Nice work. When you add grid requirements to support wind the numbers will skew even further in favor of CCGT. Scotland being a topical example at the moment. They can't utilize all the wind because it is too far from the demand and the network capacity is insufficient to transport it. Building out the network is another source of emissions and adds complexity.
you assume the turbines will not experience a “blade liberation event.” For more on that, see my substack on what happened this summer in Nantucket.
Second, wind projects are facing fierce opposition from landowners and elected officials all over the world. The economics and o&m issues are key. But land-use conflicts are constraining wind expansion.
I think 20 years is generally the best case, which does not factor in damage from bad weather and very high winds and gusts. I haven’t read any reports or studies on actual lifespan experience for the various manufacturers.
Yes, good analysis. I don't think you have made any (material) errors, not least because your calculations produce the same overall outcome as we have worked on here:
Lots of people have done the math - physicists, engineers, etc. You can find them here on Substack since it's a world not utterly dependent on utopian dreams of zero emissions. And each brings his/her own calculations based on theoretical or real world costs. Comparing numbers, basis of reasoning, and assumptions will inevitably make evident what is reality. Have a it.
As you're aware of "Lots of people have done the math", can you please point me to just one who has compared wind + backup vs. combined-cycle gas? If there is a legit study I'll happily put it at the top of my post.
If you use gas turbines as backup then gas costs will be based on spot market pricing whereas a pure gas system can negotiate forward gas contracts at a more favorable rate.
Commendable work. No energy systems are cheap, emission free or everlasting. The actual cost of your scenario (gas+wind) has the main focus on CO2 emissions vs cost. How do we find the true cost of such a combination, when the cost of grid upgrades, cabling to wind farms, mineral mining for copper, aluminium etc are not included in the calculation. Intermittent energy systems require huge peak capacities, as people will scramble to start the charging of their EVs, cooking, washing, starting their tumble driers etc when the weather conditions are favorable. So what if the entire grid system needs to be upgraded with larger cable dimensions? What about transformer capacities?
Copper mines are being depleted, as in Peru, where the quality of copper ore is decreasing, requiring more mining efforts (energy, cost) to mine the copper. This is just one example. Listen to Mark Mills from Skagenfondene https://youtu.be/sgOEGKDVvsg?si=psSGWQRtteK6gkqF covering the global mineral situation.
Before I moved into the “pipeline world” I worked in the electrical industry. It’s refreshing to see your analysis, which seems on the mark. I’ve always advocated trying different generation and storage alternatives, but they need to be applied thoughtfully and not “forced”. Kudos to your commentators adding transmission and distribution effects as well pointing out the fallacy of CO2 concerns.
60 %? In a dream maybe, here in Alberta we have what’s considered a good wind resource and our capacity factor is ~35%. Has little to do with the turbine, all geography
when considering the carbon footprint of the turbines, are all manufacturing costs considered? even down to the diesel it takes for transportation and installation? it seems to me that once all is said and done that the long term EROI/carbon emission is heavily offset by the manufacturing, transportation, installation and maintenance.
perhaps we need a reliable equation that takes all factors into consideration going forward with these “renewables”. and that includes the subsidies.
Don't forget the necessities of constant load balancing by normally rotating energy sources, small diesel gennies attached to turbines, highly toxic fluorocarbons and lubricants inside turbine mechanisms some of which inevitably then leak onto the ground (potential ground water contamination), the necessity of continued mining of a vast array of minerals and metals to support these far-flung wind farms, vastly expanded copper mining, and of course an utterly undependable power source upon which hundreds of millions of people must rely for REGULAR generation. The UK is right now going through gyrations over its lack of nukes and hydrocarbon based energy because the sun and wind don't play ball. Impossible to predict spot market electricity costs are the effect of this problem. Their little island could be coated in turbines to little effect when wind drought occurs.
good points. in our equations do we even factor in the energy and emissions created in the mining and transport of these minerals? I doubt it.
I can see a hypothetical wherein the creation and maintenance of renewables ends up creating more carbon emissions than FFs used for the same purposes.
...copper being an outsize necessity among metals that are mined. How many miles of cable must be run to get wind farm output to the grid? Thousands and thousands.
I'm limited by what I can find on the web. And then have to assume that when they say full measure, it is everything. These are good questions for anyone who's got the time & knowledge to figure this all out. Unfortunately, that's not me.
Robert Bryce is a great follow here for his knowledge on energy. I believe he restacked your post (that’s how I found it). another good follow is Doomberg here on Substack.
