-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Except you can’t crank up wind or sunshine or hydro flow in the winter! How to handle excess winter heating demand with non-nuclear renewables seems like an open question. Maybe the seasonality would be reduced with all-electric heating, but that’s not obvious to me — in the graph quoted above the natural gas heating spikes in the winter are way bigger than the electric AC spikes in the summer… for the UK anyway.
How do you reconcile this with wanting us to move away from crude?You can tell these traders don't really have the conviction the high price of WTI portends. They keep selling on Wednesday mornings before the EIA supply status report comes out.
This week's report showed flat total supplies on top of last week's +10Mb. Crude inventories up a bit, product inventories down a bit. Still treading water at what could be considered historic highs.
No clue how the market will react. Clearly the global narrative of "short supplies" is just traders and other vested interests leveraging lingering covid disruptions to look like a shortage. I guess you could argue it's kinda the same thing, but they're calling it a sort of "new normal". Which is complete nonsense. We'll be swimming in crude any minute now.
We don't move away from fossil fuels by having a world awash in them, high crude prices will lead to demand destruction and push people towards alternatives. Swimming in crude and low prices will disincentivize any transition away from them. People need to stop consuming fossil fuels, and high prices jacking up costs for end users can be a tool for achieving that.
Last edited by a moderator:
petit_bateau
Active Member
In my opinion (and professional experience) to an extent you are wrong. You can crank them up ...... if you study and build in advance.
The right way to do it is to analyse the windspeeds and solar irradiance throughout the year, and then to build a generation system that best matches the demand through the year. So (say) in a typical northern European location (i.e. north of the air conditioner belt) the primary load is during the winter, and thus the generation mix ought to be skewed towards wind. As one moves southwards (towards the equator) the optimal generation mix ordinarily shifts to include more solar so as to better match the summer aircon loads. When fully doing analyses of this nature one also calculates the optimal amount of storage to include in the system, and always these analyses are conscious of cost. If one is really clever (and has both the relevant data, and the necessary industrial heft) one also analyses the optimal amount of long distance grid transportation (500-1500 miles or more) though when I have looked at the grid aspects the size of a typical (weather) pressure cell (i.e. high being the main concern, as this is the low wind cell) is inconveniently large . Thus, for example, the typical (UK) design-driving event is a stationary mid-winter cold high pressure settling over the UK, and unfortunately this is approximately the same geographic scale as the UK, and can as a low probability event be as long as 2-weeks. This means that the UK cannot internally stretch its grid to get beyond the low wind speed region. One needs continental-scale grids to make that stretch, and really only the Chinese have been doing that as part of their exploration of this subject. (Very good clients they are too).
Clearly one also adds in the hydro, fossils, nuclear, etc. to figure out the optimal system mix. One of course studies the variability extremely carefully. I have used stochastic methods but other (cleverer) people than I use deterministic methods.
In the UK, as heat loads electrify (with ASHP) then I am expecting the mid-Winter design-driving mix to skew even more heavily towards wind. However if climate change causes significant aircon use in Summer then maybe more solar will be required.
Overbuilding wind (and/or solar) is almost always cheaper than trying to do interseasonal storage wheneve rI've analysed things. For the UK case 5-10 days storage for the mid-Winter low tends to be about right, though of course that does depend on how much demand-side response one is prepared to accept (not a concept politicians like), or how much risk one is prepared to accept (again not something politicians like to pay for themselves). Most other northern hemisphere northerly (temperate) locations have typical outcomes whenever I've modelled them.
petit_bateau
Active Member
JohnKempReuters chartbook for gas prices
===
The chartbook presents time series for natural gas prices in both nominal and real terms over various time horizons for the major consuming regions in North America, Europe and Asia: https://tmsnrt.rs/3DhpLTy
IF YOU know anyone else who might like to receive best in energy and my research notes, they can add their emails to the circulation list using this link: http://eepurl.com/dxTcl1
===
(I'm not John Kemp by the way !)
