This special interest group is for professionals to connect and discuss all types of carbon-free power alternatives, including nuclear, renewable, tidal and more.

Roger Arnold's picture
Director, Silverthorn Institute

Roger Arnold is a former software engineer and systems architect. He studied physics, math, and chemistry at Michigan State University's Honors College. After graduation, he worked in...

  • Member since 2018
  • 1,106 items added with 156,999 views
  • Jul 18, 2021

The linked article appears in PV Magazine, reporting on an MIT study published in Applied Energy. It's a straightforward and -- so far as I can tell -- an accurate report on the conclusions of the study. I have major reservations about the study itself. If it weren't for the MIT brand and the study's relevance to a critically important issue, I wouldn't be sharing this link. The published study is behind a pay wall. I haven't read it. However, I found a poster session for the study here. It gives more detail than the abstract published at Elsevier's Science Direct, but not enough to resolve my reservations about the study. A sampling:

  •  The study is supposed to deal with seasonal storage, but has lithium-ion batteries and hydrogen-fired gas turbines coming out as close competitors. From my own calculations, along with everything else I've read on the matter, lithium-ion batteries are not even remotely competitive for seasonal storage -- in California or almost anywhere else. It would take a 500-fold reduction in specific cost of energy storage to change that. 
  • Their calculations for the cost of green hydrogen presume that RE will be extensively available for powering electrolyzers at a price of only $10 per MWh.  That's an absurdly low figure. Sure, wholesale electricity prices do get that low or lower, in periods when surplus RE is being curtailed. But one can't support commercial production of green hydrogen from the occasional periods when RE is available at such low prices. Simply in terms of total energy, there wouldn't be enough MWh available to support a market. But worse, it would require a huge capital expenditure for electrolysis equipment that would only be used for a few hours a week. 
  • The study apparently assumes that hydrogen combustion turbines would be the choice for generating RES backing power from hydrogen. It ignores the likelihood that fuel cells would be a better choice. Trends suggest that high temperature SOFCs feeding CO2 Brayton cycle bottoming engines will soon be able to deliver higher efficiency with lower capital cost.

If anyone with a subscription to Applied Energy reads the full article and finds I'm wrong, please leave a comment.

Bob Meinetz's picture
Bob Meinetz on Jul 18, 2021

Thanks for your insights, Roger, your skepticism seems reasonable.

The cynic in me has a hard time understanding why comparing different types of energy storage isn't itself a false choice - that the necessity of storing gigawatthours of electricity at all isn't a simply the construction of another profit center by industrialists, one resulting in an exorbitant waste of money, effort, energy, and most importantly, time; that it isn't enabled by public fear of both climate change and an abundant, carbon-free source of energy that can be generated to efficiently meet demand in real time.

Have you had much experience in public advocacy/policy? I may be more experienced in that realm. Going up against $900/hour attorneys for Shell and Chevron before California's Public Utilities Commission, I think, would give you an appreciation for the depth of positioning, planning, and influence at work when profit available from increased consumption at that scale is at stake.

I'm more convinced than ever that the #1 obstacle to solving climate change is consumption. Until we can curb our appetite for making, selling, and buying "stuff", having any effect on climate change at all will remain an elusive goal.

Matt Chester's picture
Matt Chester on Jul 19, 2021

I'm more convinced than ever that the #1 obstacle to solving climate change is consumption.

Losing the weight before buying a new suit, if you will-- reign in consumption and maximize optimization/efficiency before continuing to build out the generation that's needed to fulfill unnecessary demand

Bob Meinetz's picture
Bob Meinetz on Jul 19, 2021

To your relevant analogy, Matt, I might add "Or not buying a new suit -  mending the one you have now. Avoiding the energy cost of harvesting and spinning cotton, weaving with synthetic threads, dyeing and other costs of textile production, trimming, sewing and shipping across the Pacific from China. Or the Atlantic from Italy, if you have disposable income (disposing income on unnecessary vanity items can even have a negative impact on your marriage)."

Peter Farley's picture
Peter Farley on Jul 28, 2021


Usually I strongly disagree with you, but on this we totally agree.

