Rerun the analysis, add gas. Remember that batteries can store gas AND solar. You’ll get as many nines as you want, and a pretty cheap system.
What’s not commented on a lot is if you want an ultra clean grid, there is a point where solar wind gas storage becomes less optimal than raw clean firm.
Path dependency is a thing, and you can get stuck!
But overall at some point at that tail that you flagged, the optimal system has a phase shift and you end up removing a lot of solar, gas, storage in exchange for clean firm!
Nice thinking. The most important lesson is that none of us know what form of energy generation is best in every possible circumstance. The answer is not to allow some "expert" agency decide, but to let the market decide. Eliminate all subsidies for all energy sources.
My takeaway on these types of analysis is something like the 80/20 rule. Except it’s closer to 80/50. Assuming your solar is firmed with gas peaker plants instead of batteries, you can do roughly 80% decarbonization for half the cost as 100% decarbonization.
A similar rule applies to sizing electric heat pumps with backup from a natural gas furnace. In a climate zone 5 (Boston, Chicago, etc.) sizing your heat pump to meet load for all but the 10% coldest hours of the year provides 90% decarbonization at half the cost.
You can even apply the rule to plug-in-hybrid electric vehicles. If 80% of your driving is in trips less than 50 miles, you can decarbonize (assuming clean electricity) 80% of your driving with a much smaller battery (at lower cost).
This rule occurs because electrification technologies have expensive capacity and low operating costs. The capacity cost of fossil fuel burning equipment (engines, furnaces, turbines) is comparatively lower, but needing to burn fuel to operate makes operating costs higher.
One detail not in this analysis is the upfront carbon emissions from equipment manufacturing. Going from 90% to 99% solar may require twice as many panels and batteries. That’s twice as much equipment that required carbon intensive materials processing and manufacturing. At some point on a lifecycle carbon perspective that includes carbon from equipment sourcing, you’re probably better off burning gas for several hours a year.
Side note: it’s a shame that nuclear got expensive in the 70s. If the build streak had lasted a bit longer, the US may have had 2x or even 4x as much clean firm generation as it has now which would make the decarb problem that much easier!
Analyzing a solar+battery only grid doesn't make sense, even if it was just renewables you'd mix wind and solar which are also anti-correlated (we often have wind when we don't have solar) plus if its the grid then you've got balances between west and east, north and south which have different patterns of wind and solar. And then you add in a little bit of gas backup - expensive to operate, cheap to build. If we are talking decentralized then similar logic applies, for example household batteries supply a lot of those looking for high reliability, as do backup generators.
What you have described is a typical monoculture > all-electric. Once you step out from under this binary logic straight jacket and into that of a biotope of contextual interrelations the picture becomes far more interesting and way more feasible.
The weight of the building, the positions and sizes of the windows, the insulation and form factor in relation to the geo-position, the occupant's behavioural consumption and the possible load reductions and enthalpic shift from high to lower forms, from electric to thermal.
One can even go further to include thermal salt battery storage and go beyond PV to include bio (CHP), wind, hydro, solar and geothermal energy generation, depending on the location. And smart to do this as a group.
Some good data. The notion that comes across to me is “no matter how good renewable energy becomes (efficiency, cost, output), it will never be enough.” That has been the tagline of the petroleum industry for decades. As renewables improve in technology and cost efficiency the demand for energy also is increasing from growing population to data centers. So the goal posts keep moving. I concur that at the present certain applications and timing make best use. But petroleum based energy escalates the costs of climate change which could be catastrophic at some tipping point—and then how do the calculations compare? And what are we sacrificing for “cheaper” non-renewable energy? It could be our planet.
You're right that getting to 100% based on solar plus batteries only in one place would be incredibly expensive, but that's not really an important question, right?
Getting to 50% is trivial and even 80% should be quite doable (I'm bullish on grid scale battery chemistries, given most research into batteries has gone into optimising size and weight for gadgets, not raw cost per watt-hour).
There are a couple of things that really help here:
1) Having a very large and interconnected grid substantially mitigates the cloud variation.
2) In many places (like the UK) there's more wind at night and during the winter, which is very convenient for these purposes.
3) A tiny proportion of gas dealers can go a long way to filling in the rest to three nines (and you can just carbon capture those few at reasonable total cost).