===
The chartbook presents time series for natural gas prices in both nominal and real terms over various time horizons for the major consuming regions in North America, Europe and Asia: https://tmsnrt.rs/3DhpLTy
IF YOU know anyone else who might like to receive best in energy and my research notes, they can add their emails to the circulation list using this link: http://eepurl.com/dxTcl1
===
(I'm not John Kemp by the way !)
well, one thing that, to me,is absolutely critical, is for the US for example, is “spin up” wind and solar by at least a factor of 10x - 16x, or more, since natural gas is used 8x as much ~32 Quads gas vs ~4 Quads wind and solar, like Tony Seba says “super power” or generating 4x the needs (I generate 1.6x my needs of solar electricity)
2/3 of energy is wasted already,
Last edited:
I certainly agree that things are different if planned in advance, and if planning to overbuild. I meant more like, you can't say "looks bad next month, let's buy some extra wind and sun" like you can buy more coal/oil/gas. Somebody has to have already put the renewable installations in place. So if you've planned and built for the worst, great. If not (and we as a species certainly have not -- yet) and if a surprise comes along, oops.
Yep - in the UK where winters are relatively mild, electrifying heating with AHSPs will be a net gain in efficiency compared to burning fuels, even if that fuel is natural gas in power plants.
Good to see someone else say that overbuilding wind/solar is a good thing - I'm tired of the utilities complaining about an excess of renewables - the goal is to build enough renewable generation capacity so that we always have the ability to crank out more renewable energy should we need more power - not to get renewables to perfectly match load. That simply isn't possible - you need to have a buffer for reliability, whether that's an excess in production or grid-storage through some beans (batteries, hydro, etc). That's what traditional power plants do - most sit idle or partially used most of the time.
adiggs
Well-Known Member
Building excess renewable power generation is going to be all kinds of difficult because its a new paradigm (and paradigm shifts are .. hard).
The main paradigm that I see that needs to change, is that in all of human history, it was important to only generate as much energy as was going to be consumed in the activity. You don't hitch up 20 horses to tow your plow around the field when 2 will do the job in a perfectly adequate fashion. That's wasted energy getting them hitched and unhitched, and the incremental work performed by the 18 horses is effectively wasted.
You don't (routinely) burn enough wood in the fireplace to raise the internal temperature of the house to 100(F) - you'll just need to open some windows to let the excess out and make it comfortable again. The extra wood is wasted, and even children know this.
And we don't burn an extra unit of natural gas, over the unit of gas needed, to produce the electricity that the plant is currently contracted to provide. It's just wasted resource. And for some of us, this is such a difficult concept to wrestle with, its hard to even put the words together to describe such a nonsensical idea.
We don't, and we've never (that I can identify), intentionally overproduced energy for some strong economic reason.
The renewable energy world not only makes this reasonable, it will make this a requirement. And even desirable as the excess production will fill in when the production is less than perfect, and when production IS good, the excess will (eventually) be used to support business models that don't and can't exist today. We don't need to know what they are today; only to know that clever engineers and business people will find technologies and business models that are dependent on nearly free and intermittent energy.
It's a new way of thinking about energy and it needs to percolate out to all of the energy production, consumption, and transportation markets.
An interesting source of resistance to this view of energy will (and does) come from the environmental world. For climate change reasons, I see a common point of view that the way to lower carbon emissions is to burn less fossil fuels (sensible), and that means lower energy budgets for everybody in the world (sensible or not, it'll only happen by motivated individuals - not a societal level choice).
I see this as an instance where the only way to solve the problem, is to go through it. Bigger and bigger energy budgets, enabled by cheaper and cheaper, renewable energy resources. Whether its the best or not in some abstract sense, its the only way that might work (MHO).
This chart is very helpful for thinking about the problem of seasonal production.
Observe the ratio of peaks to trough. For gas the peak is nearly 3X the trough of 1X (most recently about 108 to 37, or 2.92X). Likewise the for electricity, the peak is about 1.3X to a trough of 1X (most recently about 26 to 20).
So gas has a substantially higher peak-to-trough ratio than electricity, 3.0X vs 1.3X. This means that seasonal balancing is much harder for gas than electricity. Indeed, much of the seasonal balancing for electricity is drawing on gas. As the UK moves toward a lower carbon economy, they will do many of the things @petit_bateau has laid out and notably electrification will transfer more of this season imbalance from gas to the renewables-based electricity. So the peak-to-trough ratio for electricity will climb from 1.3X to something closer to 3X. For sake of discussion let's say it goes to 2.3X.
