Seeking Consensus on the Internalized Costs of Mature Energy Storage Technologies

CORRECTION? The indicated breakeven spread appears to be based only on storage used for arbitrage. It does not consider the role of storage systems for voltage and frequency stabilization or for contingency reserve. It’s not clear to me how one translates those functions to internalized cost of the storage technology, but they do strongly affect the breakeven spread and the economics of deployment.

CORRECTION?  There’s a dimension to profile cost for fossil fuel as storage that is significant but hard to quantify.  That’s the difference between using installed older technology vs. acquiring newer flexible generation technology.  

Older technology is capable of backing intermittent renewables, but at a cost.  The cost is more than just the obvious CF-related underutilization that your model captures.  It includes higher specific fuel consumption and increased maintenance / shorter lifespan.  The costs are hard to quantify because they are non-linear with degree of penetration and highly dependent on the specific equipment installed.

New fossil-fueled technology is available that largely eliminates those non-CF-related costs.  It provides a relatively flat power band that allows for a degree of load following with minimal loss of fuel efficiency or increase in wear. But acquisition is a large capital expense that is hard to justify in a steady or shrinking power market.  So utilities generally make do with what they have.  

$10/MWh is perhaps a reasonable ballpark number “on average”. But any specific “per MWh” number implies a linear cost, whereas the salient point is that the cost is not at all linear.  

At penetration levels of a few percent, the non-CF related cost for this type of “storage” is negligible, even for systems with older generating technologies.  The variations encountered will remain within the limited throttling range of the generators.  At higher penetrations, however, the increase in variability will exceed the throttling range for efficient operation.  Then it becomes necessary to operate some equipment below its efficient power band, and / or increase the frequency of stressful and inefficient start-up and shut-down of individual generators.  

At some level of penetration, the variability becomes too much for the system to handle, and the grid breaks down.  But loss of efficiency can be significant well before the point of breakdown.  The problem is that it’s invisible to the usual cost / benefit analysis for renewables.

Rebuttal: 

I think the baseline costing case should assume the grid needs new capacity to meet demand.  Global energy use is growing, so this case is more representative than the special case of flat or shrinking demand.  

DATA: round trip energy efficiency of CAES w/ natural gas re-heat is only 35.7%.

I calculated this based on the 2005 Ridge Energy Storage proposal (website, whitepaper) from Texas (which was never built), by calculating:

efficiency = (net electricity out)/(electricity in),

where (net electricity out)  is the nameplate output minus the electricity that would have been produced by a modern combined cycle gas plant with the same fuel flow rate (heat rate).

The Ridge proposal would have used 200 MW for compression, and 270 MW output, and provided 1 MWh out for each 0.8 MWh electrical input.  The heat rate was predicted to be 4500 Btu/MWh, compared to a CC plant heat rate of 5700 Btu/MWh.

This same calculation gives a more viable 68% efficient, when the natural gas is valued at 7500 Btu/MWh, as would be the case for a simple combustion turbine.  So it might make sense when very fast throttling capability is needed, and combustion turbines are the only alternative.

“…good data on the cost of fossil fuel backup at a wind/solar penetration of 10-20%..”
Those costs are highly dependent on the capability of grid management, especially its production predictions.
Normally grid management can predict that production some days in advance within a few percent, based on wheather predictions (increasing accuracy when the period shortens).

With 10-20% the changes for the remaining power plants are similar to the normal load changes.

Hydro
When there is 5-10% stored hydro, then the cost of fossil fuel backup will be near zero.
Management of hydro will utilize the stored energy capacity only when prices are high.

Assuming they run their facilitiy ~30% of the time, they will compensate the variability of ~15-30% wind/solar penetration.
So the costs of FF backup are then zero.

This presentation shows the different storage technologies and their position in the costs / storage duration (sheet 19). It also shows that:

– Load changes in Germany are >50%.
So 20% wind+solar generate only small changes.

– the electricity costs rise with only €1/MWh (from €77 to €78/MWh) when the share of renewable rises from the 2010 situation (7.8% wind+solar) to 40% (=30% wind+solar).

Note that wind+solar are very flexible;
Solar installations can be switched on/of automatic in a fraction of a second
Wind can be up/down regulated between 0% and 100% in a few minutes.
In Germany grid management made agreements so they can do that themselves directly.

If the system is optimized it may well be that 10-20% share of wind+solar decrease the total costs (incl. compensation for switching off wind+solar capacity). Because:
– wind+solar can absorb fast load changes easily, so power plants can avoid inefficiencies and continue to operate optimal.
– wind+solar production is rather accurately predicted days ahead now (weather). No sudden surprises as with a big power plant failure. So less spinning reserves needed.

