Responding to David Osmond and Peter Farley in a new post because I need to provide step-by-step demonstrations of the arithmetic for those who cannot/will not see
I may be completely misunderstanding what you are doing, but why have you included hydro in column AJ?
Surely it should be treated the same as Batteries and fossil fuels, because it can be dispatched when needed and turned off when not. Batteries and pumped hydro need to charge using other power sources. Non-pumped hydro needs to charge with rainfall, but the rain doesn't have to fall within the 1/2 hour period.
(Fossil plants may also have a limit on how fast they can be "charged" depending on how close they are to their source of fuel and how quickly it can be extracted or delivered, but it's likely that they can be charged at close to 100% of their maximum generation, so this can probably be ignored).
provides both Hydro generation and a bit of pumped.
The NEM figures don't distinguish except in that they include the category [Pumps - MW] which is always zero or negative, telling me it's the energy consumed to 'recharge' pumped storage.
My working assumption for drought-prone Aus [largely confirmed by other Aussie commenters] is that new Hydro is unlikely. Of course, new pumped hydro is energy *storage* with low water losses, so is not ruled out.
I used a future Hydro factor = 1.0 to set future Hydro = real measured Hydro every ½-hour interval.
If you see my latest post [link] you'll see in Figures 3 that I publish some relevant numbers for the last real year; max Hydro = 6,410MW; min Hydro = 190MW; average Hydro = 1,538 MW.
Just beneath Figures 3 I include the year's totals: "the summation of [total Wind] + [total Solar] + [total Hydro] ... = 30.425 + 44.644 + 13.536" TWh.
If we are to believe that future Aus could crank Hydro up to fill in gaps in 'renewables' generation while remaining within an annual total of 13.536 TWh
- we would also have to believe that future forecasting of Wind and of Solar over the next days and weeks and months, *and*
- future forecasting of rainfall / precipitation over future months and years,
- would all be so precise that future Aus could confidently dispatch more or less Hydro each ½-hour than it did throughout last year.
Weather forecasting being what it is (the attempt to predict the future behaviour of a chaotic system), I doubt that would be a sensible long-term way for AEMO to plan to use the limited resource that is Aus Hydro.
And why did you set hydro at 1. Just another of your assumptions that you so roundly criticise others for.
You are all for "sense checks". Well try your sense checks on the best and worst hours, days and weeks. In the high wind and solar periods hydro will be virtually zero, leaving water in the dams for low W+S periods.
In bad times with the completion of Humelink and Snowy II peak hydro will be at least 9 GW, Borumba and Kidston takes it to 11 GW+, If Marinus goes ahead that is another 700 MW into the mainland, while the Tarraleah and Cethana projects in Tasmania increase their peak hydro by 1,200 MW
Currently there are announced proposals for over 1,200 GWh of pumped hydro and battery storage across the NEM with 13 GW/ 400 GWh actually under construction or already commissioned.
This does not include customer batteries, V2G, load shifting from offpeak rates designed to keep coal plants running to solar soaker rates which will reduce midnight demand by 20-30% etc etc.
If solar and wind are set at 115% of a normal years demand and in drought years solar is clearly higher even if wind and hydro are lower, there has not been a year where wind and solar would not have supplied 100% of demand. That means in most years hydro will be lower so instead of ranging from 2% to 8% of annual demand, without any increase in long term output have a narrower range of 3-7%,
While you are right that we should not count on more rainfall there are still plenty of opportunities to increase hydro output from irrigation dams where less than 1/3rd of the outflow is through the hydro generators, small dams and regulating basins can be fitted with low head hydro and some dams can have increased dam heights so a higher head generates more energy for the same water flow
How about you use your brain insead of just a spreadsheet.
Can we assume there is 12GW of hydro available at any time of day if needed (although not continuously)?
Column AF (which I think is modelled future Hydro production) seems to be a copy of Column N (actual Hydro production), rather than an estimate of how much Hydro could provide if needed at that time. If Hydro was turned up at shortfall times (which it could be) it could cover some of the shortfall, although probably not all of it.
I'm fairly sure that in most places fossil fuels will be needed at certain times to cover peaks. What I don't know is the magnitude. I think that actual net zero will be almost impossible, but that doesn't mean we shouldn't try to get as close as we can.
