[JumpCrisscross]

> You object to this because it's not available, then you point to SMRs, which don't exist now except on slides

Right. Long-term large scale hydrogen manufacturing and storage (presumably as ammonia) is not a thing, not to the tune of several percentage points of primary generation. This is speculative, like SMRs.

> Hydrogen + batteries are nicely complementary and enable renewables to undercut new construction nuclear even for supplying baseload power

The math doesn’t work with current technology. Not at that scale. (The lithium alone would be orders of magnitude more than what is forecast to be needed for EVs.)

[pfdietz]

Hydrogen storage would be underground. Very cheap and more efficient than using ammonia.

If underground storage is not available (lack of proper geology), the scheme could be synthesis of methanol, with oxyfuel combustion (Allam cycle) using oxygen from electrolysis, and with the CO2 stored for recycling to produce methanol. The CO2 and oxygen would be stored as refrigerated liquids.

> (The lithium alone would be orders of magnitude more than what is forecast to be needed for EVs.)

I don't believe this is true, not even close. The US has more than a quarter of a billion motor vehicles. A Tesla has about 70 kWh of storage and about 10 kg of lithium, so that's 17 TWh of storage. An optimal "synthetic baseload" for the US from wind and solar might use 6 hours of battery storage (and e-fuels). US average power production is about 500 GW, so six hours is about 3 TWh of batteries. Even if you multiply that a bit to assume less transmission and need for peaking storage it's still not "orders of magnitude" more than for EVs.) All this is even assuming Li-ion batteries would be used for stationary storage instead of the Na-ion batteries now coming on the market.

Were you assuming many days or weeks of batteries would be needed? That's the beauty of also using e-fuels; the batteries can be reserved for the short term storage they're truly suited for.

[JumpCrisscross]

> An optimal "synthetic baseload" for the US from wind and solar might use 6 hours of battery storage (and e-fuels)

This tempers high-frequency variation. The 2% over a full year lower-frequency variation is the problem.

You’re proposing good ideas. But they’re untested and contain fundamental engineering risk. It’s similar to when pro-nuke folk assume SMRs will happen; hence, the analogy.

[pfdietz]

> This tempers high-frequency variation. The 2% over a full year lower-frequency variation is the problem.

Right. And that's where the e-fuels come in. Using batteries for that storage problem is absurd and unnecessary. The argument that renewables can't reach 100% because batteries can't do year-scale leveling is a strawman argument. It's bad engineering to try to use batteries for that use case. And because 2% (say) is such a small fraction of the total demand the low round trip efficiency of using e-fuels for grid storage doesn't hurt much overall.