25 July 2007

The Strawman Massacre

Prof. Jesse Ausubel of Rockefeller University apparently felt the need to lash out at renewable energy sources in a editorial titled "Renewable and nuclear heresies" in the journal International Journal of Nuclear Governance, Economy and Ecology (hat tip to FuturePundit). This is such an egregious piece of work that I felt a simple comment would not provide a sufficient debunking.

I cannot view the article myself as the journal is evidently not old/relevant enough for the University of Alberta to subscribe to but a press release and an older presentation (.pdf) provide some fairly outrageous strawman arguments.

First the abstract:
Renewables are not green. To reach the scale at which they would contribute importantly to meeting global energy demand, renewable sources of energy, such as wind, water and biomass, cause serious environmental harm. Measuring renewables in watts per square metre that each source could produce smashes these environmental idols. Nuclear energy is green. However, in order to grow, the nuclear industry must extend out of its niche in baseload electric power generation, form alliances with the methane industry to introduce more hydrogen into energy markets, and start making hydrogen itself. Technologies succeed when economies of scale form part of their conditions of evolution. Like computers, to grow larger, the energy system must now shrink in size and cost. Considered in watts per square metre, nuclear has astronomical advantages over its competitors.
Ah yes, hydrogen, the system whereby energy demands increase by a factor of three (.pdf) As the man says, "to grow larger, the energy system must now shrink in size and cost."

First strawman, hydroelectric power:
Hypothetically flooding the entire province of Ontario, Canada, about 900,000 square km, with its entire 680,000 billion liters of rainfall, and storing it behind a 60 meter dam would only generate 80% of the total power output of Canada's 25 nuclear power stations, he explains. Put another way, each square kilometer of dammed land would provide the electricity for just 12 Canadians.
Hmmm, but Canada already produces 58 % of its electricity from hydro and only 15 % from nuclear (source). Given that Ontario is about 10 % of the total land mass of Canada, apparently 40 % of Canada is one big lake totally devoted to hydroelectric generation. Of course, in reality, we don't build dams with only a 60 m drop. Rather one uses the natural terrain to one's advantage to channel the rainfall into a deep reservoir where it can then fall several hundred meters. If he stuck to complaining about habitat destruction, his argument would be much stronger.

Next strawman, photovoltaics:
photovoltaic solar cell plant would require painting black about than 150 square kilometers plus land for storage and retrieval to equal a 1000 MWe nuclear plant. Moreover, every form of renewable energy involves vast infrastructure, such as concrete, steel, and access roads. "As a Green, one of my credos is 'no new structures' but renewables all involve ten times or more stuff per kilowatt as natural gas or nuclear," Ausubel says.
Let's see, solar insolation is about 1000 W/m2 and has a capacity factor of about 0.2. Solar cells are, on average, about 12.5 % efficient for polycrystalline Silicon. That means we get about 25 W/m2 of continuous power for PV. If we assume a 0.8 capacity factor for the reactor, I arrive at 32 km2. Prof. Ausubel's numbers seem somewhat inflated. I've previously estimated that we would need 3.5 - 7.0 m2 of panels per person to supply all of our energy needs via solar. Much of this can be done on existing (i.e. that's not 'new') urban environments.

And the wind strawman:
"Turning to wind Ausubel points out that while wind farms are between three to ten times more compact than a biomass farm, a 770 square kilometer area is needed to produce as much energy as one 1000 Megawatt electric (MWe) nuclear plant. To meet 2005 US electricity demand and assuming round-the-clock wind at the right speed, an area the size of Texas, approximately 780,000 square kilometers, would need to be covered with structures to extract, store, and transport the energy."
And again, we have the same issue as with the hydroelectric strawman. The area a wind farm occupies can be used for other purposes. These numbers are not really fact-checkable. The actual area occupied by the turbine base is nowhere near 770 km2, which is what we should compare by this silly 'power per square meter of infrastructure' metric that the entire area is somehow tainted. Frankly, this seems to be something from the Not In My BackYard (NIMBY) school of environmental thought rather than 'how can we improve our health and restore sustainability to the biosphere' school of environmentalism. I.e. get that wind turbine out of my million dollar view.

