Today I noticed this article about EROI, or Energy Return on Investment, by James Conca in Forbes. The numbers are from this study by Weissbach et al. (2013). Here's the relevant graph:
Whoa, nuclear looks great! 75:1?? That's amazing!!
Excited by this, I looked to corroborate the numbers. Looking on Wikipedia, I found these numbers from a study by Murphy and Hall (2010):
Hmm, that's not so good. Nuclear is listed at "5 to 15" in the data, meaning from 5:1 to 15:1.
Digging a bit further, I found that the literature is very divided on the EROI of nuclear, listing it at anywhere from 1:1 (i.e., uneconomical at any price) to 90:1 (i.e., the most bountiful energy source in history).
So what is the real EROI of nuclear? What is the reason for the huge variance of numbers in the literature? Are some studies forgetting to account for the (huge) fixed costs of constructing nuclear plants? Are there big assumptions involved about how the technology will change?
Updates
Sam Wilson, on Twitter, suggests that the high numbers might be counting nuclear submarines and aircraft carriers in addition to power plants.
Colin, on Twitter, suggests that decommissioning costs may play a role as well.
"In 2006, the Australian government commissioned a comprehensive study to clarify the matter. The results were published in Lenzen (2008) who used mean values with data covering 45-years of activity and arrived at the conservative estimate of roughly 5:1 for a full life cycle analysis (from uranium extraction and
ReplyDeleteprocessing, to waste treatment and decommissioning) [88]. Most of this data is 30-40 years old and may not reflect current technology or ore grade."
I think that pretty much summarizes it. Most of the nuclear plants in the US were built 40-60 years ago, so any studies of nuclear are doing a considerable amount of extrapolation.
Nuclear EROI is high due to the energy density of the fuel. The cost of using Uranium is in mining, enriching and machining. Thorium needs no enrichment and is a by-product of rare earth mining, so that cost is minimized also. In a molten salt reactor, fuel doesn't need to be machined. The EROI might only include the fuel for the truck that delivers it from the mine to reactor. The EROI would be hilarious.
DeleteThe main explanation is rather that there is no standard methodology in EROEI studies, so numbers from different studies can't be compared with each other.
ReplyDeleteAgreed. It doesn't seem as though there are concensus numbers for all other energy forms with nuclear being an outlier, but rather, it seems there is no consensus measurement methodology for EROI at all.
DeleteNaive question...why does this matter?
ReplyDeleteSome Republican neckbeards decided to show Jihad Johny, Ayatollah Aaron and the Ruskies that they can beat down oil prices forever but choose their technology not based on economics but based on FUCK YOU LIBERALS, and Noah has been trying to show how solar is the true fuel of the libertarian neckbeard because it costs less, gets you off the evil statist grid and is basically all about you whereas nuclear power is the Sultan of Statism.
DeleteAs to why Noah is obessed about this? Who knows, probably since he realized he wanted to be a VC guy in SF he cant talk about inequality, concentration of wealth, failure of macro econ, or pick fights with economists Marc Anderseen donates to so he has to pick his fights. And beating up libertarian neckbeards is like wrestling midgets, easy and I guess sort of fun, if your options are limited.
I tried to find some modicum of intelligent thought buried somewhere in this comment but gave up. I guess it's just more yellow snow. The winter continues unabated.
DeleteEROEI is a measure of the physical efficiency of an energy source (as opposed to "an economic efficiency"): it's a measure of how much energy you get for the energy you invested in taping the source.
DeleteThe higher the EROEI, the higher the flux of energy available (you can a lot and get a lot too). And flux of energy relate well to economic growth (as everything we do consume energy).
Anon: No No No! Noah isn't interested in becoming a VC (or at least I don't think so) - he wants to join one of those super fancy quant Hedge Funds (*cough* RenTec *cough*). Why do you think he ended up somewhere like Stony Brook --> studies financial markets + possibly gets to network with people like Simons and Frey + geographic proximity...
DeleteAlso, I spend some time in random Physics blogs and you won't believe how often I see "Noah Smith" pop up in the comments. I don't even think he likes economics (not that I blame him!)