German experience shows that the distributed (less vulnerable to power line outages) and easy to switch on/off wind/solar capacity, increased reliability with a factor two; total outage now ~15min/a per customer. That also generates important value.
Seems to me the USA figure need some improvement (connections now ~2hrs/a down).

——–
Your article uses the biased Vattenfall simulation study produced by then Vattenfal employee Hirth, who now found a few others at the Potsdam. Vattenfall has big interest in power plants (nuclear, fossil), little in wind, none in solar. The study conclusions are in line with these interests.

The German Energiewende scientists (Universities, Agora, Fraunhofer) concluded that:
– no storage is needed until ~30% wind+solar (pumped storage makes losses in Germany).
– until ~70% wind+solar the extra costs to integrate wind+solar production (system costs) will be small.

Those German conclusions are in line with the results of:
– US studies, such as by NREL;
– international studies such as this study by the International Energy Agency.

How would thermal power plants reliant on fossil fuels be considered energy storage?  Unless of course the fuels used in them are somehow generated using renewable energy, though that is not what this article seems to be about.  Energy storage is about not needing fossil fuel based backup.  To include this technology in a discussion of “energy storage” is misleading and not conducive to an honest discussion of moving beyond the intermittancy of renewables.  

Storage is not that difficult, what is difficult is the expectation that it will be as cheap as oil/coal has been to this point.  We hit the jackpot in stored energy, literally millions if not billions of years of storage, and have used it in a couple of centuries.  The economic model based on this bonanza is FUBAR!  As reasoning and thinking entities one would hope to be able as a species to move beyond the baby kicking our feet tantrum of “I WANT” to a more realistic view of what actually is.  If we don’t stop with this lack of reality acceptance old Ma Nature will inevitably provide the swift kick in the rear as history tells us over and over!

Global energy use is growing, certainly.  But that’s mostly in China and a few other places whose energy policies have little to do with anything we discuss here.  In the U.S. and the E.U., the market for utility power has been shrinking.  Not dramatically, but enough to influence decisions about new capacity. That’s one reason why new nuclear plants are hard to sell.

Utilities here are reluctant to follow Germany’s lead and retire assets early so they can be replaced with new technology “flexible” thermal plants.  The result is to distort the cost / benefit analyis of renewables and the policies that support them.  The actual benefits are less than anticipated, the costs more.  Our policies are blind to the inefficiencies that follow from accomodation of “must take” RE with the currently installed equipment base.

Snce I don’t know how to account for the additional costs outside of detailed studies of individual RBAs (Regional Balancing Areas), I’ll concede your point: the costing case should apply to an assumed requirement for new capacity.  But I’d insist on a footnote about policy differences for cases where renewables are intended to replace existing capacity.

@Schalk,
Thanks for your excellent response!

Here you find the presentations of a recent German storage study day, regarding best methods with ~40%-90% renewable.
In short:

– (sheet 31) Wind+solar capacity ~140GW in 2030 = twice the max. capacity Germany needs.
– (sh.36) Development of storage costs. Pumped Storage (‘PSW’) may never become competitive.

– (sh.44) Redispatch costs  €60mln/a with 45-50% renewable (=30-35% wind+solar).
That is <0.1cnt/KWh. New storage capacity makes no sense with that share of renewable.

– (sh. 45) With renewable share of 69%, extension of Power-to-Gas installations is adviced (seasonal storage). However others think, EU-wide grid extension may be cheaper (sh.59).

– (sh 50-62) Assume a renewable share of 88% in Germany and 82% in the relevant EU countries. What would be optimal? Different scenario’s.

Scenario C; high share wind+PVsolar.
Economic storage capacity: 5-9GW (= ~10-18% of av. consumption) for 4-6hrs.

In general; DSM (Demand Side Management) has important beneficial influence on the costs. So smart grid will become really economic.

Fossil fuels aren’t storage unless they are being made not mined and thus are in themselves off topic for an article titled “Energy Storage”.  Let’s not discuss the problem as if it was the solution!

Schalk, thanks for your work, always interesting.

If fossil fuel is stored solar energy, uranium and thorium are stored solar energy from an older star that went super nova.

If we mass produce the simplest molten salt reactor design we can have thousands of years of reliable disspatchable low cost stored solar energy.

By mass producing a simple design, it can be cheap enough to be profitable at capacity factors down to 50%, eliminating the need for unreliables, wind and solar, or any other storage or source of energy. There would be no need to replace aging fossil plants with new fossil plants.