Even having cars with the same weight and performance as they did in the 80's and allowing house temperatures to vary from 50F in winter to 80F in summer would save a huge amount of energy

Matt Chester's picture
Matt Chester on Jul 29, 2021

These are good points, but they also highlight why initial attempts towards energy conservation in the 70s from President Carter faced resistance, resentment, and ultimately failure. By telling people the best way to be energy conscious was to engage in activities people saw as sacrifices-- like sacrificing their temperature of comfort-- it put them on the defensive. The goal was correct, and a less selfish people would perhaps embrace the message, but asking for sacrifice (even just perceived) is a tough hill to climb. 

Bob Meinetz's picture
Bob Meinetz on Jul 29, 2021

Energy conservation was ultimately a failure, Matt?

Those efforts are why we have an entire industry devoted to improving energy efficiency in homes and businesses today. They were responsible for one of the most influential pieces of U.S. energy legislation, the Public Utility Regulatory Policies Act of 1978 (PURPA). Its purpose was "to encourage cogeneration and renewable resources and promote competition for electric generation, and encourage electricity conservation."

The reason conservation succeeded in the 1970s was the same primal reason development of nuclear energy stalled: fear. After you've waited in a 3-hour line for gasoline, month after month, with your commute and job on the line, you gain an appreciation for how we take access to energy for granted (I know, I was there).

A parallel effort at the time was to promote "energy independence" from OPEC by drilling for more of our own oil. It's not hard to understand why these two solutions to the problem were increasingly in conflict with growing awareness of global warming.

Michael Keller's picture
Michael Keller on Jul 20, 2021

Technically, firing pure hydrogen in a combustion turbine is difficult, assuming you could consistently get enough hydrogen to actually run the machine.

Would be much easier to fire the hydrogen in duct burners to increase steam production by the combined-cycle power plant. Duct burners are routinely used with combined-cycle power plants to meet grid peaks and can increase plant output by roughly 10% to 20%. Can run this way for hours on end with no CO2 emissions associated with firing the burners on hydrogen. Hydrogen fired duct burners are already readily available.

Andrew Blakers's picture
Andrew Blakers on Jul 27, 2021

Neither batteries nor hydrogen are likely to be significant players in the long-term storage market (overnight-days-months)

Pumped hydro has 99% of the global electricity storage market because it is cheap and mature. There are unlimited off-river pumped hydro sites in most countries. Wide-area connection comprising strong HVDC links across north America vastly reduces daily and seasonal storage requirements by smoothing out solar & wind output.

Matt Chester's picture
Matt Chester on Jul 27, 2021

Neither batteries nor hydrogen are likely to be significant players in the long-term storage market (overnight-days-months)

Do you see hydrogen playing a role at all, or will any potential applications be outside the scope of a pure storage play? For example, is the potential for renewable overbuild to go into hydrogen production more of a byproduct of a hydrogen market, rather than the goal (e.g., because hydrogen to power is too inefficient to pursue as anything other than ancillary)?

Michael Keller's picture
Michael Keller on Jul 28, 2021

Ordinarily, the economics of hydrogen are a non-starter, as history has repeatedly shown. The energy required to make hydrogen is significant, the the resources has to then be distributed and finally serve some useful purpose. That cost is significantly higher than conventional approaches.

To the extent hysteria overwhelms common sense, then various colors of hydrogen may show up. However, consumer prices will significantly and needlessly rise while firms happily pocket extra profits.

Peter Farley's picture
Peter Farley on Jul 29, 2021

Smart charging of EVs, controllable loads such as hot water heating and precooling/heating of houses with solar, PCMs along the cold chain, and excess generation capacity will probably provide more backup than either pumped hydro or batteries.

But I think you are right, more flexible operation of existing hydro, combined with pumped hydro will provide more capacity than grid batteries 

Matt Chester's picture
Matt Chester on Jul 29, 2021

I'd say it's a feature, not a bug, to have these various different tools for energy storage and load shifting-- any single one of them would fall short of meeting grid needs, but by combining them in places where geography, resources, and markets allow them, the grid becomes more robust and reliable

Peter Farley's picture
Peter Farley on Jul 31, 2021

I agree, any practical system will have many technologies just as the current transport system does

Daniel Duggan's picture
Daniel Duggan on Jul 27, 2021


what you have uncovered is the tip of the renewable energy misinformation iceberg.  The narrative is something like this "all we need is lots of variable renewables plus some batteries, both of which are on an never-ending downward price spiral, and the job's done!" 

