Comment by AnthonyMouse
4 days ago
The energy density of gravity-based storage is very low, so it could work in places where you have mountainous terrain and a lot of cheap land, but what do you do after the sites suitable for it are already in use? To make that work you'd need a scalable storage technology with a low enough cost per kWh of capacity to economically scale to multiple TWh of storage in case it's cloudy for an extended period of time.
For example, if you could make batteries at a price of $115/kWh, the cost for enough capacity to sustain the US power grid for a week would be around 24 trillion dollars, and that's just for the batteries and not any of the associated electronics or the land or the photon collectors themselves. It seems like to make your plan work you'd need a scalable storage technology with a significantly lower cost per kWh of capacity.
Using batteries for week-long storage is a bad idea. Lower efficiency, but lower capex, storage technologies like hydrogen would be better for that.
Optimized combinations of batteries and longer term storage typically comes out to maybe 12 hours of batteries, sometimes less.
We wouldn't need the hydrogen until the last few percent of the grid goes renewable; until then natural gas would be fine.
Suppose you have a 65% renewable grid which is then 35% natural gas, but it's dark and the batteries are empty. Being able to generate 35% of the load from natural gas still leaves you needing to make up the other 65% from somewhere. This is obviously worse if the grid is 95% renewable and 5% natural gas, but it's the same problem.
To use hydrogen for that you'd need close to a TW of hydrogen fuel cells, and storage tanks to hold enough hydrogen to run the whole grid for a week, and enough hydrogen production equipment to refill those tanks in a reasonable amount of time if you had to use them, none of which gets used but once every year and a half. That might still be cheaper than a week's worth of lithium batteries but it's still not cheap when it's something you have to pay for on top of the generating equipment and batteries needed to run the grid on a normal day.
And it still leaves you in a bad way if you had enough hydrogen for 7 days but the clouds persisted for 9.
Hydrogen would be burned in turbines, not fuel cells. A simple cycle turbine powerplant might cost $0.60/W. Remember, the goal here is minimizing capex, not maximizing efficiency. Backing up the entire grid with such turbines is a small cost compared to powering that grid with, say, nuclear ($10/W).
A once-through steam cycle with steam generated by hydrogen-oxygen combustion heating additional water (save the oxygen from electrolysis also) might have even lower capex. The combustion would occur with water injection in a kind of rocket-like combustor, no heat exchangers needed. This is potentially cheaper because it avoids spending 50+% of the output of the turbine on the compressor.
Hydrogen, like natural gas, is extremely storable, very cheaply, in underground storage formations. Nations typically store months worth of natural gas, and the volume could be greatly extended. Europe for example has the geology to store millions of GWh of hydrogen, and that's not even touching the salt formations under the Mediterranean.