Categories
Nuclear Renewables

Energy Storage Discussion Thread

For high-penetration utility-scale wind, we'll need much bigger batteries than these...

Debate over large-scale energy storage is a regular theme in the comments on this blog. The post is intended to be a place to centralise this discussion. Some questions that might be considered in the comment thread:

1. What is the cost (per Watt hour, kWh, MWh, GWh — how does this cost scale up, and how does this scale as higher levels of reliability are required, e.g. energy delivered on demand 90% vs 99% vs 99.9% of the time)?

2. What is the energy density of the proposed storage technology currently, and what are its physical limits? (i.e., how good can it get, with perfect engineering, and how long can the energy store be held?)

3. If the storage technology becomes cheap, what is to stop baseload plants like coal and nuclear from undercutting renewables, given that they can charge large batteries in low-demand times and then sell the power during peak (high-price) periods?

4. What are the material inputs for the storage system, and how does this effect the energy returned on energy invested of the paired energy technology (e.g., what is the EROEI and life-cycle CO2 emissions of, say, a 2kW solar PV system with no storage vs the same system with 10 hours battery storage to cover nights [ignoring winter and long cloudy periods])?

5. Lifetime: how many cycles can the storage technology handle (100, 10,000, near-indefinite [e.g. conversion to hydrogen])?

6. Does the storage technology need its own power-generation system, or can it be paired to the original generating technology (e.g., a molten salt heat storage can create steam for use in the same turbine set as the solar thermal plant itself, whereas compressed air energy storage for wind requires a different generation system to the wind itself)?

(If people can propose some other general questions, I’ll add them to this list)

Anyway, to kick the discussion off, here is something sent to me by George Stanford, in response to the following missive:

Seems to me the answer to intermittency is more wind and solar power and larger grids so they have over capacity and can share across larger areas. Conversely consider energy storage devices, flywheel systems, pressure accumulators, batteries, etc. If you consider all the economic benefits of the positive environmental savings of wind and solar (when getting away from coal) then the overcapacity costs would balance out in the end.

The end benefits of ending destructive mining practices, decreased oxides of nitrogen and mercury emissions and disposal issues with fly ash probably make it worth it alone. Has anyone found a study that shows the costs of that.

George replies:

Large-scale energy storage faces hurdles like cost, scale, material availability, and environmental disruption. For example, regarding battery storage, I have found this on the Internet:

Conway, E. (2 September 2008) “World’s biggest battery switched on in Alaska” Telegraph.co.uk, 12:01AM BST 28 Aug 2003

Excerpt: The world’s biggest battery was plugged in yesterday to provide emergency power to one of the United States’ most isolated cities.

The rechargeable battery, which at 2,000 square metres is bigger than a football pitch and weighs 1,300 tonnes, was manufactured by power components specialist ABB to provide electricity to Fairbanks, Alaska’s second-largest city, in the event of a blackout.

Stored in a warehouse near the city, where temperatures plunge to -51 degrees Centigrade in winter, the battery will provide 40 megawatts of power – enough for around 12,000 people – for up to seven minutes.

This is enough time, according to ABB, to start up diesel generators to restore power, an important safeguard since at such low temperatures, water pipes can freeze entirely in two hours. . . .

The earthquake-proof contraption contains 13,760 NiCad cells – bigger versions of those used in many portable electronic appliances including laptop computers and radios. Each cell measures 16in by 21in and weighs more than 12 stone [168 lb].

Notice that this system will store 40 MW x 7 minutes = 280 MW-min of energy. So let’s do a little ball-park arithmetic. The US uses roughly 500 GW-hr of electrical energy per hour. Suppose we want enough storage to supply 1% of that energy for 10 hours (feel free to plug in your own guesses). That’s 50 GW-hr, and 3,000,000 MW-min / 280 MW-min = 10,714.

So for our postulated backup system, multiply the Fairbanks numbers by 10,700. This gives us:

Area: 21.4 million sq. meters = 21.4 sq. km.

Weight of whole system: 1,300 tonnes x 10,700 = ~ 13.9 million tonnes

Weight of NiCad cells= 13,760 x 168 lb x 10,700 = ~ 12.4 million tons = ~12 million tonnes.

Weight of Cd: ~ 3 million tonnes (this is a guess).

World reserves of cadmium: 0.6 – 1.8 million tonnes (inadequate for just US need).

In other words, current battery technology would be unable to support battery storage for any significant part of the electric supply. Equivalent considerations apply to the other storage techniques.

Without storage or hydro backup, wind power looks hopelessly impractical as a major supplier of electrical energy. Larger grid? Apart form the elaborate new transmission network needed, remember that it’s not sufficient merely to always have power being produced somewhere — you constantly have to have enough power to supply the whole grid. Think of the over-capacity needed to accomplish that! Much of the time, most of the windmills would have to be feathered to prevent over-production.

Here-s a year’s worth of daily wind energy from all the wind farms in Germany:

(The Web site I got the above chart from seems now to be inactive.)

You can see almost-real-time info about the wind contribution to the Bonneville Power Authority here:

As for ending the destructive mining practices, etc., that can’t be done more than partially (if that) by wind and solar power — but it certainly can be done, in spades, by fast reactors such as the IFR.

———————–

Then there is this comment by Tom Blees, putting recent solar costs into context (see also my recent comparison in TCASE 15):

This is being cheered on the internet by the windies and sunnies:

The Energy Department on Wednesday approved two loan guarantees worth more than $1 billion for solar energy projects in Nevada and Arizona, two days before the expiration date of a program that has become a rallying cry for Republican critics of the Obama administration’s green energy program.

Energy Secretary Steven Chu said the department has completed a $737 million loan guarantee to Tonopah Solar Energy for a 110 megawatt solar tower on federal land near Tonopah, Nev., and a $337 million guarantee for Mesquite Solar 1 to develop a 150 megawatt solar plant near Phoenix.

The loans were approved under the same program that paid for a $528 million loan to Solyndra Inc., a California solar panel maker that went bankrupt after receiving the money and laid off 1,100 workers.

These people can’t do math. A solar power plant rated at 110MW will have a capacity factor of about 18% tops (because of when the sun shines). Less in winter, of course. So that’s effectively about 20MW for $737 million. The price per gigawatt, then, is about $36 billion! And people complain about the high cost of nuclear power?! Even at the ridiculous prices for new nuclear in the states of about $6-8 billion/GW (and mind you, that’s for 24/7 availability, unlike solar), this solar investment is patently ridiculous.

Edit: It was pointed out in the comments that with Tonopah’s projected capacity factor of 55 % (with energy storage), the capital cost comes to a bit over $13 billion per GWe average power.

By Barry Brook

Barry Brook is an ARC Laureate Fellow and Chair of Environmental Sustainability at the University of Tasmania. He researches global change, ecology and energy.

215 replies on “Energy Storage Discussion Thread”

Just posted about exactly this topic in the “CEDA report” thread. I may want to repeat some of that here.

But first, let’s just all agree that having storage is useful for nuclear energy as well as for renewable energy. I recall that the world’s first pressurized air power plant at Huntorf was built to work with a nuclear power plant, switching supply to time zones with better prices and providing backup in case of a station blackout emergency.

http://en.wikipedia.org/wiki/Compressed_air_energy_storage

That kind of storage would have been extremely useful in Fukushima.

Like

//3. If the storage technology becomes cheap, what is to stop baseload plants like coal and nuclear from undercutting renewables, given that they can charge large batteries in low-demand times and then sell the power during peak (high-price) periods?//

This is a great point! I was personally quite amazed at the Nullabor seawater dam claims that a 7km diameter 20m deep dam could store enough energy to run the whole of Australia for 10 hours. The author of the paper said it might come in at $2 billion if the base is not concreted… more if it was concreted.

Click to access Australian_Sustainable_Energy-by_the_numbers3.pdf

Let’s call it $3 billion with concrete.

That’s still cheaper than many baseload nuclear power plants, and how many of those would we need to run the *whole* of Australia for 10 hours? So I was left wondering…. (in my rather scientifically illiterate manner) just how many of these hydro dams we would have to build to have true baseload power from a mix of wind and solar thermal? Add in the cost of the HVDC grid to get the power from the wind across to the Nullarbor and back again. Even so: $3 billion dollars to store power for the whole of Australia for 10 hours?

Ultimately it comes back to point 3 above. If these things are cheaper than building more nukes to cover just a few short periods of peaking power supply each day, then surely the answer is not to try and supply baseload power from intermittent renewables like wind, but shave the costs off the peaking power demands to create even cheaper electricity.

So that maybe, just maybe, we have the money to cope with manufacturing all synfuels we need for airlines etc in a post-peak oil world.

Like

EN I looked at my saved copy of the Seligman pumped seawater proposal and the key data I got out of it was 200 Gwh storage for $35 bn. The 7km diameter octagonal tank (a big version of the one in Okinawa Japan) sits atop a cliff. It was $2n for the tank and $33bn for the pumps and turbines per Table 1. I think this is the same report
http://www.ret.gov.au/energy/…/cei/acre…/029-ProfessorPeterSeligman.pdf
The table says 200 Gwh if I read it right but I haven’t checked the calcs. Other numbers in the report are for generating capacity which we can take as a given as it could just as easily store excess fossil and nuclear baseload for use in peaking.

$35 bn for 200 Gwh is $175 per kwh, similar to lead acid batteries, still the cheapest form of portable energy storage after a century. I seem to recall someone saying that some Army land on the Illawarra escarpment near Wollongong would be ideal for a trial project. Any pilot project needs to be at least a few Gwh to be serious.

Some things to like about pumped seawater include the ready made lower reservoir, no platypus to upset with changing water levels, less drama if the whole thing bursts and minimal electrochemical deterioration unlike a battery.

Like

Just to provide some perspective on what kind of scale we’d need energy storage, this excellent article on The Oil Drum is very useful:

http://www.theoildrum.com/node/8237

The financial and natural resources required are simply prohibitive on this scale. It is interesting that the author of the article puts a careful conclusion in, but any reasonable thinking person will conclude that we simply won’t do this. The risk, then, is that we’ll build some wind and solar and then just run most everything of the remainder on fossil fuels with no end in sight.

This is a risk of fossil fuel lock-in, where it doesn’t matter how cheap wind and solar are, it matters only that they can’t deliver most of the time, and fossil fuels must jump in.