So if electricity has this 2.3X demand at the peak, the UK will need sufficient renewable resources (inclusive of major transmission line and nuclear power) to hit 2.3X generation at the peak. At the trough, where demand is 1X, these renewable resources will likely produce even more than they do at the peak, say 3X. So the trough there is an excess of 2X.
The question is what to do with this 2X excess. One option is to curtail all of it. Another extreme is to have 2X capacity of electrolyzers. Near the trough this would you'd have 3X generation minus 2X electrolyzer consumption, leaving a net 1X for ordinary power consumption. This would maximize hydrogen production through the year, but a non-trivial amount of the 2X electrolyzer capacity would only be powered on for just 1 month per year. Suppose that for half the year the excess is just over 1X. This means that marginal capacity of electrolyzers added to the 1X fleet would operate for about 6 months per year (50% capacity factor). Is this the optimal point?
Well the optimal point will depend critically on the capex cost of electrolyzer capacity and the annual minimum market price of hydrogen (and derivatives). The annual minimum price for hydrogen will depend on how much hydrogen storage capacity there is, how much share of the gas market (natural gas, renewable gas, and hydrogen) is, and many other factors. So this is really hard to say, and probably depend highly on government policies. An effective carbon tax on natural gas is probably necessary. So let's just say the price for hydrogen is what it is.
The more important driver of the optimal size of the electrolyzer fleets is simply the cost of installed electrolyzer capacity. In Europe, I think this is still around $1000/kW, but the price will continue to drop substantially as production scales up and the technology advances. In Asia, the cost of electrolyzers is just a fraction of what it is in Europe. With scale up in Europe and elsewhere, we ought to see the cost drop below $400/kW. In terms of grid resources, even $1000/kW is not prohibitively expensive. For example, gas peakers have capex in range of $700 to $900 per kW for a resource that may only have a 5% to 10% capacity factor, effectively it gets used about 2 to 5 weeks per year.
Now consider the situation at the peak with not shut off of electrolyzers. Demand is 2.3X. Renewables are generating about 2.3X. and suppose electrolyzers chugging away at 1X. Thus, you'd need gas generators to provide the extra power that 1X electrolyzers would be consuming if left running. From a capex perspective, it is absurd to spend $800/kW to have a gas peaker run for the peak month just so that an electrolyzer at less than $1000/kW capex can keep running through the peak month. This illustrates the folly of those that think electrolyzer capacity is so expensive that you have to run them at 100% capacity factors. Balderdash!!! The whole issue around seasonal balancing is that you've got massive amounts of capital sitting idle most of the the year, or even most of the decade, just so that you can meet demand peaks. It would be far cheaper for the grid to pay some curtailment fee per kW-month to cover the capex when an electrolyzer must idle than to have to pay for capex for fossil peak capacity to meet seasonal peaks. As the capex of electrolyzer drops to $400/kW it will become abundantly obvious that electrolyzer are lowest capital way to do seasonal balancing on the grid.
It will take awhile for this longer-term economics to kick in. In the interim, I believe that policy makers should consider the idea of subsidizing about half the installed cost of electrolyzer capacity with clear rules around curtailing hydrogen production when net power demand is high. This would accelerate the scale up and learning curve rates, while increasing demand for more renewable resources around the year.
petit_bateau
Active Member
jhm,
I don't think one can do these sums using peak-to-trough or other simplifications, except in retrospect, i.e. one must do the detailed numbers first and only then can one back out plausible heuristics. And those heuristics might fail at another location's integrated grid (e.g. ERCOT vs UK). These analyses can be very "context-specific" to use the term I prefer.
In my experience one can either do time-domain analysis, or frequency-domain analysis. Personally I and my colleagues developed a time-domain tool and I am about to give you some screenshots from a UK model I ran in 2016 to forecast what might happen if in 2022 all coal and nuclear were taken off the UK grid and the amount of gas remained at the 2016 level. The point to appreciate is that the more the resource signature (wind or solar) is intrinsically aligned with the load signature, or can be tuned to align with the load signature by way of skewing system build out, then the less overbuilding is required. That overbuilding might be additional (battery or hydro) storage, additional transcontinental grid interconnects, additional wind & solar, or additional reserve fossil capability.
