So I see from this report from GreenTech Media that the worldwide production of photovoltaic (PV) cells is due to pass 10 GW this year. That’s a kind of random milestone, but to put it in perspective, that’s about 2 GW worth of baseload power, like that from nuclear plants. Nukes run about 90% of the time at their peak output (called the capacity factor), while solar cells only do that 20% of the time (25% in the desert). 2 GW is about 2 nuclear plants worth, or about the average output of the turbines at Niagara Falls. Also, that’s 2 nuclear plants more than have been built in the last 30 years in the US.
“Silly tree-hugger” I hear someone mutter. “There are over a hundred nuclear plants in the US alone. It’ll take 50 years at that rate to replace even those plants, never mind the other 80% of US power generation.” Yes, but it’s not the 2 GW number that’s interesting – it’s the growth curve. Production has grown from 170 MW in 2000 to 10 GW in 2010 at a compound rate of 50% per year. That’s a doubling about every 18 months – a nice match to Moore’s Law for semiconductors.
So at that rate how long would it take to get enough PV power to run the country? The US draws an average of about 400 GW, so at 20% utilization we’d need 2000 GW of cells. We’ll ignore storage for the moment. That’s 200X the world production this year, or ~8 doublings. At 18 months per doubling, that’s only 12 years.
Isn’t compound growth staggering? Surely that can’t be right. How expensive would that be? Well, First Solar is selling thin-film CdTe panels right now for $0.85 per watt. The rule of thumb in manufacturing is that every doubling of production leads to a 10% reduction in cost, so:
cost after N doublings = base_cost * (0.9)^N = $0.85 * (0.9)^N = $0.37/W
So building 2000 GW of solar panels would cost ~$700 billion. That’s less than what the US has already spent on the Iraq/Afghan occupations. Instead of killing a million people, the US could have gotten rid of fossil fuel power production. For that matter, the US just spent $700 billion to bail out Wall Street. It spends ~$500 billion a year just on oil.
How big would such a system be? Looking at First Solar again, a 70W panel occupies 0.72 m2 (4′ x 2′) and weights 12 kg (25 lbs). That would cost ~$25 at $0.37/W. That’s only about 5X as expensive as asphalt roofing shingles. 2000 GW’s worth would need 20,000 km^2 of panels, a region about 100 miles on a side.
Well, that’s a lot. But consider that the US produces about 2,000 km^2 of plywood a year. (More exactly, the USDA Forest Service Forest Products Laboratory says that the US produced ~20 billion ft2 of 3/8″ equivalent plywood in 2005). If the solar industry could produce panels like plywood, they could fill the country’s needs in 10 years.
Multiply by 5 for the whole world, or by 10 when China and India come up to reasonable living standards. Either way, it’s not infeasible.
And solar PV is actually about the most expensive form of renewable energy. Concentrating solar thermal is much cheaper, has a higher capacity factor because it can track the sun, and can store its energy in big underground steam tanks.
Wind is cheaper still. World production of wind will probably hit 50 GW this year, and it too has a higher capacity factor, more like 30%. In Europe they’ve been building more wind than natural gas plants for the last couple of years. Kansas and Texas are getting covered in towers.
I take comfort in this at a time when there’s constant bad news about energy. First it was the horrible Massey coal mine deaths, then the BP disaster in the Gulf. Behind that is the looming threat of Peak Oil, and behind that the vast catastrophe of global warming. It’s one wave after another, with a tsunami on the horizon.
Yet we can and are getting past this. Replacing the fossil fuel infrastructure looks daunting, but modern industrialism has awesome production capabilities when it gets going. It looks like it would take a lot of money and a lot of stuff, but it’s not a lot compared to quite ordinary businesses.