Those high numbers may be the marginal cost of producing one additional unit of energy, not taking into extremely high fixed costs at front & back end (i.e. building, decommissioning, maintenance). Wide differences may come from arguments about amortizing and including those as part of the operating cost, since they are known and fixed. Just a guess though
ReplyDeleteYou are not the only one to have this question. TLDR; Study quoted by Forbes was done by nuclear researchers: http://energytransition.de/2014/09/renewables-ko-by-eroi/
ReplyDeleteWhy do newspapers still bother with propaganda, when truth is just a search away?
Well, it's Forbes. Not as bad as Fox, but probably well in that direction.
DeleteThe linked article was great. The author trashes the study done by nuclear researches, which was almost certainly biased, and then goes on to discuss renewables in a clearly biased way. I guess everybody has an agenda.
ReplyDeleteTo what extent does this kind of ROI analysis attempt to incorporate the impact of externalities like global warming, the risks associated with nuclear waste, etc.?
ReplyDeleteexternalities would kill renewables even more, because of CO2 generating backup power . . . nuclear will probably benefit as only widely usable CO2 free source (great spots for wind aren't common and hydro is already used to the max).
DeleteI imagine that the efficiency varies greatly, depending on the design, how friendly the area is, how new,...
ReplyDeleteBut really solar may be the future, but we have to deal with extremely harmful carbon and other air and water pollution now. I love nuclear. Coal is massively more harmful with it's pollution and mining, so is gas. In a second I'd rather my family live by a nuclear plant than a coal plant, far greater danger to your health, life, and happiness from breathing that coal pollution everyday. Vastly more people die on average from coal pollution than from nuclear per Kw.
The Labor Theory of Value did not work and the Energy Theory of Value will not work. Go back to economic cost benefit analysis of alternatives.
ReplyDeleteThis is funny post, I thought you were promoting solar power? EROI of solar is absolutely terrible in every source . . . why bring this up?
ReplyDeleteAnyway, externalities are so hard to measure (decomissioning, CO2 pollution, required backup, worst case damage, using too rare elements . . .) that EROI is kind of pointless. It's already very hard to measure full cost, so why obfuscate things even more?
When someone brings this up just ignore him, every energy source has much clearer pro-arguments than this meaningless randomly measured value.
DeleteI think one of the most important costs is catastrophe insurance. My understanding is that there is no super-cat insurance available for nuclear the maximum liability is capped at about a billion - 5 billion approx. The rest has to be provided by the government, and that may be a cost that isn't always fully accounted for in the studies.
ReplyDeletehttp://en.wikipedia.org/wiki/Nuclear_power_in_France
ReplyDeleteThe French may assist in answering the question. They have rather a lot of nuclear power and are continuing to develop it.
If a country does not have fossil or solar resources (and technology for solar), nuclear may have much better EROI, especially if the country has exploitable nuclear material deposits. So, EROI computation is biased. India and China have large coal deposits so coal plants dominate. We are converting to NG and augmenting with solar and wind. Next Era Energy, Dominion, NRG Energy, Pacific gas and light comes to mind.
ReplyDeleteI do not think nuclear is a good choice anywhere because the mining creates extensive waste and post electric production leaves lots of radioactive material that can only be disposed off by sending to Sun (using Space X rockets!?). Better choices are solar, wind, waves, biomass, NG and increasing efficiency in our life style.
Why would what fossil resources are available in a country influence this at all? Fossil fuels are traded on world markets! Are you unaware of opportunity costs?
DeleteUranium/thorium mining leaves little waste because the volume is small so that one is crap. One of the reasons for differing EROI estimates may be how isotope separation and reprocessing is handled. Reprocessing FWIW results in smaller amounts of end product wastes but includes a proliferation risk.
DeleteEROI is a strange number. The numerator does not measure the value or quality of energy, the denominator does not measure the cost of producing it. You have to assume that marginal costs are equalized across sources by arbitrage, same for marginal benefits. Then you have to assume that these margins will be unaffected by structural changes in the uses of energy. It is a totally non-economic way of thinking.