Energy storage in hydrogen can potentially be on a scale vastly greater than what is practically possible or economic with lithium-ion batteries.  Batteries supply grid services in MWh volumes, hydrogen can be a medium which stores GWh or possibly TWh sized quantities of zero carbon fuel sufficient to supply gas turbines during days, weeks and possibly months when drought, low wind and cloudy skies inhibit the production of renewable electricity.  Lots of focus on hydrogen these days, billions being invested, it's a likely next step in the move from dispatchable fossil power plants to ever greater dependency on variable renewables, however tales of  electricity at $10 / MWh are just another example of misinformation; how can a wind farm stay in business when the output is sold at this price?  On the contrary, hydrogen is going to be considerable more expensive than fossil fuel, and that's a fact.

Gas turbines operate today on hydrogen rich fuel, and with optimised burners and control logic we can expect to see a full range of turbines capable of operation within emissions limits on both hydrogen blend, and 100% hydrogen fuel available on the market within the next few years.  Its chicken and egg; the turbines are developed in parallel with very large quantities of hydrogen becoming available, and the emergence of a market for more costly but zero carbon power.  As much of the dispatchable synchronous generation necessary to support variable renewables will be operate intermittently, perhaps frequent starts and just a few hours operation following each start, hydrogen powered open cycle gas turbines are most likely.  Hydrogen fuel cells are not yet available in very large sizes needed to power a grid, also with the ability to last 100,000 - 200,000 hours required of a power plant.  A fuel cell in a Toyota car may last just 1,500 - 3,000 hours before its junked. 

Three Mile Island, Chernobyl and Fukushima must be rank as the three most expensive events in history, without them, we would probably have almost universal acceptance of nuclear power, and enjoy low cost low carbon electricity from thousands of safe modern reactors.

Michael Keller's picture
Michael Keller on Jul 28, 2021

Gas turbines have significant difficulty firing hydrogen because of a variety of difficult technical issues, including NOx control and instability. Further, the amount of hydrogen gas required is immense and the cost of hydrogen significantly exceeds that of natural gas.

Peter Farley's picture
Peter Farley on Jul 28, 2021

We actually don't need a lot of batteries, most of the storage will be in existing hydro, perhaps reconfigured to allow higher peak output. then thermal storage from such simple things as running water heating during the day rather than at night, turning the a/c on at 3 in the afternoon to precool/preheat the house etc. Then smart charging of EVs and last of all - large quantities of grid batteries. If the US can't run its energy system with no more than the equivalent on one day's energy demand stored in batteries it is a complete failure of imagination.

As for nuclear.

Plant Vogtle will produce 17 TWh/y for $25b ( that does not include the billions lost by Toshiba and CBI) with about $35/MWh in operating costs and use 40GL of water per year. It will be offline for an average of 300 hours per year.

Two dozen wind/solar/storage installations with a total capacity of 2.5 GW of wind, 3.5 GW of solar and 1.8 GW/6 GWh of storage would never be totally offline, would also produce 17 TWh/y, use no water and would cost about $10 billion to build and $15/MWh to operate.

Which do you think is a better investment ??? 

Bob Meinetz's picture
Bob Meinetz on Jul 29, 2021

I would ask for another breakdown of how 3.5 GW of solar and 1.8 GW/6 GWh of storage would cost "about $10 billion", Peter. But this time you'd need to explain:

• How your arrangement can generate 17 TWh/yr without being totally online, 24/7/365;

• How 6 GWh of storage could possibly compensate for an intermittent source of energy (California's grid would burn through its entire capacity minutes after the sun goes down);

• Why you haven't accounted for added solar capacity necessary to charge batteries, while your 3.5 GW is powering the grid;

• Why you haven't accounted for 10-20% efficiency losses for Li-ion batteries and bi-directional AC-DC-AC inversion;
• Why you haven't tripled your figure to account for the fact nuclear plants have lifetimes 2-3x as long, or longer;
• Why your solution fails to account for the purchase or leasing of 1000x as much land and/or transmission;
• Why you believe any solar + storage facility is charging its batteries from the adjacent solar farm, and not a reliable grid connection (powered in part by fossil fuels);

• Why you believe nuclear plants with once-through cooling "use" water (Diablo Canyon returns every drop of water it borrows from the Pacific Ocean).