You simply cannot do these sums the way you are proposing, unless one runs the more complex models and backs out the right answer. Only then can you (as if by magic) show the simple sum. At least that is my opinion. That opinion is based on running many many thousands of simulated years, each of which is a 'true' scenario that faithfully represents a possible weather outcome in an area of interest (these are Markov models of wind & solar at 1h intervals, using data slices for monthly weather). The real issue and insight one obtains is how likely ( ... march of 9s) any given generation vs load system might get caught out and fail to deliver power on demand.
I have never run it but I believe you can do similar studies in HOMER. Personally we put our tool to bed in about 2015 and it will not be until I retire (maybe next year) that I resuscitate it. My understanding is that the main grid etc companies and energy ministries spray very large amounts of money at bespoke models that do similar things.
In the graphs below the variable that is being displayed is the State Of Charge of the aggregated UK national battery as calculated at 1h timersteps for a series of years. There are correspending graphs for windspeed, generation, load, cyclic life, economics, etc. There is a lot of conservatism baked into this set of runs because I assumed each hour of each day was at the upper end of the UK forecast 2022 load profiles, and various other factors.
Anyway I hope this helps illustrate the issues with some slightly more 'real' data.
Interestingly the comment I wrote to myself in my draft 2016 report was rather prescient, " The 5-years above are sequential, i.e. they are not selected. It continues to be that the real grid stressor is low resource availability periods in the Summer rather than the Winter."
I hope to return to studies of this nature, both at tool development and system analysis level, when I have more free time. It would be nice if the tSLA share price headed upwards so that I don't need to concern myself with earning a crust after this year
EDIT.
ps1. You can see how the model is correctly simulating typical frontal weather systems passing across the UK at 1-2 week intervals.
ps2. This was just blindly using the forecast demand which included some Ofgem/NationalGrid/DoE assumptions regarding EV and heat pump deployment - i.e. not much ! Clearly looking further ahead one takes more heat > grid shifting and liquuids > BEV into account. Which will almost certainly lead to a greater skew towards wind.
ps3. My draft conclusion at the time was "
Conclusion
With 36GW wind and 30GW solar allied with 30GW conventional (of which some is hydro & biomass) this
would be a deep penetration of renewables into the electricity mix, probably about 60% rather than the
envisaged 40% in the 2020s (must calc exact number) yet seems technically feasible and cost-comparable with
the nuclear option. Therefore in the 2020 period it is realistic to consider running the UK electrical system on a
non-nuclear basis with a deep renewables mix. By extension it appears that progressing beyond 60% to100%
renewables penetration in electricity generation is also possible. (Must do other things now !)
"
Attachments
Last edited:
You're thinking of this wrong because of long-standing biases in how first generation grid-tied solar/wind has been developed.
There's no wasted energy in telling your solar or wind plant to regulate it's power output, the same as there's no wasted energy in telling your coal power plant to run at 250 MW instead of 1000 MW.
What happens if you tell your solar plant to limit power output to 250 MW instead of 1000 MW? Are you wasting energy?
NO!
Does it mean that the average cost of the potential solar energy delivered by the plant goes up? YES - but same as the coal power plant! (ok - slightly less, but capacity factors are very important to the financial viability of coal, gas and nuclear plants today - every power plant operator wants to run the plant as much as possible to maximize ROI).
Solar and wind are finally getting cheap enough that it's cost competitive to build these plants, even knowing you won't be using 100% of the capacity of the plant. And that's great news! And what's also great - solar and wind continue to get cheaper, which means this gets easier and easier!
petit_bateau
Active Member
This came through my inbox today, regarding a proposed change at the IEA, and I fully support it. You may do as well.
ourworldindata.org
The International Energy Agency publishes the detailed, global energy data we all need, but its funders force it behind paywalls. Let's ask them to change it.
For energy security and progress on climate change we need open data on energy. The funders of the IEA can make this happen.
ourworldindata.org
This is very nice. It illustrates well the functioning of fairly deep storage. It looks like less than 100GWh of storage is the workhorse for daily cycling, and another 300GWh is needed for weekly cycling. Then the final 600GWh is needed when weekly production of wind and solar is insufficient to fully recharge within 7 days. Apparently, weekly deficits can can run for several weeks in a run and take over a month to recover to a full SoC.jhm,
I don't think one can do these sums using peak-to-trough or other simplifications, except in retrospect, i.e. one must do the detailed numbers first and only then can one back out plausible heuristics. And those heuristics might fail at another location's integrated grid (e.g. ERCOT vs UK). These analyses can be very "context-specific" to use the term I prefer.