He also knocks down a biomass strawman, although in this case, I happen to agree with the conclusion. Probably the worst assertion however is that renewables cannot benefit from economies of scale:
"Nuclear energy is green," he claims, "Considered in Watts per square meter, nuclear has astronomical advantages over its competitors." On this basis, he argues that technologies succeed when economies of scale form part of their evolution. No economies of scale benefit renewables. More renewable kilowatts require more land in a constant or even worsening ratio, because land good for wind, hydropower, biomass, or solar power may get used first.
The argument can be made that there's only so much good hydropower and to a lesser extent wind, but this really does not apply in the case of solar. However the assertion that, "no economies of scale benefit renewables," is quite untrue. Renewables are, by in large, made up of small incremental additions to capacity. This allows them to be produced in an assembly-line fashion. While more modern nuclear designs are often 'modular', they are still essentially one-off builds. As a result, the learning rates for solar and wind are much higher (about 20 % per doubling in capacity) compared to about 6 % for nuclear. I refer to McDonald and Schrattenholzer, "Learning rates for energy technologies", Energy Policy 29(4): 255-261. Nuclear is not going to fall in price as safety controls become more and more sophisticated folks. As I've explained previously, the big ticket item nature of nuclear puts it at a huge disadvantage compared to its smaller and nimbler competition, in terms of design cycle, capital investment, etc.

In conclusion, Prof. Ausubel seems to have developed his very own metric for greenness (square meters of stuff devoted to the production of energy), applied it in a haphazard fashion, and is now drawing some questionable conclusions from it. While I favour nuclear over coal, I really can't see nuclear competing economically with either fossil fuels or renewables. Moreover, if he feels that we already cover too much of the Earth's surface with stuff, shouldn't he preach on the demand side rather than suggesting we build more supply?

7 comments:

Anonymous said...

A quick once over on your Solar Numbers:

Average watts per square meter of 24 hours over the course of the day world average = 164 w/m2.

8 hour summer day, 40 degree latitude = 600 w/m2

8 hour winter day, 40 degree latitude = 300 w/m2

Just a bit- keep up the work!

Josh

Brian Mays said...

It is interesting to see that there appears to be a "conservation of strawmen," because while you claim to be knocking some down, you seem to be setting up some fine new ones of your own.

The various numbers used for your questionable solar calculation have already been discussed. In addition, you said:

"Nuclear is not going to fall in price as safety controls become more and more sophisticated folks."

Sorry, I have to strongly disagree with you here, and I suggest that you please do some homework before you say such silly things about nuclear energy or modern nuclear plant designs. If you had done some research, you would have found that the designs being offered for sale today are dramatically simplified from the current generation of plants operating in the US, which were designed 40 years ago. That is, they have less pipes, less pumps, less valves, and less other components, including for the safety systems. Some of the designs use passive safety features, so that they require no controls for some parts of their safety systems.

New nuclear reactors have not fallen in price only because they are bigger and produce more power. Sure, they cost more, but they are more economical. That is, you get more Megawatts for the buck. That is a good thing.

Let's face it. In terms of economics, bigger is better. That is why the designs of wind turbines get larger and larger with time. Nobody is talking about building the smaller wind turbines that were available 20 years ago. No, for wind to be able to compete in the energy market -- and it is not able to compete with coal today, but it is getting better -- many large turbines are needed.

Finally, you too hastily discount the ability for nuclear technology to improve with time. A reactor might be built for a design life of 60 years, but components get replaced during its operating life, and they can be replaced with better technology as it becomes available. Thus, it is not uncommon for reactors to get uprated to produce more power. Furthermore, experience running the plant leads to better procedures and higher efficiencies, learning from doing.

Your idle speculation about the advantages of "smaller and nimbler" renewable technologies bears no resemblance to reality. Instead, this is what experience has shown: in the time period between when the last plant in the US came online eleven years ago and the beginning of this year, the nuclear industry has added over four times the amount of electricity production per year as the entire amount of wind generation in the US today, without adding a single new reactor. In fact, several reactors were retired in this period of time.

