Tilting at Windmills, Spain’s disastrous attempt to replace fossil fuels with Solar PV, Part 2

[ Pedro Prieto presented an update on June 20, 2017 at the International Society for Biophysical Economics “Spain’s Solar Revolution Revisited: six years later” here: Pedro-Prieto_ISBPE_2017-Spains-solar-revolution-revisited.pdf.   See the pdf for more information

Losses in book overestimated and revised

Before / After / Diff / Factor

  • 0.08 / 0.12 / +0.04 / Energy spent on wafers, cells, modules: the conventional EROI (40)
  • 7.70 / 5.00 / +2.7 / Energy spent on Operation and maintenance (43)
  • 0.20 / 0.10 / +0.1 / Energy spent on washing and cleaning (44)
  • 0.50 / 0.30 / +0.2 / Energy spent on self-consumption (45)
  • 2.40 / 0.60 / +1.8 / Energy spent on security and surveillance (46)
  • 0.50 / 0.30 / +0.2 / Energy spent on fairs, exhibitions, promotions, conferences, etc. (50)
  • 0.10 / 0.33 / +0.23 / Energy spent on communications, remote control, and management (57)

Losses in book UNDER estimated and revised

Before / After / Factor

  • 2.80 / 3.00 / -0.2 / Energy spent on premature phaseout of unamortized manufacturing equipment (48)
  • 0.70 / 0.80 / -0.1 / energy spent on administration expenses (51)
  • 0.30 / 0.40 / -0.1 / Energy spent on municipality, autonomous, state taxes, levies and duties on production, etc. (52)
  • 0.40 / 0.60 / -0.2 / Energy spent on services of indirect labor (direct excluded) (54)
  • 0.10 / 0.40 / -0.3 / Energy spent on contracted generation deviations +/- 5% (55)
  • 0.00 / 0.10 / -0.1 / Energy lost on force majeure, acts of god, windstorms, lighting, flooding, hail storms
  • 3.90 / 4.70 / -0.8 / Associated energy costs to injection of intermittent loads: network stabilization associated costs (69)

The original EROI result was 2.45.  Given the adjusted positive factors, this goes up to 7.72, and then adding in the adjusted negative factors reduces the EROI to 5.92.

But Prieto makes clear that an EROI of 5.92 is still far too high because the energy for the following factors aren’t included (especially labor and energy storage).  It’s clear to me that if they were, the EROI of solar PV would be negative, but Prieto let’s the reader come to their own conclusions:

  • Energy injected to compensate for the intermittency of solar from pumped hydro storage, compressed air energy storage, battery storage, or hydrogen storage (70-73)
  • Energy invested in direct labor of employees (63)
  • Financial direct costs (64)
  • energy spent on pre-inscription, inscriptions, registration bonds and fees (58)

Other factors were probably underestimated, but not adjusted: it is likely that solar panels don’t last for 25 years (slide 25), dust loss (27), angular losses (28), non-fulfillment of power (29), losses from shadowing (31), losses from voltage and frequency sags and swells (36), over-dimensioning (37).


In Charles Hall’s latest 2017 book, “Energy Return on Investment: A Unifying Principle for Biology, Economics, and Sustainability“, he says that he has embarked on a project to discover why solar advocate EROI results are so much higher than what was found in Prieto & Hall’s “Spains Solar Revolution”.

Since fossil fuels are finite, the electric grid must be 100% renewable someday. The goal of EROI studies is to see which renewables are the most worth investing in long-term.  Someday they will be without any help from oil, coal, and natural gas, so it is necessary to have wide boundaries that includes all of the other essential infrastructure that makes solar and wind possible, especially  energy storage, the transmission system, and other renewables that can provide both millisecond balancing power and 6 to 12 weeks of energy storage, depending on size of grid and amount of renewable power in a region.

And electricity will have to power transportation, manufacturing — everything.  Electricity is just a fifth of the energy we use now (under the category Energy I explain why hydrogen, biofuels, methane hydrates and other energy resources can’t substitute for oil).

The biggest difference Hall has found so far is due to solar advocates multiplying solar electricity generation by a factor of 2.6 (BP) or 3 (IEA) because advocates claim that solar power is worth 3 times as much as fossil electricity since two-thirds of fossil generation is lost as heat.

According to Gail Tverberg, in her post “The Wind and Solar Will Save Us Delusion“, this is done by BP to account for the loss of energy when fossil fuels or biomass are burned and transformed into electricity. BP corrects for this by showing the amount of fuel that would need to be burned to produce this amount of electricity, assuming a conversion efficiency of 38%. Thus, the energy amounts shown by BP for nuclear, hydro, wind and solar don’t represent the amount of heat that they could make, if used to heat apartments or to cook food. Instead, they reflect an amount 2.6 times as much (=1/38%), which is the amount of fossil fuels that would need to be burned in order to produce this electricity.

But wait! Fossil fuels used for heat are several times more effective than heat generated with electricity, and burning natural gas at home in a 98% efficient furnace is far cheaper than burning it at a natural gas power plant and losing two-thirds of it to create electricity.

EROI simply must include all of the energy inputs required to build a solar plant as in Prieto & Hall’s study at a minimum, so we can see if it is worth subsidizing solar (or wind) in the first place.   If solar and wind can’t replace fossil fuels, because they depend on them too much, then the money/energy would be better spent building passive solar homes that would last for hundreds of years, preparing to go back to muscle power, expanding organic agriculture and departments at high schools and universities, and so on.

We have very limited time and energy left to cope with the energy crisis.  Energy transitions take at least 50 years. The 2005 DOE report by Hirsch said that you’d want to prepare at least 20 years ahead for peak oil, with time the most limiting factor, and here it is 12 years after conventional oil peaked with an energy cliff rather than a bell curve looming.

If anything, solar and wind electricity should probably be reduced considerably, because they have a very low quality energy since they can’t be counted on. Here are some reasons to reduce their EROI (see “When Trucks Stop Running” for citations and more details):

  1. All wind and solar do is add more wood to the fire. They do nothing to get rid of fossil fuels because they can’t be counted on. In the 2012 IEA world energy outlook, it was calculated that 450 GW installed capacity of wind in 2035 would only produce 112 GW of power since it’s not always blowing. But that’s no good for the grid, when it needs power, it needs power RIGHT NOW. The IEA calculated wind could be counted on only 5% of the time (the capacity credit) for just 22.5 GW at peak demand times. That means an additional 89.5 GW (112–22.5 GW) of reliable fossil, nuclear, or biomass power is needed to back up wind power. So the more you replace conventional power plants with wind, the more you depend on wind. And, the more you depend on the wind, the less you can depend on it. Great Britain’s office of science and technology estimated wind could be counted on reliably only 7–9 % of the time if the overall penetration of wind power ever reached 50%. So if 25 GW of wind capacity were built to replace 25 GW of fossil and nuclear plants that have double the lifespan of wind turbines, and the capacity credit of wind at peak demand was 5 GW, then an additional 20 GW of fossil and nuclear plants would be needed for backup, with nearly double the energy generation as before (45 GW). In regions where peak demand occurs in the winter, the capacity credit of solar power is zero, because peak demand occurs after dark. And thus the stark reality: “Investment in renewable generation capacity will largely be in addition to, rather than a replacement for power stations” (GBHL 2007). Worse yet, backup fossil and nuclear power plants must be online ready to step in immediately if wind or solar power falter, burning fuel meanwhile.
  2. So you could build energy storage to store excess wind and solar generation. Since on average the wind is blowing 33% of the time, you’d want to build 3 times more wind turbines to keep the energy storage (batteries, pumped hydro, compressed air) charged up.
  3. You’d also need to have a hugely expanded national grid, since most of the wind is in the midwest, most of the solar is in the southwest, and most of the hydropower is along the west coast.