In my experience one can either do time-domain analysis, or frequency-domain analysis. Personally I and my colleagues developed a time-domain tool and I am about to give you some screenshots from a UK model I ran in 2016 to forecast what might happen if in 2022 all coal and nuclear were taken off the UK grid and the amount of gas remained at the 2016 level. The point to appreciate is that the more the resource signature (wind or solar) is intrinsically aligned with the load signature, or can be tuned to align with the load signature by way of skewing system build out, then the less overbuilding is required. That overbuilding might be additional (battery or hydro) storage, additional transcontinental grid interconnects, additional wind & solar, or additional reserve fossil capability.
You simply cannot do these sums the way you are proposing, unless one runs the more complex models and backs out the right answer. Only then can you (as if by magic) show the simple sum. At least that is my opinion. That opinion is based on running many many thousands of simulated years, each of which is a 'true' scenario that faithfully represents a possible weather outcome in an area of interest (these are Markov models of wind & solar at 1h intervals, using data slices for monthly weather). The real issue and insight one obtains is how likely ( ... march of 9s) any given generation vs load system might get caught out and fail to deliver power on demand.
I have never run it but I believe you can do similar studies in HOMER. Personally we put our tool to bed in about 2015 and it will not be until I retire (maybe next year) that I resuscitate it. My understanding is that the main grid etc companies and energy ministries spray very large amounts of money at bespoke models that do similar things.
In the graphs below the variable that is being displayed is the State Of Charge of the aggregated UK national battery as calculated at 1h timersteps for a series of years. There are correspending graphs for windspeed, generation, load, cyclic life, economics, etc. There is a lot of conservatism baked into this set of runs because I assumed each hour of each day was at the upper end of the UK forecast 2022 load profiles, and various other factors.
Anyway I hope this helps illustrate the issues with some slightly more 'real' data.
View attachment 719225
View attachment 719227
Interestingly the comment I wrote to myself in my draft 2016 report was rather prescient, " The 5-years above are sequential, i.e. they are not selected. It continues to be that the real grid stressor is low resource availability periods in the Summer rather than the Winter."
I hope to return to studies of this nature, both at tool development and system analysis level, when I have more free time. It would be nice if the tSLA share price headed upwards so that I don't need to concern myself with earning a crust after this year
EDIT.
ps1. You can see how the model is correctly simulating typical frontal weather systems passing across the UK at 1-2 week intervals.
ps2. This was just blindly using the forecast demand which included some Ofgem/NationalGrid/DoE assumptions regarding EV and heat pump deployment - i.e. not much ! Clearly looking further ahead one takes more heat > grid shifting and liquuids > BEV into account. Which will almost certainly lead to a greater skew towards wind.
ps3. My draft conclusion at the time was "
Conclusion
With 36GW wind and 30GW solar allied with 30GW conventional (of which some is hydro & biomass) this
would be a deep penetration of renewables into the electricity mix, probably about 60% rather than the
envisaged 40% in the 2020s (must calc exact number) yet seems technically feasible and cost-comparable with
the nuclear option. Therefore in the 2020 period it is realistic to consider running the UK electrical system on a
non-nuclear basis with a deep renewables mix. By extension it appears that progressing beyond 60% to100%
renewables penetration in electricity generation is also possible. (Must do other things now !)
"
So where migh electrolyzers as dispatchable load fit into this. Be kind to my back-of-the-napkin math. The focus there was on the sort of monthly deficits for which seasonal gas storage is presently tapped.
But to integrate with your analysis it would be better to look at daily time scale to resolve weekly deficits. Apparently, about days of deficit RE generation can consume about 400GWh of stored energy. Curtailing 8GW of eletrolyzers for 50 hours a week should suffice give back 400GWh to resolve the weekly deficit. While maybe 100GWh of storage is still needed for daily cycling, the next 300GWh could be replaced by 8GW of electrolyzers plus incremental RE to produce about 1344GWh per week. In a typical week, up 50 hours of curtailment allows the electrolyzers to operate at a 70% capacity factor or higher.