Like

Regarding seawater pumped storage cost, I’d caution people from using too optimistic analysis. There is a real seawater pumped hydro station in Japan, 30 MWe, costing 30 billion Yen, which is over 300 million USD, $10/Watt. They had to spend 10% of that just for environmental remediation – this is not a low environmental impact technology. This project stores only a small amount of energy, much more would be needed to truely run on wind and solar.

Click to access Annex_VIII_CaseStudy0101_Okinawa_SeawaterPS_Japan.pdf

Like

Click to access 2010-2011_IWES_IB-HL_19%20Power-to-Gas_Erneuerbares%20Gas%20verbindet%20Energienetze.pdf

In Germany, they are proposing to use the gas grid to store variable renewable energy in chemical form.

Excess electricity would be used to produce hydrogen, which would be combined with CO2 derived from biomass to form methane, which would be fed into the gas grid.

The round-trip efficiency of the process is 35%. That’s not good, but it is the only large-scale storage option you have in Germany. According to the authors of the study the technology is supposed to become economically competitive once oil hits 220$ per barrel.

Of course baseload nuclear power plants would produce this “renewable gas” more economically than variable renewable energy sources.

Like

Molten salts are being used in CSP plants. For example, in Andasol with 2x50MW gross output with about 42% capacity factor (http://en.wikipedia.org/wiki/Andasol_Solar_Power_Station) the storage can provide 7.5 hours of power. Looking at the numbers it seems that one needs around 75 tonnes of salt per MWhe. This implies about 28000 tons for the Andasol power plant. The salts are a 60/40 mixture of sodium and potassium nitrates.

The energy density is in the 50kWh/m^3 ballpark or roughly 20 Wh/kg. This is less the energy density of Li-ion batteries, but not tremendously less than for some other batteries. This is perhaps bit surprising. Sodium nitrate costs about 800 dollars/ton (http://www.indmin.com/Article/2124417/Sodium-nitrates-prices-stagnate.html). For potassium nitrate I cannot find a good source, but in my notes I have a figure of 1000 dollars/ton. This suggests a cost of around 30 dollars/kWh from the salts. Online I can find a total cost estimate of 50 dollars/kWh which seems consistent with my estimate.
(http://www.scientificamerican.com/article.cfm?id=how-to-use-solar-energy-at-night) Another way to state this is that the cost of salts for storage for 7.5 hours adds about 1 billion/GWe delivered to the costs of the power plant. (Naturally 7.5h is not sufficient if our intent is to get rid of fossil fuels.) The amount of storage is set by the heat capacity and volume. I do not believe any major improvements can be expected in this technology.

I can also find a mention that the sodium nitrate peak production has been 3 000 000 tons/year. (http://www.scribd.com/doc/30122669/Sodium-Nitrate-and-Nitrite) If all this production were to be used for CSP storage this rate implies a maximum construction rate for Andasol type plants of 8GWe peak/year. This is almost zero so the global nitrate production would have to be ramped up drastically. I do not know about the reserves for these minerals.

Like

Jani, thanks for those calculations.

The lower energy density isn’t surprising, as sensible heat storage (molten salts) has a lower energy density than chemical change (batteries). Heat is only vibration, whereas chemical change can access the bigger energy levels of the electrons.

There are some major improvements possible.

I should point out firstly that the energy density per liter of salt can be improved by a factor of around 3, by going for a higher temperature. This is very difficult with concentrating solar for cost and thermal loss reasons. But it is easier with a nuclear plant such as an IFR or a LFTR, plus the heat storage system can be used more productively and reliably.

Another improvement is the use of an inexpensive filler, such as sand, in a thermocline system. One of the best easy to read references is this one by James Pacheco:

Click to access 21032.pdf

As we can see, the thermocline has theoretically a much lower cost, about 1/3 lower. This assumes a delta T of 84 degrees Celcius, but with a high temperature nuclear reactor up to 300 degrees Celcius temperature drop is possible. This reduces cost of the salt by a factor of 300/84= 3.6. The cost of the tank goes up slightly (need more expensive stainless steels), but the net effect is a big factor reduction in cost, down to around $10/kWh. Even with a bigger salt cost per kg that might happen from increased demand, the cost would be low enough for half a day of storage to deal with daily peaking needs.

I was hoping to use such systems in conjunction with high temperature nuclear reactors so that there would be an integrated baseload-peaker facility with a dedicated peaker turbine sucking on heat from the thermal store. This allows the reactor to operate at full throttle continuously.

Like

“3. If the storage technology becomes cheap, what is to stop baseload plants like coal and nuclear from undercutting renewables, given that they can charge large batteries in low-demand times and then sell the power during peak (high-price) periods?”

There are basically two ways to stop coal (I am not interested in stopping nuclear).

One is what is happening right now in Germany. Renewable energy has priority regardless of price. The feed-in tariff law does not only guarantee that renewable electricity is bought at fixed rates, it also guarantees that every kWh generated displaces exactly one kWh of fossil fuel electricity (Article 8, Paragraph 1 EEG).

Of course one might want to take that a step further and make any use of fossil fuel for any purpose completely illegal, once more people understand the extraordinary dangers associated with said use.

The other is that renewable energy would be just cheaper anyway than coal. Wind is almost there, and solar photovoltaic will be there in less than a decade.

Like

Now, if I’ve calculated it correctly from the Pacheco reference, the thermocline needs 11 tonnes of salt per MWh electrical.

With the higher delta T of 300 degrees Celcius possible with a high temp nuclear reactor, the amount of salt would naturally be reduced by that factor of 3.6, to about 3 tonnes of salt per MWh.

Like

Are you excluding discussion about ‘virtual’ (?) energy storage? I believe there are already companies that are selling commitments to stop using electricity at times of peak demand. I wonder how big this capacity is.

At a smaller scale smart-grid technology can also smooth demand. What are the untapped potentials of these technologies?

Like

There is an interesting suggestion from the University of Minnesota to use the electricity from wind power to generate Ammonia instead of putting the electricity on the grid. This eliminates the problem of connecting the wind farms to the grid and problems with the intermittency of wind power. I’m mostly interested in using Ammonia for vehicle power, while they want to use Ammonia for fertilzer, but it works either way.
http://renewables.morris.umn.edu/wind/ammonia/

Like

OPatrick, on 14 November 2011 at 12:30 AM said:

Are you excluding discussion about ‘virtual’ (?) energy storage? I believe there are already companies that are selling commitments to stop using electricity at times of peak demand

In the US and Canada the Aluminum smelters are for the most part on ‘interruptible power’ agreements. It’s not intended as a ‘daily peak’ but a short term ‘seasonal’ peak. I.E. Summer heat waves.
Alcoa has similar agreements in place in Australia.
http://www.alcoa.com/australia/en/info_page/Energy_Victoria.asp

Like

So Harry raises an important point…one that works against, IMO, storage for unreliables.

One is load storage for peaking and before/after peak, a place to put excess daily capacity and use it when its needed.

The second are seasonal or long term issues confronting only unreliables like wind and solar where there could be whole periods of days and even weeks where the sun doesn’t shine or the wind blows.

The BPA situation has been parsed here before so it doesn’t need review.

To get the “24 hour” CSP output, name plate capacity itself drops down tremendously. It’s only good for that ‘day’, so to speak. The same would be true for any massive molten salt storage system, which, I might add, is not being planned anywhere in the world that I know of…because it’s so damn expensive. Still, really good for only about 16hours in even really good insulated vaults.

The German hydrogen thing. The inefficiency is just staggering with the costs involved and the inability to hold a lot of H2 over a long period of time.

Long term storage means pump storage. I can tell you that there IS a lot potential for this in the United States as none of the big hydro projects, about 10 to 16% of US generation, has any form of pump storage. This could be engineered into existing hydro units.

The bottom line, however, is as someone up in the comments section noted *every* form of storage works *better* with nuclear than unreliable wind and solar. It just does. Additionally, when we get to Gen IV deployment even in the upper single digit percentages of grid generation, with their rapid loading and unloading, building to over capacity slightly will eliminate the need for storage altogether.

Like

just to add to Cyrill’s point above…indeed, molten salt storage at high temperatures can be designed integrated into LFTR and IFR simply by adding structure for high temp salt inventory in the secondary loop either at build up or via process heat bleed off and “added modularly” when ever so long as the secondary loop has the ability to off-load hot salt with piping and pumps. I think the potential for this is huge.

Additionally, if we have a “Gen IV Economy” we won’t need any storage as we wont be required to have because there will be no over generation at all. For nuclear energy generally it’s a convenience, for unreliables, it’s a necessity as they can’t control generation, or at least not well enough to make it financially viable to do so.

Like

“The bottom line, however, is as someone up in the comments section noted *every* form of storage works *better* with nuclear than unreliable wind and solar.”
David Walters, isn’t the point more that one of the big perceived disadvantages of renewables – solar and wind – is that they need back-up? If there are other concerns about nuclear then we might overcome these if there is no viable alternative, but with effective energy storage those renewables would be more viable.

Also further to my question above, what price would need to be put on *not* using a kWh of electricity for it to become attractive as an option on a large scale? Could this be managed?

Like

Cyril R. wrote:

Just to provide some perspective on what kind of scale we’d need energy storage, this excellent article on The Oil Drum is very useful:

http://www.theoildrum.com/node/8237

Yes, this is a dumb article (to put it in scientifically technical jargon). Nobody needs (or is working towards) a 7 day battery to store the entire grid. The technology is not there for doing this, it would be hugely impractical and expensive, and there is no need for doing so (is our grid really that much of a failure)? Citing that wind may not blow for a 4 day consecutive period as a justification for such a battery (or thought experiment) shows a serious misunderstanding of the nature of wind power, distributed generation, and the design and operation of energy systems in general.

A couple of sources may help us understand a few features of the problem.

Economic cost of outages? In the US, Berkeley Labs placed this cost at around $80 billion/year (to businesses and consumers, mostly industrial and commercial sectors, $1.5 billion in residential losses). 2/3 of the impact came from momentary (or less than 5 minute) losses. “Momentary interruptions were responsible for two-thirds of the cost, at $52 billion, while sustained interruptions of five minutes or more cause $26 billion. ‘This finding underscores the fact that, for many commercial and industrial customers, it is the length of the ‘down time’ resulting from a loss of power that determines the cost of interruption, not necessarily the length of interruption itself'” (here).