Except in the fevered imagination of its advocates "solar + storage" is incapable of powering an electricity grid - a fact trivially obvious to an electrical engineer. It never has, and never will (though the reasons go on, my patience does not).

Peter Farley's picture
Peter Farley on Jul 31, 2021

It should be obvious that 6 GW of generation capacity does not have to be online 24/7/365 to match 2.2GW of generation that is only on line 24/7/330. It should also be fairly clear that when the solar is generating 3.5 GW it is supplying far more power than a 2.2GW nuclear plant and the excess is used for charging the storage

2.2 GW nuclear 90% CF = average output 1.98 GW


2.5 GW of large rotor tall tower wind in the US at 45% CF = 1.125 GW average


3.5 GW of tracking bifacial solar at 28% CF = 0.98 GW average

= 2.1 GW average output,

Whenever supply exceeds 1.98 GW i.e. nominally half the time, the surplus is diverted to storage and/or used to reduce output from hydro and gas facilities.

Very little of the power actually passes through the batteries, most of it is delivered direct to loads

The storage capacity mentioned in my example does not have to supply California it has to supply with residual wind and solar the equivalent of Diablo Canyon nuclear plant. If you look through the EIA figures for California while wind and solar output can be very low at times, the combined output is never quite zero and over a day would never be less than 5% of nameplate capacity for wind and 15% of nameplate for tracking solar. i.e. worst day wind and solar output is about 0.7 GWh combined with 6GWh from storage that is 6.7 GWh. Worst day output from Diablo Canyon is zero if there is an outage on one reactor while the other is refueling eg October 2020  

You can easily make the wind and solar plants last 2-3 times longer if you have the maintenance budget of a nuclear plant. The difference is that with gradually improving technologies a 100 turbine wind farm in 30 years time would be producing 2-3 times the energy that a current farm does and solar could be doing the same if the current generation 19% panels are replaced with 27-35% multijunction cells in 15-25 years

 Most US nuclear plants are not on the sea shore and those that are are becoming increasingly vulnerable to storm activity. Even if that problem could be solved a "once through" cooling system draw in 70-100 tonnes per second per GW and raise the temperature 7-8C which is devastating for marine life.

When nuclear plants draw water from natural water sources, fish and other wildlife get caught in the cooling system water intake structures. While this is an issue for all power plants with water-cooled systems, a study completed in 2005 in Southern California indicates that the problem is more acute for nuclear facilities. The study investigated impacts from 11 coastal power plants and estimated that in 2003, a single nuclear plant killed close to 3.5 million fish--32 times more than the combined impact of all of the other plants in the study.[15] 

Evaporative cooling raws far less water but of course does not return any of it to the stream

Despite your willfull misrepresentation or at best lazy misreading of my posts I have never advocated for solar+ storage as the sole source of supply on the grid. I have always advocated a mix of technologies.

However just to show how silly your statement is, some years ago NREL calculated that 14% of US roofspace could supply 1,200 TWh, 30% of US power needs with 16% efficient solar panels. Now that east west facing roofs are economical for solar, about 30% of US roofspace is available for solar. If carparks, railway platforms etc are fitted with solar canopies that increases the area by about 20%. Within 3-5 years 25% solar panels will become available. Thus over the next 15 years rooftop solar could be expanded to 1,200 x 25/16 x 30/14 x 1.2 =  4,800 TWh, 20% more than the entire US electrical demand, with effectively zero land use or transmission infrastructure. Of that about 55% would be delivered direct to loads and the rest via storage, some of which will be thermal and some via batteries let's say an average of 82% efficiency for storage. That would mean that effective output would be reduced to 4,400 TWh, almost all of it delivered directly to the load. The 4,050 TWh generated by the current US grid, net of energy used to mine, process and transport coal, gas and uranium and HV transmission losses is net 3,400 TWh so it is clearly possible to support the entire grid on solar and storage if you were silly enough to want to.

