XLDES

 The dunkleflaute problem ( a couple of weeks with little renewable generation)  has a short term solution, but is a long term problem and I see no research or policy  addressing it. ( nuclear is not the answer as I've explained on numerous occasions)

In the short term it is possible to restart gas powered combined cycle generators ( CCGT) and handle the variability of demand using batteries that will have been added in earlier phases.  As I've commented before these need to be owned by the grid and out of the market model. They will be costed on an annual generation / power capacity basis as the generator of last resort. They will use the written off CCGT assets.

However we don't want to use natural gas due to the fugitive leakage issues in mining, pipes, ships etc. By some estimates this makes USA shale gas worse than coal.

We could make the gas by biological anaerobic digestion of waste food ,leaves and  animal waste. However unless the process is adjacent to the CCGT, and stored locally , the leakage may make this nonviable for climate reasons.

Rechargeable batteries won't work for  reasons described before eg asset cost. 

The next possibility is flow cells. Here the power is separated from the capacity  which is stored in big tanks.

I had hoped some years ago that this would fill a 2-12 hour storage gap  but the newer battery chemistries seem to be moving into that space.

There are two main types: vanadium based and zinc/bromine based. These are actually in commercial production but both are struggling to be competitive vanadium is the larger volume product.

The primary issues with flow batteries are the volumetric and stoichiometric energy density. 

For vanadium the metal is in solution and the ions shuttle between two different ionic levels. 

For zinc bromine , zinc bromide in solution is split into zinc plated on one electrode, and the bromine absorbed on the other. A general problem  with all the chloride /bromide/iodide types is that the fully charged  product eg Cl2 is not ionic and of low solubility and can appear as a gas and escape, or sneak through the ionic separator to the anode. In addition the Br- ion can carry uncharged Br atoms with it as in Br2 - ,and Br3 -  all single minus which degrades performance. This may have been solved by a recent paper  ( nature 7 nov vol 635 p80) by an electrolyte additive called a zwitter ion. This is a neutral molecule that contains a positive ion and an negative ion bound together thus its soluble in water. This controls the neutral molecules making them soluble and ionic, and stops them sneaking through the separator. ( there's more to this story)

It is possible that this may resolve some of the issues with zinc bromine and make it competitive and more reliable but it does have a finite capacity issue which makes it unsuitable for XLDES.

Organic flow batteries are quite cheap but as the relevant polar organic stable molecules are not too soluble, the tanks are large and the volumetric density is terrible. Progress on these is incremental and slow.

Metal sludge flow batteries have a potentially high volumetric density, but sludge management is not easy though the materials are cheap.

Thus I don't see flow batteries delivering a two week storage solution.  

FORM is talking about a 30 hour iron particle/air battery  . The materials will be cheap  but it seems to fit in the daily , or low renewable day rather than dunkleflaute.

Thus there are currently no technical directions that might lead to a dunkleflaute solution for renewables.

Its time for a silly suggestion that might generate some thoughts about a more sensible one!

See diagram .

The idea is to use Aluminum, a trivalent metal that holds a great deal of energy ( 17 megawatt hours per tonne), is easy to store as it doesn't rust,, is already produced in huge quantities, by sites with massive grid connections, and is safe.

The aluminium is piled up at the aluminium foundry when energy is cheap . (the Al box is the factory , the S box on the diagram). 

When energy is expensive it is fed into a reactor/battery, consumes air , turns into aluminium hydroxide which is removed, and generates electricity.

It generates DC electricity. This can be used directly to run the aluminium plant ( resulting  in the grid demand of the plant going hopefully to zero and releasing power for other users) or , if the plant is taken down to minimum flow rate most of the electricity can be converted to AC and delivered to the grid .Any heat generated in the process can be fed to the aluminium electrolysers to keep them molten and reduce the electricity demand further

The aluminium hydroxide produced is the input to the aluminium smelter ( bauxite is impure aluminium hydroxide)

The additional equipment that must be funded is the gigawatt level aluminium discharger/reactor /battery, G on the diagram, and the transformer X on the diagram.

For the UK a two week dunkleflaute at around 30 GW demand , would need  about 20 Twh of storage, or 1.17 megatonnes of aluminium, For scale the world produced 70 megatons of aluminium last year. However note in this system the aluminium is a one off stock as its completely recycled at the aluminium plant. Its also possibly incremental since it can act for shorter term periods of perhaps two days where batteries are not effective 

The volume of aluminium is the reason that I suggest it's kept local to the aluminium plant to avoid transport!

This is the highest energy density I can think of by essentially burning a metal, that is stable enough to be stored, in a battery 

Any higher density  solution is going to involve carbon or nitrogen  hydrogen bonds reacting with oxygen. It would  require making fuel such as methane hydrogen ethane, ammonia  or methanol ( the last two can easily be stored at normal temperature as liquids so easier to store in volume) . These could all then be burnt in existing CCGT, with no additional global warming effect ( if the carbon comes  from CO2 captured from the air in the same year, the hydrogen comes from electrolysis of water, and the nitrogen from the air). Burning them all has the issue of heat loss inefficiency ( around 30-40% of the energy is lost in burning , in addition to the energy lost in making the fuels) with a  probable overall efficiency around 20-30%. There is little additional capital equipment for t generating , but a lot for the fuel creation.

If we are willing to accept the aluminium route in principle  on the argument of tight closed loop and end to end efficiency, there are a few little technical problems.

No one has yet made a continuous flow aluminium consuming battery.

There are startups with single-use 1000 mile aluminium batteries ( send the battery back to the aluminium plant), batteries where you manually insert a new aluminium plate when the old one has dissolved, and send the sludge back to the aluminum plant.

This suggests that some of the problems are solvable and what is needed is a big Musk type concept to put it all together at massive scale. 

It might be feasible to build initially on a process like this  to stabilise electrical input to an aluminium plant supplied by a wind farm. Direct connected wind is probably going to be the cheapest source of electricity, but its variability is a problem for a continuous aluminium process. Due to the enormous electricity demand batteries may not be the solution except for very short term power conditioning, and varying the smelters only gives about 30% variability with loss of production. It just might make commercial sense to operate something like this for both daily and seasonal purposes?

That's my current thoughts on Dunkleflaute.

We have perhaps 30 years before we have a significant excess of renewable energy sufficient to store it for dunkleflaute however it's stored.

If electricity then becomes very cheap , the efficiency of the overall process may not matter too much and some drop in fuel will be the solution. 

If electricity is still limited , then efficiency will matter, and the metal cycle described might be much better

Ian

jan 2025.