vanesch
Staff Emeritus
Science Advisor
Gold Member
- 5,109
- 20
Andrew Mason said:50W average represents around 270W peak. In daytime with no clouds and overhead sun, the solar irradiation would be the full amount (1367 w/m^2 less a small amount that does not make it to the surface). So the output at 20% efficiency would provide 270 watts/m^2.

These are all good points. It was a ball-park figure. The $1 trillion for infrastructure may be a little low. If you distribute the solar panels over a large geographic area and keep them in the lower latitudes in areas that have more sunny days, you can avoid many of these problems.
Well, you can attenuate them: the peak/year average can be as low as 3 (while in moderate regions, this is rather 5-6), and the summer/winter variation is probably smaller too. But you will still have at least a factor 2 or 3 over simply "yearly average", if solar is to provide a *large fraction* of the provided power (say, 70% or so of consumption) in a reliable way.
As long as solar (or wind or other erratic renewables) is a minority contributor (say, 15% or 20%), then this doesn't play a role, and the price per KWhr delivered will be much lower, as we can just use "yearly average". What renders this expensive is when we need reliability (which is not needed when it plays in the 15% ballpark, because reliability is then provided by the other technologies).
We are talking about providing all of the power needs for the largest consuming nation on earth. No one is going to do that with one single technology, of course But the cost appears to be competitive with nuclear, and the fuel is free.
I really don't think that, even as a minority contribution, at actual prices, solar PV is competitive with nuclear (you have 25c/KWhr for solar, while this is ~8c/KWhr for nuclear/coal). But even then this comes about because solar is not providing for an essential function in power delivery: reliability and load following.
That is low. But with large demand, you may be amazed what kind of efficiencies and cost reductions might become available. I have used the $2 figure based on today's figures. Where I may be wrong here is in assuming that this represents the price for average wattage and not peak wattage.
Well, current average retail price per peak W is ~$4.5 or so. So assuming this to be $2,- is already assuming a serious drop in price (for instance upscaling). In moderate regions, you have to multiply this with 6, and in sunny southern regions, with 3 to go from peak to yearly average.