In recent posts here, here and here Euan Mearns and I have published estimates of the amount of storage needed to integrate intermittent renewable energy with the UK grid in meaningful quantities. All of them point to the same conclusions:
1. The volume of storage needed to convert intermittent renewable energy into dispatchable energy is very large, with estimates running in the 1 to 5 terawatt-hour range even at modest levels of renewables penetration. (Note that these estimates are confirmed by an independent estimate from David Mackay detailed in this comment.)
2. Pumped hydro is the only large-scale, commercially-proven technology that has the potential to handle such storage volumes, but there’s no realistic prospect that the UK could add anything like this much new pumped hydro (total UK pumped hydro storage capacity is presently only 0.03TWh).
In short, the UK will have very considerable difficulty integrating large amounts of intermittent renewable energy into the grid if a solution to the storage problem can’t be found.
And most other countries are in the same position.
Yet there are studies which claim to have developed scenarios that allow the UK and other countries to be powered largely or entirely by renewable energy by or before the middle of the century. The authors of these studies are aware that an energy storage problem exists, so clearly they believe they have found a solution to it. What might it be? To find out we will briefly review the results of two such studies, one for France and the other for the UK:
In April 2015 the French Environment and Energy Agency ADEME completed a report entitled “Vers un mix eléctrique 100% renouvelable en 2050” which purports to show how France can generate 100% of its electricity from renewable sources by 2050 without the lights going out. The report has not yet been issued because ADEME apparently wrote it without authorization and its conclusions are ruffling feathers, but a leaked copy (in French) is available here. ADEME developed the 100% scenario by juggling 14 separate sources of renewable energy, three different types of energy storage, interconnector flows with neighboring countries, demand management, hourly generation data based on five different future meteorological scenarios and four different energy mixes in 21 individual French regions.
I began the review by making an independent estimate of energy storage requirements for ADEME’s 100% renewables scenario, which I did by applying ADEME’s 2050 mix of 63% wind, 17% solar, 13% hydro and 7% thermal plus ADEME’s assumed 14% reduction in demand, to actual 2014 generation data (details in the Addendum at the end of the post). I came up with 20TWh, a truly enormous amount, and went away feeling confident that ADEME would not have allowed for anything like this much in its study.
Which just goes to show how wrong you can be. A more detailed reading of the report revealed that ADEME in fact allows for at least 46TWh of storage, well over twice my estimate.
And where does all this storage come from? ADEME provides the following details:
Short-term (diurnal) storage: Batteries, adiabatic compressed air energy storage, others. 12GW installed capacity, 6 hours duration. Implied storage capacity = 72GWh
Short/medium term pumped hydro storage (STEP in French): 7GW capacity, 32 hours duration. Implied storage capacity = 224GWh.
Long-term (seasonal) storage: Power-to-gas (methane)-to-power. 17GW capacity, storage capacity ~46TWh (implied ~2,700 hours duration).
The 224GWh of pumped hydro capacity by 2050 is probably achievable (according to Eurelectric France already has 184GWh) but the 72GWh of short-term battery, ACAES and “other” storage probably isn’t without a major breakthrough in battery technology (72GWh = 10 million Tesla 7kWh units) and rapid commercialization of CAES. These, however, are bit players compared to ADEME’s long-term storage, which I quantified at ~46TWh by summing the bars on Figure 34 of the ADEME report:
ADEME Figure 34: Inter-seasonal storage (red) and releases (blue) from gas-to-methane storage (note that the plot begins in June)
This huge amount of storage is achieved entirely by storing renewable energy in the form of methane gas. Surplus renewable electricity is used to make hydrogen, the hydrogen is combined with CO2 to make methane and the methane is stored in tanks before being fed to gas-fired plants to fill demand during periods of renewable energy deficit.