Bob Meinetz's picture
Bob Meinetz on Aug 3, 2021

"It should be obvious that 6 GW of generation capacity does not have to be online 24/7/365 to match 2.2GW of generation that is only on line 24/7/330."

Peter, a common mistake among those inexperienced with principles of grid energy is the belief that 2.2GW of average power is equivalent to (is a "match" for) 2.2GW of constant power. It isn't - sometimes at a windfarm there's no wind at all. Sometimes there's very little wind across tens of thousands of square miles. Sometimes, it lasts for days. And just because it's more windy at other times of the year doesn't "make up for" depriving millions of residents and/or businesses of access to all electricity, even for a day or two.
Another common mistake is that all outages at nuclear plants are unplanned. More than 99% of shutdowns are planned refueling outages, and 59% of U.S. plants have never had an unplanned automatic shutdown. At this moment, nuclear plants in the U.S are running at 97% capacity, with 49 plants running at 100% capacity. I would challenge you to show me 49 individual wind turbines that are simultaneously producing 100% of their rated capacity - but I know you can't.

"the combined output is never quite zero..."

Never say never:

"NREL calculated that 14% of US roofspace could supply 1,200 TWh..."

Of course the National Renewable Energy Laboratory (NREL) "calculated" that. They're paid to promote renewable energy (it's part of their Mission Statement).

Michael Keller's picture
Michael Keller on Aug 3, 2021

Bob is exactly right. 

Peter Farley's picture
Peter Farley on Aug 7, 2021

You show a 2012 video of a single windfarm that is not generating. It is also a very sunny day where any nearby solar farm would have been generating at near full capacity. The combined resource would have been well above zero, 

On the other hand Your favoured nuclear technology was offline for weeks. That is the equivalent of 3,000 US wind turbines being offline simultaneously 24/7 for 3-4 weeks. Never happens.

Surprisingly the grid does not demand constant power, so even in France there are times where nuclear supplies only about 50% of demand, other times where plants are shut down altogether and exports to Switzerland, Italy etc are ramped up to use the excess power.

For the cost of plant Vogtle, approaching $30 bn including the losses by the original contractors, you can build 7 GW of wind 7 GW of tracking solar and 4.5 GW/ 60 GWh of storage. Even in a dull low wind year, the plant will have a minimum 250% more output than Vogtle and cost half as much to run. Minimum output will be always be positive. In contrast at 95% availability, (high for a nuclear plant), the nuclear plant will be limited to a maximum of 50% capacity for about 450 hours per year and occassionally for week or three at a time (see Diablo Canyon above) will produce nothing. 


Bob Meinetz's picture
Bob Meinetz on Aug 9, 2021

Peter, Diablo Canyon experienced one (1) unscheduled shutdown in July '20. The plant was taken down deliberately in the fall to perform needed maintenance - because demand was lower. When wind farm owners can tell the wind to blow slower and faster to match customer demand, you'll have a point.

"For the cost of plant Vogtle, approaching $30 bn including the losses by the original contractors, you can build 7 GW of wind 7 GW of tracking solar and 4.5 GW/ 60 GWh of storage."

No, no, no. Solar power in Georgia has a capacity factor of 23%, wind's CF in the Peach State is a pathetic 6%. Maybe you missed it above, but the average capacity factors of all U.S. nuclear plants is currently 97% - it's not even close. Your costs for intermittent wind, and solar, and storage, and wind to charge the storage, and solar to charge the storage, are at least six times too low - and we haven't even begun to take the limited lifetimes of solar and wind farms into account.
Independent of weather, independent of time-of-day, Vogtle is an investment in serious energy for serious applications well into the future. I know it will take some adjustment, but nuclear is the carbon-free energy source of the future.

Michael Keller's picture
Michael Keller on Jul 30, 2021

Not your two dozen wind turbine or solar installations, which could not remotely replace a large nuclear unit because of renewable energy’s dismal capacity factor and unreliable operation. The nuclear unit easily operates at full power at a +90% capacity factor.

You really need to familiarize yourself with the difference between energy (megaWatt-hours) and power (megawatts).