DOE targets for bulk storage in the US (to improve grid reliability, support anticipated expansion of renewables, and prevent economic losses from outages)? In a Feb 2011 planning document, they suggest 10 to 100 GW of bulk storage over the next 5-10 years is a good estimate. For current EES grid targeted batteries (redox flow, Na-based, lead-carbon, Li-ion) they see the following price and performance targets: $1,750/kW capital cost, $250/kWh or less system capital cost, 0.20/kWh levelized cost, 75% system efficiency, 4,000 cycle life. They envision higher targets for new technologies based on previously discovered materials and chemistries: $1,250/kW capital cost, 5,000 cycle life, 80% system efficiency, $150/kWh system capital cost, and 0.10/kWh levelized cost (which would represent some 30% reduction on current costs for many technologies).

Conway, E. (2 September 2008) “World’s biggest battery switched on in Alaska” Telegraph.co.uk, 12:01AM BST 28 Aug 2003

A 2009 Arpa-E energy storage workshop has a pretty good summary of current grid energy storage technologies, costs, and benefits. Among those topping the list with detailed descriptions: pumped hydro, CAES, vanadium redox and zinc-bromine flow batteries, lead acid, sodium sulfur batteries, metal air batteries, flywheels, ultracapacitors, and superconducting magnets (SMES). There are also several high-risk concepts mentioned: fuel as storage media, electrochemical conversion to methanol, ericsson cycles, algae farms growing carbon electrodes, multiple platforms V2G, molten salt systems, and more.

Let’s take DOE benchmarks and apply them to economic costs of outages (Berkely Labs study):

– 100 GW bulk storage in 10 years (high end)
– capital cost (best case): $175 billion (over 10 years)
– capital cost (pumped hydro alternative, at full 100GW): $225 billion ($2250/KW, 100-1000MW range, discharge 4-10 hours).
– Benefits of storage (low cost compared to outage costs in Berkeley study): 22% (best case) or 28% (if 100% from pumped hydro at current cost).
– Approx physical dimensions 100GW NaS battery (based on current production NGK 8MW units for daily load shifting): 9,000 square miles (14,500 sq km) or 75,000 cubic km. This is about an area equivalent to Real Energy’s northern Cooper–Eromanga Basin oil and gas and tight gas leasing tenements in Queensland, if units built to 5m height.
– Approx physical dimensions 100GW pumped hydro reservoir (based on multiples of Ludington Storage Plant at 1872 MW): 70 square miles (113 sq km).

There are numerous ways to cover this additional cost of storage for utilities and consumers. DOE energy storage planning document projects EES energy costs in best case will be $0.10 – $0.20/kWh USD (levelized retail rate). At this price, energy arbitrage is an effective way to cover this cost (reducing the overall retail price of electricity at times of peak demand).

Like

Barry Brook or Tom Blees wrote?

These people can’t do math … The price per gigawatt, then, is about $36 billion! And people complain about the high cost of nuclear power?! Even at the ridiculous prices for new nuclear in the states of about $6-8 billion/GW (and mind you, that’s for 24/7 availability, unlike solar), this solar investment is patently ridiculous.

Have we started comparing apples to mexican jumping beans (and don’t even have an orange on hand to make a faulty comparison)? Solar has no fuel cost, which significant lowers it’s levelized cost of energy when compared to coal, natural gas, or nuclear. Capital costs are reported as a cost for a unit of capacity (not energy delivered). The Crescent Dunes or Tonopah solar tower project is projected to have a 52% capacity factor. It has already secured a power purchase agreement from NV Power for a reported $0.135/kWh (same source). This blog is going to become totally irrelevant if you don’t start giving “fair” representations of competing technologies to nuclear?

Like

I think it is important to understand that there are several different scales of energy storage, and they serve different purposes.

1) Very short time scale storage (circa 1 hour). This serves to smooth out rapid fluctuations in renewable generation (such as wind gusts or clouds). This poses engineering difficulties because the fluctuations are quite fast and large: physical systems like pumped storage may well find themselves putting energy into the reservoir at the same time as they are taking it out. I would expect this smoothing to be done with batteries/capacitors which might, just maybe, be doable for a large renewable contribution (see above battery materials estimates). If you have this, then renewables can be used above the perturbative level. Currently provided by the fuel tanks of peaker plants.

2) Day-length storage (18 hours). For solar, this serves to eliminate the day/night cycle, and an absolute requirement for large solar contributions. For wind, this averages over individual storms and the like. Right now, this can only be provided by pumped storage, or, being honest, fuel tanks. But the duration is short enough that pumped storage can be contemplated. Probably with a shudder. If you don’t have this, the maximum renewable contribution will be really low.

3) Week(s) length storage. This serves to smooth out weather systems. Basically only possible through fuel tanks. Here there is an IMPORTANT thing to realize: my weather right now is correlated with your weather in the past, and that other guy’s weather in the future. In other words, we can claim week(s) length storage, OR geographic smoothing. We can’t claim both at full value. Geographic smoothing is cheaper (unless you use fuel tanks, which is what we do now). If you don’t have this, the maximum renewable contribution will be merely low.

4) Month length storage. This serves to smooth out long term signals, like the factor of near 10 in Germany’s summer/winter solar output. Required for solar to be useful at any significant latitude. Only possible with fuel tanks. This, finally, allows for a high renewables contribution. Pity we can only do it through fuel tanks…

Funny, I said “fuel tanks” a lot. Non-fuel tank solutions are really only available for (1) and (2).

Like

DOE…In a Feb 2011 planning document

In a Feb 2011 Research and Development Goals Document

Click to access OE_Energy_Storage_Program_Plan_Feburary_2011v3.pdf

Under ‘Challenges’ on Page 17
Challenges—Most of current EES technologies are not competitive in capital cost and/or life cycle cost for broad market penetration.

Challenges—As a utility asset, storage technologies are required for a life at least 10 years, often longer, and a deep cycle life (e.g., more than 4,000 cycles). Minimum or no maintenance is preferable. Long term reliability and durability are not typically proven for technologies entering the market.

Challenges—The non-vehicle energy storage sector is still, generally, commercially immature.There is not enough manufacturing capability, particularly in the U.S., to deliver the projected quantity of EES systems to meet future demands.

Then on page 28 as a ‘goal’
A national assessment of the role of energy storage: This will quantify the role of energy storage technologies as the grid transitions to a greener, reliable, and highly efficient energy infrastructure. The assessment will estimate the potential market size of energy storage for various grid services as storage competes against other generation technologies (such as gas turbines), demand response, and transmission assets.

Page 33 – an improved cost modeling system
Detailed cost and “state of health modeling” of energy storage: To provide transparency of the cost composition of an entire battery system, component-based cost models will be developed for
stationary energy storage systems.

Here is a US DOE discussion of ‘fusion’
http://science.energy.gov/fes/
With the support of FES, a devoted, expert, and innovative scientific and engineering workforce has been responsible for the impressive progress in harnessing fusion energy since the earliest fusion experiments over sixty years ago. As a result, we are on the verge of a new age in fusion science during which researchers will undertake fundamental tests of fusion energy’s viability.

Like

EL, you’re not listening to what I’m saying.

First and foremost, running with less than at least a week of full power storage on a national level whilst relying on unreliables, means burning lots of fossil fuels. In a growing world economy and energy demand, even 25% fossil isn’t acceptable.

Second, the article is good in that it tries to do this. It is not a dumb article, it is extremely useful limit analysis. It does not matter if you insert pumped hydro, the conclusion is the same since pumped hydro has not much lower cost than lead acid.

Third, if your notion of a wind and solar powered world is one with less reliable electricity and longer outages, then just be honest about that.

Here in the Netherlands, a single 5 minute outage per year is considered a national scandal. Good luck eating away some nines in reliability – my reliability last year was 100%, same as the year before. Not 99.9999 percent, 100 percent. Zero outage. Not one second. I’m willing to pay for this. I’m not willing to pay a lot of money for greenwashing fossil fuels with unreliables. I have little patience for energy plans that don’t add up.

Like

Barry,

I haven’t caught up with all the comments on this thread, so apologies if someone else has already pointed this out.

A solar power plant rated at 110MW will have a capacity factor of about 18% tops (because of when the sun shines). Less in winter, of course. So that’s effectively about 20MW for $737 million. The price per gigawatt, then, is about $36 billion!

Tonopah has molten salt storage and expects to generate 485,000 MWh per year http://www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=60 . So the expected capacity factor is 50%. Therefore, average power is 55 MW. On these figures, the capital cost per average power delivered is:

$737,000,000 / 55,000,000 W = $13.3/W.

Like

Also, EL, uranium costs 0.3 cents per kWh. This is not a substantial advantage for PV. PV being a marginal energy source that is not there 80% of the time in the desert, and not there 90% of the time in Germany, gets little capacity credit. So PV has to compete with 0.3 cents per kWh nuclear, or 2-3 cents per kWh coal fuel.

This is why significant buildout of PV only occurs with massive subsidies, several times market rates for an energy source that isn’t worth 1% of that.

Like

By the way, $13.3/average Watt is twice the cost of the most expensive nuclear power plant being constructed right now (Olkiluoto).

Has anyone else noticed these massive internal inconsistencies of the pro unreliables anti nuclear folks? Ask them about nuclear they’ll say, oh the waste, its such a problem. They never look at the waste of the unreliables. They’ll make vague statements about innovation. When asked about nuclear innovation in pyroprocessing and next gen nuclear power they’ll say it can’t be done. But they’ll happily make breakthrough technologies in energy storage part of their energy plans.

Like

The fact that the whole energy system of the world (not only electricity) can be powered for decades from stored energy follows clearly from the fact that it is already done right now.

Fossil fuel is nothing else than stored energy, though admittedly stored by an extremely inefficient process taking hundreds of millions of years and provided by nature.

Since there is technology available to make many different kinds of synthetic fuel (hydrogen, ammonia as mentioned above, quicklime, methane gas, methanol, gasoline) it is just a question of how much of your carbon free capacity (nuclear or renewable, doesn’t matter for storage) you want to dedicate to making those fuels. That is a question the market will figure out.

I recall that the “Energy from Thorium” blog just discussed some such schemes in connection with nuclear energy:

http://energyfromthorium.com/2011/11/07/nuclear-cement/

If you take the CO2 necessary for making synthetic fuels right out of the air as the “Green Freedom” project described at the link proposes, there is nothing to stop you from taking twice as much CO2 as needed and dispose of half of it (by weathering peridotite or olivine), making CO2 negative fuel.

Like

The fact is that most hydrocarbon fuels had several hundred million years preparation time. As that carbon was buried the atmosphere’s oxygen content increased. Very late on the scene humans worked out how to exploit this. Now we want fuels produced in real time, not in batches over millions of years. It’s as if we were fleas infesting a dog and now we’re saying ‘this dog is sick let’s find another one’.