Instead of tripling the energy credited to electricity, since wind and solar aren’t always available, the EROEI to build, maintain, and operate energy storage facilities, backup fossil and nuclear power plants, and an expanded national grid ought to be subtracted from the energy generated by intermittent renewables.  Energy storage, a national grid, and fossil fuel electricity generation plants are not separate entities which can be ignored in EROI studies, because solar and wind and the electric grid itself can’t exist without them.

Alice Friedemann   www.energyskeptic.com  author of “When Trucks Stop Running: Energy and the Future of Transportation”, 2015, Springer and “Crunch! Whole Grain Artisan Chips and Crackers”. Podcasts: Practical Prepping, KunstlerCast 253, KunstlerCast278, Peak Prosperity , XX2 report ]

Part 2: Critiques and rebuttals of Spain’s Photovoltaic Revolution. The Energy Return on Investment”, by Pedro Prieto and Charles A.S. Hall

Part 1 is an introduction and overview, followed by a book review of “Spain’s PV revolution” 

Below are 5 rebuttals of criticism of Prieto & Hall’s book:

  1. 2017.    Hall, Charles A.S. Energy Return on Investment: A Unifying Principle for Biology, Economics, and Sustainability. Springer.
  2. 2016-5-26. The real EROI of photovoltaic systems: Professor Hall weighs in. Ugo Bardi’s blog: Cassandra’s Legacy.
  3. 2015-4-1: Stanford Net Energy conference
  4. 2015-4-11 Pedro Prieto responds to criticism (private communication)
  5. 2015-4-11 Ted Trainer responds to criticism of Prieto & Hall

This book was only available electronically at the University of California, not in print form, and not available to the public. It’s a shame libraries are putting more and more journals and books into electronic versions only.  Especially this book.  Microchips, motherboards, and computers will be among the first casualties of declining fossil fuels, because they have the most complex supply chains with many single points of failure, dependence on rare metals, and so on (see Peak Resources and the Preservation of Knowledge for details). Nor is it guaranteed the electric system can be 100% renewable, as it must be some day, or that transportation, especially trucks, can be electrified (see my 2015 Springer book ““When Trucks Stop Running: Energy and the Future of Transportation”.

I encourage you to get your (university) library to buy a hard copy of this book, so that future scientists, historians, and the public, who only have access to hard-copy books even though our taxes pay for the University, will understand why our society didn’t replace fossil fuels with “renewables” even though we knew oil couldn’t last forever.

On an energy forum in March 2014, Prieto said: “Since we wrote the book, I have been able to experience a few more incidental factors: mice delightfully gnawing the cables and covers and optical fiber communication color cables, and storks excreting on modules with about 6 inches size -one cell- per excretion. Real life has many factors that they are not accounted in organized studies in labs, universities with particular technologies and plants in perfect irradiation places.”

Alice Friedemann   www.energyskeptic.com  author of “When Trucks Stop Running: Energy and the Future of Transportation”, 2015, Springer and “Crunch! Whole Grain Artisan Chips and Crackers”. Podcasts: Practical Prepping, KunstlerCast 253, KunstlerCast278, Peak Prosperity , XX2 report ]

Hall, Charles A.S. 2017. Energy Return on Investment: A Unifying Principle for Biology, Economics, and Sustainability. Springer.

[ Mostly verbatim, sometimes cut or paraphrased. See the book for cited references, tables, and graphs]

In the first decades of the 21st century a number of studies gave EROIs of 6–10:1, though rather than EROI studies more often used the energy pay back times of just one or two years for photovoltaic (PV) systems (e.g. Fthenakis et al. 2011; Raugei et al. 2012). These numbers were used by solar advocates to argue for the importance and economic viability of solar PV systems, and in some cases that solar PV systems were comparable to fossil fueled systems.

But in 2013 Prieto and Hall came out with a much lower estimate of EROI of 2.45:1 for sunny Spain with sophisticated engineers, which caused a great stir amongst solar advocates and initially greeted with disbelief by many in the industry.

This book provides the most comprehensive assessment of all of the energy costs of solar PV.  It differs from many earlier analyses in that

  1. it attempts to include (nearly) ALL energy costs actually used, not just the costs of the modules and some related hardware
  2. it uses measured rather than estimated energy output.
  3. It uses actual data from Spain which has a much higher insolation than Switzerland, Germany or the Netherlands.
  4. Of particular importance is that Prieto and Hall attempted to calculate the complete energy used to support the PV system by “following the money”, i.e. by attempting to assess all the money flows necessary for the system to operate (understood by Prieto because of his extensive on-site experience as Project Director, Project Designer, Consultant and Director of Development of a solar PV company).

They assigned an energy cost to each monetary cost using specific energy intensities: the mean energy use for the Spanish economy (7.16 MJ per Euro), and twice that for manufactured or engineering items, and one third that for business services as given in the protocol paper by Murphy et al. (2011). They derived about the same energy cost when they took all money spent times the national mean (7.16 MJ/Euro, similar to the global mean) as they found when they did a very detailed analysis of 24 categories of items, including such things as energy costs of roads and cleaning, surveillance, business services, meetings attended by engineers as well as modules.

This is consistent with the view of Herendeen and Bullard (1975) that when one purchases a complex product from final demand all the different energy intensities tend to “come out in the wash”. Raugei and Leccisi (2016), for example, did not calculate any energy cost for which they could not get a direct energy measurement for in their assessment of PV and fossil fuel derived energy for England. To me this seems to miss some costs.

Since then similar results were published by Palmer (2013) for rooftop PVs with battery back up in Australia, and Weissbach et al. (2013), for Germany (see also Raugei 2013; Weissbach et al. 2014; Raugei et al. 2015). In 2016 Ferroni and Hopkirk published an estimate of a negative EROI for cloudy Switzerland and Germany.

Yet Leccisi et al. (2016) and Raugei and Leccisi (2016) came back with estimates of values of 9:1 or higher. How could two different groups of competent investigators get such different estimates?

Why do some solar PV studies have a low EROI and others a high EROI?

1. The largest difference in EROI between these investigators is based on corrections for quality between fossil fuels and electricity.

Raugei et al. (2012) were very critical of comparing the apples of fossil fuels (where EROIs at the source were generally higher) with the oranges of higher quality electricity. They said that a number of summaries (e.g. the “widely cited ‘balloon graphs’ (Hall et al. 2008; Murphy and Hall 2010) and bar charts (Hall and Day 2009) have compared many technologies simply in ‘heat equivalents’, i.e. the energy values are given in terms of their abilities to heat water with no correction for energy quality”. The fundamental issue is that since we are willing in society to trade about 3 heat units of coal, oil or gas to generate one heat unit of electricity, the EROIs of the electricity derived from PVs or wind turbines (or nuclear power plants) should be weighted by a value of some three times that of a heat unit of fossil fuels.

2) Theoretical versus actual electricity output

EROI values in many studies are too high because they used “nameplate” values (1,800 kWh/M2-year) for assessing electricity outputs from PV facilities rather than the actual output. [My comment: this is because private solar facilities usually won’t give researchers this information.  But Prieto & Hall had 3 years of government data from all the facilities in Spain].  Nameplate is inaccurate since the actual electricity output is reduced by clouds, bird droppings, overheating, dust accumulation, lightning, equipment failures, and degradations over time to less than “Nameplate” value.  Also, too much output can fry electrical components at various locations in the grid.