Your basic error is to treat this as a side-by-side comparison for a fixed system. This is not reflective of reality, in which we have distributed demand served by a network of different supply points. None of the supply points has 100% reliability, and therefore to serve all load conditions to some given level of LOLP (loss of load probability) the system maintains reserves. How much of a reserve margin is required depends on the characteristics of the load and the characteristics of the supply network, with each supply point having its own particular reliability parameters. In a real-world system, you do not have and do not need 100% reserve margin.
Put another way, your CCGT case has a higher LOLP than your wind + SC. Whenever your CCGT is down for maintenance and whenever it fails to run (which it will with non-zero probability), you are completely unable to serve load in the CCGT case. Thus the side-by-side comparison is a false one.
Since every real-world system has to maintain reserves, those reserves are available to fill in the gaps when one or more CCGTs (or coal or nuclear plants) is unavailable to run and also to cover shortfalls from expected generation from wind farms. They are a shared cost for the entire system, and you can only evaluate relative cost-effectiveness at the system level.
What this means in practice is that the answer to your question is: it depends. Introducing a small amount of wind into a well-diversified generation system can lower overall costs and emissions. As the proportion of wind in the system rises it becomes relatively less valuable. If the wind is concentrated in large chunks, or if there is high correlation between the available wind resource at different geographical points across the system, wind requires relatively more reserves. You have to understand the entire system to know whether or not adding incremental wind will increase or decrease total system costs.
One final point. At $3 gas, it becomes very hard for any other resource to compete, but gas is not always at $3. At $10 the picture looks very different. In the U.S. we have been fortunate overall to have had policy somewhat supportive of gas production. Even so, we have seen big swings in gas prices periodically. Your view of the relative economics of gas-fired generation relative to every other type will thus depend in part on your expectations for future policy with respect to natural gas exploration and production as well as transportation (i.e. the ability to build pipelines).
You are correct on the gas price issue. But with single-cycle having an efficiency of 35% and combined cycle having an efficiency of 60%, if the wind turbines are spinning 45% of the time or less (quite common), the gas price is a push.
You are correct that they don't match a set of wind turbines to a specific gas plant. But at the end of the day, if the wind stops somewhere, there will be a gas turbine spun up. It is a fair measure.
The same is true for the CCGT plant in your comparison. When it does not run (whether for planned or forced outage), another gas turbine will be spun up. This is simply a reflection of the fact that in most modern systems gas turbines perform the balancing function. But it does not follow that every resource needs 100% back-up to achieve a given level of system reliability or that all resources require the same level of reserves or that this factor itself is solely dependent upon generation technology. All of these are a function of the overall system and without detailed knowledge of the system it is very hard to generalize accurately.
To get your side-by-side comparison to be closer to a fair model of reality, you would need to factor down the cost of the SC (or other) capacity you are assuming for back-up to the wind plant and add to your CCGT cost the cost of the SC (or other) capacity needed to back it up when in outage. The reserve factor will be higher for the wind plant will be higher than for the gas plant, but the former will not be 10% and the latter will not be zero. Exactly what they should be, however, can only be determined by understanding the total system. Even this comparison will still be a simplification. In fact, the back-up to wind will generally not be SC but CCGT. CCGTs are nowadays very flexible, and in most well-managed systems SCs are only rarely used - typically on hot days when a major generating source (coal, nuclear) has a problem and trips.
I can assure you from my experience in the business that gas prices matter a lot to your optimal resource plan!
What I have read is the for each 1.0 MW of intermittent power, they need 1.1 MW of backup power. Why the extra 0.1, I haven't seen that explained in detail. But it's definitely not the less than 1.0 that you say.
Can you share your source for that, as it seems way off for anything other than a system that is near 100% wind with no resource diversification?
You can get a very rough, real-world check by looking at ERCOT (Texas), which has accepted a lot of renewables over the past couple of decades.
In 2000 their peak load was around 55GW, implying 63-63 GW of installed capacity at 12-15% reserve margin. At that time they had very little wind on the system.
In 2024 peak load was about 85GW with 104GW of installed capacity, about 38% of which is renewables (roughly 2:1 wind:solar) - say, 40GW.
So non-renewable (primarily gas with some coal and minor other) is 64GW.
In other words, over a quarter of a century they have added 30GW of demand and met essentially all that incremental demand through the addition of renewables. The non-renewable fleet is about the same size as it was back then.