Like

Karl-Friedrich Lenz,

it is just a question of how much of your carbon free capacity (nuclear or renewable, doesn’t matter for storage) you want to dedicate to making those fuels.

If one isn’t careful the Energy Returned on Energy Invested ends up being less then one.

Like

More pumped hydro and other storage technologies

Click to access Doty-90377-Storage-ASME-ES10.pdf

http://www.electricitystorage.org/technology/storage_technologies/technology_comparison

In particular, the Doty Energy study suggests that even ordinary pump hydro for diurnal cycling is fairly expensive. Of course, for that the geography (and other factors) has to be appropriate. Underground pumped hydro can be built almost anywhere, but the costs are then much higher.

Unless the unreliables become nearly free, using those to energize a storage scheme is not economic.

Like

I had a chance to review several storage schemes as part of a DOE evaluation team. Most of the schemes had approximate costs of $1 per watt for the equipment and $0.40 per watt hour. Redcently I went to a luncheon talk on storage and noted the price of their equipment went up to $1.2 per watt. The battery cost was not stated. I think these numbers are good for rough estimates.

Like

@harrywr2

I don’t understand that. Obviously, any form of storage will come at a loss compared to just using electricity as it is produced. But if you want to time-shift your energy, or use it in a car where you have no way of delivering it except carrying it with you in some way, that can’t be helped.

Like

This is not about EVs but about “time-shift” energy (i like the term) for purposes of using it later either:

1. Because you *need* it later or
2. You have too much now and can’t use but could use it later.

Yes there is a loss for storage. So the question is, then, why are we storing generation? Is there any financial, resource utilization advantage to doing this? If we add storage to CSP do we not have to knock down the actual name-plate capacity by a factor of 4 or 5?

DW

Like

David Walters — The nameplate (MWe) is provided by the actual generator; the thermal source provides MWt which with some efficiency and possible delay via thermal stoarage then runs the generator.

Like

To compare the cost of unreliables + storage with nuclear, do the comparison on an LCOE basis. For unreliables add the capital cost of the unreliable generator plus the capital cost of storage.

You may need to replace the storage several times during the life of the unreliable generator, so you need to include multiples of the capital cost of the storage.

You will need to replace the unreliable generator and generator several times during the life of the nuclear plant. So you need to include multiples of the capital cost of the unreliable system.

The capacity factor for nuclear is normally quoted is the expected life time capacity factor for the system, allowing for planned and unplanned outages. However, the capacity factor for the unreliables is normally quoted as the best that could possibly occur over a short term, rather than the average over the plant life. Therefore, the unreliables’ energy output and capacity factor should be the expected life time values (to be valid, these figures should be based on actual experience from actual plant lifetimes and performance).

In calculating the LCOE of the unreliable system, you also need to add the cost per MWh of energy storage capacity.

If the proponents of unreliables would do these calculations themselves, they might start to realise why most people why have already done such calculations realise that unreliables are a massive waste of time and money. They are a distraction.

Like

I just read the link given upthread by harrywr2 on the views of heavy industry like aluminium smelting. They are a bit disingenuous saying they got brown coal fired electricity cheaply so as to use up capacity. That’s a bit like eating bad food so as not to waste it. They seem as bamboozled by carbon tax as everybody else. Note hydro + nukes = no problems for aluminium apart from some perfluorocarbon emissions.

Their views on intermittent power are revealing. Again the link is
http://www.alcoa.com/australia/en/info_page/Energy_Victoria.asp

Like

@David Walters

Storage is very much about electrical vehicles. That is where a lot of the deployment will happen and will be necessary. Regardless of your carbon free energy source, you will always need some way of storage to carry the energy with you in a car. Same with airplanes.

This might also be a good time to quote David McKay, who has shown that the amount of storage needed per person “beautifully matches” what will be available for free anyway once everybody switches to electrical cars:

http://www.inference.phy.cam.ac.uk/withouthotair/c26/page_195.shtml

The economic case for adding storage for time-shifting right now is that this is the only way to use electricity if you generate more than supply. If you don’t have hydrogen storage, as the Enertrag hybrid wind park recently opened in Germany has, you would need to just turn off your generation whenever the grid can’t handle the extra load.

Reference for Enertrag hybrid wind park:

http://k-lenz.de/1004 (Lenz Blog)

The same is true for running a nuclear based system like in France. If demand drops, you need to throttle or completely turn off your nuclear power plants if you don’t have storage. Again, storage is useful and for cars essential regardless of what flavor of carbon free generation you prefer.

People want storage right now because there is too much electricity in certain time slots and they don’t want to throw that away. Alternatively, they want time-shifting since it enables them to sell their power when prices are highest.

Like

KFL

The role of EV’s will most likely be restricted to limited use of batteries for occasional high demand days. Given the high cost of EV batteries and their cycle-limited life, consumers will need an exceptional premium from network operators to justify limiting the life span of their batteries, providing they are available at peak times. The use of EV’s will be primarily for converting baseload to peak load, rather than converting intermittent to baseload. This seems to be a widely misunderstood by renewable advocates. The IEA Smart Grid Roadmap briefly discusses these issues:

The ownership strategy of the vehicle battery will have a significant impact on whether using vehicle batteries for grid storage is realistic, as this may reduce the life/reliability of vehicle batteries for not much financial return for the vehicle owner.

Click to access smartgrids_roadmap.pdf

Like

Barry:

A solar power plant rated at 110MW will have a capacity factor of about 18% tops (because of when the sun shines).

Peter Lang:

Tonopah has molten salt storage and expects to generate 485,000 MWh per year http://www.nrel.gov/csp/solarpaces/project_detail.cfm/projectID=60 . So the expected capacity factor is 50%. Therefore, average power is 55 MW.

The capacity factor is not such a simple metric when storage is involved. For a CSP project like this you can meaningfully talk about the capacity factor of the solar field, or the capacity factor of the turbine. If you try to define the capacity factor of the integrated facility, it can only be relative to some nominal “expected” output, which may be arbitrarily chosen, and therefore not very meaningful.

Lets look at this Tonopah facility. From Peter’s link it has a 110 MW turbine which is expected to produce 485,000 MWh/y. So the turbine is operating at 50% capacity factor (55 MW continuous).

But what’s the solar field? From SolarReserve’s FAQ,

The plant will generate between 100-600 megawatts of electricity, depending on the configuration of power load that the utility or customer requires.

So it looks like the solar field has ~600 MW capacity. If Barry’s 18% capacity factor applied that would be an average 108 MW continuous equivalent.

But the expected continuous equivalent is only 55 MW. So the solar field is operating at only 9% capacity factor. Thats not much.

I assume the drop from 18% might include poor winter output, round trip efficiency from collector to store to turbine, and thermal losses from storage, particularly if more heat energy is stored than used in a single night – that excess is probably lost.

So you can argue this facility has a capacity factor of 9%, or 50%, or anywhere in between. Its a configuration tradeoff. You could give it a 100% capacity factor if you liked, just give it a really small turbine. A high capacity factor is meaningless by itself.

Like

John Morgan, yes, excellent points. Still, the original estimate of the cost over average delivered power was clearly wrong… putting aside issues of the quality/dispatchabiliy of the electricity generated from Tonopah.

Like

@Graham Palmer

McKay doesn’t address questions of how the market will play out, he only points out that 30 million electrical vehicles will provide 1200 GWh of storage.

Even if they won’t feed energy back to the grid, this is clearly a massive amount of storage. If that can be done, then 50 GWh storage as assumed to be impossible in the post above should be not much of a problem as well.

And even if you want to build your system with nuclear only (good luck with the politics of that), you would still need batteries or synthetic fuel for all the cars. There is no way around building these massive storage capacities.

Like

Actually, those capacity factors suggest a useful metric:

Generator CF / Collector CF = Overbuild required for realized output

ie for Tonopah to realize 485000 MWh/y (55 MW equivalent) from a 110 MW turbine it had to build solar capacity of 5x the turbine rating.

That would be a more meaningful way to compare different CSP facilities and possibly other stored energy systems. Smaller overbuild is better, as it reflects storage efficiency and appropriate turbine sizing.

Like

Karl, yes electric vehicle charging is a big energy sink. However it is mostly a day to day energy sink, most cars being driven daily. So it is useful for nuclear, that has excess nighttime capacity for people to charge their vehicles when they sleep. It is not useful to deal with longer term wind lulls and not useful to deal with solar (which isn’t there in the evening and nighttime when people come home and want to charge).

Electric vehicles are an important enabler for a close to 100% nuclear grid, as most mismatch is diurnal and this is taken care of with the controllable charging rate of electric vehicles.

Clearly electric vehicles are a much better match to reliable nuclear power output than unreliable wind and solar.

Like

Hi John Morgan,

The capacity factor is not such a simple metric when storage is involved. For a CSP project like this you can meaningfully talk about the capacity factor of the solar field, or the capacity factor of the turbine. If you try to define the capacity factor of the integrated facility, it can only be relative to some nominal “expected” output, which may be arbitrarily chosen, and therefore not very meaningful.

The ‘solar multiple’ is set so the solar field provides sufficent energy over the day time to charge the storage for the number of hours of generation at full power stated. I don’t recall if this is based on annual average charging hours or winter charging hours.

Let’s assume we have 8 hours when the solar plant can generate electricity at full power. If we want 8 hours of storage as well as generating in the day time we need a solar multiple of 2. If we want 16 hours storage we need a solar multiple of 3. Because the distance from colletor to power tower of power block increases as the solar multiple increases, the cost increases faster than the storage capacity. You may recall the NEEDS analysis explained this well.

Click to access RS1a%20D12.2%20Final%20report%20concentrating%20solar%20thermal%20power%20plants.pdf

I think the capacity factor refers to the number of hours per year the plant can generate at full power.

However, I’d add for the benefit of other readers, the capacity factor of an ‘unrelable’ is not equivalent to the capacity factor of a fossil fuel, hydro or nuclear plant. The capacity factor of the reliables is determined by their availability and the amount they are dispatched by the operator. The capacity factor of unreliables is generally the best you could hope for and is not a realistic indication of their average life time capacity factor.

Like

Jani Martikainen

Sodium nitrate is rarely mined anymore; it is made.

There are several industrially important processes for making sodium nitrate.

The nitrate ion invariably comes from haber-bosch in all of these processes(either as nitric acid or ammonium nitrate, depending on which process is used). Haber-bosch uses any source of hydrogen, heat and pressure to make ammonia(usually ‘stranded’ natural gas is used), which can be oxidized further to nitric oxide, which react with water to produce nitric acid.