I think that solar (and wind) make a decent off-grid power supply for applications that, as you note, are generally good with coping with intermittent availability. I suspect that a 2x max load solar panel plus sufficient battery to smooth (say 1-2 hours of maxload ) will work well for a lot of use cases. You'd get an average of perhaps 6-8 hours of availability per day which could be quite enough if planned well.
I note that plenty of developing nations where power is intermittent adapt to doing stuff that needs power when the power is available and doing other stuff when it isn't.
A lot of temporary lighting for road works etc. is totally solar + battery powered. LED lights don't consume much power so you can get away with enough panels to charge a small battery to run ~24 hours in one hour of full sun
I hope at some point the global community can get trustworthy data from China on how green power supplies affect their grid. They are so far ahead of the rest of the world that it seems like the best place to see how it's likely to look everywhere else if we manage to catch up.
I'm skeptical of the cost increase analysis of Cembalest. Doug Sheridan at Energy Point Research has done an analysis that shows that costs almost double when you get to deeply decarbonized grids.
It would be worth comparing base assumptions. These are unabated costs as well so maybe that's part of it.
As you've rightly pointed out, it is a very bad assumption to simply think that costs go down over time so solar will eventually catch up. It won't go down for any installations already in place.
I think the other thing that is often overlooked is the capital intensity of alternative fuels. Even if you got to a breakeven cost, solar requires almost all of the capital up-front. These are nowhere near the same financing profiles and risk profiles. That is something that is often lost. It's literally as if you offset the fossil alternative by a solution which requires you to building a plant and buy all the fossil fuel UP-FRONT. Another reason which LCOE is a rotten measure.
Interesting, thanks. Hmm, is Doug estimating 100% solar? Because if so then I think a 2x cost increase is an *underestimate.* I must be misunderstanding something.
Note that Cembalest's estimate of 15–35% cost increase is only for increasing 30 points from a 40% non-carbon grid to a 70% one! Nowhere near 100%.
You're right that all the solar costs are up front, but financing costs are accounted for in all these estimates, I think. The fact is that the capex for solar is now pretty cheap per watt of nameplate capacity—just over $1/W in the US. Even if you divide by the capacity factor, I think it's competitive with many other energy sources.
When you bring in aggressive measures for Building Automation (controlling lights, HVACs and other equipment very tightly) and add good design, good windows, good insulation, you can knock down the power consumption of a typical building by 25%. This isn't cheap or easy, but it's a whale of a lot cheaper and easier than building more solar plants.
We waste an enormous amount of energy. But, no politician gets to put their name on a conservation project and the opportunities for graft are far lower, so nobody talks about doing the easy thing.
The obvious way to get more uptime is to allow the cost to vary. The vast majority of the time, the price of electricity will be low, in line with current prices, with some predictable variation over the course of the day and some predictable variation over the course of the seasons. Then there will be a small amount of time (a few hours a year up to maybe a few days a year) where prices spike quite a bit. People (and industrial customers) will use less electricity during those times, and backup sources like gas peaker plants, long-distance transmission lines, etc., will be incentived to provide a lot of extra power. We just need to clean up the regulatory environment to let the power of the free market to set market-clearing prices.
Yes, good point, and we should definitely do that. But IMO it's still a failure if electricity often spikes to prohibitively expensive levels. We should have energy that is affordable all the time.
The daily consumption of the US grid is about 12,000 GWh. The output of the world's largest lithium ion battery factory is 100 GWh. So with 120 years of output we could back up a whole day.
California already has enough batteries to back up several hours of the total state’s electricity use, and they haven’t been buying the majority of all batteries produced. So I suspect this is possible, even if we need a whole day of backup (which is unlikely).
We are conditioned to suppose that the cheapest solution will be the one with the lowest cost to produce instantaneous power sufficient to meet peak demand. And that, at utility scale, battery storage is too expensive to sufficiently address intermittency to significantly change the optimum configuration.
However, if you start with an entirely different set of assumptions....
Rerun the analysis, add gas. Remember that batteries can store gas AND solar. You’ll get as many nines as you want, and a pretty cheap system.
What’s not commented on a lot is if you want an ultra clean grid, there is a point where solar wind gas storage becomes less optimal than raw clean firm.
Path dependency is a thing, and you can get stuck!
But overall at some point at that tail that you flagged, the optimal system has a phase shift and you end up removing a lot of solar, gas, storage in exchange for clean firm!
And wind.
Nice thinking. The most important lesson is that none of us know what form of energy generation is best in every possible circumstance. The answer is not to allow some "expert" agency decide, but to let the market decide. Eliminate all subsidies for all energy sources.