I used a future Hydro factor = 1.0 to set future Hydro = real measured Hydro every ½-hour interval.
If you see my latest post [link below] you'll see in Figures 3 that I publish some relevant numbers for the last real year; max Hydro = 6,410MW; min Hydro = 190MW; average Hydro = 1,538 MW.
Just beneath Figures 3 I include the year's totals: "the summation of [total Wind] + [total Solar] + [total Hydro] ... = 30.425 + 44.644 + 13.536" TWh.
If we are to believe that future Aus could crank Hydro up to fill in gaps in 'renewables' generation while remaining within an annual total of 13.536 TWh
- we would also have to believe that future forecasting of Wind and of Solar over the next days and weeks and months, *and*
- future forecasting of rainfall / precipitation over future months and years,
- would all be so precise that future Aus could confidently dispatch more or less Hydro each ½-hour than it did throughout last year.
Weather forecasting being what it is (the attempt to predict the future behaviour of a chaotic system), I doubt that would be a sensible long-term way for AEMO to plan to use the limited resource that is Aus Hydro.
I understood that most hydro, particularly pumped storage schemes, can be dispatched as needed on a minute by minute basis and respond to live demand. You don't need to work out in advance what the weather forecast and electricity demand will be and then carefully match it - you can do the matching in real time, not only at a 1/2 hour level, but even second by second if you control the valves with the grid frequency.
I can't see a reason that what hydro did in any particular half hour in your reference has any relevance to your model for the future. Hydro can fill in any unmet demand as long as it's not above the maximum (10Gw ish), and as long as there is still water in the reservoir (approximately 1 TWh / month).
I genuinely believe that you probably have a point and it will not be possible and certainly not safe to switch of gas entirely any time soon, but your calculations seem off.
Pumped Hydro is energy storage. Yes, flexible. Once it's empty it has to be 'recharged'.
No, pumped Hydro doesn't rely [very much] on precipitation, but dammed or run-of-river Hydro does. Which is why I adopt David O's approach (after all, I *am* offering a critique of his "simulation" results so I should do what he says he does even if he doesn't publish his numeric results to back it up) which is to keep precipitation-dependent Hydro, by and large, within existing parameters.
In my later post [link] I used a different algorithm for flexing Hydro:
Depending on whether the last ½-hour interval has a RE Shortfall (Hydro flexed up by 'up factor') or a RE surplus (Hydro flexed down by 'down factor')... [pumped Hydro assumed part of energy storage].
I've now put that different algorithm into my extrapolation tool with 1 year+ of ½-hourly NEM data:
- Over a year, existing NEM Hydro generation = 13.51 TWh / 365d
- using up-factor = 1.6, constrained to max 6,410 MW, down factor = 0.8, for future year 2044 allowing for 1½ times Demand, for example, I match Future Hydro = 13.50 TWh / 365d.
My point re: accuracy of forecasting over days/weeks/months relates specifically to forecasting of precipitation. If NEM / AEMO dispatched higher proportions of Hydro now to fill a hole in RE production in anticipation of future precipitation, but that future precipitation never arrives, then oh drat!
So I think it's prudent to design for around the same Hydro TWh/365 days as current. In that I agree with David O's approach.
I agree. I won't want a global government even if it can "work" because the risk of increasing state intervention in the economy will go up exponentially.
This analysis fails to recognize the difference between AVERAGE Nat Gas Generation and PEAK Generation. While it may appear that you can reduce the Nat Gas Fleet because the average is low, in fact, you must still keep the majority of your Gas Fleet ready to generate on almost a daily basis.
If batteries are used alongside gas turbines for firming the number of gas turbines you needed to meet peak demand goes down.
But I understand your point that as renewable penetration goes up the cost of firming goes up as you need firming power plants to stay off-line for months leading to high LCOE costs.
Below is another article that does some modelling with solar, batteries and gas for California. Let me know what you think.
Reading through that California piece I note that Brian Potter includes a couple of 7-day charts from CAISO, but then seems to switch to annual charts for 'CAISO daily max load' and 'maximum solar capacity factor'.