Although the latest figures that I can find for wind generation come from 2004, my estimate here includes the large amounts of capacity that were added in 2005 and 2006. Using the 2004 numbers would be a little unfair of me, since the nuclear industry has added over seven times the amount of electricity that was produced by wind in 2004. Those extra wind turbines (and the land that they occupy) help.

Robert McLeod said...

brian wrote:
"Nuclear is not going to fall in price as safety controls become more and more sophisticated folks."

Sorry, I have to strongly disagree with you here, and I suggest that you please do some homework before you say such silly things about nuclear energy or modern nuclear plant designs. If you had done some research, you would have found that the designs being offered for sale today are dramatically simplified from the current generation of plants operating in the US, which were designed 40 years ago. That is, they have less pipes, less pumps, less valves, and less other components, including for the safety systems. Some of the designs use passive safety features, so that they require no controls for some parts of their safety systems.


You claim that I put up a strawman to 'knock down' which isn't really the case. I cited a referred publication that shows the learning rate of nuclear at 5.8 % as opposed to around 20 % for both wind and solar. Furthermore, both wind and solar have much further to grow before they saturate the market. Since industry financial speculation isn't likely to be a good indicator of future performance, I'll stick with the historical trends, thank you.

On the design side, I am primarily familiar with the CANDU designs as opposed to the pressurized light water reactors of Westinghouse et al. I do have a passing familiarity with the various safety systems of the CANDU, and I understand that the gravity emergency cooling, water cooling, etc. may be applied to some of the new US reactors.

Generally speaking whenever you violate the KISS principal it costs money. Your attempt to twist a single statement I made to that effect does not lend credence to your argument. The new designs you speak of are all 3 1/2 generation plants. They represent an incremental evolution in reactor design rather than a revolutionary one, and I do not see them offering an ahistorical modification to said ~ 6 % learning rate.

Finally, you too hastily discount the ability for nuclear technology to improve with time. A reactor might be built for a design life of 60 years, but components get replaced during its operating life, and they can be replaced with better technology as it becomes available. Thus, it is not uncommon for reactors to get uprated to produce more power.

I would think that would be difficult to accomplish without rearranging the fuel bundles to increase the neutron flux or adding more bundles. I suppose the turbines are often replaced with more efficient units.

Perhaps you intend to speak of reliability, i.e. capacity factor?

Your idle speculation about the advantages of "smaller and nimbler" renewable technologies bears no resemblance to reality. Instead, this is what experience has shown: in the time period between when the last plant in the US came online eleven years ago and the beginning of this year, the nuclear industry has added over four times the amount of electricity production per year as the entire amount of wind generation in the US today, without adding a single new reactor. In fact, several reactors were retired in this period of time.

Please provide a citation. You make it impossible to fact check your claims. That is a very 'sock puppet' like action, to make unverifiable claims.

If you are perhaps referring to the growth of capacity factor in nuclear plants, I'm not sure you really want to go there. I'm fully aware that many of the PWR reactors were only able to meet 60 % of their nameplate capacity for many years. Capacity factor can only increase so much before you run into unity.

I would certainly question why the nuclear industry needs its lobbyists to go inserting billions loan guarantees into power bills in the dead of night:

http://www.nytimes.com/2007/07/31/washington/31nuclear.html

Either the nuclear industry is mature or it isn't. If it is a mature industry, it shouldn't need said subsidies. Wind clearly doesn't need subsidies to grow in North America, at the least. Solar still does, but offers the lowest ultimate cost. I've already estimated the total cost to drop the cost of PV down to $0.05/kWh at US$90-154 billion over 16-18 years:

http://entropyproduction.blogspot.com/2007/05/glittering-future-of-solar-power.html

Brian Mays said...

Oh my. Sorry, you are so wrong on so many points. Where to begin?

First off, let me state that what I am saying here is not an attack on renewable energy or anything like that. I support subsidies for "renewable" generation, both for R&D and for actually building and running plants/farms. My only purpose here is to correct what you have misrepresented. Since you have felt the need to lash out at Ausubel's article, I have decided to give what you have written the same treatment. Please note that I am not defending a single word that Ausubel has said. I'm just putting your words through the same grinder. Turnabout is only fair play.

So, let's get started playing, shall we?