For sure
My takeaway on these types of analysis is something like the 80/20 rule. Except it’s closer to 80/50. Assuming your solar is firmed with gas peaker plants instead of batteries, you can do roughly 80% decarbonization for half the cost as 100% decarbonization.
A similar rule applies to sizing electric heat pumps with backup from a natural gas furnace. In a climate zone 5 (Boston, Chicago, etc.) sizing your heat pump to meet load for all but the 10% coldest hours of the year provides 90% decarbonization at half the cost.
You can even apply the rule to plug-in-hybrid electric vehicles. If 80% of your driving is in trips less than 50 miles, you can decarbonize (assuming clean electricity) 80% of your driving with a much smaller battery (at lower cost).
This rule occurs because electrification technologies have expensive capacity and low operating costs. The capacity cost of fossil fuel burning equipment (engines, furnaces, turbines) is comparatively lower, but needing to burn fuel to operate makes operating costs higher.
One detail not in this analysis is the upfront carbon emissions from equipment manufacturing. Going from 90% to 99% solar may require twice as many panels and batteries. That’s twice as much equipment that required carbon intensive materials processing and manufacturing. At some point on a lifecycle carbon perspective that includes carbon from equipment sourcing, you’re probably better off burning gas for several hours a year.
Side note: it’s a shame that nuclear got expensive in the 70s. If the build streak had lasted a bit longer, the US may have had 2x or even 4x as much clean firm generation as it has now which would make the decarb problem that much easier!
Analyzing a solar+battery only grid doesn't make sense, even if it was just renewables you'd mix wind and solar which are also anti-correlated (we often have wind when we don't have solar) plus if its the grid then you've got balances between west and east, north and south which have different patterns of wind and solar. And then you add in a little bit of gas backup - expensive to operate, cheap to build. If we are talking decentralized then similar logic applies, for example household batteries supply a lot of those looking for high reliability, as do backup generators.
Keep up the good work!
Simply put, rich deposits of coal and oil will eventually run out and using inferior types and sources of fossil fuel will become very expensive.
We must keep using them now and begin a worldwide process of developing nuclear fission reactor technology before that time comes.
"Net zero" and "carbon neutral" and civilisation are mutually exclusive.
There is no real man-made global climate crisis outside of partisan political and economic thinking.
What you have described is a typical monoculture > all-electric. Once you step out from under this binary logic straight jacket and into that of a biotope of contextual interrelations the picture becomes far more interesting and way more feasible.
The weight of the building, the positions and sizes of the windows, the insulation and form factor in relation to the geo-position, the occupant's behavioural consumption and the possible load reductions and enthalpic shift from high to lower forms, from electric to thermal.
One can even go further to include thermal salt battery storage and go beyond PV to include bio (CHP), wind, hydro, solar and geothermal energy generation, depending on the location. And smart to do this as a group.
Some good data. The notion that comes across to me is “no matter how good renewable energy becomes (efficiency, cost, output), it will never be enough.” That has been the tagline of the petroleum industry for decades. As renewables improve in technology and cost efficiency the demand for energy also is increasing from growing population to data centers. So the goal posts keep moving. I concur that at the present certain applications and timing make best use. But petroleum based energy escalates the costs of climate change which could be catastrophic at some tipping point—and then how do the calculations compare? And what are we sacrificing for “cheaper” non-renewable energy? It could be our planet.
Petroleum is basically irrelevant for electricity!
Talking about propane, natural gas, fuel oil, coal— all carbon based. I should have been more specific.
You're right that getting to 100% based on solar plus batteries only in one place would be incredibly expensive, but that's not really an important question, right?
Getting to 50% is trivial and even 80% should be quite doable (I'm bullish on grid scale battery chemistries, given most research into batteries has gone into optimising size and weight for gadgets, not raw cost per watt-hour).
There are a couple of things that really help here:
1) Having a very large and interconnected grid substantially mitigates the cloud variation.
2) In many places (like the UK) there's more wind at night and during the winter, which is very convenient for these purposes.
3) A tiny proportion of gas dealers can go a long way to filling in the rest to three nines (and you can just carbon capture those few at reasonable total cost).
*gas peakers
I think that solar (and wind) make a decent off-grid power supply for applications that, as you note, are generally good with coping with intermittent availability. I suspect that a 2x max load solar panel plus sufficient battery to smooth (say 1-2 hours of maxload ) will work well for a lot of use cases. You'd get an average of perhaps 6-8 hours of availability per day which could be quite enough if planned well.
