The problem is, the only place where you can make that worth the while is by building out solar in Northern Africa. Unfortunately, the countries in that region are a combination of failed states, governed by dictators, under threat of war or terrorism or pissed off after hundreds of years of Western colonial powers coming in, taking natural resources and leaving no meaningful income and perspectives to the locals.
There's no easy solution for a single one of these problems, much less for all of them.
No it could work in other places. You buy 1 unit of energy for £50 per unit. And then sell at £150 per unit, or £90 for your remaining 0.6 units. The profit is from the price difference. You make money from arbitrage. And this is the kind of price difference you would expect in a grid with lots of intermittent sources. And the cheaper the source commodity becomes the less efficiency really matters.
This is essentially the principle of a virtual powerplant - household tariffs like Octopus Agile (price fluctuations every 30 minutes) with a home battery make it possible to do this at quite a small (eg household) scale. A bigger operator could find bigger returns if they could make it work.
I think there's something to be said for storing excess energy like this - it beats paying into fossil fuel economies.
It's difficult to do these kinds of arbitrage deals with energy, at least electric energy - it's simply incredibly expensive and requires a lot of upfront cash to build out serious storage, whereas it's easy to store and ship oil - in a pinch, you can just buy an old tanker truck for a couple thousand dollars. Not to mention that in order to build such a project you would need a lot of transmission capacity in the grid so that you could e.g. have a battery in Bavaria buy up surplus power from the North Sea wind farms.
The one area where the principle works currently is ultra-short storage times like Tesla did in Australia... but that's not much of actual energy being stored, the service Tesla provides here is smoothing out demand - it has 194 MWh @ 150 MW, so barely one and a quarter hours of runtime at full load, with the complete grid having something like 60 GW peak capacity [1].
Another idea that is being floated is to repurpose old electric vehicle batteries or electric vehicles themselves as decentralized storage. That is a very charming idea, but again it requires large upfront investment for the batteries as well as for expanding the grid to keep up with the demand... and there are no standards yet for all the "smartness" needed for such a system to work, many chargers and electric vehicles are not capable of running in reverse, and many people are skeptic of putting a difficult to extinguish fire risk into their basement.
And yet another idea for arbitrage comes from the consumption side - basically have large consumers with storage such as a heating or warm-water system enabled for remote control so that oversupply spikes can be mitigated. But again, the current grids lack the "smartless".
Technically it is exactly the same as a generator, you sell power at a particular price. From the grid perspective it is exactly the same. The fact that you also consume energy is coincidental. And it can work well with existing grid systems by turning on as frequency drops or you get signals from a control room. It uses the same kind of approach as a peaker plant (the tanker truck filled with diesel and a generator).
Also knowing the usual imperialist power games played there even if we would with huge cost stabilise the area, an other power coming in destabilizing it again is quite a big risk...
About 60-70%. Possibly it could be improved with better electrolysis techniques or large-scale facilities. Some Swedish companies are pushing for that solution: https://www.hybritdevelopment.se/en/a-fossil-free-future/ For now it's only for the steel industry but could in theory be used for other parts of the grid too.
A 60-70% round trip efficiency would be very impressive for round trip efficiency of electricity to hydrogen and back. Its in the same ballpark as pumped storage. I can only assume it's assuming using fuel cells which are 90% efficient for the conversion to electricity. I'm not convinced these have really been deployed at that scale before, the biggest fuel cells I'm aware of are used in non-nuclear submarines where cost is not a priority.