The sodium is used as either sodium hydroxide, sodium bicarbonate or sodium carbonate depending on process. Sodium carbonate is mined or made from brine and limestone in the solvay process.

The raw materials are abundant but it takes A LOT of energy to make sodium nitrate, mainly because of the nitrate.

Like

@ EL, on 14 November 2011 at 3:22 AM.

In a couple of dozen paragraphs about energy storage, EL mentioned only power (kW, MW), not energy (kWh), until the final paragraph. No link between these two measures was made.

Discussions about time-shifting energy must account for both the energy recoverable from storage and the rate at which it is recovered, ie the power. Ditto input power and energy, but the energy figure represents the size of the battery or flywheel or dam or whatever the storage system is, and is thus the first consideration.

In terms of Barry’s questions 1 – 6, Energy is Q1. Power, the rate in kW or MW that energy can be fed back to the grid in times of need, is just as relevant (Q7?). Connections between generation and storage systems, ie transmission capital and operating costs, are worthy of becoming Q9. The results from Q8 and Q9 are needed as inputs for calculation of system (not simply generating unit) LCOE’s. As Peter Lang mentioned above, the system is what matters, not components. He cited the possible need for several sets of batteries to span the life of longer-lived components.

However, this thread is about storage, so I will return to it with a simple rule-of-thumb example.

The power of a hydro system which is adequate as a storage and reserve generation component of a coal or nuclear powered notional 20GW base loaded system might well be about the size of a couple of the largest generating units, say 2GW total (10%). If the same system was adapted to back up a series of wind farms, the maths changes hugely.

Capacity factor considerations suggest that, with wind at say 25%, there must be 80GW of wind turbines to harvest that notional 20GWx8660 hours load, averaged over a notional year to allow for seasonal effects.

When generating, up to 60GW will be sent to storage. The pumped hydro storage system needs 60GW of pumps and use all of that wind… or even more wind turbines, or some dirty GT’s or… this is the point at which the wind sales person may deflect the discussion to considerations of HVDC links circling the earth, or charging batteries on 30 million EV’s at 2kW, or some other equally unachievable non-solution.

Summarising, for a notional 20GW steady load system.

Reliable baseload requires:
Backup plant – say 20% => 24GW installed.
Hydro backup generation: say 2GW.
Hydro pumping back uphill: allow 30% cycle losses = 2 x 1.30 = 2.60GW.
Water storage uphill of say X Gigalitres.

The equivalent wind setup will require:
80GW nameplate plus pump plus 30% cycle losses = 100GW installed.
Hydro backup of 19GW generating.
Pump capacity of 80GW.
Water storage capacity >> 20X gigalitres, depending on how long the summer doldrums need to be fed using water stored during winter’s windy season. I suspect that 100X will not suffice in the real world.

The wind system will also need 80GW connection to the pumped storage facility, compared to 2.6GW for reliable baseload. Both systems require 20GW connection to their loads.

So, it becomes evident that reliability of input energy is the single most valuable factor in any dependable electricity network. I suspect that the same, more or less, applies to other energy types, eg natural gas or batteries, sun, wind, liquids and all the rest. If input energy is less likely to be there when needed, the cost of unreliability rapidly becomes huge due to the need for storage.

Like

@ Graham Palmer,
thank you for providing that link on EV’s. It makes sense. Using EV’s to backup an intermittent grid would probably kill the battery years earlier. We’re going to have enough problems just asking EV’s to reduce even *some* of our demand for oil, let alone asking them to solve a messed up intermittent grid. I’ll stop spreading such rubbish until I actually see some super-battery perform this miracle over decades.

Like

@ harrywr2, on 14 November 2011 at 4:45 AM:

Great citation from the US Dept of Energy. The table of 5-year goals on Page 2 includes target costs which could be taken as a current upper bound answer to Barry’s Questions 2 and 5 and as inputs to indicative LCOE calculations.

The text provides an easily digested overview of much of this thread’s topic.

Bravo!

Like

I’ve got some questions that I’ve not the time to answer. Sorry for that, I’m a commercial scientist with tight deadlines and demanding customers.

I’ve heard talk of compressed air storage using geological structures or tunnels to hold energy. Does this storage hold a future?

Like

Yes, I’ve also run the numbers on the cost of cycling batteries of electric vehicles for grid storage, and concluded it is too expensive for today’s li-ion and LiFePo batteries; it would be cheaper to use lower cost dedicated deep cycle lead acid (which is still prohibitively expensive for bulk electric storage, by the way). If the lithium titanate batteries deliver on cycle life, this vehicle to grid idea could work.

Personally I don’t see the need for storage; simply having the electric vehicles on scheduled demand – controlling only the charging rate without actually discharging for storage – is good enough to integrate close to 100% nuclear.

What Soylent says is very interesting. I thought most nitrate was still mined. If it isn’t then this would indeed be a very big energy sink for an energy storage application. Electrolysis of NaCl and production of nitrate component is very energy intensive. I wonder how it affects energy return on investment for a very large CSP thermal storage that would be needed to cover several cloudy days…

Like

@ Karl-Friedrich Lenz, on 13 November 2011 at 10:52 PM:

“Renewable energy has priority regardless of price. The feed-in tariff … guarantees that every kWh generated displaces exactly one kWh of fossil fuel electricity (Article 8, Paragraph 1 EEG).”

When a kWh of renewable energy is produced, it may displace a kWh of FF electricity, but that is not the end of the matter.

The carbon dioxide emissions which are thus avoided means that, under a capped ETS, another similar amount of carbon dioxide may be released elsewhere. Somebody else’s marginal additional carbon emissions will fit under the cap to fill the space vacated by the zero-carbon production. The nett effect of adding extra low- or no-carbon generation in a capped scheme is thus precisely zero. What happens is that there is one more kWh of power available.

Jevon’s Paradox ensures that this additional energy will find a user.

The same is true when Australians opt to purchase Green Power. The green customers pay more for the privilege of helping their retailer to offer a cheaper mix of more carbon intensive electricity to its other customers, whilst still meeting the legislated green energy targets.

The green power users are paying money to support the carbon-emitters, at least until the cap is further reduced, thus forcing the retailer to purchase more green power for sale to the non-green customers. The green power customers have wasted their money.

Like

@John Bennetts: It is kind of funny that your argument has been used both to oppose wind power as well as by the anti-nuclear crowd to oppose nuclear. The problem is there and is related to the stupid design of the ETS. I guess that the cap should automatically be lowered when ever new carbon free power source comes online, but that is not the case today. The caps are set by politicians who naturally feel also other pressures than those caused by the climate change. Also, only some GHG sources are under the ETS in the European Union. I think that for example district heating is not in the ETS system. So if you have a CHP plant and you need heat, CO2 is still emitted from the plant no matter what renewable sources have a priority access to the grid.

Like

@John Bennets

I have argued against this understanding of ETS in another thread. With exactly the same reasoning you can say that shutting down nuclear in Germany doesn’t matter for the climate since carbon emission levels are fixed anyway. And the government did exactly that.

This does not factor in that the amount of permits in the system are not fixed forever, but may and should be adjusted by legislation, looking at how much carbon emissions are actually needed. I just called for doing exactly that a couple of days ago on my blog, since prices for carbon permits have fallen under 10 Euro per ton, indicating that there are way too many of them around.

I don’t buy the idea that a fixed emissions level makes any efforts at carbon emission reductions unnecessary. If it did, that would be rather counterproductive.

Like

John Bennetts, on 14 November 2011 at 10:21 PM said:

The table of 5-year goals on Page 2 includes target costs which could be taken as a current upper bound

Click to access OE_Energy_Storage_Program_Plan_Feburary_2011v3.pdf

I think we have to take them as ‘goals’ that are not yet realized.
For example one of the goals for 2012 was to complete evaluation of the Beacon Flywheel Storage System.
Beacon went bankrupt 2 weeks ago. In the US Bankruptcy doesn’t necessarily mean ‘the end’.

Like

Cyril R.

“I’ve also run the numbers on the cost of cycling batteries of electric vehicles for grid storage, and concluded it is too expensive for today’s li-ion and LiFePo batteries.”

V2G is not likely to be built around today’s high cost, low energy density, and high material demand battery technologies. The wishful thinking about a 300-500 mile range EV battery is not going to come from conventional Li-ion chemistries, but from metal air batteries. These have much higher specific energies, are inexpensive to produce, raw materials are more abundant than lithium, have low toxicity (for easy recyclability), and are now feasible with technological fixes for cycle life (which was previously limiting factor for this very old and established technology). Car manufactures now are racing to patent their game changing metal air or lithium air battery designs, and technology companies like Google, IBM, and GE want in on the game too. “Eos’ proprietary rechargeable zinc-air battery will initially be sold for $1000/kW and $160/kWh, and is electrically rechargeable with a life of over 10,000 cycles (30 years) with a full duty cycle and at full depth of discharge.” Musk at Tesla believes ultracapacitors will likely win out in the end.

Like

harrywr2 wrote:

I think we have to take them as ‘goals’ that are not yet realized.
For example one of the goals for 2012 was to complete evaluation of the Beacon Flywheel Storage System.
Beacon went bankrupt 2 weeks ago. In the US Bankruptcy doesn’t necessarily mean ‘the end’.

The business model for flywheel storage rests heavily on electricity market regulation, and contrasting costs between “fast” frequency regulation and “slow” fossil-fuel generation. New FERC rules in late October were targeted to address these issues, but came too late for Beacon. “The ruling could have doubled Beacon revenues but time ran out” (here).

One could say much the same for Solyndra. They were undercut by low cost solar panels from China (with some accusing China of unfair trade practices and dumping solar panels on US markets for a loss). Not sure either of these cases has anything to do technological goals “not yet realized,” but more to do with the highly competitive nature and rapid fire business markets for these emerging technologies. And flywheel storage is expensive when compared to more cost effective battery storage, a market that is likely to see $122 billion in investment over the next 10 years (according to
one research report).

Like

@ Paul Whyte:

I’ve heard talk of compressed air storage using geological structures or tunnels to hold energy. Does this storage hold a future?

I’ve seen this proposed before with regard to storing wind-generated energy. I understand it would have some issues, one of which is the massive temperature drop inevitable with depressurisation as the energy is extracted. This would cause ice to form in the exhaust stream which poses threat of damage to the turbine. The proposed solution to this problem was reheating of the air stream with natural gas combustion, which seems to defeat the purpose of low CO2 generation.