I note that plenty of developing nations where power is intermittent adapt to doing stuff that needs power when the power is available and doing other stuff when it isn't.
A lot of temporary lighting for road works etc. is totally solar + battery powered. LED lights don't consume much power so you can get away with enough panels to charge a small battery to run ~24 hours in one hour of full sun
I hope at some point the global community can get trustworthy data from China on how green power supplies affect their grid. They are so far ahead of the rest of the world that it seems like the best place to see how it's likely to look everywhere else if we manage to catch up.
I'm skeptical of the cost increase analysis of Cembalest. Doug Sheridan at Energy Point Research has done an analysis that shows that costs almost double when you get to deeply decarbonized grids.
https://www.linkedin.com/posts/sheridandoug_renewables-solar-energytransition-activity-7295064061501681665-XBkm
That's for a high-solar load region. He extended it to most of the lower 48 below. Obviously it gets worse.
https://www.linkedin.com/posts/sheridandoug_energy-renewables-solar-activity-7307732540495212544-uk7g
It would be worth comparing base assumptions. These are unabated costs as well so maybe that's part of it.
As you've rightly pointed out, it is a very bad assumption to simply think that costs go down over time so solar will eventually catch up. It won't go down for any installations already in place.
I think the other thing that is often overlooked is the capital intensity of alternative fuels. Even if you got to a breakeven cost, solar requires almost all of the capital up-front. These are nowhere near the same financing profiles and risk profiles. That is something that is often lost. It's literally as if you offset the fossil alternative by a solution which requires you to building a plant and buy all the fossil fuel UP-FRONT. Another reason which LCOE is a rotten measure.
Interesting, thanks. Hmm, is Doug estimating 100% solar? Because if so then I think a 2x cost increase is an *underestimate.* I must be misunderstanding something.
Note that Cembalest's estimate of 15–35% cost increase is only for increasing 30 points from a 40% non-carbon grid to a 70% one! Nowhere near 100%.
You're right that all the solar costs are up front, but financing costs are accounted for in all these estimates, I think. The fact is that the capex for solar is now pretty cheap per watt of nameplate capacity—just over $1/W in the US. Even if you divide by the capacity factor, I think it's competitive with many other energy sources.
When you bring in aggressive measures for Building Automation (controlling lights, HVACs and other equipment very tightly) and add good design, good windows, good insulation, you can knock down the power consumption of a typical building by 25%. This isn't cheap or easy, but it's a whale of a lot cheaper and easier than building more solar plants.
We waste an enormous amount of energy. But, no politician gets to put their name on a conservation project and the opportunities for graft are far lower, so nobody talks about doing the easy thing.
The obvious way to get more uptime is to allow the cost to vary. The vast majority of the time, the price of electricity will be low, in line with current prices, with some predictable variation over the course of the day and some predictable variation over the course of the seasons. Then there will be a small amount of time (a few hours a year up to maybe a few days a year) where prices spike quite a bit. People (and industrial customers) will use less electricity during those times, and backup sources like gas peaker plants, long-distance transmission lines, etc., will be incentived to provide a lot of extra power. We just need to clean up the regulatory environment to let the power of the free market to set market-clearing prices.
Yes, good point, and we should definitely do that. But IMO it's still a failure if electricity often spikes to prohibitively expensive levels. We should have energy that is affordable all the time.
BTW this echoes a post of mine from last year - https://ombreolivier.substack.com/p/batteries-and-energy-storage
The daily consumption of the US grid is about 12,000 GWh. The output of the world's largest lithium ion battery factory is 100 GWh. So with 120 years of output we could back up a whole day.
How many factories are there? What's the output of the entire global battery industry?
California already has enough batteries to back up several hours of the total state’s electricity use, and they haven’t been buying the majority of all batteries produced. So I suspect this is possible, even if we need a whole day of backup (which is unlikely).
We are conditioned to suppose that the cheapest solution will be the one with the lowest cost to produce instantaneous power sufficient to meet peak demand. And that, at utility scale, battery storage is too expensive to sufficiently address intermittency to significantly change the optimum configuration.
However, if you start with an entirely different set of assumptions....
https://www.youtube.com/watch?v=6zgwiQ6BoLA
https://23227526.fs1.hubspotusercontent-na1.net/hubfs/23227526/Energy%2BReports%2B-%2BMethodology-1.pdf