Yes, 20% is substantial. I was trying to point out the fact that the Sun shines only (roughly) 20% of the time. If one wanted solar to play a larger role then storage becomes an issue.
Wind and solar are likely to end up having very similar LCOEs. Somewhere at or just below five cents per kWh.
Storage will have some sort of a price. If Ambri's liquid metal battery works then it might be on a couple of pennies. If it's something like Eos's air-zinc battery then closer to ten cents. Or pump-up hydro somewhere between the two.
The cheapest electricity will be electricity straight from the generator rather than stored electricity. The wind blows more than 20% of the time, far more in good wind sites. Wind, even if it turned out to be slightly more expensive than solar, will probably play a dominate role because it is available more hours per year.
The best idea, IMO, is to use grid power for trains. The more we combine supply sources and demands the more variability is smoothed. Trains and non-train demand can share generation and storage.
Trains might even be a way to do some load-shifting. We might, for example, run many more trains late at night when other demand is down and the wind typically blows harder. Or if supply was stretched thin and brownouts were threatening we might halt a few trains or run them at a slower speed for a few hours until the crisis has abated.
--
I've had the opportunity to ride on some of Europe's rail system the last couple of years. After many years of flying, cramped into airplane seats, and years of driving highway speeds, in the company of speeding 18-wheelers, I'm in love with trains.