Germany seems to move toward towards electricity-to-gas/fuel conversion as an important factor to overcome the seasonal variability of wind+solar.

Ideas are that power plants burning this renewable generated gas/fuel in combination with biomass & waste, will fill the gaps that wind+solar+batteries leave (especially in winter).

So the combination will deliver the 100% renewable electricity during the whole year.
Even if the country cannot use hydro or geo-thermal.

“…storage will not become economic until very high wind/solar penetrations…”
At grid scale; yes.
Especially since the general conclusion in Germany is that grid extension & improvements can solve most of the intermittency issues, while being much cheaper and having other advantages (more redundancy, etc).
Thought pumped storage would become economic with ~35% wind+solar penetration. Now I doubt that. May be the minimum wind+solar penetration is ~45%.
And when that 45% is reached in ~2030, grid batteries may be cheaper than pumped storage.

Different situation for customers.
For customers that have PV-solar, batteries may become economic at ~2020. Those store the cheap PV-solar electricity for use in the evening/night (so 4-6hrs storage), avoiding the 28cnt/KWh electricity from the grid.
The subsidy program to install those batteries, 30% investment subsidy only for small rooftop owners, is already a big success. So those battery installations only have to become ~30% cheaper.
I estimate that in 2030 most (90%) PV-installations will have such batteries.

“…€60/MWh balancing costs (primarily profile costs from underutilization of thermal plants and curtailment of wind/solar…”
Check sheet 72 (my translation). Conclusion 3
Also with renewable share of 90% in Germany and 80% in Europe, supply and demand can be matched without much extra storage capacity and only 1% curtailment of wind/solar.

That €60/MWh: May be with old base load power plants only, and the excellent predictability of wind+solar production not used, etc. (German grid management has special weather forecast, and can see wind flaws passing through their area in 8hrs as they see the production of the windturbines).

Thanks to the predictabilty of wind+solar, power plants that take 24-48hrs to come online can be switched off. Considering also that less spinning reserve capacity is needed (>1000x more generators, dispersed over the delivery area), the average load of the running power plants won’t be much lower.

And Germany’s power plants are now far more flexible. Utilities replace old baseload plants with new flexible plants that keep their high efficiencies at low load, require less staff, etc.
Thanks to the 50years long scenario, utilities could wait until the old plants are written off and then replace those so little losses. Even the hated NPP’s can run until they are written off (30-35years).

Energiewende experts will laugh about that €60/MWh figure, their estimations are at ~€10/MWh with ~50% renewable (=35% wind+solar).

Indications on how DSM will work in practice?
At the moment German grid operators have some agreements with electricity intense facilities (factories, etc) that they switch part of their machines off when asked (often by phone). Only few in which grid management can control the consumption.

Smart grid is coming in The Netherlands & Germany, there is a program to install smart meters, so everybody will have them in a few years. You get one if you install rooftop solar. However those are still rather primitive. Those allow for changing tarrifs, etc.
But automatic switching of equipment (dish washer, dryer, rooftop solar, etc) will take another 10years I think. There are issues regarding standards, etc.
We (I too) have a primitive time-of-day tariff; Electricity is cheaper in the night and in weekends. So I set the timer of my dish washer such that it starts at 2 AM.

At the supply side, grid management itself can reguate the output of some wind parks; that is faster and more accurate. But little to regulate if there is no wind.
I’m not aware of similar regarding PV-solar parks. Regulating PV-solar output is really fast; similar as with batteries & flywheels.
(owners get compensation for missed production)

“… efficiency … base power consumption drops by almost 20% over the next four decades…”
Many folks on the forum here have the strange idea that you need more energy to live more comfortable. My experience is that living in a better isolated, well ventilated house delivers more luxus, while delivering lower energy bills.

The trend is set by Denmark (~15years ahead of Germany).
Regarding new houses; only 100% energy neutral houses are allowed!
Germany may follow that example at ~2030 (some debate in Germany about it).

“…how much energy intensive industry Germany manages to maintain… new legislation…”
German energy intense industry gets lower rates than our energy intensive industry, and that in UK, and and USA; they pay whole sale rates. And Germany has lowest whole sale rates now (~€35/MWh).
We in NL recently lost our aluminum smelter because German electricity is cheaper.

The new legislation creates some marginal increases only for less energy intense industries. It is a compromise with Brussels in which Merkel got what she wanted. Normally Brussels would give a huge fine such as the ~$2billion to Microsoft, and ban the system.

Language
Sorry, the most interesting stuff is nearly always in German.
I found an only slightly older English interim report, that spends some attention to system costs (Agora, Consentec and Fraunhofer).