Prieto and Hall found that the actual output for a facility in Spain with a nominal output of 1,800 kWh/m2-yr was measured at an actual 1,375 kWh/m2-year.  Ferroni and Hopkirk (2016) also found measured values considerably less than nameplate values.

3) Solar facilities probably don’t last for 25 to 30 years

A related issue is the assumptions about how long the facility will last.  Most investigators have applied a life span of 25 years for PV facilities, and the IEA guidelines suggest 30 years.  Since most solar facilities are new, this is hard to measure, but in reality it may be less.  Ferroni and Hopkirk (2016) came up with an estimate of a mean of 18 years for Switzerland.  Prieto (personal communication) believes it is much less than 25 years in Spain because many companies have declared bankruptcy and thus do not honor their warranties.  Without warranties or the specific parts to fix failures, many PV facilities in Spain have been abandoned or completely reconfigured.

4) Boundaries and Comprehensiveness of the cost assessments

Carbajalis Dale et al. (2015) state in a footnote in reference to Prieto and Hall’s study that they are inconsistent in their definition of system boundary and arbitrary in the inclusion of a large number of non-energy inputs. In their online Commodities and Future trading where they claim that “Renewables have a higher EROI than fossil fuels” they state “Prieto and Hall add every incidental energy cost they can think of, like the energy costs of building fences around the solar farm, and so on. They even add energy costs for things like corporate management, security, taxes, fairs, exhibitions, notary public fees, accountants, and so on (monetary costs are converted into energy by means of a formula)”.

We respond: “As if these were not legitimate energy costs to build and operate PV plants?” In fact they are. For example, fences and security are necessary, given the high value of things like scrap copper, so plants are very susceptible to thieves stealing electrical components (a cost, incidentally, Prieto and Hall included).  Without fences and security, the EROI goes to zero.

Nor can facilities exist without roads, module washing, and financial institutions.

Based on earlier studies (e.g. Hannon 1981), it is clear that all services and goods require substantial amounts of energy, about a third to half per dollar compared to societal means to undertake. In order to make a comprehensive assessment we “followed the money” and assigned a very conservative one-third the national mean energy cost to all service expenditures, which Prieto knew because as chief site engineer he signed for every penny and activity at a large gigawatt plant in Spain. The services we mentioned are not incidental but necessary and should be included in energy costs, and we’ve yet to hear a good reason why we should have excluded any of them.

Prieto and Hall found that the construction of modules and basic electronic components such as inverters were only about a third of the total energy cost of building and operating a solar facility in Spain.  We assume this is true elsewhere, yet our critics have yet to do a study that includes much of the real energy costs that we did.

Raugei et al. (2013) have argued that we included costs such as site preparation and environmental issues that are not included in our assessments of oil or coal.  This is not true since all such costs are included in our indirect energy assessments which are based on total “upstream” expenditures by industries. We agree with them that the boundaries should include all energy costs that any energy gathering activity experiences.

5) Technological changes over time

Another issue raised by Raugei and other solar analysts is that the monetary and energy cost of making solar PV modules has been declining for decades and will continue to do so, although perhaps at a declining rate. They criticize the Prieto and Hall study for using technology appropriate in 2008 (actually we used 2009-2011 technology) when there has been a 10 to 20% decline in the energy cost to make modules since then (some of which, in terms of money if not energy, is attributable to such things as subsidies by the Chinese government). As far as we know, there has not been a similar decline in the other inputs to PV systems.  I agree with them that one should do costs and benefits for particular years.

The EROI of Storage for Solar Energy

Swenson (2016) argues that at large scale, solar PV technologies will become more efficient.  But that adds to the energy and monetary costs to build storage and integration into the grid, lowering EROI.

Since sunshine and wind are dependent on nature’s only partially predictable whims, and can’t be programmed in advance, meeting the demand load can be very difficult.  A day might be sunny or cloudy (with half or less of the insolation), and wind blows on average only about 30% of the time (closer to 20% in Germany), and there can be periods of two weeks or more with no wind at all.  Although PV systems are slightly more predictable, storage is required to compensate for these intermittencies. Yet even if we used all of the batteries in the worldwide, they’d store less than one minute of global electrical output, nor is this cost effective on a massive scale.

And not just storage, but some other kind of readily dispatchable power. Right now the only storage option feasible at large scale is elevated storage of water in existing facilities or specially constructed pumped storage, but:

  • there’s an electricity of loss of 25-35% in the pump up and later release systems
  • the availability of such sites are limited
  • the intermittent release of water harms fish and aquatic ecosystems (Ward and Stanford 1979)

Carbajalis Dale et al. (2014) estimate that adding a relatively small amount of storage to PV systems would quickly put them into energy deficit.

Palmer (2013) found that batteries doubled the energy cost of rooftop solar systems.

These energy costs tend to be ignored by PV and wind advocates, who also argue that coal and nuclear facilities have their own problems with responding to variable loads (which however are being met readily now).

Future EROI assessments should include the large energy costs of storage, which will only grow larger as more intermittent renewables are added to the grid.

Exponential Growth of energy production

Many say we must grow these systems very rapidly and indefinitely.  But Neumeyer and Goldston (2016) found that an initial EROI of 10:1 quickly dropped to 2:1 as most of the power output went to generating new plants.  Carbajalis Dale and Benson (2013) and Kaufmann and Shiers (2008) found a similar very sharp drop in net power output if growth were large.

Thus a large exponentially growing PV system will have to be constructed using fossil fuels.

But there may not be enough materials to exponentially grow PV facilities.  Fizaine and Court (2015) and Gupta and Hall (2012) found that an exponentially growing PV system might run out of copper in a very few decades.  Hertwich et al. (2014) found that PV systems can use 11 to 40 times more copper than conventional fossil generation systems, though they thought there was enough copper to build a large renewable system.

Business Services and Taxes

All investigators agree that direct (on site) and obvious indirect energy costs should be included (Murphy et al. 2011).

But what about the energy to support business services used?  These require energy to build brick and mortar buildings, which need to be heated, cooled, and lighted with electricity and fossil fuels.  What about the energy to support the taxes paid?  Taxes, when spent, require energy also, such as the energy to build and maintain roads, provide schooling, and so on.  Oil and gas fields as well as PV facilities require considerable construction and maintenance costs that are paid for by governments which in turn operate from taxes.

For example, Pennsylvania has found there are very high costs associated with the new “fracked” gas wells due to the heavy trucks full of water driven over low quality roads during all seasons of the year, creating damage that has to be fixed with heavy equipment.  In addition the driller’s children require schooling, and there’s also an increase in the need for policing and health services (Dutzik et al 2012; Food & water watch 2013).  Therefore, tax expenditures and the energy required to generate these governmental services should be included in the energy cost of a project.


Perhaps most controversial is whether to include the energy required to support labor. There are various kinds of energy costs that might be included, such as the 1.8 MJ/hour a hard working person requires to do heavy labor.  People are basically machines that operate at about 20% efficiency, requiring food with 9 MJ/hour.  This is trivial compared to the machines most laborers use, such as a diesel engine burning 135 MJ per hour.

Labor is not available without pay, so the energy to support the worker’s paycheck might be included. Assume a worker is paid $50,000 a year.  Energy must be spent within the economy to produce the goods and services demanded by the worker or his/her family spending the paycheck.  In 2015 the U.S. economy used roughly 5.6 MJ per average dollar of GDP.  Thus, assuming that our worker’s family spends their money on “average” goods and services, it would take about 280,000 MJs of energy, equal to 46 barrels of oil, to support their paycheck.

