.
[1] just look at wind production in the UK for this year or last year https://gridwatch.co.uk/
>This is an important point, because some renewable critics say that about a week’s worth of storage is needed
https://reneweconomy.com.au/much-storage-needed-solar-wind-p...
Something a lot of people forget or don't notice is that solar and wind anticorrelate. The model takes this into account and doesn't assume a 100% wind grid. That would require a lot more storage, but isn't realistic.
There will be some differences between Australia and the UK but it's not going to be hours vs. weeks.
It currently is a cold winter day in the UK and solar is right now providing 1.8GW of power - almost as much as two small nuclear power plants.
Do you have a more realistic mathematical model than the one above to cite?
If I can find examples just looking at one year's worth of data, how is this going to survive a once in 25 or 50 years weather event with no wind for something like 3 weeks?
[1] https://gridwatch.co.uk/
There is definitely a need for a model. Even more so if you start making wild claims about the weather.
>how is this going to survive a once in 25 or 50 years
You think the sun going out and the wind stops blowing for an entire month once every 25 years? That's quite a brave claim.
>And last winter was an unusualy warm winter.
I remember. Solar power was unusually productive last winter, heating requirements were a lot lower than usual. The wind didn't stop blowing for a month though.
The article says energy consumption (which includes electricity and also fossil fuels burned for heating/transportation) will go down, but electricity consumption will go up.
But even the rise in electricity consumption will be tempered by the fact that electrical end uses are so much more efficient than fossil fuel burning equivalents, in particular for heating and transportation.
Heat pumps can move 3-4 times more energy than they consume, and EVs are likewise far more efficient at turning energy into motion.
Last year I was renting a house built in the early 2000s. It was poorly insulated and inside we had the thermostat set to 18c, yet our gas bill for November was €400. Over Christmas I moved into a newer house built in 2018 that was the same size, the thermostat was set to 22c and during February (which was much colder than November) the gas bill was under €200.
For residential, batteries are being added to appliances. Much cheaper (not panel upgrades) and much larger potential market size (apartments, condos) than "powerwall" style storage. These will follow the standard replacement rate of appliances, accelerated by incentives.
Sodium batteries are just starting to hit scale. Of course they'll replace lithium for stationary storage. I don't know enough to predict how big of a role they'll have in grid scale systems. Thermal batteries will be super competitive.
And of course, there's many more aspirants. Like flow batteries.
https://shkspr.mobi/blog/2021/03/1-year-of-edent_solar-we-ar...
Rooftop solar is extremely viable in the UK.
Renewable capacity factor isn't predictable on a daily basis. You need a backup - which basically means having an entire grid's worth of gas turbine/coal/hydro plants in storage for a few hours a year, or having grid-scale storage. That's not really a solved problem (even the biggest batteries are a couple of orders of magnitude too small to provide an entire country's electricity supply).
Electric vehicles (with vehicle to grid) will help. But even then, we need way, way more long term storage than we currently have.
You burn green hydrogen in turbines. A simple cycle turbine power plant is maybe 5% of the capital cost of a nuclear power plant of the same rated power output. And because its used so infrequently, the fuel costs are quite tolerable.
The hydrogen is stored in salt formations, which the UK has in sufficient quantity.
There are other more complicated e-fuel schemes with the same effect.
The UK should be exporting green energy. Which means installing two or three times the capacity projected, plus as many nuclear power stations as we can throw into the mix.
For a nation that relies upon imported food, it is vital that our exports are also required by the rest of the world. Being a major green energy exported - particularly if we can bottle it as synthetic liquified methane - is the way the UK needs to go.
You don't need storage if you build big. Those who import from you need the storage.
There is also quite a high correlation with the weather across Europe, and when it is windy in the UK, chances are it is also windy in most of the major countries the UK could export energy to.
As for nuclear, again, it doesn't make sense to use it to compensate for wind. Just make it run whether the wind blows or not, and save the cost of a wind farm.
Europe has the geology for cheaply storing many petawatt hours of hydrogen.