Can power-to-methane storage be made to work? We will discuss that shortly. But the 100% renewables scenario won’t work if it can’t, because without it France will be unable to meet demand when the wind doesn’t blow. ADEME’s Figure 112 (modified by the author) shows methane released from storage providing no less than a quarter of France’s total electricity consumption – and up to half of it for short periods – during a simulated windless week in February 2050:
ADEME Figure 112: Contribution of storage to France’s electricity supply during a simulated windless week in late February 2050. The red-shaded area shows combined releases from all three categories of storage
(Figure 112 also shows France obtaining another ~10% of its power from imports during the week. This isn’t a storage-related issue, but where these imports will come from given that the wind probably won’t be blowing anywhere else in Western Europe isn’t made clear.)
The latest version of the Centre for Alternative Technology’s study “Zero Carbon Britain” can be accessed here . The study claims to show how the UK can totally decarbonize its electricity sector by 2030 while at the same time more than doubling its electricity generation, which is a bit of a stretch, but we will review it anyway. Here’s CAT’s proposed 2030 generation mix (note that the UK generated 359TWh in 2013):
The CAT report doesn’t address the question of the feasibility of installing the additional ~250GW of renewable capacity needed to meet these totals by 2030 nor does it discuss the cost, which would probably exceed £1 trillion. But we’ll assume it all gets installed. And having installed it, what does the report have to say about storage? First it correctly identifies the problem:
As most of the energy in our scenario is from variable sources there is often a mismatch between supply and demand, with both large surpluses and shortfalls. Adding more electricity generating capacity would increase surplus electricity production without significantly reducing the problem of shortfalls. (Demand management) and combining different renewable sources of energy helps but doesn’t completely solve the problem.
And presents supporting graphics:
Then it touches on the shortcomings of some other storage technologies:
- Pumped hydro: Not enough of it
- Batteries: Not cost-effective
- Hydrogen: Explosive, hard to store, no infrastructure
Then it gives specifics:
Our scenario combined various short term energy storage mechanisms (hours to days) with the capacity to store up to 60TWh of carbon neutral synthetic gas for months or years.
And once again the carbon-neutral synthetic gas is methane.
So CAT comes up with the same long-term energy storage solution as ADEME – power-to-methane – and with a comparable storage requirement.
I didn’t discuss power-to-methane storage in previous posts because it seemed much too far away from large-scale commercialization and also probably too inefficient and uneconomic to be given serious consideration. On the other hand it’s not limited by the availability of topographically- or geologically-suitable sites, as is pumped hydro, CAES and FLES, or by the prohibitive cost of battery storage, and as such it represents effectively the only option for storing surplus renewable energy in large enough quantities to support high levels of renewables penetration. So now we will take a closer look at the question:
Will power-to-methane work?
Using the ADEME scenario as an example the following things must happen before it does.
The technology must be commercialized: This will be a major undertaking because at present there is only one small commercial power-to-methane plant in operation in the world – the 6.3MW ETOGAS plant in Germany, which produces methane by chemically combining hydrogen from electrolysis with CO2 from a neighboring biogas plant. The electrolysis process works with intermittent renewable energy – an advantage – but while the plant consumes 6MW of electricity it outputs only about 3MW worth of methane. The graphic below, reproduced from the ETOGAS article linked to above, summarizes system performance over a 20-hour period of operation with intermittent power delivery:
And when this energy loss is allowed for we find that over five thousand ETOGAS-sized plants will be needed to fuel the 17GW of gas-fired capacity ADEME considers necessary to support 100% renewables generation (which itself appears to be an underestimate. 17GW of capacity would have to operate at a 166% capacity factor to generate the 19,000GWh projected in February – see ADEME Figure 34 above.) An energy storage system that works at only 50% efficiency – less when downstream losses are taken into account – is of course inefficient but still preferable to having no storage system at all.
CO2 feedstock for the methane plants must be obtained: This is another potential roadblock. If ETOGAS consumes 2,800 tons of CO2 a year then roughly 15 million tons of CO2 will be needed to support 2050 methane production. ADEME claims that this CO2 will be obtainable from “numerous possible sources” but there are actually only two –biogas and carbon capture & storage. By 2020 France is projected to have 740 biogas generation plants with 315MW of installed capacity, but these would still supply only about 150,000 tons/year of CO2 even if all of it were captured, and CCS isn’t applicable even if it could be commercialized because ADEME’s 2050 scenario doesn’t allow for any.