Like

CAES has been around a long time – in Germany. The idea is to use low incremental cost base load generation at night to charge up the compressed air. Then during the peak of the day the compressed air replaces the compressor stage on a CT. The CT is able to put out a lot more electrical power because it does not have the work of the turbine compressor stage. This is the concept. However when implemented in Germany, the operators wanted to hold the generation in reserve because of its quick start and black start capabilities rather than use it ot move low cost energy from off peak to the peak period so the original concept was not realized in practice. I guess if Germany does not have nuclear plants and coal is being retired, the CAES can now be charged with wind energy. It could have a nice leveling effect for wind. All the CAES I know of use salt domes and very high pressure air. This is necessary to keep the salt dome from slowly shrinking. Houston Light and Power (now Reliant) once looked at CAES but decided against it because the salt domes were more valuable as natural gas storage than as compressed air. So there are many variables to consider when deciding to use CAES or not. The good news is that its a mature technology.

Like

EL

“Eos’ proprietary rechargeable zinc-air battery will initially be sold for $1000/kW and $160/kWh

There seems to be some legal discussion about the ‘propriatery’ silicon technology company that shares street address and management as EOS Energy Storage, formerly known as Grid Storage.
http://guntherportfolio.com/2011/03/solar-grade-litigation-dow-corning-vs-rsi-silicon/

I find it difficult to believe that in an industry where 100’s of millions of dollars are being spent on R&D that a single individual working for a privately held firm in Easton, PA could come up with breakthroughs in both Silicon and Battery technology in the space of less then 4 years.

Like

Note about ammonia syntheis for energy storage:
Ammonia is already moved by tanks & pipelines so it might end up generated by stranded hydroelectric & geothermal plants & moved elsewhere in the world rather than being used to even out the swings of wind & solar production.

This would especially be the case if this:

Click to access SSAS_Oct2007_Final.pdf

becomes as cheap as is claimed in the pdf.

Like

@Jani Martikainen

You can get a lower bound by just looking at the haber-bosch process. The molar weight of ammonia is 17 g/mol; and the molar weight of sodium nitrate is 88 g/mol; so 17 grams of ammonia can be used for 88 grams of sodium nitrate.

Making a metric tonne of ammonia with haber-bosch from natural gas requires ~10 MWh of mostly natural gas. Making a metric tonne of ammonia from electricity with electrolysis and haber-bosch requires ~12 MWh of electricity.

At an absolute minimum you need 1.9 MWh of natural gas or 2.3 MWh of electricity per tonne of sodium nitrate. That’s ignoring transportation, purification and the sodium side of things altoghether.

The generation cost of the gemasolar plant(a solar thermal plant with molten salt storage) is estimated at 27 eurocents per kWh by its owners; If solar thermal had to “eat its own cooking” the absolute minimum cost for just the energy(not to mention capital, labour, transporation…) for sodium nitrate would be €600 per tonne.

Like

o1finrod wrote:

The proposed solution to this problem was reheating of the air stream with natural gas combustion, which seems to defeat the purpose of low CO2 generation.

Adiabatic CAES (or “advanced” CAES) removes the need for a natural gas burner or heated metal mass, and is expected to have a roundtrip efficiency of around 70%. The following peer-reviewed study envisions artificial systems (where natural reservoirs are lacking) in 1-10 MW range, and from components “widely available on an industrial basis” (p. 11). Currently, the costs of such a system are prohibitive when compared to natural storage plants (such as the 200 MW Adiabatic CAES plant in Staßfurt Germany, which will utilize salt caverns).

Like

The aqueous sodium battery from Aquion looks promising. They’re aiming at a lead-acid price-point per Wh, but with 10,000 cycles or more (currently tested to 5000 cycles). So the long-term capital costs would be 20x less than lead-acid. Materials are cheap and plentiful: water, sodium sulfate, manganese oxide, and carbon.

Like

Urban and rural biowaste has the largest capacity for contributing to renewable and and even nuclear energy production– if its converted into methanol.

Converting biowaste into methanol would allow practically any community on Earth to use it for base load or peak load energy production in addition to the added versatility of converting methanol into gasoline through the MTG process for automobile transportation.

Adding nuclear hydrogen to the mix could potentially increase biowaste methanol production five fold. If this were done in America, the US could completely eliminate the need for foreign and domestic petroleum.

Like

@EL

The German pressurized air plans are described in detail in this blog post at the “Smart Grids” blog:

http://k-lenz.de/1005

They say that RWE has succeeded in raising the efficiency of the process from 50 to 70% with the simple idea of storing the heat that gets released when pressuring air. There are plans for 20 compressed air storage plants, with the first 360 MW project to start construction in 2013.

The nice thing about this approach is that the sites suited for pressurized air storage are located in the north of Germany, near to the wind parks. No need to transport the electricity to the mountains in the south for water pump storage.

Like

Karl, your reference project is for a 4 full load hour storage system (360MWh/90MW).

Looking at Germany’s wind turbine output, they need more like, at least 4 days.

http://uvdiv.blogspot.com/2010/03/uptime-downtime_07.html

Looking at Germany’s PV output, 4 months wouldn’t be enough (Germany needs more energy in winter when its cold dark and not sunny).

http://energyfromthorium.com/forum/viewtopic.php?f=39&t=2689

What’s the plan for the longer term backup/storage?

Let me guess: burn fossil fuels.

Karl, you go ahead and be politically correct and physically incorrect. I’ll take my chances with turning over public opinion and politics. If we go against physics and nature, we’ll lose for sure, whereas with politics we have a fighting chance.

Like

John Bennetts, on 14 November 2011 at 8:53 PM said:
You seem to be misunderstanding how pumped hydro works:

Hydro pumping back uphill: allow 30% cycle losses = 2 x 1.30 = 2.60GW.
The losses mean more GWh storage but not larger pumping capacity. Remember that only a portion of total consumption needs to be stored.

The equivalent wind setup will require:
80GW nameplate plus pump plus 30% cycle losses = 100GW installed.
Cycle losses only apply to the fraction of yearly GWh consumption that needs to be stored. Wind turbines connected to the AEMCO grid in SE Australia are generating slightly less than the average capacity factor(0.33) to slightly more than this most of the time. Thus most of the time only a small portion of the 20GW demand will be stored or released via hydro. Pumping losses only apply to this portion.
<Hydro backup of 19GW generating.
this seems correct, but since demand is not a steady 20GW would need additional hydro capacity for peak demand(30-35GW)
Pump capacity of 80GW
In Australia’s case would need 60GW wind capacity plus enough to cover pumping losses plus losses due to wind spill during high wind output(say above 70% capacity). An additional 10% generating capacity should cover these two losses, or 66GW for Australia ( or 88GW if using the 0.25% capacity factor).
Returning to what pumping capacity would be required, if demand is 20GW and >70% is spilled would need (66×0.7=46GW minus 20GW=26GW pumping). If peak demand is 35GW would need 33.5GW hydro generation so could probably have 27GW pumping available. With 0.25 CF would need 40GW pumping, so it would be better to install wind turbines with a higher CF.
The actual GWh needed to be stored would be much higher than that needed for meeting daily peaks using coal or nuclear. With large hydro storage dams this is not an additional cost, the cost is for the pumps and tunnels( ie GW capacity not GWh).

Like

Neil Howes,
(Personal comment deleted)
Pumped-hydro cannot be used economically for back up for wind energy at the scale that would be needed. If divert the little pumped storage potential Australia has for unreliables, it would not available for where it could be used for much higher value. If it was diverted to trying to make unreliables a little less unreliable, that woud be a serious misuse of Australian resources.

Like

@Cyril R.

They are talking about 360 MWh at 20 sites, which adds up to 7200 MWh. Of course that is probably not enough, but it is also only one of many options.

Your question about “my plans” for long term storage gives me a good opportunity to mention the quicklime cycle, as proposed in the recent paper by Benjamin Müller and colleagues at the University of Erlangen quoted here:

http://k-lenz.de/1006 Lenz Blog September 2

They only propose that as a method of transporting energy from the Middle East and North Africa (MENA) Desertec project to Germany, but there is nothing to keep people from building a stockpile of the stuff.

Also, since there needs to be much more electricity generation than demand (to make synthetic fuels for cars, trucks, airplanes and industrial processes, as well as to make and transport water in a hotter world), it follows that one would just burn carbon neutral synthetic fuel on the occasions it is needed. On a small scale that is exactly what Enertrag has started doing right now.

I would also like to recall that I am strongly in favor of keeping existing and aggressively adding new nuclear capacity, so in my energy scenario I have nuclear power stations running all the time at full capacity and mostly making fuel or water. They could be used anytime to fill in. While I think it is certainly possible to go renewable only (Iceland shows that it is already done), I don’t see any merit in excluding any non carbon source, especially a heavy hitter like nuclear.

My plan would certainly not include any use of fossil fuel.

Like

Cyril R. wrote:

Looking at Germany’s wind turbine output, they need more like, at least 4 days.

http://uvdiv.blogspot.com/2010/03/uptime-downtime_07.html

Jani Martikainen wrote:

Tom Murphy has again a wonderful posting in his “Do the Math” blog. This time on pumped storage. http://physics.ucsd.edu/do-the-math/2011/11/pump-up-the-storage/ [proposing 7 day storage for the entire grid]

We’re going to have to try and get a better handle on storage needs for modern energy systems now and in the future with greater degrees of renewable energy penetration. The two sources mentioned above are entirely off base, and have no credible basis in any research done to date (in peer reviewed literature, by technical advisory groups, governments, private utilities, independent operators, ancillary service providers, and the like). Let us try and not side track the discussion into power systems operation and the merits (pros and cons) of variable generation, but just stick to what has been documented, and where the needs stand today and in the forseeable future. Kray has a very useful comment above on scales of energy storage (and their application): 1 hour or less (frequency regulation), 18 hour (load following and backup for variable generation), week and month long (seasonal demand basis).