When first presented in around 1970 as a potential factor in EROI, economists said this was inappropriate since this was consumption which shouldn’t be added to production.   But I think the energy to support worker’s paychecks is a legitimate part of the cost of production, but it is so controversial that we do not include it.


2016-05-26 The real EROI of photovoltaic systems: Professor Hall weighs in. Ugo Bardi’s blog: Cassandra’s Legacy.

May 8, 2016: a recent paper in Energy Policy on the EROI of photovoltaic solar systems came up with a NEGATIVE EROI of 0.85:1 (Ferroni and Hopkirk 2016).

They found that “at today’s state of development, PV technology cannot offer an energy source but a NET ENERGY LOSS, since its ERoEIEXT is not only very far from the minimum value of 5 for sustainability suggested by Murhpy and Hall (2011), but is less than 1 [0.85].

Prieto (private communication) notes that “Ferroni/Hopkirk calculate the failure rates of PV modules as per the public statistics of PV Cycle (a recycling European entity), that do not count modules that are simply abandoned.  The data contradicts the IEA PVPS program, calculating 30 years of life span for PV systems: the failure rates give a figure much closer to 18 years.”  

Prieto also says that a criticism of their paper is that Spain is a lousy country, but Germany is the “flagship of perfect workmanship”. Yet after 10 years, the reality is that Spain is producing twice as much energy as Germany in per MWp installed basis, not only because Spain is better irradiated (about 50%), but also because of the more efficient and better maintained utility scale installations in Spain versus the scattered rooftop individual home installation in Germany.” He also notes that the Ferroni (2016) paper “is dramatic and is raising some blisters.”

Ferroni, F., Hopkirk, R. J. 2016. Energy Return on Energy Invested(ERoEI)for photovoltaic solar systems in regions of moderate insolation. Energy Policy 94:336–344


2015-4-1: Stanford Net Energy conference 

Notably, the founder of EROI, Charles A. S. Hall, wasn’t invited.

At this Stanford University conference the goal was to start a new net energy think-tank that would standardize net energy by having a specific way researchers ought to conduct their studies, with the most up-to-date life cycle and other data, boundaries and assumptions, and so on.  If researchers strayed from this format or added additional material, they’d need to say why. The lack of standardization is one of the many reasons policy makers don’t take EROI studies seriously.

In my opinion, this makes it easy for proponents of various renewable solutions to calculate the EROI to be much higher than it actually is.  Without standards, it is easy to increase EROI by not counting the energy to make steel because the researcher claims it was 100% recycled, or cherry-picking the best performing wind or solar farms over the best performing time period, and so on.  Policy makers can’t be expected to make policy decisions or recommendations when EROI studies of wind ranges from 4 to 115.

Meta-studies can’t be done either because there is too much missing data, and/or unstated assumptions, and/or different models used, and rarely is real data available, since private companies don’t have to, and often don’t want to reveal their true performance, operation, and maintenance costs or they’d get less investment and lower stock prices.

Yet even at the conference, several EROI presentations were not clear about their boundaries.  Long after the artificial photosynthesis presentation which would combine hydrogen with CO2 to make liquid fuels (with a spectularly low EROI of only 1.66), I found that the outside boundary was set at 300 feet outside the factory gate and didn’t include storage or delivery to the customer.  Probably not calculated because the EROI would be less than 1, an energy sink.

By the end of the conference I was a bit frustrated at the lack of discussion of boundaries, because this has been a problem for 40 years and is the main problem to be solved to get policy leaders to pay attention, and more importantly fund such studies, since the researchers often have to pay for these studies out of their own pocket.

So at the end of the conference, with this issue rarely referred to the entire time, I asked the panel what they thought should be done about the boundary issue. For example, ethanol studies using narrow boundaries found higher EROI values than those with the widest boundaries, which often found a negative EROI.  I recommended Spain’s solar PV revolution by Prieto and Hall which used real production data over several years rather than theoretical data used in 99.9% of other studies as a good way to decide what to include or not include, since it made sense to make the boundaries wide, not narrow. Also, since nearly every presentation was on renewables that generated electricity, perhaps new standards for a fossil-free world should include how much electricity it would take to transport the 8,000 pieces of a wind turbines supply chain, the electricity to mine iron and make steel, cement, fiberglass, copper, and electric trucks and electric grid and batteries or a catenary system for trucks to deliver goods and the final wind turbine to its site.

I had the strong impression this was not a welcome question. No one leaped to answer, and finally one of the panelists said that the boundaries ought to be wide but that this question was best talked about over a glass of wine.

After this session one of the speakers, Marco Raugei, at Oxford Brookes University, came over.  He was very upset by my question because he thought Prieto and Hall’s book was awful. He told me it was so bad that several scientists had tried to prevent Springer from printing it.

I told Raugei that I had looked very hard for any criticism of the book but had not been able to find any rebuttals, so what exactly was wrong with it?  Raugei replied that the book wasn’t peer-reviewed. Well hello, books aren’t peer-reviewed, surely he knew this…and I pointed out that Farrell in 2006 had used non peer-reviewed papers in his famous ethanol EROI study. So I asked why someone didn’t write an analysis to refute the book? Raugei replied that since it wasn’t peer-reviewed, why bother.

When I asked Raugei to tell me more about what was wrong, he said that it was inconsistent in so many ways, not defensible the way economic inputs were converted from money to energy such as the insurance figures, some air travel expenses, too haphazard, inconsistent in method and goal, not clear enough in stating that this is just one snapshot moment in time in Spain and that it used an ill-advised subsidy scheme, that the EROI is not the same in other countries and parts of the world, and that the goals should have been more explicitly explained. I thought: What goals? Did he think Prieto and Hall had a goal of a low EROI figure?

Prieto has strong motivation to find a high EROI, since he built some of the solar plants he writes about in the book. He could make more money by exaggerating solar PV EROI.  Hall certainly has no dog in this fight.  In general, the scientists who are funded by industry produce the most problematic research.  For example National Corn Growers Association funded scientists found the highest EROI results for ethanol in their non-peer-reviewed papers.  Recently it was discovered that several Harvard scientists were paid by the sugar industry to blame fat, not sugar, for obesity.

It was ironic that Steven Chu was the opening keynote speaker at this net energy conference, since Tad Patzek once wrote me that “Steven Chu decided not to fund my Laboratory Directed Research and Development (at Lawrence Berkeley Laboratory) project whose goal it would have been to arrive at a consistent thermodynamic description of all major energy capture schemes bio and fossil, so that we compare apples with apples. What I did not appreciate is that no one wants to know that they may be working on a senseless project, such as industrial hydrogen from algae. I despair seeing the rapid corruption and sovietization of American science (without the Soviet strengths in basic sciences), but can do little about it. … It is not easy to get funded on the subjects I have proposed.  …In fact, my LDRD proposal to develop the comprehensive thermodynamic language to talk about the different energy resources was just not funded…”

Someday when a future history of science author attempts to write about the history of EROI, I hope that Patzek, Hall, and others have written memoirs that discuss how hard it was to get funding, get published (did scientists really try to prevent Spain’s solar revolution from being published?!), the criticism they received, and so on, because I think it will be of great interest to the grandchildren and further generations down the line.  Understanding why renewables have such low EROI might prevent cargo-cult like behavior to spend huge amounts of resources and time to build them after the dark age that may ensue at some point on the downslope of Hubbert’s curve.


2015-4-11 Pedro Prieto responds to criticism (private communication)

(Bold is my emphasis):

Alice, as promised, let’s start answering and commenting on some of your wise comments.

The first thing is to confirm that no EROI studies can be taken seriously if the range of results varies so wildly. So it is quite a sensible approach to try to reconcile the different studies and methodologies.