So it would be interesting to take your model and include 8GW of electrolysis (and incremental RE) to see how low the SoC gets over your 4 years. The weekly cycling is a bit easier to get over because it is based more on social behavior driving demand, not so much the weather. This means that electrolyzer curtailment happens with weather that is more comparable to surrounding days. But when the weekly deficit is particularly severe, the extra 1344GWh/week of RE capacity is actually generating a fraction of a typical week. So even curtailment of 8GW for an entire week might still be insufficient to resolve the deficit. At this point the option remains to add another 600GWh of storage or more electrlyzer capacity plus incremental RE. I don't have the tools or data to calculate just how electrolyzer capacity would be need to avoid the final 600GWh of storage, but my hunch is that it is around 24GW.
Optimizing this might not even be necessary. If demand for hydrogen is high enough, then more electrolyzer capacity could be demanded than is required to do seasonal balancing for the grid. Oddly enough as demand for green hydrogen exceeds the requirement to balance the grid, electrolyzer capacity factors will increase. For example, in a typical week where upto 400GWh of electrolyzer consumption must be curtailed, this is at least 70% CF for 8GW of electrolyzer capacity, but at least 90% CF for 24GW capacity. The point here is that massive demand for natural gas suggests an enormous addressable market for green hydrogen. If this addressable market is large enough compared with ordinary demand for power, then you have a favorable long-term state where most seasonal balancing is done economically by a massive electrolyzer fleet and storage capacity is mostly limited to only what is required for daily cycling.
Let me know when you're ready to dust off your model. We can explore this more rigorously. I'll be retiring too in a year or two. We could have a bit of fun with this. Cheers!
You are not wasting fuel when curtailing power generation with either a coal or solar plant, but there is an opportunity cost. The question is what alternatives are there for curtailed power? If there is a battery than can be charged, it is clearly more valuable to charge a battery than to curtail zero marginal cost solar power. The problem is that demand response is often lacking to take advantage of the surplus. For example, EV owners would gladly charge at near zero cost if this was available to them. So inefficiencies in retail pricing mechanisms prevent utilization of curtailed power. Additionally, there are many situations where fossil generators continue to generate power on a must run basis while solar and wind get curtailed. This is also why market prices can go negative. With sufficient storage, however, it may be possible to avoid must-run generation. When we get to that sufficient level of storage capacity, there will be much less curtailment, and when curtailment does happen it shuts down all generation. Specifically, fossil fuels will not be wasted while paying a negative price for the privileged of dumping unneeded power on the market.
A little bit of curtailment is not a terrible thing, but surely there are opportunities to make better use of apparent surpluses.
Absolutely. I spend lots of time looking for useful data. I would love to have access to IEA data.This came through my inbox today, regarding a proposed change at the IEA, and I fully support it. You may do as well.
The International Energy Agency publishes the detailed, global energy data we all need, but its funders force it behind paywalls. Let's ask them to change it.
For energy security and progress on climate change we need open data on energy. The funders of the IEA can make this happen.
petit_bateau
Active Member
The hydrogen industry you are describing is the holy grail of big industry and big government. Personally I think it is a low probability outcome. At the limit it is the Japanese objective, but more concrete steps are being taken in Germany/Denmark/etc. The UK talks big but so far has always been talk without action. I have found it difficult to get reliable economic data out of the hydrogen people, and based on the limited data I have gotten over the years, and the round-trip efficiencies, it seems unlikely to my eyes at this stage. But I watch with interest.
(You can get a sufficient storage outcome only with lithium batteries as a result of millions of small adoption decisions of only $10k or so each, using technology solutions we already know work and are economic. But to get a sufficient storage outcome with hydrogen you need to place - and succeed in - some $10bn adoption bets in order to get a viable economic & technical ecosystem arising. Given the cost-ineffectiveness of hydrogen, and the P(success) of mega projects in big industry, I think I know which pathway is more likely).
The models I have built/etc over the years allow for good studies of intermittency at the 1h level which we found to be a good practical timestep. From there one can slice & dice to lower frequency stuff - daily, monthly, etc. And importantly one can generate milions of years of runs. In contrast the Tony Sheba work suffers from the deficiency of only being based on one year of data in those studies I have seen.
If you want to contribute I'll look you up in a year or so. I need to go back and teach myself Python, then agent-based models and despatch models in order to go down the pathway that I know will be needed. That's where the youngsters in my team went and have done good things since, so it will be soon time fo rme to catch them up.
