Physicist Anders Carlsson, at Washington University in St. Louis, and Sid Redner of the Santa Fe Institute have created a new mathematical model to describe the most reliable, efficient and cost-effective way to harness solar power.
It’s always important to read to the end of an article!
Carlsson said the math of renewable energy points to another important lesson: The search for perfection might be counterproductive. A hypothetical system that runs exclusively on renewable solar and wind power would be significantly more expensive than a renewable system that used small amounts of natural gas as a backup, he said.
He estimated that, with current technology, a 100% renewable system that powered St. Louis could cost $130,000 per household. A system that was 95% or 99% renewable, however, could be in the range of $80,000 to $90,000.
“Extremely highly renewable systems are very expensive,” Carlsson said. “If we can get to 99% renewable in 10 years, versus 100% renewable in 30 years, we'd better figure out how to get to that 99%.”
What does the actual paper say?
For an annual failure rate of less than 3%, it is sufficient to have a solar generation capacity that slightly exceeds the daily electrical load at the winter solstice, together with a few days of storage.
The definition of “annual failure rate” appears to be the chance that, in a given year, a day will come, on or about the winter solstice (when insolation is least), that the storage will be exhausted and power demand cannot be met. This does not necessarily imply 263 hours of blackout in an average year, which would indeed be poor value for money!
There are two glaring flaws in the analysis. The first, which mostly affects price calculations, is that only the present–day electrical demand is considered. No allowance is made for the likely doubling or tripling of system loads as a result of promoting electric cars, electrification of home heat, and so on. The second, which appears to completely invalidate the analysis, is that, while great effort is made to simulate the variation of solar energy input, no allowance whatever is made for variation of system load, which is assumed to be a constant 4·6 GW, all day long, all year round.
The paper quotes the cost of a solar installation “just sufficient to supply the daily electrical load of the St. Louis region during an average insolation day at the winter solstice” at $75 billion (covering a land area of 16×16 km, out of the 270×270 considered as the “region”), and the cost of storage for one day worth of load at $22 billion. The minimum–cost result of their simulation calls for about 1·2× the minimum solar installation, and 2 days of storage, for a total of about $134 billion.
Let us consider the alternative that is not mentioned. A 100% nuclear energy system is commonly assumed to be far too expensive. Three EPRs, generating 4·9 GW continuously (excepting refueling outages, once every 18 months per reactor or 6 months for the plant, which can usually be scheduled during periods of low demand), at the price of Hinkley Point C, would be about $60 billion. Four AP–1000s, generating 4·5 GW, would be something like $60 billion at the price of Vogtle 3&4. And three Korean APR–1400 units, generating 4·0 GW, would be about $18 billion at the price of Barakah. These figures should give us some kind of basis to work from.
For this price, even at the exceptionally high prices of Hinkley Point C and Vogtle 3&4, we could buy some 10 GW of nuclear generation, which would be adequate to meet, under virtually any conditions, a system load of twice the average. At Barakah prices, 30 GW could be had, which would be more than adequate to handle any foreseeable load escalation.
The above calculation does not even consider the possible use of nuclear heat. Nor does it account for the cost differences due to lifespan of facilities ― storage batteries will probably need replacement in 6 to 10 years, PV panels in 12 to 20, and nuclear steam plants in 40 to 60 years (with major refurbishment after 20 to 30 years).
We are often told that wind, solar, and storage are cheaper than nuclear, but this hardly seems to be the case. We are also often told that they are constantly coming down in price, so that even if they aren’t cheaper this year, they will be next year, and there is no reason to make investments in nuclear. We wonder. People in the industry seem to think that even Barakah costs are much above those possible, given the kind of learning and replication involved with the kind of large global commitment which a real effort at decarbonization would require.