Having said that, the prevailing methodology is what fails, specifically in the case of Solar PV analyses, but also in others. Experts in solar PV will have more and more available data as time passes from global installations.

Until now, we had seen many studies on different solar PV technologies with different typologies and topologies. Even before our book “ Spain’s Photovoltaic Revolution. The Energy Return on Investment” (Prieto & Hall. Springer 2013) appeared, there were already many variances and divergences.

Even works of Fthenakis or Raugei have contemplated significant variances in the EROI results over time and with different studies of solar plants.

But they all had a methodology in common: they generally used, as you have correctly pointed out, the best material recovery, the best theoretical solar PV system in each case, the best irradiated areas, the assumption that systems will operate in full along the lifetime with no problems. In summary, a methodology that has helped or served as documentary support or reference to many to reach global conclusions on the long term ability of modern renewables to replace, take over or substitute fossil fuels, from a given particular plant analysis extrapolated massively. That was the case, for instance, of Mark Jacobson and Mark Delucci in their studies on how modern renewables could replace fossils and supply the present global consumption. This is a traditional bottom-up approach.

After my experiences of several years in the field with different technologies, typologies, topologies, latitudes and state of development countries and confronting with the real world results, Charles Hall and myself, after having had a pint of beer in an Irish Pub in Cork commenting these issues, in the ASPO International Conference held there in 2007, decided to embark in a study on solar PV. But we tried to do it in a radically different form. It took us several years of back and forth, discussions, checks and double checks, consulting with other experts and so on.

The study, as many of you may already know, was on a real world installed plant in the best irradiated country in Europe (Spain), with the official and very accurate energy production records of the Ministry of Industry (read by telemetry to more than 40,000 digital sealed meters in each of the respective individual plants) over a period of three complete years (2009-2011). That was the main innovation: a top-down analysis and the huge scope of the solar PV plants working in the real world, rather than theoretical academic bottom-up approaches.

With more than 140 GW of installed plants worldwide, and several complete yearly cycles of operation of many of them, it is going to be increasingly difficult for some authors to continue with the academic approach, to verify real behavior of the EROI.

Now, about the energy input boundaries.

Of course, if we focus only on the energy inputs of the solar modules and their composition (glass, aluminum frame, connection box, copper or silver soldering, doping materials, silicon, ingots, wafers, cells, etc.) and perhaps inverters or metallic structures orienting and tilting the arrays, then we may come with spectacular results in a very good irradiated area with the theoretical module yield. This is what has been generally considered in most of the studies carried out to date and what is proposed by some authors as the recommended methodology.

But this is just one of the factors we looked into when we decided to analyze the energy inputs of a complete solar PV system, not just what appears in the marketing pictures of the solar plants.

After many years working in the field, one can appreciate the number of activities that are indispensable (sine qua non conditions), for a solar PV plant to work and operate as some of the authors of several EROI/LCA/EPBT studies consider they are going to work.

We differentiate some 24 factors and additional analysis that was not absolutely complete nor exhaustive, but proven and existing. None of these factors had been considered or hardly appeared in but few of the analyses made by the most renowned solar PV EROI authors. Your study of our book already identifies some of them and I have mentioned them on many occasions.

One of the factors, “a7” (the energy input required for modules, inverters, trackers (if any) and metallic infrastructures, labor excluded), was precisely the EROI as usually calculated by many authors. We decided not to judge the different results of this universe of conclusions but to accept a sensible average of the range of many publications that gave us an EROI in itself for this concept of 8:1; that is, for 25 years of lifespan an Energy Pay Back Time (EPBT) of 3.1 years. Or an energy input cost equivalent to 0.125 of the total generation along the lifespan of the system.

But then we started to consider the rest of the factors (boundaries or extended energy input boundaries) and discovered that conventional EROI studies were ignoring two-thirds of the energy inputs indispensable to get the solar PV plants into operation.

The list calculated the energy inputs, based on the experience of several plants in Spain and extrapolating to the 4 GW installed power studied in the book, to road accesses to the plants, foundations, canalizations, perimeter fences, evacuation lines, rights of way, O&M module washing or cleaning self consumption, security and surveillance, transportation — sometimes as far as from China, premature phase-out or un-amortized manufacturing and other equipment, insurances, fairs exhibitions, promotions or conferences (like the one you had in Stanford –to whom to attribute the involved energy expenses?), administration expenses, municipality taxes, duties and levies, cost of land rent or ownership, circumstantial labor (notary publics, public officers, civil servants, etc.) agent representative or market agent, equipment stealing or vandalism, communications, remote control and plant management, pre-inscription, inscription and registration bonds and fees as required by the authorities, electrical networks and power lines restructuring ass a consequence of the newly injected 4 GW in a national network with about 100 GW, in unexpected and not previously planned nodes of the grid, faulty modules, inverters or trackers, associated costs to the injection of intermittent loads: network stabilization associated costs (only referred to combined cycle gas fired plants, well known costs).

Some of these factors may certainly have diminished with time. Many others, have certainly increased over time. Taxes, for instance, have raised sharply. Stealing in Spain, for instance, is not relevant, but in many countries of the world it is a problem.

We mentioned and developed a little of the associated energy costs of the injection of intermittent loads, by pump up or other massive electric energy storage systems, because we knew it was going to be fundamental and relevant and did not want to open any more the old wounds in an already meager EROI. These costs are still today in a fierce debate in Spain and in many other countries, but they are certainly relevant, should the modern renewables have to replace the present fossil fueled global societal functions.

As you can see, the BOUNDARIES are of essence to determine the real life EROI, rather than an academic EROI. No one critical of our book, could say, to the best of my knowledge, that any of these briefly listed factors was not a real one and was not needed to have (at least in Spain) a solar PV system up and running along its lifetime. But for some strange reason they had never considered them.

Once they recognized the facts of real life, then this battlefield was rapidly abandoned and shifted to the “comparison” with other energy sources, namely the fossil fuel sources. Some authors were claiming that if fossil fuels were treated with these ‘extended’ energy input boundaries and factors, their EROIs should obviously go down in a similar proportion.

What they did, then, was to use a multiplying factor on the order of 3 for solar PV, arguing that it has a logic, when comparing equivalent systems and using an equivalent methodology. I fully disagree and I have shown in several occasions the reason why:

The world uses (mostly burns) about 13 BToe/year of primary energy or more than 510 EJ/year.

Of that, approximately 170 EJ of fossil + nuclear go to produce an equivalent of 40 EJ of clean and useful electricity, this making the point of Raugei valid to some extent, if the solar PV systems would entirely go to replace electricity produced by fossil fuels, because of the losses of about 2/3 of the primary energy in the conversion process.

But the world is not behaving in this way, as scientists like Raugei and Fthenakis must know. New renewables just enter into the energy equation to simply provide more energy to the global system.

Above all, the most important flaw in this assumption is that the world also consumes about 285 EJ in non-electrical uses, like aviation, civil works, mining, transportation, merchant fleets, armies or agriculture (eating fossil fuels, Dale Allen Pfeiffer). And it happens that if we would pretend to use electricity from renewables to replace the fossil fuels used for these global activities, likely through an energy carrier like the eternal hydrogen promise, the pretended multiplication factor used by Carbajales et. al, would immediately operate in the reverse form and become a division factor, probably in the order of 3, with respect to the direct use of fossil fuels of today. That is why we did not employ this “correction factor” used by Carbajales et al.

I will not enter into this debate further, because I find it futile. I do not care if when treating the EROI of coal, oil or gas with these extended boundaries may go down two-thirds from already published studies, now ranging with the old methodologies, for instance, from 100 to 12:1 for oil, depending on the period and places, or 60 to 20:1 or coal or gas in similar levels.

Taking down these two-thirds of present EROI studies will not change the fact that this society is now operating on 80% fossil fuels.  And makes it possible to move it. This is the final proof.

An important part of the rest (excluding perhaps a part of biomass in underdeveloped countries) is also being produced because the energy subsidies given by fossil fuels to the other sources, like nuclear, or hydro, that we could not have dreamed of having if a well endowed fossil fueled society and its related machinery and technology weren’t available. Nuclear, hydro, solar PV, solar thermal or wind energies are absolutely underpinned by a fossil fueled society, not the vice versa. The global society has been making its growing economic, industrial and technological life basically without those energy sources. But we could not imagine these sources working and feeding themselves in all the complex value chain, plus providing an important net energy surplus to the global society. Not now, nor in a foreseeable horizon.

We can not ignore this crucial fact: biomass helped initially to coal to develop, but 60 years from the first massive use of coal, this fossil fuel had already passed biomass in volume and versatility of use and became quite independent of biomass.

This happened circa 1900, at the level of 800 MToe/year of global primary energy consumption and with about 1.6 billion inhabitants.

Then came oil, much more dense and versatile than coal. It took oil again about 60-70 years to pass coal and biomass as the main global energy source. This happened circa 1960, but then, in a consumption level of 3,000 MToe/year and with 3 billion people on Earth.

Now, we move in the level of 13,000 Mtoe/year of global primary energy consumption and with about 7.2 billion people. But gas or nuclear have not passed oil as the prime energy source. And we have to wonder why, if they were discovered and used massively more than 60 years ago.

Quite the contrary, we are moving fast, because of peak oil, back to the possibility of coal surpassing oil again in a decade or so, as the main energy contributor, but this time, probably at a lower global consumption level and probably with a world population still growing in numbers and in poverty.

The first two big energy transitions (biomass to coal and coal to oil) were made with the surpassed energy source still growing and helping to initially boost the coming one, but soon proved to be quite self sufficient to feed a growing and demanding global society, well after paying for their own energy inputs in the exploration, mining or drilling, extraction, transporting, refining and distributions processes WITHOUT ANY DOUBT, because nobody will doubt the evolution of the last century and the role of the fossil fuels on it. Now, we have to face the third big energy transition, in the highest level of energy consumption and population and with the main energy fuel, oil, in depletion.

Of course, one has to accept that in this complex world, all energy sources are somehow interrelated, but, as Orwell said in The Animal Farm, ‘all animals are equal, but some animals are more equal than others’. This is exactly what is happening with the energy sources and its properties and qualities: they can all be measured in EJ or in TWh or whatever, but some are more equal than others. Meaning that there is an obvious ASYMMETRIC interdependence of energy sources, since in the last century, fossil fuels (and oil in a very first place), were responsible for our present global status.

To me, then, there is a non sequitur to shift the EROI battlefield to try to extend the boundaries in the fossil fuel EROI studies, to lower them and favor renewables by comparison, because whatever the EROI and boundaries are considered, it is obvious that the present global society spending 13 BToe/year of primary energy (80% fossils), has been able in the last century (we shall see for how long) to pay their own energy expenses, AND provide a huge net energy surplus at the disposal of 7.2 billion humans who have grown at a spectacular rate for more than a century.

For instance, when the IEA mentions in their WEOs the costs of ‘subsidies’ to different energy sources, it always calculates much bigger subsidies for fossil fuels than for the modern renewables. It is a sort of energy fallacy, from my point of view.

If the global society has resources to subsidize anything, it is because it has previously gotten a surplus of resources from somewhere. And this ‘somewhere’ is obviously a global society that has created them using mainly fossil fuels at discretion. I can ‘subsidize’ my son to go to the cinema, but I cannot ‘subsidize’ myself from the salary I  earn by myself and saved in my left pocket, by changing it to my right pocket.

I understand that some fossil fueled activities may certainly be ‘subsidized’ in certain forms. For instance, kerosene for aviation in the airports, which is tax exempted in many countries, when compared with gasoline. Or ‘subsidized’ coal prices paid to depleted coal basins in Spain to continue producing low quality brown coal, to keep the social peace in the region and avoid the miners revolting. But it is a fallacy to conclude that ‘somebody’ is ‘subsidizing’ fossil fuels globally speaking, when fossil fuels are 80% of our global activities creating surplus. From a strict energy point of view, fossil fuels are subsidizing basically all world activities. Period.

What in reality the OECD watchdog does is a mystifying operation. When digging up the IEA figures of ‘subsidies’ of fossil fuels, one discovers that they are really talking about ‘prices’ or ‘price levels’ of fuel in the producing countries that are selling them domestically at prices lower than those the IEA  would wish they had, to leave more ground to the big OECD importers to buy this fuel from producers at prices OECD can afford.

Coming back to the energy input expenses in extended boundaries, we also left out the financial costs, despite knowing that they were quite large and generally also a sine qua non factor. Most of the plants have been financed in an 80% of the total turnkey projects at about 10 years term, with interest, that ranged from 2% to 5% per year. I firmly believe that finance is a form of using a pre-stored available resource (in a fossil fueled society, coming from fossil fuel related activities) to erect or put in place and operate a given system. In that case, an energy system. So, when one asks for credit or leasing and has to pay back this resource both the principal and the interest to the bank, in let’s say a 10 year term, this is energy evaporating into the system through the bank.

Labor energy input costs were also left aside, even though we had a very good set of data from industry in Spain, classified by categories, skills and full time and part time employees in the sector. The reason was that some of our factors may have had already included part of this labor in, to avoid some limited duplication.

If we had included these financial (even just the additional money created and having to pay back in the form of interests by the requested credits or leasing) and labor energy input cost, the solar PV EROI would have probably plummeted to  <1:1.

In fact, it is very surprising how they criticize the methodology we used to evaluate the financial data (which they did not question basically in numbers), by stating that the conversion of monetary into energy units is not adequate and do not conform to conventional input-output methodologies. Our methodology is clear in these conversion units and reflects a quite direct relation between GDP and total primary energy spent in Spain or between active labor and energy spent per laborer or any given and specific related industrial activity or service rendered. This despite we mentioned that Spain hasn’t published, for years, any input-output tables for the economy (Carpintero, Oscar).

However, it seems remarkable how some are incapable of detecting any anomaly in describing EPBT’s of solar systems recovering the energy spent in them in a question of few months for a life time of 30 years (EROI’s of 40:1 !!) and the astounding divorce with the economic reality, of a world or promoters that look for about 10 years economic recovery, this including heavy premium tariffs (Germany, Spain, Italy, now UK or France) or tax holidays or exemption (US and others) or economic recoveries that last more than the expected life time, if no economic incentives are given.

Without these incentives, the rest of the world is a renewables wasteland. Promoters are virtually not investing (with few exceptions in volume worldwide) in modern renewables, if there are no such incentives. The 140 GW world installed base so certifies, with about 70% of the global installed base made in developed countries with incentive schemes and some 25% made by emerging countries, like China or India (now Brazil or South Africa in a much lesser amounts), also with strong political incentives to cope world markets, leaving a meager 5% for the rest of the world. Doesn’t this crude reality show anything in their conversion of monetary units to energy units methodologies, to the ones giving EPBTs of few months and financial recoveries of many years?

So, I am not surprised, Alice, that some experts, having in their records tens of papers published with high solar PV EROI results, would have shown some annoyance at your question on our book. I would humbly ask from here that when somebody mentions that we work with some methodological ‘inconsistencies’, -a term to which they are so fond of to disqualify other disturbing views- they should rather look into the above explanations and facts of the real world.

I have kept silent until now on what I consider a very regrettable behavior now made public by Raugei, as per your comments. It is true that they dared to write our publisher asking him to stop publishing the book when it was in a draft version in a sort of censorship I had not seen since several centuries in medieval Spain. The recommendation came after somebody took the draft from our publisher without our consent some time before the release and they tried to stop the publication, even threatening that they would discredit it, as they have been doing since it was published, if it were published.  I have never seen such a type of behavior, even less in the academic instances.

The reason they gave first is that we missed our final EROI (2-3:1 being quite conservative and I reaffirm myself more and more as years are passing) by an order of 3. That was precisely the Raugei view on the penalty to be imposed on fossil fuels, if a clean electricity source could replace every kWh of fossil fuel origin, considering that in conventional fossil fuel (or nuclear plants for the case) we need about 3 units of primary energy to get out 1 unit of electric energy. We tried to clarify this in some posts, but unsuccessfully.

Fortunately, the publisher did not consider this a direct threat and the book was finally published.

As for the Raugei comment that the book was ‘awful’ because it had not been ‘peer reviewed’, he qualifies himself. Just look at the acknowledgements of the book. Two professors in Physics from different universities did review the book and produce sensible comments. Charles Hall, the coauthor, is an institution in EROI, that is here questioned with superficial comments. Besides, I understand that publishing a book is a free decision, that does not necessarily require peer revisions, yet despite that, we did have our work reviewed. Perhaps what Raugei wanted to say is that the peer review was not made by the usual reviewers in an inbreeding game.

I have been observing that in the academic world, things are getting unfortunately tougher. Some of the technical papers have sometimes more pages of references than pages of content (see more of my comments on the article below). In the case of solar PV systems, and the references in published papers, it seems there is an excess of ‘selfies’ which were a fashion in the academic papers. And secondly, it appears that credits are gained or given by the number of references that a given person is quoted and this has started a race for a sort of interbreeding cross-quotations, affirming Tadeusz Patzek’s fears about the ‘Sovietization’ of the American science. Perhaps what disturbed Raugei about our book is that we also skipped somehow from these habits and did not leave to the usual teams a review that, with all probability, would have ended up in the garbage.

Of course, Raugei is right when he presumes that our case is perhaps valid for Spain and for the 4 GW installed within the period 2009-2011. Because if we had considered Germany and its public production of solar PV systems within the same period, the Energy Return in terms of MWh per MWp installed would have been less than half of those of Spain.

I am now retired and happily growing my organic farm. Not now or since 2001, when I left working for a telecom corporation, have I had any interest in discrediting or crediting solar PV systems. I am not making my life by publishing papers and trying to gain credibility on a given subject. If anything, I should have defended, as you very well stated, the solar PV systems, because I own 50 kW within a 1 MW plant that I manage and I have helped to design, develop and done some consulting (including what we call here ‘permisología’, the intricate paperwork to get all permits and licenses to the the solar PV plants) of more than 30 MW that are working with different technologies, typologies, and topologies in different latitudes in Spain. I have also cooperated with projects in some Latin American and African countries and I have worked as director of Development of Alternative Energies for a listed Spanish company for a couple of years within the period.

Just a final nota bene, with additional comments on the paper Energy return on investment (EROI) of solar PV: an attempt at reconciliation. Michael Carbajales-Dale, Marco Raugei, Vasilis Fthenakis, Charles Banhart Journal of Latex Class Files. Volume 11 No. 4 December 2012

1) http://www.researchgate.net/profile/Michael_Carbajales-Dale

2) select the thumbnail picture on the left and then download

I can’t get the link provided to work, but maybe it’s my computer settings: http://www.researchgate.net/publication/271699871_Energy_return_on_investment_%28EROI%29_of_solar_PV_an_attempt_at_reconciliation:

The title of this paper, is a supposed attempt to reconcile different views on solar PV EROI, but I have never been informed by the authors of it, even though I have the dubious honor of being cited several times in it.

I did not know that I had formed a so-called “Prieto group in Madrid”, in second place, after Fthenakis group in Brookhaven and before Weissbach group in Berlin or Brandt group in Stanford.

Also surprising is that the document is dated December 2012 and our book was not published until the spring of 2013. Even more surprising, is that the book is criticized several times with the wrong citation:  P. Prieto and C. Hall, “Eroi of Spain’s solar electricity system,” 2012 rather than the correct Prieto & Hall. Spain’s Photovoltaic Revolution. The energy Return on Investment”. Springer, 2013, in the bulky references, that occupy almost as much space as the article in itself. This does not seem to be a very edifying example in referencing others.

Then, the paper comments that “an average energy payback time (EPBT) of 3 years and lifetime of 25 years are used to calculate the EROI subscript PE-eq = 8.33 value for this part of the system. No references are given for any other input data; though it appears that anecdotal worst cases of installations were generalized by the authors”.

Well, a brief look to the a7 factor (page 78) of Energy derived from Conventional Life Cycle Analysis Studies and Calculated as an Inverse Factor of EPBT”, comes out with an EROI of 8:1 for the energy content in modules, inverters, trackers and metallic infrastructure, quotes some works of Fthenakis, Alsema and Kim among others not cited, not to make too boring the EROI publications ranging around 8:1 in their conclusions and with these parameters analyzed (without extended energy input boundaries). Some more could be found in many places. In fact, these levels of EROI for solar PV were quite common in the early years of 21st century. See, for instance, Bankier and Gale in its Energy Payback of Roof Mounted Photovoltaic Cells. Energy Bulletin. June 16. 2006, where they come out with a number of EROI’s ranging from EPBT’s from 1 year (EROI 25:1) to 25 years (EROI = 1:1)