ReplyDeleteJust a small math. PV now convert only 18% of solar energy to electricity. However, there are recent technology for make cheap PV that convert 45% of solar energy to electricity (the different scientists teams are running for be the first to make 50% cheap conversion PV). Just google "perovskites", they are cheap and permit to build PV and LEDs. So, EROI of PV will go from 4 to 10, or from 2 to 5 with energy storage (that will be need only if 80% of all electric energy come from PV). Anyway, both numbers are well above the real nuclear EROI, that is really close to 1:1...
ReplyDeleteTake note, when these new photovoltaics (cheap PV with 50% energy production) enter market around 2020-25, everyone will run for invest at PV and will be no future for nuclear or coal electricity.
You had me trying to pay attention until "real nuclear EROI that is really close to 1:1." But if PV, which is just capturing extraterrestrial nuclear energy, can become economical enough to replace local nuclear, I'm all for it.
ReplyDeleteWell, nuclear power too need count the energy costs for decommissioning the nuclear reactors. It will need a lot of energy and work maintaining plutonium at a safe place for some thousands years... and if we count eventual accidents like Chernobyl and Fukushima, that will sooner or later happen again, the energy cost is greater...
DeleteWith relation to the Sun extraterrestrial nuclear power, we will don't pay its decommission because we will not exist when the "decommission" happens at 1 billion years from today.
But if we use current monetary market prices and not "malleably to corporate interests" EROI, wind power is cheaper than gas termo thanks to China. And thanks to China and to new technologies, solar PV will be the cheapest of all energies around 2020. Currently, at 2015, thanks to China, solar PV is cheaper than gas termo at the countries around the equator, and that is the reason why South Africa and India are going "gung ho" for it. Brazil is currently going huge for wind energy, it is like one or two Itaipus being build now and my country is not trying off shore wind because the potential for on shore (a big windy litoral) is far away from total use. Brazil is starting to go solar PV too, but maybe slower than the necessary, because Brazil is close to equator line and solar PV market prices, thanks to China, are competitive with termo electric with no government incentives.
tobernite na urgeiriça inda está espalhada por lá né e no gabão e em minas gerais ou generais ...
DeleteOne of the most energy intensive actions in nuclear fuel cycle is enrichment. In the early days, when most of US NPPs were built, enrichment was done using gaseous diffusion, and now is mostly done using gas centrifuge, much more efficient process. From expert point of view, Weissbach et al. study was well done. The only big question is how to treat wasted resources while politicians change their minds how and where to dispose of high-level waste.
ReplyDeleteJust to clarify, since a lot of people seem confused:
ReplyDeleteThe author is a Forbes freelance outside blogger, not a Forbes journalist. Most articles on the Forbes site now are by outside bloggers. Many of them are flogging things related to their day job. This guy seems to be a kind of de facto pro-nuclear spokesman.
EROI as used in this article and study is just an alternative spelling of EROEI, which is energy produced per energy spent. It's not energy produced versus money spent as it might appear.
EROEI is extremely difficult to accurately calculate and so far nobody's really trying. There are all sorts of small estimations involved as energy is consumed in every operational and capital input of any energy-producing system, all the way down to every little piece of material used including the food the labor eats. With nuclear a true EROEI calculation would have to include every little thing that goes into building the plant, digging and refining the fuel, operating the plant, decommissioning the plant, disposing of waste, and protecting the decommissioned plant and waste for as long as that has to be done (a really, really long time).
What the high number for nuclear here reflects is simply that the authors of study had no idea how to do such an extremely complex calculation, so they took a very limited subset of energy costs and pretended those were all there are.
Note though that all EROEI calculations suffer from that problem. They are all wild guesses. Nobody really calculates the EROEI of any energy system, that would be way too hard.
Let's analyze the sources:
ReplyDeleteSolar and wind energy are "outside" energy (they are results of Sun's energy, not Earth's) with very low density. For their energy efficiency low energy density is the most important thing - we need to build lots of collectors (wind turbines or solar plants) over lots of ground to get meaningful amount of energy, so we have to spend lots of energy in advance and lots of energy to clean and supervise power plants (more important for solar). Low energy factor is expected, in fact, some 15 years ago solar and wind power plants needed close to 10 years of operation just to return energy spent while building them (without taking into account maintenance and recycling/decommission).