I don’t know if this fits your screened criteria, but there is an emerging consensus on stored energy in Minnesota worth mention. A good starting reference is the Department of Natural Resources scientist, Anna Dirkswager. Anna was among several presenters at a Grand Rapids, MN. meeting held April, 2014.

Grand Rapids is where Blandin Foundation (Finland) has offices. The idea is to use enormous amounts of stored energy in American forests. Other groups mentioned include the US Forest Service, US EPA, USDA, other state agencies, and several non-profit advocacies. Sweden has had a group here.

I don’t know where it stands. But I had the opportunity to speak with Anna and suggested including a comparison with fuel oxygenate success of gasoline/ethanol; coal combustion efficiency and EPA emissions concerns. I also urged others to consider biomass railroad transport, given that the trucks now used are VERY heavily loaded.

Since North American forests are roughly the size of Europe, and growing at unprecedented rates due to new CO2, and as Anna said “will burn one way or another,” we can’t completely ignore it.

Having benefitted from Minnesota Department of Natural Resources experts for well over half a century, I’m delighted by their leadership. Please refer to them for numbers and many further references.

I believe that the intent of listing fossil fuel under storage was to be able to account for the cost associated with using them to cover the external cost of added variability due to solar and wind power.  We need to account for this extra cost when comparing solar and wind to energy sources like geothermal and nuclear which do not produce additional variability.

This is very important, as it helps to explain that even when solar and wind match the cost of conventional generation, adding them to the generation mix will make the total cost of electricity go up! (i.e. the annectodotal reports of wholesale power costs going down when wind power is added must be ignoring some costs somewhere in the system).  

Then it needs to be dealt with under a separate heading.  It is NOT a storage option.  To have it as a primary option in a storage discussion is entirely misleading.

Just a little bit of molten salt storage could be the MSR’s way to deal with the daily grid fluctuations. I believe it is the most efficient storage option (in this case) since the generator already relies on thermal for generation.

It is only misleading that storage options will need FF backup because of costs of said storage. Note that the graphs only go up to about 40% capacity. If they were to back up solar and wind 100%, then, there would be no need of the mention of FF’s. I believe these costs are based on renewable mandates rather than a sincere plan to save the biosphere (that RE forcings arise from the love of money). The best way is to find the cheapest storage for the application (for example, molten salt for excess thermal generation and pumped storage for excess electrical generation).

The “open market” you describe, wherein electrcial energy is bought using the lowest bid for each hourly or sub-hourly interval, was created to generate profits for merchant fossil fuel powered plants by taking market share from utility owned power plants.  It is not the only type of market possible, nor is it optimal for other situations.

In the US, the market design in which most wind and utility solar power is purchases is using a 20 year power purchase agreement.  In this type of market, the buyer agrees to buy all of the delivered output from a plant for an agreed-upon price schedule.  A similar market design is being used in the UK for new nuclear power plants.

In Canada, some nuclear power is purchases as a public-private partnership.  The plant is owned by the public utility, and companies bid the contract to operate the plant for a period of time. 

In Europe, there is some talk of splitting the electricity market into energy and capacity markets (the energy market is the conventional purchase of sub-hourly MWh energy units; in the capacity market, the supplier agrees to keep a certain number of MWatts of power plant available to sell energy, regardless of whether any is sold).  In the past, when fossil fuel merchant producers dominated the market, they effectively gave away the capacity for free, assuming there were able to sell enough energy to stay in business.  With the growth of energy-only merchants (i.e. wind and solar PV), utilities may need to buy the capacity separately.

In China, merchant power plant (typically coal-fired) owners are guaranteed a minimum energy sales per year.

All of these are valid market designs, and the choice of market design will clearly be driven by what energy sources are preferred by society.

As described in this article on the American Nuclear Society blog, it is not really true that nuclear plants can’t load follow.  But it is almost always the case in a mixed generating portfolio, energy sources with low fuel cost (i.e. nuclear, wind, solar, geothermal) get dispatched first, and high cost fuels will only get used when demand hits peaks.  A big difference is because of their high capacity factor, nuclear and geothermal can provide all of the baseload demand, whereas wind and solar must split the baseload demand with fossil fuel backup.

You can’t make solar’s and wind’s integration costs go away by proclaiming it to be so (no matter how many solar enthusiasts side with you).  Please read some of Schalk’s references.

I’m not sure how biomass energy use will develop, Schalk.

One interesting possibility is beginning to appear in national advertising. Koch Industries includes their Georgia Pacific wood processing giant, as well as their large energy portfolio, in a PR effort on broadcast TV channels. I have not seen this effort until recently.

