Friday, March 19, 2010

Archdruids and Net Energy

This week's Archdruid Essay is an improvement on the ones I discussed last week, but although he's retreated a little onto higher and more defensible ground, I still think his position has some poorly defended salients that he should abandon.  In particular, I think he's still too hung up on energy concentration, and is still trying to argue that there's something physically difficult or impossible about transitioning to a renewable powered civilization.  In my view, there's nothing physically impossible about that (though I'm perfectly willing to concede that the total cultural inertia of western civilization is enormous and that worries me a lot).  The Archdruid makes two main arguments in his post, and in this post I'm going to take on the first of them (hopefully coming back to the other at some point in the future when time and interest allows).

That first argument is that solar power has been tried in the marketplace before, and didn't succeed, and therefore this is evidence that civilization as we know it cannot be powered by renewables:
Yet it’s at least as instructive to pay attention to what hasn’t worked. The approach central to today’s large-scale solar plants – mirrors focusing sunlight onto tubes full of fluid, which boils into vapor and runs an engine, which in turn powers a generator – was among the very first things tried by the 19th century pioneers of solar energy. As discussed in last week’s post, these engines work after a fashion; that is, you can get a very, very modest amount of electricity out of sunlight that way with a great deal of complicated and expensive equipment. That’s why, while solar water heaters spread across rooftops on three continents in the early 20th century, solar heat engines went nowhere; the return on investment – measured in money or energy – simply didn’t justify the expenditure.
He doesn't mention it, but of course it's a-fortiori true that wind-power had been used for many centuries before the nineteenth century and also largely disappeared from the economy at about the same time that the solar experiments he refers too were being tried and failing.

However, this doesn't prove the proposition "twentieth century renewables can't power civilization". It is strong evidence for the proposition that "nineteenth century renewables couldn't compete with nineteenth century fossil fuels".  But those two propositions are very far from being the same thing.  Let's take this apart.

Let us, as a first approximation, neglect the various forms of government intervention in the marketplace, and just consider two energy sources competing as commercial enterprises.  The key issue in business success is profit - the difference between revenues and expenses.  After paying for labor, materials, energy, and all the other inputs the business requires, the revenues from selling the product or service had better cover the costs and then some, in order to reward the investors and entrepreneurs that make the business happen.  Generally speaking, in most normal markets, competition will drive profits of successful firms down to somewhere in the range of 5-10% of revenues.

The general character that this gives to business beyond the startup stage will be familiar to many readers with commercial experience - small differences matter a lot.  Something that causes only a few percent difference in the revenues or expenses of a business causes a few tens of percent difference in the profits.  Needless to say, a business that is consistently significantly more profitable as another one in the same general line of work will be vastly more successful.  For example, if one business has twice as large profits as another, by providing twice as large a return on investment to investors, it will attract much more investment capital and be able to scale up much faster, etc.  This is why businesses expend a huge amount of effort trying to understand and control how they apply their costs to generate their revenue - they live and die by the efficiency of that process.

So, as a very simple hypothetical, let's consider the difference between a company, Renewable Co.,  introducing a renewable with an energy return on energy invested (EROEI) of 10 (for 1 unit of input energy, 10 units of energy are output), and another company, Fossil Inc,  introducing a fossil fuel energy with a EROEI of 100.  Let's suppose their labor and other costs are identical, their energy outputs are equally useful, and let's suppose after the initial phase, the market adjusts the price of energy such that the fossil fuel company is making a 10% net profit.

In this situation, we can calculate the net profit of the renewable company.  We know that the input energy cost of Fossil Inc must be 1% of their revenues (because they have EROEI of 100), and since Renewable Co has a EROEI of 10, it must require ten times as much input energy, so 10% of it's revenue will go on input energy.  So the costs of Renewable Co exceed those of Fossil Inc by 9% of revenues.  Therefore, the net profit of Renewable Co. will be 1%, while the net profit of Fossil Co. will be 10%.

Clearly, investors are going to massively prefer Fossil Inc. Renewable Co is barely profitable at all, and doesn't stand a chance.