Are the “millions of small adoption decisions” basically individuals or groups attaching aggregating batteries into VPP’s (Virtual power plants)? like the experiment in South Australia where up to 50,000, ~10kilowatt PV systems with ~14kwh battery systems are proposed to software aggregate into a 1/2 gigawatt plant with minimum of 700 megawatts of storage, distributed over the 50,000 placements?
The green Hydrogen may have come along some.
I remember when i attended the Solar Decathlon on the Washington DC mall.
one of the zero energy houses used the excess Photovoltaics on the roof to electrolyze water to store H2 for a fuel cell “battery”. they were there a bit under 2 weeks and had the 100% power the house and a small neighborhood electric vehicle, a GEM.
the only comment i could get out of one of the students was “highly inefficient” i regret not asking more questions.
(this was a few years before the Germans from Darmstadt covered 5 sides of a cube shaped house with PV panels including individual louvers of shades individually glued and connected)
(I’ve been following Tony Seba for years and also Amory Lovins of Rocky Mountain Institute)
petit_bateau
Active Member
Anything from kWh scale householders, operated either as singletons or as a VPP, through to MWh and GWh scale decisions by utilities and industry. Each of these decisions is far smaller than the $bn-many required as a huge leap-of-faith required to make a valid entry into the hydrogen game. And you are not alone in struggling to get financial truth out of the hydrogen lobby, and the hydrogen pilot projects that have been kicking about for 40-years or so.
My worry is that demand for hydrogen will be too high for green hydrogen to supply and we'll be stuck with lots of gray hydrogen for a long time to come. We need alot of hydrogen already for non-energy uses, ammonia, steel, petrochemical of all sorts. I am actually opposed to using hydrogen in ground transportation because it will require decades to build up a big enough electrolyzer fleet to supply displace all the hydrogen now produced from gas and coal.
Additionally, I view the round trip efficiency of hydrogen as a red herring. As long as there is not enough hydrogen to eliminate gray hydrogen, it is perfectly fine to source natural gas to backup power grids. That is, what is the logic of generating hydrogen from so that you can use hydrogen to backup the grid? The only rationale I know of is so called blue hydrogen, which is gray hydrogen plus CCS to limit emissions. So in theory if you use blur hydrogen to generate power that is comparable to using natural gas plus CCS to generate power. CCS, of course, will always be an incremental cost and energy waste when added to either a hydrogen or power generation plant. So I'm very skeptical that the economics will ever workout well for CCS.
So my outlook is that even if all the power grids of the world were to generate enough hydrogen to do seasonal grid balancing, the globe would still need places like Australia and the Middle East to export green hydrogen, and that still would not be enough to eliminate gray hydrogen. I am not opposed to using massive battery and hydro storage to balance a grid. I am optimistic that battery costs can continue to decline quite a bit. I am also optimistic that the cost of electrolyzers can decline quite a bit as well. It just turns out that when an electrolyzer is optimized for maximum return, it draws power only when it is marginally profitable to do so. This is important to keep in mind when building out an agent based model. When the power price goes above the marginal breakeven, typically between $20 and $30 per MWh, generation shuts off. Battery storage is unlikely to discharge until the price exceeds a higher threshold. So microeconomics of electrolyzers is that they compete with batteries for cheap power to charge on and do not create demand for the discharge from batteries. This microeconomic interplay means that electrlyzers will undermine the fragile economics of low frequency storage. When you do your agent-based modeling, I think you'll see this play out.