Of course, the fossil fleet is much more efficient and flexible than it was back then, and they also have some 7GW of batteries for short-term support. But clearly, this is nowhere close to requiring 1.1 MW of additional back-up per incremental MW of renewables.
You could argue that the reliability of the system is not as good as it was back then, but that is debatable.
I am not suggesting that the ERCOT resource mix is the low-cost mix - I'd guess that by now they are indeed over-penetrated with wind and solar. But wholesale prices remain generally lower there than in most other regions of the country, so they are certainly not high cost.
In any case, it strongly suggests that given the right system planning and decent wind resource you can absorb a lot of wind onto a system without driving cost way up, though of course the value of the tax credits have been passed through to the market there.
All worth mentioning to disabuse the novice reader on such topics (me), but still, I find the analysis useful and compelling. Many kinds of systems are not closed, and yet we learn a lot from analysis with some imperfect boundaries and assumptions. Thanks for your comments and to the author for the work to share his work.
Nice work. When you add grid requirements to support wind the numbers will skew even further in favor of CCGT. Scotland being a topical example at the moment. They can't utilize all the wind because it is too far from the demand and the network capacity is insufficient to transport it. Building out the network is another source of emissions and adds complexity.
Very good work here. Attaboy.
I’d suggest two more points:
you assume the turbines will not experience a “blade liberation event.” For more on that, see my substack on what happened this summer in Nantucket.
Second, wind projects are facing fierce opposition from landowners and elected officials all over the world. The economics and o&m issues are key. But land-use conflicts are constraining wind expansion.
Again, good work.
Great breakdown! Just one small thing to add: the lifespan of wind turbines is typically estimated at 20 years, at least in Europe.
I think 20 years is generally the best case, which does not factor in damage from bad weather and very high winds and gusts. I haven’t read any reports or studies on actual lifespan experience for the various manufacturers.
Nice analysis. Question: wouldn’t a higher wind at full capacity % (e.g., 70% v 50%) shorten the lifespan of the windmill system?
That's a good point and my guess is likely.
Yes, good analysis. I don't think you have made any (material) errors, not least because your calculations produce the same overall outcome as we have worked on here:
https://sites.google.com/view/the-lpf/home
Also here: https://thenewreformer.uk/2024/10/13/not-shooting-the-lights-out-energy-policy-that-might-actually-work-part-i/
Wonder why no one has performed this kind of analysis before 🤔. Thanks for putting in the work on it.
Forget fusion - we already have fission and it ROCKS. Check it out:
https://xkcd.com/1162/
Lots of people have done the math - physicists, engineers, etc. You can find them here on Substack since it's a world not utterly dependent on utopian dreams of zero emissions. And each brings his/her own calculations based on theoretical or real world costs. Comparing numbers, basis of reasoning, and assumptions will inevitably make evident what is reality. Have a it.
As you're aware of "Lots of people have done the math", can you please point me to just one who has compared wind + backup vs. combined-cycle gas? If there is a legit study I'll happily put it at the top of my post.
See my post further along with several references.
Can I add one further thought to the analysis?
If you use gas turbines as backup then gas costs will be based on spot market pricing whereas a pure gas system can negotiate forward gas contracts at a more favorable rate.
Presumably that will swing the balance further.
You didn’t get it wrong and your analysis was very generous to the wind assumptions
Commendable work. No energy systems are cheap, emission free or everlasting. The actual cost of your scenario (gas+wind) has the main focus on CO2 emissions vs cost. How do we find the true cost of such a combination, when the cost of grid upgrades, cabling to wind farms, mineral mining for copper, aluminium etc are not included in the calculation. Intermittent energy systems require huge peak capacities, as people will scramble to start the charging of their EVs, cooking, washing, starting their tumble driers etc when the weather conditions are favorable. So what if the entire grid system needs to be upgraded with larger cable dimensions? What about transformer capacities?
Copper mines are being depleted, as in Peru, where the quality of copper ore is decreasing, requiring more mining efforts (energy, cost) to mine the copper. This is just one example. Listen to Mark Mills from Skagenfondene https://youtu.be/sgOEGKDVvsg?si=psSGWQRtteK6gkqF covering the global mineral situation.
Nicely done. Everyone's better off with honest and forthright research such as what you have done and explained. Thank you.
Before I moved into the “pipeline world” I worked in the electrical industry. It’s refreshing to see your analysis, which seems on the mark. I’ve always advocated trying different generation and storage alternatives, but they need to be applied thoughtfully and not “forced”. Kudos to your commentators adding transmission and distribution effects as well pointing out the fallacy of CO2 concerns.
60 %? In a dream maybe, here in Alberta we have what’s considered a good wind resource and our capacity factor is ~35%. Has little to do with the turbine, all geography
Several (oldish) references on this topic:
"CO2 Emissions Variations in CCGTs Used to Balance Wind in Ireland"
http://euanmearns.com/co2-emissions-variations-in-ccgts-used-to-balance-wind-in-ireland/
"Cost and Quantity of Greenhouse Gas Emissions Avoided by Wind Generation" By Peter Lang
https://bravenewclimate.com/files/2009/08/peter-lang-wind-power.pdf
"Does wind power reduce carbon emissions?" by Barry Brook, references Lang above.
https://bravenewclimate.com/2009/08/08/does-wind-power-reduce-carbon-emissions/
"Why solar and wind won’t make much difference to carbon dioxide emissions"
https://blog.oup.com/2017/10/solar-wind-energy-carbon-dioxide-emissions/
"Wind Integration: Incremental Emissions from Back-Up Generation Cycling (Part V: Calculator Update)" By Kent Hawkins
https://www.masterresource.org/wind-power/wind-integration-incremental-emissions-from-back-up-generation-cycling-part-v-calculator-update/#more-7271
There was also a Bendix study of wind energy with Coal backup in Colorado, but I can't seem to find the link to it.
Edit: Bentek, not Bendix. https://docs.wind-watch.org/BENTEK-How-Less-Became-More.pdf
Thank you!
great write up.
when considering the carbon footprint of the turbines, are all manufacturing costs considered? even down to the diesel it takes for transportation and installation? it seems to me that once all is said and done that the long term EROI/carbon emission is heavily offset by the manufacturing, transportation, installation and maintenance.
perhaps we need a reliable equation that takes all factors into consideration going forward with these “renewables”. and that includes the subsidies.
Don't forget the necessities of constant load balancing by normally rotating energy sources, small diesel gennies attached to turbines, highly toxic fluorocarbons and lubricants inside turbine mechanisms some of which inevitably then leak onto the ground (potential ground water contamination), the necessity of continued mining of a vast array of minerals and metals to support these far-flung wind farms, vastly expanded copper mining, and of course an utterly undependable power source upon which hundreds of millions of people must rely for REGULAR generation. The UK is right now going through gyrations over its lack of nukes and hydrocarbon based energy because the sun and wind don't play ball. Impossible to predict spot market electricity costs are the effect of this problem. Their little island could be coated in turbines to little effect when wind drought occurs.
good points. in our equations do we even factor in the energy and emissions created in the mining and transport of these minerals? I doubt it.
I can see a hypothetical wherein the creation and maintenance of renewables ends up creating more carbon emissions than FFs used for the same purposes.
...copper being an outsize necessity among metals that are mined. How many miles of cable must be run to get wind farm output to the grid? Thousands and thousands.
I'm limited by what I can find on the web. And then have to assume that when they say full measure, it is everything. These are good questions for anyone who's got the time & knowledge to figure this all out. Unfortunately, that's not me.
Robert Bryce is a great follow here for his knowledge on energy. I believe he restacked your post (that’s how I found it). another good follow is Doomberg here on Substack.
Yeah - his books was one of my inspirations for this post.
This is imports t scientific thought and enquiry
Your basic error is to treat this as a side-by-side comparison for a fixed system. This is not reflective of reality, in which we have distributed demand served by a network of different supply points. None of the supply points has 100% reliability, and therefore to serve all load conditions to some given level of LOLP (loss of load probability) the system maintains reserves. How much of a reserve margin is required depends on the characteristics of the load and the characteristics of the supply network, with each supply point having its own particular reliability parameters. In a real-world system, you do not have and do not need 100% reserve margin.
Put another way, your CCGT case has a higher LOLP than your wind + SC. Whenever your CCGT is down for maintenance and whenever it fails to run (which it will with non-zero probability), you are completely unable to serve load in the CCGT case. Thus the side-by-side comparison is a false one.
Since every real-world system has to maintain reserves, those reserves are available to fill in the gaps when one or more CCGTs (or coal or nuclear plants) is unavailable to run and also to cover shortfalls from expected generation from wind farms. They are a shared cost for the entire system, and you can only evaluate relative cost-effectiveness at the system level.
What this means in practice is that the answer to your question is: it depends. Introducing a small amount of wind into a well-diversified generation system can lower overall costs and emissions. As the proportion of wind in the system rises it becomes relatively less valuable. If the wind is concentrated in large chunks, or if there is high correlation between the available wind resource at different geographical points across the system, wind requires relatively more reserves. You have to understand the entire system to know whether or not adding incremental wind will increase or decrease total system costs.
One final point. At $3 gas, it becomes very hard for any other resource to compete, but gas is not always at $3. At $10 the picture looks very different. In the U.S. we have been fortunate overall to have had policy somewhat supportive of gas production. Even so, we have seen big swings in gas prices periodically. Your view of the relative economics of gas-fired generation relative to every other type will thus depend in part on your expectations for future policy with respect to natural gas exploration and production as well as transportation (i.e. the ability to build pipelines).
You are correct on the gas price issue. But with single-cycle having an efficiency of 35% and combined cycle having an efficiency of 60%, if the wind turbines are spinning 45% of the time or less (quite common), the gas price is a push.
You are correct that they don't match a set of wind turbines to a specific gas plant. But at the end of the day, if the wind stops somewhere, there will be a gas turbine spun up. It is a fair measure.
The same is true for the CCGT plant in your comparison. When it does not run (whether for planned or forced outage), another gas turbine will be spun up. This is simply a reflection of the fact that in most modern systems gas turbines perform the balancing function. But it does not follow that every resource needs 100% back-up to achieve a given level of system reliability or that all resources require the same level of reserves or that this factor itself is solely dependent upon generation technology. All of these are a function of the overall system and without detailed knowledge of the system it is very hard to generalize accurately.
To get your side-by-side comparison to be closer to a fair model of reality, you would need to factor down the cost of the SC (or other) capacity you are assuming for back-up to the wind plant and add to your CCGT cost the cost of the SC (or other) capacity needed to back it up when in outage. The reserve factor will be higher for the wind plant will be higher than for the gas plant, but the former will not be 10% and the latter will not be zero. Exactly what they should be, however, can only be determined by understanding the total system. Even this comparison will still be a simplification. In fact, the back-up to wind will generally not be SC but CCGT. CCGTs are nowadays very flexible, and in most well-managed systems SCs are only rarely used - typically on hot days when a major generating source (coal, nuclear) has a problem and trips.
I can assure you from my experience in the business that gas prices matter a lot to your optimal resource plan!
What I have read is the for each 1.0 MW of intermittent power, they need 1.1 MW of backup power. Why the extra 0.1, I haven't seen that explained in detail. But it's definitely not the less than 1.0 that you say.
Can you share your source for that, as it seems way off for anything other than a system that is near 100% wind with no resource diversification?
You can get a very rough, real-world check by looking at ERCOT (Texas), which has accepted a lot of renewables over the past couple of decades.
In 2000 their peak load was around 55GW, implying 63-63 GW of installed capacity at 12-15% reserve margin. At that time they had very little wind on the system.
In 2024 peak load was about 85GW with 104GW of installed capacity, about 38% of which is renewables (roughly 2:1 wind:solar) - say, 40GW.
So non-renewable (primarily gas with some coal and minor other) is 64GW.
In other words, over a quarter of a century they have added 30GW of demand and met essentially all that incremental demand through the addition of renewables. The non-renewable fleet is about the same size as it was back then.
Of course, the fossil fleet is much more efficient and flexible than it was back then, and they also have some 7GW of batteries for short-term support. But clearly, this is nowhere close to requiring 1.1 MW of additional back-up per incremental MW of renewables.
You could argue that the reliability of the system is not as good as it was back then, but that is debatable.
I am not suggesting that the ERCOT resource mix is the low-cost mix - I'd guess that by now they are indeed over-penetrated with wind and solar. But wholesale prices remain generally lower there than in most other regions of the country, so they are certainly not high cost.
In any case, it strongly suggests that given the right system planning and decent wind resource you can absorb a lot of wind onto a system without driving cost way up, though of course the value of the tax credits have been passed through to the market there.
Sorry, it is said in passing numerous places but never with a why. It seems to be so basic everyone just uses it.
All worth mentioning to disabuse the novice reader on such topics (me), but still, I find the analysis useful and compelling. Many kinds of systems are not closed, and yet we learn a lot from analysis with some imperfect boundaries and assumptions. Thanks for your comments and to the author for the work to share his work.