It seems obvious to me that seasonal load variability is best handled with adequate resource planning and not with storage (especially since we have no storage alternatives to handle large fluctuations in seasonal load). So this leaves us with the daily, hourly, and minute by minute side of the scale. I’ve provided sources elsewhere showing that 10h storage (here) is sufficient for systems with high renewables (to 70%) to assure sufficient levels of reliability and meet LOLE targets for loss of load of 8 hours in 10 years (a standard measure used in power system operation and planning). Two more sources may help to frame the issue: the economic benefits of storage (especially for intervals less than 5 minutes), and national resource estimates from experts in energy planning and power systems design (who recommend 10 – 100 GW of storage for the US over the next 10 years). This appears to be the scope of the challenge that we face in the forseeable future. IT would be great if we could focus on the best way to meet this challenge in this thread (economic impacts, alternatives, technological limits, scalability, readiness of new technologies in the lab, all the issues identified by Barry in the lead article), rather than dithering over faulty assumptions, exaggerated claims, very basic misunderstandings of renewables and power system operations, and sources that nobody takes seriously. So unless anybody else has credible (and fully researched) sources to the contrary, here is how I understand the major dimensions of the problem:

Demand load: there are many seasonal load variables that are important to consider for specific locations in planning for a robust grid, adequate capacity, and energy storage. We are mainly concerned here with daily, hourly, and smaller time periods.
Generation mix: modern energy system utilize baseload, intermediate, and peak resources. How might energy storage change the picture for this resource mix, and any other tools we may use to better match electricity supply and demand.
Time scale: economics of anything beyond 10h storage are not attractive or helpful to making good use of available resources, so we’re pretty much stuck with this time scale. Anything in the 4 day to 7 day range for storing the entire grid is a red herring.
Capacity range: for a system as large as the US, it appears we are looking at 10 – 100 GW of storage capacity added to the grid over the next 10 years.

There is nothing in this picture that looks particularly challenging to me (from a technological point of view), and the solutions available to us will only improve over time. Above, I looked at commercially available NaS batteries and pumped hydro to meet this need. For the NaS battery option currently being sold by NKG for daily load shifting, it looks like an investment of $175 billion over 10 years will get us to our goal (and yield a potential economic benefit saving $625 billion in industrial, commercial, and residential losses from outages). For the pumped hydro option, using the current storage plant in Ludington as an example, it looks like $225 billion will get us to our goal.

On the experimental side of the NaS option (which is a very good battery for load shifting), let’s see if we can improve on the cost and environmental footprint of the conventional NaS battery developed by NKG, which is a modular design based on hundreds and thousands of individual cells. This is not a particularly good use of space, and also has very high material and O&M costs. Prof. Sadoway at MIT has a different approach, and is getting a lot of attention for his single cell reversible ambipolar electrolysis battery, or liquid NaS battery. LMBC is working on a commercial prototype (with funding from DOE, Bill Gates, and Total). Instead of an aggregate of thousands of individual cells (assembled on a modular basis), it’s a “scalable” single cell design with a self-assembling molten anode and cathode metal, and molten sodium sulfide electrolyte (with no separator that degrades over time). While keeping much of the design and materials a secret, the initial prototype will most likely be constructed from molten magnesium (Mg), sodium sulfide (Na2S), molten antimony (Sb-121), some zinc telluride (ZnTe), and a refractory lining (materials that are commonly available, recyclable, and relatively non-toxic). He suggests the size of the battery will be 50 to 100 times smaller than the conventional NHK approach, and have an energy storage cost as low as $50/kWh (with full development). For a detailed description of this battery (it’s design basis and background), I recommend the first half of a 48 minute video on the MIT World website. More here from MIT Tech Review. Where does this stack up with respect to environmental impacts:
60,000 sq meters for 13 GW of load shifting capacity, or 462,000 sq meters for a full 100 GW. No info available on total capital costs, but target storage costs are $50/kWh.

Some here are merciless in focusing on the glass half empty critique, and only seek to look backwards at established technologies. I prefer to see a world full of opportunities (and exciting new development challenges).

Like

Karl, Iceland has little wind and solar. It has geo and hydro – both useful zero carbon energy sources, like nuclear, unlike wind and solar. Iceland is a small country with lots of geothermal and hydro potential nad little electric demand (compared to say the US or continental Europe). Iceland is wonderful, and non-replicably. We need solutions for the world, for 7 billion people, and 10 billion in the future.

What is the efficiency of the quicklime cycle?

Like

Some commenters are neglecting to comply with the BNC Citation Policy, reproduced below for your information:

Citing literature and other sources: appropriate and interesting citations and links within comments are welcomed, but please DO NOT cite material that you have not yourself read, digested and understood. As a general rule, please introduce any and every link or reference with a short description of the material, your judgement on its quality, and the specific reason you are including it (i.e. how it is relevant to the discussion).

Like

@Cyril R.

The efficiency of the quicklime cycle is discussed in detail in the paper I referenced earlier. The authors say that one advantage is that the “Carnot efficiency” will go up with their proposal since it works at 1000 degrees Celsius at the desert site, compared to the around 400 achievable when working with molten salt in concentrated solar power plants.

The proposal also gains efficiency because waste heat can be used in Germany, while it will be just wasted in the desert.

They say their proposal will deliver 11.9 percent of the solar irradiation energy to Germany, as compared to 10.8 percent when generating electricity and delivering it over a 3000 kilometer power line.

For storage purpose the energy density is probably the most interesting value. They give about one third of coal as an estimate.

Like

@EL

I looked at the paper you cited, and I’m pretty sure that they aren’t saying that 10 hours of storage will net you only 8 hours of brown/blackouts per year. Which would be a bold, bold statement given the plentiful time series with multi-day null-generating wind floating around. That being said, I’m not entirely sure what they are claiming. I think that they expect only 8 hours per year where storage (at 30 euros/carbon ton) could find buyers over other generation mechanisms, but doesn’t have any reserves. In short, they are assuming adequate generation and trying to figure out how much thermal generation gets displaced by storage, rather than worrying about whether the generation is adequate in the first place.

I could be wrong… but it isn’t a detailed study of the statistics of wind intermittency (non-Gaussian!), which is what most of us here are worried about.

Like

EL wrote:

The two sources mentioned above are entirely off base, and have no credible basis in any research done to date (in peer reviewed literature, by technical advisory groups, governments, private utilities, independent operators, ancillary service providers, and the like).

Wow, you’re claiming that statements such as gravitational accelleration is about 10 m/s “have no basis in science”? And that the numbers by Uvdiv are wrong without providing a detailed authoritative source yourself?

I believe I’m quite done talking to you, EL. You’re not taking any reference seriously and trash them with bold statements without providing credible references yourself. The ones you cite don’t support your optimism.

You’re one of those “impress with many references and HTML” kind of people, without having read your own or others references.

I’m quite sure this is against site policy, EL.
MODERATOR
Thank you Cyril. It is indeed against site policy but hard to pin down without meticulously checking all references which is an impossible task. Please notify me of individual instances which you may come across and I will take action.

Like

@ Karl. No one has ever made CSP with 1000 degrees Celcius. Even 500 degrees Celcius is proving too much to handle. Making receivers at higher temperatures is not only challenging, it is also extremely lossy, Boltzman-Stefan’s radiative heat losses rising to the fourth power of the absolute temperature difference…

No one has ever made even a power cycle that operates at 1000 degrees Celcius. Currently the best we’ve got is 600 degrees Celcius (supercritical steam) and we have good prospects for 700 degrees Celcius in one or two decades (superduper steam and Helium Brayton cycles). We have open power cycles (gas turbines) that operate at or above this temperature but this is very different technology (combustion) than non-combustion technology.

Even with those developments the best hoped for efficiency is 50%. This is a very poor cycle efficiency; you need twice as much solar to make up for the losses. If you get that both ways you’re down to 25% cycle efficiency, and you need 4x as much solar.

This is clearly not affordable or practical.

Like

Kray wrote:

That being said, I’m not entirely sure what they are claiming. I think that they expect only 8 hours per year where storage (at 30 euros/carbon ton) could find buyers over other generation mechanisms, but doesn’t have any reserves. In short, they are assuming adequate generation and trying to figure out how much thermal generation gets displaced by storage, rather than worrying about whether the generation is adequate in the first place.

You are correct, they run the scale of variable generation up to 68%, and find that storage costs have the potential to be economical (i.e., competitive with CCGT) with a € 30/ton carbon cost in the mix and a reliability performance measure matching LOLE targets of 8h in 10 years (a standard measure used in power system operation and planning). What this means is that storage can replace fossil fuel reserve capacity (on an economic basis) at current costs for pumped hydro (+ carbon costs) and up to 68% variable generation (and still be able to deliver electricity at a performance standard consistent with what we are used to today for their modeled system). They cite Doherty and O’Malley (2005), Ensslin et. al. (2009), and Hasche et. al. (2011) for their wind modeling reference points.

I realize that this is an emerging area of study, and we can anticipate new findings as storage alternatives get more attractive from an economic vantage point, regulatory hurdles are removed, and new technologies start moving out of the prototype phase and into deployment and commercialization. If you have any detailed and substantive research to provide on this front (how much storage is needed per share of variable generation), I’m all ears and would enjoy looking at it. I haven’t done a thorough review of the literature, but what I have seen conforms pretty well with my current understanding. Time scale is important (particularly with respect to economics), but even more important is share of installed capacity for storage, and amount of variable generation in resource mix. Since we have few storage alternatives longer than 10h, few studies look beyond this time scale, and yet many are successful in modeling for energy systems with high degrees of renewables (here, here, and I could go on … which would likely rise the ire of moderators on BNC).

From a technical standpoint, none of this is too complicated (and can be done with current technology). From an economic standpoint, the bar is set pretty low trying to match the price and performance target set by fossil fuels. I’ve indicated where I think some of the benchmarks are located, and some of the price and environmental considerations for getting there. I do admit, matching the price performance targets of “fuel tanks” (as you put it above) are not going to be easy. I think the picture changes quite dramatically with a battery that can be run on the scale of an aluminum smelter (at some 5 GWe capacity per city block), a projected $50/kWh storage cost, and with materials as common and recyclable as dirt (as Prof. Sadoway likes to put it). Until that day, we’re stuck working with commercially available NaS, CAES, PSH, flow batteries, lead-acid, lithium-ion, flywheels, double layer capacitors, and numerous other technologies at bench and pilot scales (and a great deal of heavy lifting from the grid, inefficient fossil fuel capacity reserves, reliable baseload, and demand response).

Like

Cyril R. wrote:

Wow, you’re claiming that statements such as gravitational accelleration is about 10 m/s “have no basis in science”? And that the numbers by Uvdiv are wrong without providing a detailed authoritative source yourself?