Another thing you may find is that the as a grid is dominated by zero-marginal generation with sufficient storage, you no longer have fuel-based generation setting or dominating the price for generation. So if generation is only weakly determining the market price for power, what is? I believe marginal industrial-scale consumers like electrolyzers will be really key in determining and stabilizing the wholesale price of power. Batteries a great as market makers as they switch from charge to discharge to take advantage of price volatility, but they are not so good when there is a surplus of RE generation. In your charts there are many hours where SoC is near 100%. Near 100%, batteries will find very little price spread to trade on. If you were modeling power prices, you'd find that the price of power is nearly zero. This means that your batteries and all generation capacity are producing almost no economic value at all when the market is oversupplied and SoC is nearly at 100%. But if you've got even a small electrolyer fleet, they can feast off of these low power prices. Indeed suppose the marginal breakeven price for electrolyzers is $30/MWh, then they are when RE is in surplus to what even the batteries need to charge on, they are enjoying a surplus profit of nearly $30/MWh just to be running. This bigger this surplus profit is over the course of the year, the more eagerly investors will add capacity to the fleet. Basic microeconomics anticipates that the electrlyzer fleet will grow, bidding up the price of cheap power, until some equilibrium is reached. At this equilibrium, electrlyzers set the prices at which batteries charge. The spread from the average discharge price for batteries to average charge price will determine the economic value of battery capacity. If the battery fleet is too small, this average spread will be high, and investors will add more capacity. Thus, the battery fleet grows to some equilibrium. If you do agent-based modeling, on the choice to added more battery or more electrolyzer capacity to the market, you'll see this dynamic play out.
There is a market equilibrium between electrolyzer and battery capacity such that it is unprofitable to add one more unit of either to the market. At this point, if marginal wind or solar have low enough LCOE below the marginally profitable power price for electrlyzers, then that unit of wind or solar will find adequate demand. So basically, wind and solar can be added to grid based solely on marginal demand for green hydrogen. Indeed, the marginal breakeven for electrolyzers is proportional to the price of green hydrogen. Green hydrogen is competing with gray hydrogen, which derives its breakeven price from natural gas. This pits grids against natural gas as suppliers to the hydrogen market. This is a long way around from the current situation where grids are strictly consumers of natural gas.
The way I see deep decarbonization working with market economics is for the grids of the world to transition from being consumers of natural gas to being producers of green hydrogen. There is a interim phase grids are seasonal consumers of natural gas and seasonal producers of green hydrogen. At some point, a grid displaces as much natural gas via green hydrogen as it consumes for back up power. At this point, the grid is a net zero consumer of gas and quite close to net zero carbon emissions. Note also that it is not necessary for the grid to generate any power from hydrogen to get to this point so long as there is sufficient global demand for hydrogen. But getting to net zero carbon in the grid is not going nearly far enough.
Deep decarbonization requires that no fossil fuels can be used in a manner that emits carbon dioxide. So grids can be net zero, but we still have a huge amount of fossil fuels used for heat energy. We'll certainly want to lean on net zero electricity as much as possible for heat, but there are still going to be some hard to decarbonize uses for natural gas outside the grid. Green hydrogen can still have enormous demand growth well past the point grids reach net zero. In the past I've tried to estimate how much the global grids must grow just to satisfy the current demand for gas, excluding the portion used for power generation. My crude math indicates the globe needs 3X or more power generation to quit natural gas. At 1.5X, electrolyzers would largely become net zero. So at 3X, electrolyzers are consuming about 2/3 of all power generated. This is why I don't worrying about decarbonizing the grid as much as I used to. Instead, I worry about how do we get off of natural gas.
The scale of RE that is required for this is truly mind blowing. Current global power genation is about average 3TW. To triple this we are looking at close to 36TW of solar and wind capacity along with 6 to 12 TW of electrolyzer capacity. The natural reaction many have to this is that the scale is ridiculously large. But I'd humbly submit that one does not have a firm grasp of just how much gas it takes to run our global economy. Even that chart on UK gas and UK electricity showed that gas demand was about 3X that of electricity. So even tripling tripling the grid on quenches gas consumption if the alternative are 1/3 more efficient on average. So the prospect that the UK may need 3X power generation to achieve deep decarbonization should not be shocking to anyone. Now maybe the UK can't scale up to that with domestic RE resources. That's why massive power lines to Morroco and Norway are necessary and economical. Also this is why MENA and Australia can become global exporters of green hydrogen. As Europe and Asia are currently tapping LNG imports to balance their energy needs, we should probably expect that some day green hydrogen or some more easily transported derivative of it will be exported to regions that are running a deficit on renewable energy. Yep, hydrogen might have a low round-trip efficiency, but that matters little when you've got to import massive amounts of backup energy. LNG is certainly not the most energy efficient way to get natural gas, but this inefficiency matters little when there are regional energy shortages.
At any rate, that's my outlook. Cheers!
Edit, 24GW of electrolyzer would only be required with 25% CF, but 50% or more is reasonable.
Last edited:
Similar threads