You are entitled to your opinions about learning rates. I really don't know what you are talking about. The paper that you cite seems to be rather vague -- or biased to say the least. Let's compare.

Nuclear energy has been around for about half a century, right? The theoretical idea of getting energy from solar and wind have been around for much longer than that. We can cite practical examples of gathering (low amounts) of energy from these sources that go back centuries to millennia. What? Strawman? OK, let's return to modern times and modern technology.

It might surprise you to learn this, but the solar cell is over half a century old, and it has been a phenomenal success at providing power for satellites, which was one of its first indented applications. The first 1 MW wind turbine was built even earlier, in 1941. It was not much of a success, however.

So exactly what have solar and wind learned in the last 50 years? How not to generate power? Even today, they contribute a trivial amount to the US's electricity production.

Compare that with nuclear, which provides 20% of the electricity for the US, and provides close to 80% of the electricity for France. Can you show me any country in the world that gets 80% of their electricity from these quick learners?

So much for learning rates.

Germany gets more energy from wind than any country in the world, and there is reason to suspect that it has reached the saturation point, because in spite of all of its wind generation, Germany plans to build new coal plants.

OK, on to reactor designs. Again, I'm sorry, but you made a stupid statement and I'm calling you on it in the spirit of the "massacre" that you seem to enjoy inflicting on Ausubel's article. Thus, I exposed your statement as obviously ill-informed and flat out wrong.

Here is a hint about new reactor designs, which I hope that you'll keep for future reference: The term "3 1/2 generation" or "3+ generation" is pure crap. It was invented by the marketing people at the various companies who sell reactors, as an edge against their competition. The true situation is that all currently proposed plants are third-generation plants (i.e., the current generation plants are second-generation plants, and this is the next generation that will be built). Some of these designs are evolutionary, as you mention, but not all.

The primary contenders in the US right now are:

ABWR - either an evolutionary design or a revolutionary design, depending on whether you consider this design to be an evolution from previous designs and the ESBWR to be a change from this or whether you consider the ABWR to be the radical change and the ESBWR to have evolved from this design. Perhaps the main contender for a "half-generation" design.

ESBWR - see above.

AP1000 - a revolutionary design that incorporates many passive safety features.

EPR - an evolutionary design -- i.e., improvements to current technology -- developed in Europe.

APWR - a design from Mitsubishi, which is an evolutionary development of currently operating reactor technology.

In any case, all of these designs feature simplified safety systems with less components. Your statement is simply flat-out false, which was my point.

Reactors get uprated for several reasons. It usually has nothing to do with what is going on in the core. Sometimes they are uprated because the knowledge that comes with operating the reactors means the the very large margins that imposed for safety can be reduced. These margins are very large because, historically, little was known about how the components would perform. With decades of experience comes new knowledge, and hence, less reason for ridiculously large margins. They also can be uprated because of replacements of components, for example, the steam generators. In the end, the result is the same: you have the same plant producing more electrical power.

My claims about generation come from the EIA. Since you cite them, I had assumed that you would be able to look up the numbers for yourself.

Nevertheless, here goes:

I am considering the time period between 1996 (when Watt's Bar 1 came on line) until 2006. In this time, no new reactors were added to the US's nuclear capacity (although a couple were retired in this time). Sure enough, the nameplate capacity reflects this -- as one would expect with power uprates canceling out the closures of old plants. Thus, the capacity in 2006, 100.1 million kW, is not very much different than the capacity in 1996, 100.8 million kW.

Nevertheless, the amount of energy produced by nuclear per year has increased dramatically. It has risen from 674.7 billion kilowatt-hours in 1996 to 787.2 billion kilowatt-hours in 2006. That's an increase of 112.5 billion kilowatt-hours, without adding a single reactor in a decade.

All of this comes from Table 9.2 of the EIA Annual Energy Review 2006.

Okay, so that covers nuclear, what about wind? Well, the EIA says that wind generated 17,810,549 thousand kilowatt-hours in 2005, with a nameplate capacity of 8,706 MW. Now, if we take the word of the American Wind Energy Association that wind's capacity has increased to 11,603 MW in 2006 (which is impressive -- it is equivalent to the addition of about two or three nuclear reactors), then assuming that the same amount of energy would be generated by each unit of capacity (i.e., assuming similar capacity factors as the previous year), I get that wind should be expected to generate 23,737,200 thousand kilowatt-hours in 2006.