However, this does not prove that, had Fossil Co. not existed, Renewable Co could not have succeeded.  All else being equal, if energy prices were just 10% higher, it could have made the same level of net profit that Fossil Co. ended up making and had the basis of a perfectly successful business.  It's hard to see how 10% higher energy costs would destroy civilization.

Of course, I concede this is a very simplified example, but I think it illustrates something important about the situation - renewables in the nineteenth century weren't an experiment in isolation, they were an experiment in competition with fossil fuels.  And of course, it's true to this day that fossil fuel EROEI's are as good or better than renewables, and renewables are just barely (wind) or not yet (solar) cost-competitive with fossil fuels, and hence have required government incentives to get them deployed.

Of course, those fossil fuel costs would probably look a bit different if the coal and oil companies were required to take out insurance against the risk of places like California and Spain turning into deserts.


Bill Pulliam said...

I think you might need to consider the possibility that diffuse solar energy does have thermodynamic limitations to it on beyond its low density. If a thin cloud of photons is a more entropic system than a dense one (which seems likely; that is true for gasses and other mass-energy systems), then you really can't extract as much work from it. The work used in concentrating it will of necessity more than offset the extra yield. This distinction between concentration and quality might not be so absolute as you suggest; they might in fact be strongly linked, and Greer might have actually stumbled closer to the mark than you recognize.

Stuart Staniford said...


I hope to come back to the point about concentration and quality in a post shortly - just didn't have time to cover everything this morning.

DC said...

I think you missed JMG's point of exergy. He's referring to an age of diffuse energy because our activities over the last 200 hundred years has brought us to this point. Solar PV is just an example of the latest "wishful thinking" that will fuel the next all out desecration of our natural world. He is stating that we wasted far too much of the only extremely dense energy resource we got to make a complete transition to a diffuse energy culture with the same "growth" mindset in place. There are physical limits to everything we do, because we are material heterotrophic beings that occupy space and time. I am sure you have seen this:

I enjoy your blog, though...

Stuart Staniford said...


I know it's terribly old fashioned of me, but I think if folks want to argue that something is physically impossible, they ought to back up their argument with, you know, numbers and logic, stuff like that. They ought to identify with some precision which particular physical principle would need to be violated, and then explain exactly how and when it would have to be violated for civilization to continue. All I'm getting in response is vague circular reasoning.

For example, if you feel that thermodynamics prevents civilization continuing, could you explain which law of thermodynamics would need to be violated? And when and why exactly?

Manolo said...

Stuart, your comment " the total cultural inertia of western civilization is enormous " must be the understatement of the century. :-), but I am with you on the theoretical side: "...there's nothing physically impossible..."
We humans need to be back to the wall, completely, desperately, cornered, to kick our brains into creative panic mode. Interesting times ahead.

Bill Pulliam said...

I don't think you'll find that Greer argues that civilization will not continue; far from it. He does argue that it will transform, as does anyone with half a brain looking at our current position in the historical arc. He also argues that some major features of it will not continue, such as global industrialization, rampant ubiquitous high tech and telecom, centralized imperial-style governments, and energy-intensive agriculture. I don't believe he has ever claimed that massive large-scale solar energy generation is impossible; I believe he argues that it will prove expensive enough that it will turn out to be impractical, therefore will likely not be the foundation of a new industrial era. Those sorts of prognostications are really more judgement calls than matters of hard numbers, clear thermodynamic limits, and absolute facts. We know people will use solar energy somehow (we always have, back to drying strips of mammoth meat), and we know we won't be using dragon power and magic beans. In between that, well, it's all educated guesswork.

"Exergy" does have a specific meaning, and he did get sloppy with the term. Interestingly, the more I read, the more I discover that calculating the "exergy" (useful energy) of solar radiation at the surface of the earth is rather challenging, and different approaches have different results... Stuart's 94% and my 25% seem to be the extremes on both ends, by the way.

Stuart Staniford said...


John says things like:

Industrial civilization is a complicated thing, and its decline and fall bids fair to be more complicated still, but both rest on the refreshingly simple foundations of physical law.