There are certainly many good scientists with many good ideas. In simplest physics terms I do know trees have gravity on their side compared to digging in the ground for energy. Nothing seems cheap or easy to me. Yet somehow an amazing energy system works.

“…not really true that nuclear plants can’t load follow….”
The experience based graph below shows the inability of nuclear to load follow.
Nuclear continues to produce at >70% of max. even while prices are negative!

Waste to fuels is a good idea at the dump. Nevertheless it can’t power much (because it is negative EROI). Once technology is around to convert trees into methanol, that’s end of story, like a trillion times worse than nuclear (and even worse than coal) because greed will set in and devour what’s left of the forests (and the dead stuff that needs to be left in the ground to make good soil for CO2 retention!

High temp nuclear could be developed in line with molten salt (heat) storage. Such could then be of a lower nameplate capacity and still deal with daily grid fluctuations as long as the turbine itself can be throttled up and down.

“…How on Earth do you get 3-6c/kWh for rooftop solar in the US?”
Big solar in the US is already within that range, as confirmed by the Austin PPA.

Assuming:
– US installers will learn, and migrate towards similar efficiency as the German installers;
– the big import tariffs (~40%) will be off in 10years (China complained at the GATT);
– progress in efficiency continues (Sunpower plans to deliver 23% efficient panels next year);
You can expect that price levels in USA in 2030 or so

Remember that the present Feed-in-Tariff for rooftop solar is 13cent/KWh in Germany (going down with 1%/month). As av. insolation in USA is ~30% better, this translates to 13$cnt/KWh.

With the long term decrease of 8%/a that price level will be reached in a decade.
Taking into account the fast technology improvements the labs find, I estimate it will go down further towards <3$cnt/Kwh for rooftop and wall-covers in ~2035 (using thin-film).

Nearly all new houses will get a roof with integrated solar. Last year I saw the first of those in Italy. This spring the first in the Nethelands.

I don’t see the relevance of data on German nuclear plants which are not representative a plants which would be built today (and which apparently lack the load following features of their French competitors which were built in the same timeframe).

Also, you are mistaken in citing the examples of Scotland and Denmark.  These are not complete grids, but small pieces of larger grids with much lower renewable penetration.  

German policians can promise whatever they like for 2050; it will someone else’s problem to actually make it work (and as I have said, Germany already has a very large amount of energy storage on their grid:  enough for 15% of their average demand; you’re counter claims that their storage is not cost effective under the  current market design does not prove that it is not needed).  

There is a difference between doing and talking. Since I’m sweaty and need a break from the wood chipper, allow me to try restrain my disappoinment with the ignorant.

I have 50 more poplars laying on the ground (for next year) to make room for DNR supplied trees (like white cedar) to enhance biodiversity. Tree swallows are everywhere. Raspberries ripening, ready to give away (some Indians once traded walleye). I’m a Biophysicist by training, my wife a food scientist. I got “solar biofuels” on the national reasearch agenda. I can’t keep up as it is, so please take a hike.

I don’t believe we should use the forests as industrial energy supply source because it is not sustainable (at the large scale). Other than building materials, it should only be used for small scale (wood burners, etc). If there is “too much”, then we should somehow sequester the CO2 from the wood and isolate it “forever” to try to reverse the excess CO2 that is causing warming and chemical changes. If possible, this should be done to the bark beetled wood before it burns from wildfire.

I definitely agree that we should not use virgin crops for energy. However, although a good idea (being that you addressed the more important concerns about good soil) wastes can only be a rather small part of the overall energy picture. And thanks for the link, I want to learn more about biochar!

“…forests as industrial energy supply source … not sustainable (at the large scale). Other than building materials, it should only be used for small scale (wood burners, etc)…”

Small scale wood burners create big scale pollution, which is worse than large scale wood burners.

Of course it is sustainable with good wood management as shown in NW-Europe (nealy only production woods) since centuries.

I agree with Udy on some of his comments here. We shouldn’t be talking so much on consensus and the like rather than actually getting things done. Storage solutions for the future energy sources really require us to focus on getting battery units or whatever things implemented not theoretically argued about.

I’d like to know just what is the cheapest storage. If Highest efficiency, then less collection capacity needed. But really cheap storage might afford a little more capacity build up. 

However, what really counts is if the system can energy pay for itself. For example,  the whole world could be powered by solar and batteries but if they were lead acid, it would require more energy than it’s worth (and way to much land). I do believe that other kinds of batteries will require far less energy input.

It might be easier once we figure out aneutronic fusion by use of picowatt scale lasers… (or simply deal with molten salt nuclears for the time being).