And then seems to draw all his conclusions from the **annual data** which EXCLUDES sufficiently granular information to tell us anything about what happens under the most difficult (winter, night-time or high summer high Demand low Wind night-time) conditions.
His chart of "Solar Energy Supplied" confuses installed Solar *capacity* MW [horizontal axis] with Solar Power... because 100% of CAISO power at its highest is 50,000 MW as he states just above that chart. Little things like that reduce my confidence in the article's general value.
His "Costs to meet CAISO Demand with Solar PV" chart has several discontinuities in it. To me that implies changes of assumptions as %Solar is increased. Again, that doesn't give me a warm fuzzy feeling about the material.
As others have pointed out, his use of Lazard LCOEs likely introduces a bundle of inaccuracies into his conclusions.
Other than those points, it's great!
I agree that, in principle, a pragmatic, fully-costed assessment of the options will today result in lower % 'renewables' with fossil backup for the difficult times.
IF new better lower-cost lower-minerals energy storage tech is developed, that will push the pragmatic solution to a higher % 'renewables' but possibly still with fossil backup for the most difficult times. BUT we currently do NOT have that grade of energy storage tech, a fact that Brian Potter glosses over.
In any case, societies have to find funding mechanisms to pay for electrical energy security as 'decarbonisation' continues.
In the beginning, I had the same thought about using Nat Gas to charge storage. I actually ran multiple model scenarios and what I found was it did not provide any benefit.
This is because if you have storage being charged by wind and/or solar, there is always surplus available. This means that storage typically operates from "State of Charge High Limit [80%]" That said, there may be some limited opportunity for this during some of the transition period, but it will be a rare event and not a standard procedure.
I went through Brian Potter's article. His modelling, although directionally correct does not consider seasonal effects.
I summarized my comments in the following .PFD as I needed to show a number of graphs to illustrate my assessment.
Re: "I use a detailed Grid Model (GridSim V3.0) of the California Grid to investigate various scenarios.
It utilizes real-time 5-min data over a 1-year period downloaded from CAISO."
I don't have access to that model. But it's excellent that it uses granular real data.
Was that year especially benign or especially troublesome, weatherwise? Because that year's results may be better or worse than we might expect from a "typical" year.
But anyway, any well-engineered power generation system should have design factors / design margins / safety margins / however you wish to label them, so that the lights reliably stay on whatever the conditions.
I won't comment in any detail on your financial modelling results, other than to say directionally they look sensible.
You are sort of right, but all grids have some low utilisation assets. For example the US has 20 GW of liquid fueld plants that operate at 2.1% capacity. The combustion turbine gas fleet operates at 8-10%.
It was getting to the point as gas prices rose that operating a CC gas plant with a a 2MW/4 hour battery battery for every MW of generator capacity was cheaper than running 3MW of OC gas turbines, but with gas prices easing and CC plant costs exploding, that is probably not the case.
I may be completely misunderstanding what you are doing, but why have you included hydro in column AJ?
Surely it should be treated the same as Batteries and fossil fuels, because it can be dispatched when needed and turned off when not. Batteries and pumped hydro need to charge using other power sources. Non-pumped hydro needs to charge with rainfall, but the rain doesn't have to fall within the 1/2 hour period.
(Fossil plants may also have a limit on how fast they can be "charged" depending on how close they are to their source of fuel and how quickly it can be extracted or delivered, but it's likely that they can be charged at close to 100% of their maximum generation, so this can probably be ignored).
Hi Jamie, thank you for your Qs.
Australia has a mix of Pumped Hydro Energy Storage and Hydro.
For example, the original (1974) Snowy River scheme
[ https://www.snowyhydro.com.au/generation/the-snowy-scheme/ ]
provides both Hydro generation and a bit of pumped.
The NEM figures don't distinguish except in that they include the category [Pumps - MW] which is always zero or negative, telling me it's the energy consumed to 'recharge' pumped storage.
My working assumption for drought-prone Aus [largely confirmed by other Aussie commenters] is that new Hydro is unlikely. Of course, new pumped hydro is energy *storage* with low water losses, so is not ruled out.
It will be but the magnitude will change, it will more often run at very low values so that occasionally it can run at 10-12 GW.
Hi Peter,
How about, instead of pulling numbers out of your backside, you engage with a year of data:
https://open.substack.com/pub/chrisbond/p/one-year-in-australias-nem?r=om40y&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
As I said to Jamie,
I used a future Hydro factor = 1.0 to set future Hydro = real measured Hydro every ½-hour interval.
If you see my latest post [link] you'll see in Figures 3 that I publish some relevant numbers for the last real year; max Hydro = 6,410MW; min Hydro = 190MW; average Hydro = 1,538 MW.
Just beneath Figures 3 I include the year's totals: "the summation of [total Wind] + [total Solar] + [total Hydro] ... = 30.425 + 44.644 + 13.536" TWh.
If we are to believe that future Aus could crank Hydro up to fill in gaps in 'renewables' generation while remaining within an annual total of 13.536 TWh
- we would also have to believe that future forecasting of Wind and of Solar over the next days and weeks and months, *and*
- future forecasting of rainfall / precipitation over future months and years,
- would all be so precise that future Aus could confidently dispatch more or less Hydro each ½-hour than it did throughout last year.
Weather forecasting being what it is (the attempt to predict the future behaviour of a chaotic system), I doubt that would be a sensible long-term way for AEMO to plan to use the limited resource that is Aus Hydro.
And why did you set hydro at 1. Just another of your assumptions that you so roundly criticise others for.
You are all for "sense checks". Well try your sense checks on the best and worst hours, days and weeks. In the high wind and solar periods hydro will be virtually zero, leaving water in the dams for low W+S periods.
In bad times with the completion of Humelink and Snowy II peak hydro will be at least 9 GW, Borumba and Kidston takes it to 11 GW+, If Marinus goes ahead that is another 700 MW into the mainland, while the Tarraleah and Cethana projects in Tasmania increase their peak hydro by 1,200 MW
Currently there are announced proposals for over 1,200 GWh of pumped hydro and battery storage across the NEM with 13 GW/ 400 GWh actually under construction or already commissioned.
This does not include customer batteries, V2G, load shifting from offpeak rates designed to keep coal plants running to solar soaker rates which will reduce midnight demand by 20-30% etc etc.
If solar and wind are set at 115% of a normal years demand and in drought years solar is clearly higher even if wind and hydro are lower, there has not been a year where wind and solar would not have supplied 100% of demand. That means in most years hydro will be lower so instead of ranging from 2% to 8% of annual demand, without any increase in long term output have a narrower range of 3-7%,
While you are right that we should not count on more rainfall there are still plenty of opportunities to increase hydro output from irrigation dams where less than 1/3rd of the outflow is through the hydro generators, small dams and regulating basins can be fitted with low head hydro and some dams can have increased dam heights so a higher head generates more energy for the same water flow
How about you use your brain insead of just a spreadsheet.
Can we assume there is 12GW of hydro available at any time of day if needed (although not continuously)?
Column AF (which I think is modelled future Hydro production) seems to be a copy of Column N (actual Hydro production), rather than an estimate of how much Hydro could provide if needed at that time. If Hydro was turned up at shortfall times (which it could be) it could cover some of the shortfall, although probably not all of it.
I'm fairly sure that in most places fossil fuels will be needed at certain times to cover peaks. What I don't know is the magnitude. I think that actual net zero will be almost impossible, but that doesn't mean we shouldn't try to get as close as we can.
Hi Jamie,
I used a future Hydro factor = 1.0 to set future Hydro = real measured Hydro every ½-hour interval.
If you see my latest post [link below] you'll see in Figures 3 that I publish some relevant numbers for the last real year; max Hydro = 6,410MW; min Hydro = 190MW; average Hydro = 1,538 MW.
Just beneath Figures 3 I include the year's totals: "the summation of [total Wind] + [total Solar] + [total Hydro] ... = 30.425 + 44.644 + 13.536" TWh.
If we are to believe that future Aus could crank Hydro up to fill in gaps in 'renewables' generation while remaining within an annual total of 13.536 TWh
- we would also have to believe that future forecasting of Wind and of Solar over the next days and weeks and months, *and*
- future forecasting of rainfall / precipitation over future months and years,
- would all be so precise that future Aus could confidently dispatch more or less Hydro each ½-hour than it did throughout last year.
Weather forecasting being what it is (the attempt to predict the future behaviour of a chaotic system), I doubt that would be a sensible long-term way for AEMO to plan to use the limited resource that is Aus Hydro.
https://open.substack.com/pub/chrisbond/p/one-year-in-australias-nem?r=om40y&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
I understood that most hydro, particularly pumped storage schemes, can be dispatched as needed on a minute by minute basis and respond to live demand. You don't need to work out in advance what the weather forecast and electricity demand will be and then carefully match it - you can do the matching in real time, not only at a 1/2 hour level, but even second by second if you control the valves with the grid frequency.
I can't see a reason that what hydro did in any particular half hour in your reference has any relevance to your model for the future. Hydro can fill in any unmet demand as long as it's not above the maximum (10Gw ish), and as long as there is still water in the reservoir (approximately 1 TWh / month).
I genuinely believe that you probably have a point and it will not be possible and certainly not safe to switch of gas entirely any time soon, but your calculations seem off.
Hi Jamie,
Pumped Hydro is energy storage. Yes, flexible. Once it's empty it has to be 'recharged'.
No, pumped Hydro doesn't rely [very much] on precipitation, but dammed or run-of-river Hydro does. Which is why I adopt David O's approach (after all, I *am* offering a critique of his "simulation" results so I should do what he says he does even if he doesn't publish his numeric results to back it up) which is to keep precipitation-dependent Hydro, by and large, within existing parameters.
In my later post [link] I used a different algorithm for flexing Hydro:
Depending on whether the last ½-hour interval has a RE Shortfall (Hydro flexed up by 'up factor') or a RE surplus (Hydro flexed down by 'down factor')... [pumped Hydro assumed part of energy storage].
I've now put that different algorithm into my extrapolation tool with 1 year+ of ½-hourly NEM data:
- Over a year, existing NEM Hydro generation = 13.51 TWh / 365d
- using up-factor = 1.6, constrained to max 6,410 MW, down factor = 0.8, for future year 2044 allowing for 1½ times Demand, for example, I match Future Hydro = 13.50 TWh / 365d.
My point re: accuracy of forecasting over days/weeks/months relates specifically to forecasting of precipitation. If NEM / AEMO dispatched higher proportions of Hydro now to fill a hole in RE production in anticipation of future precipitation, but that future precipitation never arrives, then oh drat!
So I think it's prudent to design for around the same Hydro TWh/365 days as current. In that I agree with David O's approach.
https://open.substack.com/pub/chrisbond/p/another-28-days-and-conclusions?r=om40y&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
https://open.substack.com/pub/mliebreich/p/decarbonizing-the-last-few-percent?utm_source=share&utm_medium=android&r=o2bbq
What do you think of this analysis?
Thank you, Md.
I absolutely agree with the concept of diminishing returns which seems to be summarised in the opening chart.
I disagree with globalist Liebreich's carbon tax approach because we don't have a World Govt and we never will have one, humanity being what it is.
I agree. I won't want a global government even if it can "work" because the risk of increasing state intervention in the economy will go up exponentially.
This analysis fails to recognize the difference between AVERAGE Nat Gas Generation and PEAK Generation. While it may appear that you can reduce the Nat Gas Fleet because the average is low, in fact, you must still keep the majority of your Gas Fleet ready to generate on almost a daily basis.
Having a low average gas generation with a high peak generation means that the gas generation is very expensive: https://wrjohn1.substack.com/p/capacity-factor-cf-vs-lcoe
If batteries are used alongside gas turbines for firming the number of gas turbines you needed to meet peak demand goes down.
But I understand your point that as renewable penetration goes up the cost of firming goes up as you need firming power plants to stay off-line for months leading to high LCOE costs.
Below is another article that does some modelling with solar, batteries and gas for California. Let me know what you think.
https://open.substack.com/pub/constructionphysics/p/can-we-afford-large-scale-solar-pv?utm_source=share&utm_medium=android&r=o2bbq
Thank you again, Md.
Reading through that California piece I note that Brian Potter includes a couple of 7-day charts from CAISO, but then seems to switch to annual charts for 'CAISO daily max load' and 'maximum solar capacity factor'.
And then seems to draw all his conclusions from the **annual data** which EXCLUDES sufficiently granular information to tell us anything about what happens under the most difficult (winter, night-time or high summer high Demand low Wind night-time) conditions.
His chart of "Solar Energy Supplied" confuses installed Solar *capacity* MW [horizontal axis] with Solar Power... because 100% of CAISO power at its highest is 50,000 MW as he states just above that chart. Little things like that reduce my confidence in the article's general value.
His "Costs to meet CAISO Demand with Solar PV" chart has several discontinuities in it. To me that implies changes of assumptions as %Solar is increased. Again, that doesn't give me a warm fuzzy feeling about the material.
As others have pointed out, his use of Lazard LCOEs likely introduces a bundle of inaccuracies into his conclusions.
Other than those points, it's great!
I agree that, in principle, a pragmatic, fully-costed assessment of the options will today result in lower % 'renewables' with fossil backup for the difficult times.
IF new better lower-cost lower-minerals energy storage tech is developed, that will push the pragmatic solution to a higher % 'renewables' but possibly still with fossil backup for the most difficult times. BUT we currently do NOT have that grade of energy storage tech, a fact that Brian Potter glosses over.
In any case, societies have to find funding mechanisms to pay for electrical energy security as 'decarbonisation' continues.
> In any case, societies have to find funding mechanisms to pay for electrical energy security as 'decarbonisation' continues.
Security doesn't need to be an societal level decision, insurance markets already exists. But otherwise great analysis. Thank you.
In the beginning, I had the same thought about using Nat Gas to charge storage. I actually ran multiple model scenarios and what I found was it did not provide any benefit.
This is because if you have storage being charged by wind and/or solar, there is always surplus available. This means that storage typically operates from "State of Charge High Limit [80%]" That said, there may be some limited opportunity for this during some of the transition period, but it will be a rare event and not a standard procedure.
I went through Brian Potter's article. His modelling, although directionally correct does not consider seasonal effects.
I summarized my comments in the following .PFD as I needed to show a number of graphs to illustrate my assessment.
https://www.goinggreencanada.ca/Response%20to%20Md%20Nadim%20Admed.pdf
Thank you, Bill.
Re: "I use a detailed Grid Model (GridSim V3.0) of the California Grid to investigate various scenarios.
It utilizes real-time 5-min data over a 1-year period downloaded from CAISO."
I don't have access to that model. But it's excellent that it uses granular real data.
Was that year especially benign or especially troublesome, weatherwise? Because that year's results may be better or worse than we might expect from a "typical" year.
But anyway, any well-engineered power generation system should have design factors / design margins / safety margins / however you wish to label them, so that the lights reliably stay on whatever the conditions.
I won't comment in any detail on your financial modelling results, other than to say directionally they look sensible.
You are sort of right, but all grids have some low utilisation assets. For example the US has 20 GW of liquid fueld plants that operate at 2.1% capacity. The combustion turbine gas fleet operates at 8-10%.
It was getting to the point as gas prices rose that operating a CC gas plant with a a 2MW/4 hour battery battery for every MW of generator capacity was cheaper than running 3MW of OC gas turbines, but with gas prices easing and CC plant costs exploding, that is probably not the case.
Like you, I have done a critique of David's Model. It comes up short in many areas with multiple omissions and errors.
You can read it here: https://www.goinggreencanada.ca/David_Osmand_near_100pc_Grid.pdf
Hello Bill,
Thank you, but my malware protection gives the following message when I clicked on your link:
"Your connection isn't private
Attackers might be trying to steal your information from www.goinggreencanada.ca "
So I haven't read it.
The website cert needs to be changed to one that is supported. I placed the 37 page pdf on my Google Drive, you should be able to download it from here https://drive.google.com/file/d/1kiNjQC8wi75JVjOhtdaiQ2Weh8zXWsJi/view?usp=sharing
Many thanks, Todd.
I've downloaded it ok, skimmed through it, I'll need to read and digest.
Bill Johnson please note.
Well, Chris, don't say I didn't warn you its 37 pages long! :)