One of the common ways to avoid thinking about our predicament, as I mentioned last week, is to cite the quantity of energy that arrives on Earth by way of sunlight every day, and note that it’s vastly greater than the quantity of energy our civilization uses in a year. That’s true enough, but it misses the point, which is that the energy in that sunlight has very modest amounts of exergy by the time it crosses 93 million miles of space to get to us, and it can therefore do only modest amounts of work. Strictly speaking, we don’t face an energy crisis as fossil fuels run short; what we face is an exergy crisis – a serious shortage of energy in highly concentrated forms. That’s a problem, because nearly every detail of daily life in a modern industrial society depends on using highly concentrated energy sources.

The 94% is just the straight second law efficiency limit based on the temperature of sunlight. Practical limits are lower as I indicated. If you don't like my calculation, take it from folks who actually do this stuff full time:

Existing terrestrial photovoltaic modules rarely achieve solar power conversion efficiencies in excess of 15%, yet the absolute thermodynamic efficiency limit for solar power conversion is 93.6%. At present, no known practical approach can reach this high figure, but device models exist for solar cell devices that can theoretically attain an efficiency of 86.8%. Indeed, many commercial solar thermal (hot water) systems operate at efficiencies around 70%. However, there are inherent limitations with the common photovoltaic device designs that limit the thermodynamic efficiency to no more than 31%. There are three broad approaches, illustrated below, that can overcome the limitations of the present devices.

Source: Quantum Photovoltaics group at Imperial College

Bill Pulliam said...

We've already established that JMG was misusing the term "exergy," and I believe he has conceded that point. Replace it with "the practical costs of concentrating diffuse sunlight and converting it into forms that can be stored and transported." There it becomes an economic and EROI argument, and an argument that these are frequently overstated for solar. Harping on the fact that someone misused a term or misapplied a concept, after they have already conceded this and their argument is still cogent when it has been restated without the misunderstanding, is not good rhetorical form. Fundamentally, you and he just disagree on how easily these concentration and storage issues will work out. He thinks they will be expensive enough to make solar an ineffective large-scale substitute for fossil fuels, you disagree. History will tell. I have to say, I do agree with him that the boosters of technologies tend to vastly overstate their economic value early on; nuclear that would be "too cheap to meter" and the EROI of corn-based ethanol are two good examples of this. The fact that the published stuff about economics and EROI of PVs generally don't include address storage question (without which PV loses a large part of its usefulness) is another example.

Really, what it comes down to is that he is approaching this as a historian and you are approaching it as an engineer, and both of you are deeply skeptical of the other's approach.

As for the entropy of sunlight -- I have come across that several places, but I have not seen why it is correct. When hydrogen atoms expand into a vacuum do they lose temperature and gain entropy? If so, why don't photons? I did find a series of papers that looked at the exergy (their term) of sunlight at the bottom of the atmosphere, not the top, and concluded that it has dropped to a range of 50-77% (because of dispersion and polarization, primarily, I believe), depending on atmospheric conditions. So here we have it again: even if 93.6% is correct for the top of the atmosphere, it overstates the case by as much as a factor of 2 for sunlight at the places we can actually use it. So when you say someplace averages 300W/m2 of sunlight over the year, you might need to point out that this is only 150 W/m2 of useful energy. It might also apply to published efficiencies of PVs -- if they are tested in the lab against simulated full-spectrum sunlight versus actual "wild" sunlight, you might be overstating their efficiency again by as much as a factor of 2. How are these tests generally done?

I still need to see a simple explanation of why the entropy of an expanding cloud of photons remains constant. FYI -- incoming diluted solar radiation at the earth does not actually appear to have a definable temperature per se, as it is not in thermodynamic equilibrium (its spectral distribution and its energy concentration are out of synch). From what I read, blackbody radiation only has a defined temperature when it is contained within a blackbody cavity and its spectral characteristics reach equilibrium.

Oh, and another point -- heating water is not work, entropy/exergy/2nd Law issues do not apply. That is just heat flow, not mechanical energy. So saying that a solar hot water system operates at 70% efficiency has no bearing on the utility of sunlight for generating electricity or mechanical work. Once again, it's a misrepresentation/overstatement of the case for solar.

Stuart Staniford said...


For future reference, if you want to persuade me of something, such as that the net energy calculations of renewables in the literature are all wrong, you need to either link to, or directly present, at least a rough quantification of the main points of your argument - what factors did the authors miss, and why are they large? Vague unquantified assertions that the literature is all wrong are completely unpersuasive to me.

Bill Pulliam said...

As you know perfectly well, there is a vast literature on EROI calculations and much disagreement over the methods and values. That's like asking me for a reference on the statement "The Health Care plan currently in congress is highly controversial." Besides, that was a very small part of my last comment, which was mostly continuing the entropy-of-sunlight discussion.

Interesting (and maybe revealing?) that you assume I am trying to persuade you of something or promote a viewpoint or agenda. I'm actually trying to achieve understanding. The large number of question marks in my writing might suggest this. Take the scientific literature at face value? Not hardly, no competent scientist does this, not until you understand what it presents and why (I was an academic researcher for many years, have taught upperclass undergraduate science classes, and have a Ph.D. as well, though I see little reason to promote this fact on my profiles). Some of the derivations used in these entropy/exergy/etc. calculations do not yet make sense to me; since this is not my primary field I am probably just missing something but until I can see what it is and understand it I'm not going to just believe that "10 papers say it therefore it is correct." For instance, the one reference you quoted and linked to about the usable energy in sunlight, in the same paragraph contains a substantial scientific misapprehension (that using solar energy to create heat is equivalent to using it to create work). But I must not question the earlier statement in that same quote that the entropy of sunlight is equal to the entropy of 5900K blackbody radiation at equilibrium in a blackbody cavity, or at least ask why this is so?

The studies finding a 50-77% exergy/energy ratio for sunlight at the surface of the earth under the atmosphere are by Stephan Kabelac, published in various places. One abstract is here:

His work the only quantitative treatment of the subject I have yet found, but I don't have ready access to a proper scientific library. I believe he also starts with the 94% number above the atmosphere, but still finds a significantly lower number at the surface.

Stuart Staniford said...


The reason that thinking solar radiation is basically black body radiation at the temperature of the surface of the sun is due to the fact that the energy of photons is set entirely by their wavelength, so the energy content (ie effective temperature) of sunlight is just associated with the frequency spectrum of the light (since frequency and wavelength are deterministically related), and that doesn't change at all in passing through space, and not too much in passing through the atmosphere. To a first approximation, the main effect of the atmosphere is just to absorb a fraction of the light. If one wanted to be more precise, one could take account of the fact that the atmosphere scatters blue light more than red, so lowering the effective temperature a little, especially at low angles (the reason the sun gets orangish at sunset and distant mountains look bluish). This effect is probably not a very big deal when the sun is overhead at low latitudes but could be significant at high latitudes.

I can't really tell what the abstract you link to is doing, not having the paper.

Also, I do agree with you that hot water and electricity are not the same end product, and the conversion efficiencies of the two processes are not directly comparable. I think it's a lot more likely to be an inadvertent error than a deliberate misrepresentation on the part of the Imperial College website, however.

Stuart Staniford said...

Looking around a little more, I found this really nice figure at the Wikipedia which shows the actual spectra at top of atmosphere and sea level, as well as the ideal black body case. It looks to me like it gets chopped up a bit by the various absorption bands, but the effective temperature is not massively changed.

Bill Pulliam said...

Found the derivation of why the expansion of blackbody radiation away from the source is a reversible (i.e. entropy neutral) process. Noticed that this derivation violates the uncertainty principle, expansion is not reversible, and entropy of radiation must increase with expansion. Don't believe I am equipped at present to calculate how much, but seems necessary that it must, even if it interacts with nothing else on the way. Details on my blog at:

Stuart Staniford said...

Bill - the laws of thermodynamics are part of classical physics, and can be applied only to bulk properties of materials. At the quantum scale, their status is much more complicated, and things can happen which would violate energy conservation thinking classically (eg the famous quantum tunneling), although such violations become less and less likely the more violation they involve, or the longer the go on for.

So the spherical mirror thing you found is a reasonable proof of reversibility in the classical context.