Now the numbers work them out for themselves. This value (total generation from wind) is less than 1/4 of the 112.5 billion kilowatt-hours that were added to nuclear power's generation since the last reactor was built.

That's where my numbers come from. Do you have anything to compare?

Finally -- and my goodness, it takes a lot of effort to correct your mistakes, whether intentional or not -- your claim that "wind clearly doesn't need subsidies to grow in North America" is contradicted by the American Wind Energy Association, who wisely continue to argue (see the bar graph on the first page) that federal subsidies are essential for the growth of wind in the US (or do you not count the federal production tax credit as a subsidy?!).

I support wind's production tax credit, since I think that it gives this technology a chance that it otherwise would not have in a free market. I'm just honest and knowledgeable about it.

Robert McLeod said...

Alright, I call uncle and retract my statement that, "Nuclear is not going to fall in price as safety controls become more and more sophisticated folks."

I still think you are taking the wrong lesson from the growth in capacity factor of nuclear plants in North America. Plants are producing more energy by running more, not producing more peak power. While light-water reactors were only able to run about 60 % of the time in the 1980s they've been able to streamline refueling procedures to the point that they can run 90 % of the time now. I wish I could find the white papers that discuss this but I cannot. Annals of Atomic Power and the Journal of Nuclear Energy appear to be silent on this issue. I was able to find a comment to that effect from a nuclear power proponent in a debate sponsored by the Bulletin of Atomic Scientists:

http://www.thebulletin.org/roundtable/nuclear-power-climate-change/

See the essay, "Nuclear is more reliable, safer than before." by R. Stephen Berry, 4th paragraph

Now, certainly, 50 % growth over twenty years is fairly impressive. But it also highlights the fact that there is no further room for significant growth. One cannot run the power plant 100 % of the time, nor increase the thermal output of the reactor vessel itself, so those gains are essentially consolidated now. It would be duplicitous to suggest that nuclear can continue to add energy production in this fashion.

On the issue of economies of scale, I don't see your point. Nuclear reactor designs get bigger because the marginal costs of adding more power aren't that large. You don't need to pay any more operators, for example. Wind turbines are getting bigger because the wind blows faster and longer with increasing altitude.

On the issue of subsidies, I truly feel that the marginal cost of wind is low enough to compete on the marketplace, as is. E.g.:

http://dx.doi.org/10.1016/j.enpol.2004.06.007

That said, I have yet to see an industry turn down an opportunity to feed at the public trough even when it may be in their best interests to do so. That applies to wind, nuclear, coal, as well as big oil. All of these technologies are subsidized (in the USA anyway) by the federal government in one way or another. Wind is still growing at a great rate (~ 30 % per annum) against amortized thermal plants. Wind does need regulatory assistance, to ensure existing electricity providers don't use red tape on ancillary services as a bludgeon against their emerging competitors.

Krassen Dimitrov said...

the notion of 3.5-7 sq.m. of PV cells per person that points to an older post is a real head-scratcher... I posted a comment to the older post...

Unknown said...

Brian, I don't know about the veracity of the rest of your post as it's not in my area of expertise.

However, the fact that the solar cell is over 50 years old is not a barrier to its rapid cost reduction. Experience curve methodology states that the cost of production declines by a constant learning factor for every doubling of production. Initially, the production volumes were tiny (according to the main industry sources, the EPIA, IEA and PV News) and the costs astronomical. But data - particularly work done by Paul Maycock of PV News - shows that until April 2004 it was following an experience curve with learning factor 18-20%.

In April 2004 the Germans put a whacking great subsidy (EUR 0.45/kWh) on PV and the silicon shortage prevented price decreases, but that's another story. It will return to trend when silicon manufacturing capacity catches up.

Robert - your work on the experience curve looks perfectly sound to me, although I'm not sure governments will subsidise 33% growth for the next 8 years. I've been professionally fiddling with what the experience curve looks like if you remove the effect of the silicon shortage, which will have an effect in the opposite direction.