Peter Farley's picture
Peter Farley on Jul 31, 2021

You seem to have some difficulty with arithmetic.

Many modern windfarms are exceeding 45% Capacity factor a few more than 55%, similarly some solar farms are exceeding 30% but to be conservative we will use 45% x 2.5 GW for wind + 28% x 3.5 GW for solar = 2.1 GW average = 18,400 GWh/y


90% x 2.2 GW for nuclear = 1.98 GW average = 17,344 GWh/y

Michael Keller's picture
Michael Keller on Aug 3, 2021

You have difficulty with reality.
The capacity factor of renewable energy assets is highly dependent on where the devices are located. Cherry picking ideal locations is disingenuous; what counts is the average.

Peter Farley's picture
Peter Farley on Aug 7, 2021

45% is a typical capacity factor for modern windfarms in the US not the exception, 31% is the capacity factor for the best US solar farms so I used 28%. However new coatings on solar panels eneable them to collect more off axis energy, lower thermal co-efficients allow more hot day production and bifacial panels collect 8-15% more energy per year so there will be solar farms in sites with lower insolation will still hit 28-30%.

Similarly the efficiency of wind turbine blades, is improving slowly, towers are getting taller and the ratio of swept area to generator size is increasing. All of these improvements are moving wind farm capacity up towards 55% in the best cases and commonly 48-50% 

But lets be even more conservative and redo the above calculation with 3 GW of wind at 40% and 4 GW of solar at 25%, it still produces 15% more energy than the nuclear plant. And lets triple the duration of the storage so the investment now becomes $13 bn less than half the nuclear plant

Michael Keller's picture
Michael Keller on Jul 29, 2021

As for hydro, you need water and large elevation differences. Those two items are generally in short supply. Also, perhaps you missed it, but the environmentalists are hell bent on tearing down dams.

Andrew Blakers's picture
Andrew Blakers on Jul 29, 2021

Michael, there are 35,000 off-river pumped sites in the USA with heads in the range 100-800m which is 100X more than needed to support 100% renewables. Water requirements are 10X smaller than equivalent coal plant.

Please read

Michael Keller's picture
Michael Keller on Jul 30, 2021

In the Midwest, even a 100 meter elevation difference is not realistic. In the West and East, elevation differences are more plentiful. However, the West is running out of water.

Thermal power plant water requirements are a function of whether or not the plant uses cooling tower. Using once-thru cooling does not dump water into the atmosphere as is the case with cooling towers. Some thermal plants use air cooled condensers, drastically reducing water requirements, but a few percent drop in output and efficiency occurs.

I generally do not have a problem with dams, but the available locations are being increasingly diminished by those opposed to dams and those advocating removal of dams.

Bob Meinetz's picture
Bob Meinetz on Aug 4, 2021

On Friday in California, the Helms Pumped Storage Facility reservoir - from where water was to be pumped uphill to save clean energy from CA wind and solar farms - was empty. The pumps were pumping air uphill, and storing very little potential energy in the process.

With the grid on the verge of collapse, Gov. Gavin Newsom shortly thereafter declared a state of emergency.

Renewables + storage = worse than useless.

Peter Farley's picture
Peter Farley on Jul 29, 2021

Most of the backup will be provided as it is now, by excees generating capacity. The US fossil fuel fleet at safe operating capacities can provide about 3,600 TWh/y. It actually provides about 2,400 TWh. If you build enough new wind, solar, geothermal, waste to energy etc to provide 3,600 TWh in an average year, for about 50 weeks per year there will be enough electricity. In the worst weeks without nuclear you would be 20% short. However by managing hydro more as a backup resource and using existing nuclear, the shortfall would be about 10-12% of demand over the week. In turn well over half of that will be provided by thermal storage of some sort, smart charging of EVs etc leaving the need for the equivalent of between 1 and 2 days total energy use for backup

 Of course as Bob implies below, if you were as energy efficient as Italy or Spain you could halve your electricity use. Peak demand in Italy works out at 1 kW/person, In Spain about 0.8kW. In Texas it is 3 kW. In Southwest Power Pool 2.6 kW

Roger Arnold's picture
Thank Roger for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member

Get Published - Build a Following

The Energy Central Power Industry Network® is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »