ERCOT renewable energy: reality check

by Energy Meteorologist

A local example of the penetration problem for renewable energy in Texas

The Electric Reliability Council of Texas (ERCOT) operates Texas’ electric power grid that serves 25 million customers in Texas.  ERCOT’s sources of generating are natural gas (51%), wind (24.8%), coal (13.4%), nuclear (4.9%), solar (3.8%), and hydroelectric or biomass-fired units (1.9%).  Power demand in the ERCOT region is typically highest in summer, primarily due to air conditioning use in homes and businesses. ERCOT region’s all-time record peak hour occurred on July 8, 2022, when consumer demand hit 78,204 MW.

This article describes an extended lull in wind and solar power during the period 8/1/22 – 9/12/22.  I then describe what it would take in terms of a system with 100% renewable generation plus storage to produce sufficient electric power for Texas during such a lull.

For background, read these previous articles by Planning Engineer:

Assigning blame for blackouts in Texas

The Penetration Problem. Part I:  Wind and Solar – The More You Do, The Harder it Gets

The Penetration Problem. Part II: Will the Inflation Reduction Act Cause a Blackout?

August 2022 temperatures for Texas were fairly typical – hot (daily high temperatures frequently broke 100oF), but not exceptional.  Texas temperatures during the first two weeks of September were slightly cooler than normal, with daily maximums in the high 80’s as a wet pattern enveloped most of the state.  What was exceptional about this period was an extended lull in both wind and solar power in the ERCOT region.

This chart plots the actual electricity demand (load) for ERCOT during the period 8/1/22 – 9/12/22 vs combined wind and solar production of electricity.  During this period of lull in renewable energy production, the demand loads were pushing into the mid and upper 70 GW range.

Screen Shot 2022-10-30 at 9.25.15 AM

What would it take for a hypothetical electric power system to meet such a demand solely with wind and solar power?

Texas saw wind generation fall to 600 MW or lower with demand loads of 41-43 GW in the early AM hours when solar was zero, so it’s either 75 times wind capacity or a crapton of batteries. Either way, massive amounts of battery storage are needed in ERCOT for a 100% wind/solar/hydro grid. 400 – 450MWh of batteries would be needed to cover the extended wind lulls around Aug 22nd and Aug 30th, and during both events the batteries would be drained to 5-15% of capacity. Come the next day, you need excess energy to charge them so they can be used again the following night. The deepest discharge period (Aug 30th – Sep 2nd) during this time lasted for 63 hours and required a discharge of ~940 MWh while only charging ~75 MWh over that time period.

Clearly, a massive buildout of wind and solar would be needed to cover that gap.  Two scenarios are considered here, using back-of-the-envelope calculations.  Scenario #1: if you increase wind and solar 8 times current capacity AND add 900 GWh of battery storage, you would have been able to cover that month-long period with 100% renewables; this turns out to be much more cost efficient than the second scenario.  Scenario #2: wind at 4.05 times current capacity and solar at 8.45 times current capacity with 5000 GWh of battery storage.

Screen Shot 2022-10-30 at 9.42.50 AM

Yes this whole scenario is rather sketchy, but how much would it actually cost?  For Scenario #1, 8X the current wind buildout and 8X the current solar buildout (364,000 MW) plus 900,000 MWh of battery storage, the cost would be $800 billion + transmission.  This calculation assumes that the charge/discharge of the batteries is 100%; if you take into account losses, you need another 10-15% of battery storage.  Additional transmission lines cost $2 to $5 million per mile.  Scenario #2 is much more expensive than Scenario #1.

The scenario costs are  based on the following cost assumptions (see References at end of post):

  • Cost of Wind $1.35 Million /MW
  • Cost of Solar $1.5 Million/MW
  • Battery cost $385,000/MWh

Compare this with an estimated cost of nuclear power at $9 million /MWh, whereby that same amount of money for renewables ($800 billion) could build over 90 GW of nuclear power vs the 230,000 MW of renewables.   The figure below shows the current renewable generation stack plus the buildout of nuclear power (red line). Using that money to build out nuclear power instead would yield more power than ERCOT would need for the coming few years even with steady load growth AND would use <1% of the land area of the renewables. If you site the nuclear power plants at old coal facilities and outdated gas plants, the transmission interconnection is already there and costs would decrease as more nuclear capacity is built.

Screen Shot 2022-10-30 at 9.52.14 AM

The recent month long period was unusually light for wind and solar BUT that is what electricity grids need to be designed for.  Assuming the 8xWind and 8xXolar + 900,000 MWH of batteries, this is what the summer would have looked like from a supply standpoint with batteries fully charged.  A little overkill, IMO.  Remember this assumes the charge and discharge of the batteries is 100% efficient, if you use a more realistic estimate you need to increase the battery storage by 10-15%

Screen Shot 2022-10-30 at 10.00.30 AM

Power production from wind and solar hardware typically decays at 0.5% per year. Not much at first, but half way through their lifecycle it adds up to BIG numbers. Load growth as well needs to be taken into account, with 82 GW possible in the coming summer for peak demand. ERCOT has been averaging ~2 – 3% growth per year.

Since ERCOT is more or less an isolated grid, it is a good example for an academic/economic exercise such as this.  With such an overbuild of wind and solar for ERCOT, there would be a great deal of curtailed/wasted power once the batteries were fully charged. This chart shows the hypothetical wasted power for the recent Aug/Sept period with the 8xWind, 8xSolar and 900,000MWh batteries.  The grid would have wasted/lost a total of 37.54 TWh to serve a total load of 63.17 TWh. If you want to transport that power elsewhere, remember it costs $2 to $5 million per mile for new transmission lines.

Screen Shot 2022-10-30 at 10.02.02 AM

Wind and solar are cheaper to build, but not when you take into account the overbuild and storage to fully serve the grid.  When total costs are considered, nuclear power is the cheapest option while also having the smallest environmental footprint overall.

Here is a link [ERCOT Load vs Renewables ] to the spreadsheet, you can play around yourself with different scenarios.


Click to access 20190212%20PSC%20Item%2005a%20Transmission%20Cost%20Estimation%20Guide%20for%20MTEP%202019_for%20review317692.pdf



61 responses to “ERCOT renewable energy: reality check

  1. “When total costs are considered, nuclear power is the cheapest option while also having the smallest environmental footprint overall.”

    I agree with the point described above, with one important limitation:
    Nuclear Electric Plants should be situated in some remote from traditional human habitat regions.
    Also, the seismic active regions should be strictly avoided.

    • We need nuclear in each region, so that a compact grid can keep the region on. Any power from far away can easily be lost when any part of the transmission lines are taken out by man or nature.
      It is now known how to build nuclear power plants that can be safely shut down. In Japan, the Radiation killed very few, the earthquake and tsunami killed thousands.
      Yes, do not build nuclear power plants that cannot withstand whatever is going to adversely affect it.

    • Joe - the non climate scientist

      The only issue I have with Nuclear is the electric power generation lacks the ability to increase or decrease rapidly to reflect the daily & hourly fluctuations in electricity demand. Electric demand in the summer mornings is very low relative to the demand in the late afternoon.

      Nuclear is great for a solid base load and definitely should be used. while gas is excellent for the late afternoon peaks.

      • Aplanningengineer

        I was informed here long ago by a knowledgeable poster, “safety” regulations prevent nuclear from helping with load following more than the inherent capability of nuclear power. Economically it makes sense to tour it full out all the time as well. Thinking out of the box for the future, nuclear could have more capability to vary output. I agree that gas will stay as the best for short duration load as it is now but nuclear could help more and that may be more appropriate with the economic considerations of the future power supply situation,

      • If all you are worried about is daily peaking, then much smaller battery packs can handle that.

    • Put them in the ocean which has an artifical freshwater lake.

    • thecliffclavenoffinance

      In many places the cost of nuclear power is infinite — green zealots won’t allow new nuclear power plants to be built

      • The cost of solar power at night is infinite as well. It can’t be obtained at any price.

  2. Consider the risk from natural and man-made disasters. The wind farms and solar farms and the transmission lines are at risk from natural disasters and domestic and foreign terrorists and war with even a minor country who could use drones from Iran, as Russia is using in Ukraine. A statewide grid is a very fragile thing, a hacker could hold it for ransom.
    Each region should have a grid with enough generation that it could operate independently in difficult times.
    During the February 2021 freeze, my water and electricity were off. Hospitals in downtown Houston were without water, employees transported water in buckets on carts and they had porta-pots in the halls. People died in Texas from lack of energy. In years past, when we had more 24/7 fossil fuel power, we weathered much worse freezes with out major outages such as this. Never before a statewide outage. More wind and solar, even with hugely expensive and environmentally harmful batter backup will steadily make this problem worse. Our grid will taken out much quicker than Ukraine’s grid. Electric vehicles will quickly become useless.

  3. Joe - the non climate scientist

    The cordova power plant in Granbury Texas produces approx 260MWh and has 260 MWh battery storage added in the summer of 2022 which covers approx 5 acres. Note that the 260MWh is only for One (1) hour. After one hour, the batteries need to be recharged.

    As noted above, the peak power usage in ercot was approx 73GWh. The average hourly usage in the summer is approx 50GWh (averaged over the entire 24 hour day ).

    As noted in the article above, there are frequent lulls in wind and solar lasting 5-7 days (9 day lull during the Feb 2021 freeze).

    My calculation is the land area needed to cover 7 days with little or no wind would be approx 150-200 square miles of land.

    • Leslie MacMillan

      Joe, you are confusing power (watts) and energy (watts x time, usually hours). The Cordova plant produces approx 260 MW (power) and the battery storage is 260 MW-hr (energy), which is enough to last one hour (theoretically) at a discharge rate (power) of 269 MW.

      Peak power in ercot is 73 GW, not GW-hr, and the summer usage averages 50 GW, also power so not GW-hr.

      Your calculations break down if you confuse instantaneous power with energy delivered during a specified time.

      • Joe - the non climate scientist

        I am happy with better explanation especially if I am interpreting the data incorrectly. I am getting my numbers from

        If I understand, the peak power usage during summer of 2022 was approximately 73GW for one hour, and the daily usage is approximately 60Gw per hour x 24 hours (low of 47gw per hour and 73gw per hour).

        If I understand the capacity of the Cordova plant – It rated capacity is 260 Mw and if operating at peak capacity producing 260mw of electricity for How long?

        thanks for any clarification

      • Leslie MacMillan

        *260, not 269. Typo. Apologies.

    • Leslie MacMillan

      The problem comes when you don’t know how long each power requirement lasts. You need to know the power integrated over time if you are going to calculate how much battery storage you need. If peak demand is 73 GW and your storage capacity is 73 GW-hr, you can meet that peak demand for an hour. To meet it for 10 hours, you need 730 GW-hr. And so on. More than that because battery charge-discharge is not 100% conserved but you get the idea.
      Demand and supply are expressed as power. Storage is expressed as energy, which is power integrated over time.

  4. Prof. Michael Kelly FRS and Fellow of the Royal Academy of Engineering has calculated that the battery installed at a cost of £45m in Adelaide, Australia, would power the three emergency wards at Addenbrookes Hospital Cambridge (UK) for 24hours on a single charge.

    The current emergency back up is supplied by two diesel generators which can run for as long as there is fuel and cost £0.25m.>content>uploads>2022>03>Kelly-Net-Zero-Progress-Report.pdf

  5. Everyone knows nuclear power is very good at providing abundant and affordable energy…that’s the problem for some.

    • During both the winter storm of 2011 and the 2021 storm Uri Texas lost at least one nuclear power plant.
      You can’t ignore the Achillies heel of all thermal power plants, their massive water use. Both intake and discharge temperatures have to be within very specific ranges or efficiency drops. Add in periodic severe droughts which can completely shut down water cooled power plants and even nuclear plants will go offline.
      There are no perfect solutions.

      • The NuScale Power Module is an advanced light-water small modular reactor capable of generating 60 megawatts of electricity. Each power plant can house up to 12 modules, which will be factory-built and about a third of the size of a large-scale reactor. Its unique design allows the reactor to passively cool itself without any need for additional water, power or even operator action.

      • Texas should replace solar plants with SMRs on the same site. That would free up some land for productive use.

      • When they install a SMR under Washington DC then we will know it’s safe.

      • From jim2 post above, re Nuscale, site says
        “As the hot water in the reactor system passes over the hundreds of tubes in the steam generator, heat is transferred through the tube walls and the water inside the tubes turns to superheated steam.” Am I right in assuming then that the reactor water is at some quite high pressure?

        The similar module in the link puts one of the benefits as ‘load following’. Will the reactor vessel then by pressure and temperature cycled? LCF of the material may be quite onerous.

  6. thecliffclavenoffinance

    While extreme heat has become a problem for the Texas grid in recent years, there another problem that began in the 1980s.
    It led to a 3.2 million person blackout in February 2011 and a 5 million person blackout in February 2021. that problem is the Texas energy infrastructure does not function properly in extremely cold weather. The problem has not been solved and apparently no fix is in progress — too expensive.

    • Define “extreme” …

      Else it’s just pointless arm waving.

      • thecliffclavenoffinance

        If there are blackouts in very cold weather, that would be extreme cold

        If there are brownouts rolling blacks in very hot weather, that would be extreme heat.

        You practice pointless tongue waving.

  7. Beginning to think, if batteries are a good idea and in fact required for a renewables grid, it’s probably also a good idea for a fossil fuel-powered grid but not required.

  8. If you do hour by hour analysis the storage requirements get much larger. Either way it is impossibly expensive. The interesting question is how this green impossibility will manifest itself? Price spikes are most likely, then protracted blackouts. In either case the end is in view.

  9. Leslie MacMillan

    So if I wrap myself in a down duvet and thereby stay warm on a cold night, I am violating the Second Law of Thermodynamics?

  10. Here are some back-of-the-envelope wind-only results I came up with for ERCOT a couple of years ago. They suggest that with no curtailment you’d need over a month’s worth of battery storage:


    The Clausius statement of the corollary of the Second Law of Thermodynamics correctly quoted reads:

    “Heat can never pass from a colder to a warmer body without some other change, connected therewith, occurring at the same time.”

    Hence the one-way passage of GH gas back radiation from a cold region of the troposphere to a warmer region of the surface cannot cause an effective transfer of thermal energy (ie heat) into the surface that would raise the surface temperature. I say that because “No other change, connected therewith” is “occurring at the same time.”

    Ill-informed climatologists think they can excuse this obvious violation of the Second Law by claiming there is net energy out of the surface, but that surface cooling (some by evaporation or convection) is not “connected therewith” nor necessarily “occurring at the same time.” In fact the cooling generally occurs at a later time such as in the afternoon after morning warming, or in winter after summer warming.

    So that is the last nail in the coffin of the so-called “Radiative Forcing” conjecture which, when you study their energy diagrams, really boils down to their implied warming of the surface by such back radiation.

    The claim about slowing surface cooling is easily debunked because the surface is not a blackbody but they apply blackbody physics. Let me explain: radiation can only slow that portion of surface cooling which, in a small section of Earth where there has been a hot summer day, is itself cooling by radiation. But there is plenty of non-radiative cooling which simply accelerates to compensate so that the overall rate of surface cooling is barely, if at all, affected. Everywhere else, especially at night, it is the temperature of mostly nitrogen and oxygen in the atmosphere that slows surface cooling.

    For every one degree of warming or cooling of the troposphere carbon dioxide can only affect 0.0004 of a degree this calculation being based on percentage composition and the specific heat measurements for each of the gases.

    Now, since the greenhouse garbage for the gullible is thoroughly refuted by the Clausius statement, that leaves climatologists with no explanation as to why the surface is warmer than direct solar radiation can make it.

    The correct explanation depends upon those words “without some other change, connected therewith, occurring at the same time” because there can be such an effective transfer of thermal energy (ie heat) by non-radiative molecular collision processes. This is because, in the process I described in my 2013 paper “Planetary Core and Surface Temperatures” there is another change occurring, namely a change in the gravitational potential energy as gravity redistributes molecules. This is easily understood in regard to gravity forming the density gradient which is also the Second Law operating.

    You see, science has come a long way since the days of Clausius and that is why the Second Law of Thermodynamics states that “in a natural thermodynamic process, the sum of the entropies of the interacting thermodynamic systems never decreases.” You will note that the word “interacting” takes care of the Clausius condition “without some other change, connected therewith, occurring at the same time.” Scientists (around the 1980’s) began to realize that the concept of “entropy” was of vital importance and it extended the application of the Second Law to processes such as ice melting, balls rolling down planks, chemical reactions and in fact every single natural thermodynamic process occurring anywhere in the Universe.

    For more on this I recommend, especially to silent readers, that they visit my third climate website, read my papers and also note the compelling evidence written in my second site that is linked from my website.

  12. “estimated cost of nuclear power at $9 million /MWh” – think that the cost should be per MW

    • Leslie MacMillan

      If he’s talking capital cost to build it, yes, that’s likely what he means: $9 million per MW rated power capacity. But if you want to know capital + operating expenditure (including cost of capital and investment in the decommissioning fund) discounted over the expected lifespan, then the proper measure is Levelized Cost of Energy in $ per MW-hour, an energy unit. For nuclear, this is estimated at $160 with wide confidence bands around it, about 4X that of wind and solar. LCOE for wind and solar incorporates the fact that the generators work at their stated capacity only 1/8 to 1/4 of the time. LCOE for wind and solar does not include the external costs of compensating for that intermittency which, as the post demonstrates, are large.

  13. Pingback: ERCOT Renewable Energy: Reality Check – Watts Up With That?

  14. Pingback: Reality Check – Watts Up With That? - Lead Right News

  15. Back of the envelope calculations like this are useful to consider extreme cases and boundaries. From observation of your figure 4 it is obvious that this is a suboptimal scheme. A different combination of generation and storage would be optimum.

    You assume batteries for storage, but they are not always the best option for prolonged discharge. A long term storage option is often pumped hydro. Here is a global atlas of potential pumped hydro sites including Texas.
    Unfortunately Texas is not home to many potential sites other than in the north and west which are not close to the main load centers.
    Unfortunately you also want to stay isolated from other grids which takes away the option of diversifying geography of sources which is a great strength of the USA otherwise

    Another alternative if there is significant excess generation most of the time is to generate green hydrogen, some of which could be burned to generate electricity when required and the balance sold in growing domestic and international markets.

    And of course there is no reason that a combination of renewables and nuclear which would be much closer to the optimum. So my guess at a combination which could be close to optimum for Texas would be about one third of the wind/solar in your scheme, less batteries and more hydrogen, plus a fair amount of nuclear, perhaps 20GW ?. Of course connecting to other grids would help.

    A useful exercise

    • If you want to store electricity in batteries in order to use it as electricity later you get a round trip efficiency around 85%. In other words, you need to put in close to 18% more electricity than you are planning to use. This holds for short term storage. For long term storage you have to consider loss of charge, this varies with the battery chemistry chosen.

      You suggest hydrogen as a better solution. I don’t know why you think it is better, but here is why I do not think that is true.

      If you store electricity as hydrogen in order to use it as electricity later on, the best round trip efficiency you get is 30%. This comes about as 70% conversion efficiency to hydrogen at 6-7 bar ambient temperature. You get around 90% efficiency to store it as a liquid, slightly better stored as compressed gas but I’ll get to that in a bit. Then you have about 45% efficiency converting back to electricity, including the power electronics needed. These efficiencies are in series: 0,7 x 0,9 x 0,45 = 0,28.

      That is the best you get. It translates into the need for much, much more generating capacity in order to throw away most of the energy produced for a little bit of hydrogen.

      Then, if you need to store for more than a week you can not rely on liquid hydrogen. After a week you need to start handling boil-off, either just let it go or re-liquify.

      You don’t want to store it as a gas, since it will diffuse through the storage tank. Hydrogen is the substance with lowest energy density, so storing a lot of energy takes up huge volumes.

      The only viable way of storing huge amounts of hydrogen over long periods of time is to use metal hydrides or liquid organic carriers (heating oil or similar). The best density you get then is around 54 kg hydrogen per cubic metre, but you pay a huge price to do so. Release of hydrogen requires 10-13 kWh of heat per kg hydrogen released. Hydrogen holds 33,3 kWh per kg.

      Long term storage means your round trip efficiency closes on 10-15%.

      This is not considering safety.

      Hydrogen is not a solution unless you have abundant energy, otherwise known as nuclear. And if you have abundant energy, you don’t need storage.

      • Leslie MacMillan

        >Hydrogen is not a solution unless you have abundant energy . . .

        That’s the nub of the problem, isn’t it. The idea of “electrifying everything” really runs on the premise that nearly unlimited electricity will be available nearly free. Fusion, anyone?

        Consider what we want our prime generators to do: serve base load, respond to peak load, store electricity for later if there is an intermittency defect, cope with new demand such as electric cars and electric steel smelting and the mining and manufacture of things like car batteries and cement, AND still have some (a lot!) to remove net CO2 from the atmosphere to atone for all the emitting we’ve done since 1750. Remember that the climate activists don’t want us to get to just net zero, they want net negative. That means we have to stop burning fossil fuels AND do use non-emitting sources to drive direct atmospheric CO2 capture and storage or irreversible mineralization.

        Where ever is all that electricity going to come from?

      • The option of hydrogen is only suggested because according to Figures 4 and 5 there is a massive amount of electricity generated which is wasted.
        Otherwise I agree with all the comments about the shortcomings about hydrogen

    • There are not many efficient ways to use hydrogen as a fuel or energy source.

  16. Pingback: Fact Check - Increase with that? - News7g

  17. Pingback: ERCOT Renewable Energy: Reality Checkon November 1, 2022 at 10:00 am - Always Bet On Black

  18. Pingback: ERCOT Renewable Energy: Reality Check - The Crude Truth

  19. Building nuclear power plants at old coal facilities seems to make more sense than tearing down a wind farm to expand a coal mine such as is now happening in Germany.

    • Well, the Germans, who may be getting rather cold soon and already losing industry, may tend to disagree. Reality bites.

  20. The option of hydrogen is only suggested because according to Figures 4 and 5 there is a massive amount of electricity generated which is wasted.
    Otherwise I agree with all the comments about the shortcomings about hydrogen

  21. Off topic but fun.
    UN COP27 — It’s a gas, gas, gas
    By David Wojick

    The beginning: “Who would have guessed that a UN COP would turn into a promotion point for fossil fuel development? COP27 is looking that way, thanks to the energy crisis and Africa’s determination to develop itself. After all, these are developing countries, right? In this case some of them want to develop, use and sell their abundant natural gas resources. Imagine the green horror! Poor countries actually making money from fossil fuels.

    How, or even if, this rapidly emerging issue will get on the COP 27 table remains to be seen, but it is certainly a lively topic of discussion on the side. Europe needs gas and Africa has it; green agenda be damned.”

    Lots more in the article. Please share it.

    Go Africa!

  22. Pingback: ERCOT erneuerbare Energie: Realitäts-Check | EIKE - Europäisches Institut für Klima & Energie

  23. Pingback: ERCOT erneuerbare Energie: Realitäts-Check -

  24. This ERCOT essay may be a little too optimistic about renewable energy grid penetration.

    2 November 2022
    US clean-power build tumbles to three-year low:
    “US third quarter utility battery storage, onshore wind, and solar PV installations declined to a three-year low of 3.43GW, as project interconnection and supply chain challenges together conspired to hobble growth, according to last figures from the American Clean Power Association (ACP).

    Quarterly installs of renewable energy technology fell 22% compared to a year earlier and were also down 18% in the first nine months to 14.2GW versus the same 2021 period.”

  25. “Compare this with an estimated cost of nuclear power at $9 million /MWh, whereby that same amount of money for renewables ($800 billion) could build over 90 GW of nuclear power vs the 230,000 MW of renewables.” I find this sentence confusing. Presumably the gist is that nuclear is a bargain compared to renewables but the numbers seem to imply the opposite. Please revise!

  26. Where does the spent nuclear fuel go?

  27. Here’s the most interesting energy interview I’ve seen in a long time. B.F. Randall argues for using process heat from nuclear power for making diesel and jet fuel. Well worth a listen:

  28. Matthew R Marler

    Energy Meteorologist,

    Thank you for your essay.

Leave a Reply