Biomass/gas/coal - energy is result of chemical processes (burning). Using E=mc2 chemical processes can turn at most 10 ppb (parts per billion) of the original mass into energy, often much less. So, lots of energy needs to be spent to acquire enough of the fuel and to transport it to the power plant (note that if biomass power plant is built somewhere where biomass if freely available, like next to sawmill or local landfill, energy calculation changes, because biomass that is used would otherwise still be brought there and then thrown away, so energy needed to acquire that biomass and transport it can be removed from calculation).
Hydro power uses, basically, the same source as wind power, but much more concentrated. Unfortunately, almost all the best locations are in use worldwide. Hydro power plant needs plant and dam, and than minimum expenses for up to century, so its ratio should be huge.
Nuclear fission and fusion use nuclear power, much more dense than chemical power. Standard fusion reaction turns around 0.1% of starting mass into energy, while fusion turns several times more, standard reactions turn 0.2% - 0.7% of starting mass into energy. In fission reactor some 2% of starting mass of the core undergoes fission in a year, so fission turns around 20 ppm of the starting mass into energy, or 10 000 - 100 000 times more than most chemical reactions. In other words, fission and fusion have very, very high energy density, and need very little fuel. So, all the energy costs of extraction of uranium ore, making fuel and shipping it, except enrichment using old method, are not significant compared to the energy received. Calculation of waste disposal is where the problem lies: if one repository for the whole US is considered, then energy costs of such repository compared to the energy produced from all the US NPPs is not that significant. If, on the other hand, full cost of the repository is added to each power plant, very low EROI values can be reached.
Fusion power currently has problems, depending of technology used, with inefficient lasers, too much power needed for cooling to achieve superconductivity, too great escape of energy produced, or any combination of the above, so it still has to reach break-even point where reaction itself has EROI over 1, not taking into account energy costs of the facility, obtaining fuel or other costs. I'm very skeptical about these reports that we will have fusion in five years. Or even 50 years (commercially viable fusion).
Even if nuclear produced 1000X more energy output for each unit of energy input, it will NEVER be worth it, because the cost of contamination from a single meltdown & catastrophic contamination of air, water, soil, for MILLIONS of years, puts the cost higher than the entire plant's economic outputs for those same millions of years.
ReplyDeleteHow do you measure the "worth" or "valuable energy" of something that can destroy all DNA based life on the planet?
Another "energy cost" (and money & labour cost) of nuclear power, is the fact that there is NO TECHNOLOGY on this planet that can eliminate or cancel radiation, nor is there any technology that can ever separate tritium from any aquifer or from rainfall, until after it is in some test-tube, etc.
How much "energy", labour, time & "money" would it take to remove EVERY radioactive particle and DNA damage from the planet, just from any SINGLE nuclear accident? Fukushima, Hanford, Chernobyl, TMI, and this is not even counting the "normal & ordinary" releases of radiation that are deemed "ok"?
If the "nuclear industry" or supporters of nuclear power, had to pay the full cost of curing every person, animal and plant from radiation induced cancers and damage for the next billion years, I think the answer would be obvious.
For you "religious folks" out there, consider that a contaminated world with U238, the half life is about 4.4B years (the age of the entire planet), Thorium232=14 Billion years, Calcium48= well over a TRILLION years.... my point?
Is that not even JESUS will be able to ever come back to such a world, even if he could, there would be NO LIVING THING to talk to, or to reward or punish. We are producing things that will stay deadly as long as the universe exists.
There has never been, nor will there ever be, any way to make nuclear safe or viable as an energy source that will outpace the lifespan or energy needed to babysit the stuff for longer than God will exist.
Give me all that "krappy" solar & wind stuff any day compared to nuclear in any form.
If you read the Weissbach et al article (or even just the wikipedia page of that graph) you'd notice they explictily talk about this asymmetry.
ReplyDeleteIt all depends whether you consider the 3/4 of century old diffusion enrichment method, or you use numbers from centrifugal one.