[akira2501]

Aren't those cost factors based upon the type of load curves we currently see? Isn't there some reason to suspect that the efficiency rating will drop if we experience much greater offsets between time of generation and time of demand with the types of peaks that EV charging might bring? Wouldn't it be nice to have all this without having to engage with the daunting prospect that is the "smart grid?"

[nostrademons]

There's a common misconception that thinks of electricity like a fluid or supply chain. If you generate more of it locally, it reduces the load on the grid and you need to spend proportionally less on distribution. This is false: the actual electrons in AC move a small fraction of a millimeter, electrical potential travels at a speed that is effectively the speed of light, electric conductivity is all-or-nothing, and if you need the grid at all, you need the full expense of building and maintaining that segment of grid. You probably need some form of grid if only to even out load spikes (running your clothes dryer often takes 20x the power as all the lights and electronics in your house, but not everybody runs their clothes dryer at once) and manage seasonal variations (solar power, particularly in northern latitudes, can be 5x higher in summer than winter, which is not a problem when you're powering southerly residents' air conditioning but is when your cold house doesn't need A/C).

The load curve over time only matters to the extent that you can entirely remove remote consumption. You can use batteries to smooth out night and day. You can reduce the use of batteries by sponsoring V2H EVs and workplace charging, so that you charge your EV when solar is abundant in the day, and then drive it home to power the rest of your house. But this does nothing for summer vs. winter, it does nothing for wanting to run a clothes dryer or space heater (many of which actually exceed the max power draw of a whole-home battery), it does nothing for wanting to charge your EV up to full before a long road trip.

I am in favor of microgrids, but this is more a statement that we should rationalize our distribution infrastructure rather than that get rid of the grid entirely. When power plants were large centralized industrial buildings that needed a steady supply of fossil fuels delivered by road or rail, it made sense to just build a few of them and then have a huge grid that distributed the electricity everywhere. When you can put solar on every rooftop, it might make more sense to have the smaller remote communities all invest in rooftop or community solar, wire them up in a microgrid of ~1000 homes, put in a big utility-scale battery, but otherwise disconnect them from the main grid so that power lines don't go through tinder forests. And then the big cities draw from big utility-scale solar and wind farms in the desert, connected by conventional power lines along major transportation arteries. But there's still some grid there, it's just a smaller, cheaper grid where you make the connections that are easy to maintain and distribute generation to the remote communities that can run their own self-sufficient grid.

[thinkcontext]

> Aren't those cost factors based upon the type of load curves we currently see?

No, that would effect the price the electricity would fetch, not the cost to buy panels and put them on the grid. Home rooftop panels are so much more expensive because of economies of scale.