An extensive pipeline network must be constructed: The CO2 needed to make the methane, which would come from thousands of small biogas plants, must be transported to the methane plants and the methane must then be transported to the gas-fired power plants. ADEME proposes an extensive pipeline network that doesn’t presently exist but which it assumes will exist (donc on présuppose l’existence) by 2050. But France’s transition to 100% renewables would have to begin long before that, so work on the pipeline network would have to start soon. And it would be embarrassing if there wasn’t enough gas to fill it once it was finished.
Incentives to build energy storage capacity must be provided: Anyone who has studied renewable energy knows how important energy storage is, but governments still haven’t caught on. They pay generous subsidies to producers who deliver non-dispatchable renewable energy to the grid whether the grid needs it or not but none worth speaking of to energy storage projects. As a result the storage capacity needed to make the renewable energy dispatchable isn’t getting built, and it won’t get built until governments begin to recognize the need for it and apply appropriate market stimuli. At this time no progress is being made towards building the infrastructure necessary to support large-scale power-to-methane operations.
Clearly the chances that power-to-methane storage can be deployed on the necessary scale within the XX-percent-renewables-by-20YY time frames that many governments have committed themselves to are not good. Power-to-methane is clearly not an option the world can hang its renewable energy future on.
But what about power-to-hydrogen? This is something I haven’t looked into, but CAT doesn’t think it will work:
In principle, hydrogen can be stored and then used directly to produce electricity using gas turbines or fuel cells. However, hydrogen is a very light gas that needs to be highly compressed for storage. It is also quite corrosive and can even corrode metal. It is possible to store relatively large amounts of hydrogen (a few 100GWh) over long periods of time, for example in salt caverns. However, compared to natural gas (primarily methane) hydrogen is difficult to store and transport and there is almost no existing infrastructure suitable for it.
And ADEME doesn’t mention it at all.
So if power-to-methane isn’t feasible, what’s left? We are back to the only approach that can be guaranteed to work – keeping enough conventional backup generation to fill demand when the wind doesn’t blow and the sun doesn’t shine. With this approach the lights stay on even at high levels of renewables penetration, but the system would be highly inefficient and probably ruinously expensive.
ADDENDUM: Estimation of storage requirements for the ADEME 100% Renewables by 2050 Scenario:
2050 hourly generation was estimated by factoring actual 2014 hourly generation from the Paul-Frederik Bach data base to match ADEME’s generation mix (63% wind, 17% solar PV, 13% hydro, 7% thermal – assumed to be biomass operating as baseload and delivering a constant 7GW). Factors used were wind = 16, solar = 15, hydro = 0.81.
2050 hourly demand was estimated by factoring actual 2014 demand by 0.86 to match ADEME’s assumed 14% demand decrease.
The Gridwatch data for France between November 2014 and the present were used to approximate average hourly hydro generation in 2050 (the Bach data base gives no data for hydro).
2050 generation by source obtained by factoring 2014 generation is shown in Figure A1:
Figure A1: Hourly generation by source for ADEME 2050 100% renewables scenario
Figure A2 sums generation from all four sources and superimposes it on 2050 demand.
Figure A2: Total generation versus demand, ADEME 2050 100% renewables scenario
Figure A3 plots generation surpluses and deficits relative to demand:
Figure A3: Generation surpluses & deficits, ADEME 2050 100% renewables scenario
And Figure A4 shows net storage requirements through the year, estimated by accumulating the surpluses and deficits shown in Figure A3. An 80% overall efficiency is assumed, with 10% of the energy lost during the storage cycle and 10% lost during the release cycle.
Figure A4: Storage requirements, ADEME 2050 100% renewables scenario
Figure A4 shows that 17TWh of storage will be needed if the renewables mix is to fill demand at all times during 2050, but this is still only just enough to keep the lights on in December. 20TWh of storage would probably be safer.