Sure enough, energy storage, a mix of generation technologies (renewable and otherwise), and intelligent load management are integral to Germany’s future renewable energy targets as proposed for 2050. I don’t see where the the blogger “uvdiv” has accounted for operation of capacity reserves (renewable or otherwise), correlation with demand curve, peak shaving potential with storage, curtailment of wind (when storage capacity is not available), forecast errors, or anything else used in modern power systems operation and management to come up with his result. While it may be a good visual illustration of the variability of variable energy resources, intended to “wow” the casual reader with the appearance of lines squiggling all over the place, I don’t consider his or her “aggregate” representation of the data from “four major utilities” to have any serious thought behind it. It remains to be seen how this energy is being used, to what extent “flexible” generation is an asset or a deficit, under what conditions, and how this picture changes when energy storage is thrown into the mix.

If you’re still looking for my own credible references to the contrary (drawn almost exclusively from the peer reviewed literature), you’re not reading my posts. I’ve made one effort to look at and quantify the problem (in terms of infrastructure need for storage, rising penetrations of renewables in the US, and available technologies to meet this need). If you think it is technically impossible to operate an electricity grid with high renewables (40 – 60%), stable baseload, a modest degree of inefficient fossil fuel peak capacity reserves, and 10h low carbon energy storage at your disposal (at some share of installed capacity), please point me to the research that shows me this can’t be done (because I am pretty sure it doesn’t exist). I believe such a system would yield significant global carbon savings at a reasonable price, and be scalable nearly everywhere in the world (regardless of political conditions, local resource availability, human resource expertise, or waste storage capacity). And it may even help with rising rates of over-consumption and non-renewable resource utilization (driven by low cost carbon-emitting fuel sources). I’m willing to concede storage needs in the US (to take one example where we have a planning document) may be greater than 100 GW over 10 years, and longer than 10h in some locations (where the seasonal variations are the greatest). But to me, this sounds like a pretty fair place to start. And I also concede a better battery (something along the lines of an aluminum smelter in reverse) will pretty much guarantee a shift of development activity along these lines. Since I know you are a strong advocate for nuclear, you’ll also note I don’t leave reliable baseload generation out of the picture.

Like

@Cyril R.

The fact that there is no concentrated solar power electricity generation over 400 degrees is actually the point here. A solar kiln to make quicklime operates at 1000 degrees, which beats the efficiency of making electricity. If people were able to build concentrated solar power electricity plants working at 1000 degrees, that relative advantage of the quicklime system would disappear.

Like

The German government “experience report” of May this year has a chapter on storage which makes a couple of relevant points.

http://k-lenz.de/1009 (Bundesumweltministerium)

First off, they note that storage is not only necessary to assure stability of supply when there is less available. It is also necessary to avoid throttling generation when there is not enough demand to meet supply, a situation that will occur much more often with more renewable capacity.

I comment that this is also true for France with their nuclear fleet. They should use storage instead of load balancing. Run the plants at full capacity all the time and use excess energy to make some fuel or other.

The government report’s estimate for necessary storage in 2050 is between 10 and 30 TWh. That is not possible with pumped hydro in Germany (they say it might be possible in Norway). Therefore, Germany will use gas (wind to gas).

For that reason, that is the main topic of discussion of that report. Some of the recommendations have been enacted in the law reforming the feed-in tariff adopted on June 30.

Once you have wind gas, probably the most efficient way to use it is in 19 kW small power plants built in the basement of buildings, as Volkswagen and Lichtblick are doing right now. They get 90% efficiency by cogeneration and startup times of less than a minute. I have blogged about that yesterday:

http://k-lenz.de/1008

Like

@EL,

If you think it is technically impossible to operate an electricity grid with high renewables (40 – 60%), stable baseload, a modest degree of inefficient fossil fuel peak capacity reserves, and 10h low carbon energy storage at your disposal (at some share of installed capacity), please point me to the research that shows me this can’t be done

When I was in university 30+ years ago we designed a 100 MPG car. There is no research that shows a 100 MPG is technically impossible. All you need is a ceramic internal combustion engine and some carbon fiber composites.

In 1985 Isuzu announced it would have a production ceramic engine by 1990
http://www.chron.com/CDA/archives/archive.mpl/1985_46146/ceramic-engine-for-isuzu-car.html
Here is the 1988 status of the Isuzu ceramic engine
http://papers.sae.org/880011/
And the year 2000 status

Click to access 06_07.pdf

Here is a youtube video an actual working prototype

There is no research that shows a ceramic engine is impossible.
A working prototype was been demonstrated 20+ years ago.
There also isn’t a production ceramic engine despite decades of ‘promising’ R&D.
There is a big difference between what you can manufacture under laboratory conditions and what can be manufactured under real world ‘mass production’ conditions.

Like

Here is an interesting idea by a German, Eduard Heindl.

http://www.authorstream.com/Presentation/heindl-958931-hydraulic-energy-store-system/

It is a huge rock cylinder, much like pumped hydro under pressure or pressurized air storage with water as working medium.
He claims 1/100th the space, 1/4th of water usage, 1/10th the cost compared to pumped hydro.
It`s circulated widely at the moment in Germany. Some of my engineers have brought it to my attention some weeks ago. It looks like he is forming a team and a group of companys around the idea. The physics seems to be very sound. The pumps/generator might pose a problem. I will ask Heindl about that.

Like

Heindl has his own (German) blog about the topic onhand.
He discusses cavern, pumped hydro, redux flow batteries and H2 energy storage options. His conclusion so far is that only hydro and pressurized air storage can provide the needed storage capacity today. He is citing a lot of German studies on the topic.

Like

Jan Mrosik — Others are working on undergroudn pumped hydro (UPH) schemes, for example
http://gigaom.com/cleantech/a-new-energy-storage-option-gravity-power/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+earth2tech+(GigaOM%3A+Cleantech)
which is all very well until one starts putiing some prices on such schemes:

Click to access Doty-90377-Storage-ASME-ES10.pdf

Basically, at current price ranges, UPS can make its way in the diurnal power market provided there is an assured supply of low cost energy overnight and an assured market for high priced energy during the daytime. This will also be true for the cylinder schemes as well, but possibly the differential between the nighttime and daytime prices might not have to be as great.

Such schemes are most unlikely to be able to offer vast reserves of available energy unless, by some luck of geography, both the upper and lower reservoirs are exceptional.

Like

Sorry I am a bit late into this thread but I thought you might like to see a table that will be in my new book (The Power Makers’ Challenge) which will be out in June next year. [Yes, I admit it is a flagrant plug for my new book but that’s what us poor authors have to do these days.]

This table has been drawn from many sources and shows properties of varies energy media including natural resources and man made storage systems. You might find the information useful.

I will try to add it as an image but if that fails you can download it from my website.

Click to access storage%20properties.pdf

Like

Martin Nicholson — Thanks for sharing. By the way pumped hydro only rarely can acheive 90% efficiency. More typical is 80%.

Like

@Jan Mrosik

Thank you for that interesting link. I just checked out some of the material. He claims to be able to deliver 1.7 TWh of storage with a VERY BIG ROCK 500 meters in diameter and 1000 meters deep, which would be a very conveniently small sized installation with quite the punching power.

While we’re at sharing links to new concepts, Bill Gates just posted at his website a patented scheme to store energy by moving objects uphill:

http://www.thegatesnotes.com/Topics/Energy/Taking-Energy-Storage-to-a-Higher-Level

That description is rather vague. I could not understand how they plan to reclaim the energy or what kind of setup they have in mind.

I also rather strongly disagree with the idea of someone patenting gravity.

Like

Martin Nicholson, Thank you for that information. I’d like to know how to post tables lihe you have done here.

I would like to see average cost data added to your list. The thing I would most like to know is the $/kW and $/kWh of storage capacity.

I wonder why you say the storage time PHES is 100 h? Is that definded by economic considerations that if you don’t use it sufficiently it can’t pay the capital cost of the plant? If not, why is there a limit of 100 h for storage of water?

Like

@Jan Mrosik and Karl-Friedrich Lenz

The “solid rock cylinder” jacking system for energy storage has been discussed previously on BNC. The people proposing it have clearly never discussed their proposal with anyone who knows anything about rock engineering.

Some questions you might to ask are:

1. How would excavate it?

2. How would you hold the cylinder together in one sold block? How will you stop it collapsing into a pile of rubble?

3. Are you aware the not only will the rock in the cylinder and in the walls of the excavation collapse into a pile of rubble, but if it could be kept intact somehow, the gap will close due to the stress relief – unless you are planning on cutting a wide slot.

4. If you are intending to excavate a wise slot, how will you do it/

5. Have you had a mining engineer calculate the cost of the excavation and of whatever means you are going to try to use to contain it as a solid rock cylinder?

Like

@Peter Lang

Would you have a reference at hand for the previous discussion? Thanks in advance.

Just to be clear, I have no idea if this idea (or that Bill Gates just posted) has any merit. The increases in storage per area of two orders of magnitude are interesting at first glance.

Like

Karl-Friedrich Lenz wrote:

Thank you for that interesting link. I just checked out some of the material. He claims to be able to deliver 1.7 TWh of storage with a VERY BIG ROCK 500 meters in diameter and 1000 meters deep, which would be a very conveniently small sized installation with quite the punching power.

We don’t really need huge numbers like this … we have a grid do we not? We have capacity planning for seasonal variation. I don’t really see the need to stretch the intermediate and peak side of the scale to fit into the baseload box (and build such a thing like we were running our electricity grid in full peak mode 24/7 and 365 days a year). Xcel has some performance data on their pilot NaS and wind turbine integration project. The conclusion: “the technology works.” They see a range of 5:1 – 10:1 for battery capacity scale, in order to:

– Effectively shift wind energy from off-peak to on-peak availability

– Reduce the need to compensate for the variability and limited predictability of wind generation resources

– Support the transmission grid system by providing voltage support, which contributes to system reliability

– Support regional electricity market by responding to real-time imbalances between generation and load

Detailed report here. At 10:1, this amounts to about 100GW of storage for a place like the US (which I have already suggested is a reasonable target, or a high target according to one planning document).

Anybody following developments in isothermal compressed-air storage (for above ground, deploy anywhere, storage tanks in the 4 MW range)? These do not require natural gas to run, but utilize pistons that work at a much broader pressure range than turbines. SustainX is one developer in the States (with DOE funding), and touts a 7x price reduction in storage costs “compared to classical CAES,” 30 year operating life, modular design for scaling to a variety of applications, no geological constraints, and more. Air products is another company breaking away from geological constraints with their cryogenic liquid air (LAES) system for energy storage in 100MW range, 12 hour discharge time, modular configuration, 75-85% efficiency, and more. Adsorption-enhanced compressed air storage is another advanced CAES concept.

Like

Leave a Reply (Markdown is enabled)