Author Low Estimate (years) Low Estimate Key Assumptions High Estimate (years) High Estimate Key Assumptions
Alsema (2000). 2.5 Roof mounted thin film module 3.1 Roof mounted mc-Si module
Alsema. & Nieuwlaar (2000) 2.6 Thin film module 3.2 mc-Si module
Battisti & Corrado (2005) 1.7 Hybrid photovoltaic / thermal module 3.8 Tilted roof, retrofitted mc-Si module
Jester (2002) 3.2 150W peak power mc-Si module 5.2 55W peak power mc-Si module
Jungbluth, N. (2005) 4 mc-Si module if emissions are not taken into account 25.5 sc-Si module if emissions are taken into account
Kato, Hibino, Komoto, Ihara, Yamamoto & Fujihara (2001) 1.1 100MW/yr a-Si, modules including BOS 2.4 10MW/yr mc-Si module including BOS
Kato, Murata & Sakuta (1997) 4 Sc-Si module. Excludes all processes required for micro-electronics industries. 15.5 sc-Si module. Includes all processes required for micro-electronics industries.
Kato, Murata & Sakuta, (1998) 1.1 a-Si module. Excludes all processes required for micro-electronics industries. 11.8 sc-Si module. Includes all processes required for micro-electronics industries.
Knapp & Jester (2001). 2.2 Production thin film module 12.1 Pre-pilot thin film module
Lewis & Keoleian (1996). 1.4 36.7 kWh/yr frameless a-Si module located in Boulder, CO 13 22.3 kWh/yr a-Si module with frame located in Detroit, MI
Meijer, Huijbregts, Schermer & Reijnders (2003). 3.5 mc-Si module 6.3 Thin-film module
Pearce & Lau (2002). 1.6 a-Si module 2.8 sc-Si module
Peharz & Dimroth (2005). 0.7 FLATCON (Fresnel-lens all-glass tandem-cell concentrator) module – 1900 kWh/(m2 yr) insolation 1.3 FLATCON (Fresnel-lens all-glass tandem-cell concentrator) module – 1000 kWh/(m2 yr) insolation
Raugei, Bargigli & Ulgiati (2005) 1.9 CdTe module including BOS 5.1 mc-Si module including BOS
Schaefer & Hagedorn (1992). 2.6 25 MWp a-Si module 7.25 2.5 MWp sc-Si module
Tripanagnostopoulos, Souliotis, Battisti & Corrado (2005). 1 Glazed Hybrid photovoltaic / thermal 4.1 Unglazed Hybrid photovoltaic / thermal
Alsema E. (2000). Energy Pay-back Time and CO2 Emissions of PV Systems. Progress in Photovoltaics: Research And Applications, 8, 17-25.
Alsema. E. Nieuwlaar, E. (2000). Energy viability of photovoltaic systems. Energy Policy, 28, 999-1010.
Battisti, R. Corrado, A. (2005). Evaluation of technical improvements of photovoltaic systems through life cycle assessment methodology. Energy, 30, 952–967.
CSIRO, Advanced Gasification Research Facility, Queensland Centre for Advanced Technologies, http://www.cat.csiro.au/3_4.htm
Jester, T. (2002). Crystalline Silicon Manufacturing Progress. Progress in Photovoltaics: Research and Applications, 10, 99–106.
Jungbluth, N. (2005). Life Cycle Assessment of Crystalline Photovoltaics in the Swiss ecoinvent Database. Progress in Photovoltaics: Research and Applications, 13, 429–446.
Kato, K. Hibino, T. Komoto, K. Ihara, S. Yamamoto, S. Fujihara, H. (2001). A life-cycle analysis on thin-film CdS/CdTe PV modules. Solar Energy Materials & Solar Cells, 67, 279-287.
Kato, K. Murata, A. Sakuta, K. (1997). An evaluation on the life cycle of photovoltaic energy system considering production energy of off-grade silicon. Solar Energy Materials and Solar Cells, 47, 95-100.
Kato, K. Murata, A. Sakuta, K. (1998). Energy Pay-back Time and Life-cycle CO2 Emission of Residential PV Power System with Silicon PV Module. Progress in Photovoltaics: Research and Applications, 6, 105-115.
Knapp, K. Jester, T. (2001). Empirical Investigation of the Energy Payback Time for Photovoltaic Modules. Solar Energy, 71, 165–172.
Lewis, G. Keoleian, G. (1996). Amorphous Silicon Photovoltaic Modules: A Life Cycle Design Case Study. National Pollution Prevention Center, School of Natural Resources and Environment, University of Michigan.
Meijer, A., Huijbregts, M., Schermer, J. Reijnders, L. (2003). Life-cycle Assessment of Photovoltaic Modules: Comparison of mc-Si, InGaP and InGaP/mc-Si Solar Modules. Progress in Photovoltaics: Research and Applications, 11, 275–287.
Odum, H. (1996). Environmental Accounting: Emergy and Environmental Decision Making. John Wiley & Sons, New York.
Pearce, J., Lau, A. (2002). Net Energy Analysis for Sustainable Energy Production from Silicon Based Solar Cells. Proceedings of Solar 2002 Sunrise on the Reliable Energy Economy June 15-20, 2002, Reno, Nevada
Peharz, G., Dimroth, F. (2005). Energy Payback Time of the High-concentration PV System FLATCON. Progress in Photovoltaics: Research and Applications, 13, 627–634.
Raugei, M. Bargigli, S. Ulgiati, S. (2005). Energy and Life Cycle Assessment of Thin Film CdTe Photovoltaic Modules. Energy and Environment Research Unit, Department of Chemistry, University of Siena, Italy.
Schaefer, H. Hagedorn G. (1992). Hidden Energy and Correlated Environmental Characteristics of P.V. Power Generation. Renewable Energy, 2, 15-166.
Tripanagnostopoulos, Y. Souliotis M. Battisti R. Corrado A. (2005). Energy, Cost and LCA Results of PV and Hybrid PV/T Solar Systems. Progress in Photovoltaics: Research and Applications, 13, 235–250.

As can be seen from the above, we were far from using as an EROI for modules+inverters, plus metallic infrastructure in a sort of anecdotal worst cases of installations generalized by the authors. On the contrary, we were more in the low estimate in years (high estimate EROI), than using worst cases.

Now, for the record, it should also be very convenient for all the prolific authors on solar PV EROI to revise the figures given in papers published several years ago, to double check how are they performing (Energy return statistics). We are very anxious and expectant to learn how it has gone with, for instance, the hybrid PV/Thermal promising analysis, or even better, the results, years after publication, of the Fresnel lenses combined with high efficiency cells in concentration mode.

I recall specifically in this respect  the V.M. Fthenakis and H.C. Kim paper, titled “Life Cycle Assessment of High-Concentration PV Systems”, in which they analyzed the estimated EPBT of the Amonix 7700 PV high concentration system with Fresnel lenses in operation at Phoenix , AZ, and found 0.9 yrs for its EPBT. I wonder if they could still support this analysis, just five years after their study and how the promising system has contributed to the grid parity worldwide, considering they recovered the energy spent on it in less than one year.

Scientific authors should be more careful when accusing to others of using ‘anecdotal worst cases’, especially for the expected Energy Return along a life time, when they are probably using ‘anecdotal best cases’, instead of basing their research on real life 3 years cycle proven and official statistics of production for 4 GW of installed parks.

Talking about the life time (directly involving the Energy Return), it is very interesting to see how some papers have changed the estimated life time of solar PV Systems from 25 years to 30 years. It is curious that virtually all manufacturers give a maximum of 25 years of power guarantee of their modules (with the corresponding degradation process over the years) and 5 years of material guarantee (the latter superseding or prevailing on the former in case of failure) and we find scientists happily granting 30 years for the EROI studies. In my opinion this is a clear attempt to produce higher EROI’s and lower EPBT’s with no rational grounds.

The fact that the Carbajales et al paper ends recommending “that the conventions outlined by the EIA PV Systems Program Task 12 (Environmental, Health and Safety) be followed in conducting EROI calculations, considering that the IEA methodology has easily swallowed the 30 years life time for solar PV modules, gives us a very clear clue of what is going on with these recommendations.

In our discussions on this topic a couple of years ago, an editor came to say that if our factors were really sine qua non (indispensable) for the system to be up and running and the IEA methodology did not considered them, it was time to change the IEA methodology.

I would just recommend the IEA tour Spain (it is not the worst country in solar PV systems; on the contrary, it is one of the most efficient in terms of MWh produced per Mw installed). The IEA should come and check and double check how many solar PV plants have not lasted, for a variety of reasons, the 25 year life time of the manufacturers or the 30 years of the IEA backed by some scientists. Just in 2015 alone about 40 MW have been dismantled, with a lifetime averaging about 5 years. Trials are the delight of reputable and expensive law firms, which earn quite a lot of money preparing lawsuits against promoters, manufacturers, banks and the government. That is real life, far beyond the academic instances. I am following now a demand of a promoter that has decided to buy 2/7th of the modules he originally bought for his 500 kW plant, because the manufacturer (not Chinese), he originally bought from 6 years ago, has disappeared, as have most of the European manufacturers in the last 5 years. One wonders what is the value of a technical guarantee on power, if the life time of the manufacturers becomes much shorter than the one of the power of the promised modules. This is, of course, ‘anecdotal’, although not for the interests of the affected promoters.


After a couple of years from the publication, I have much more data to reaffirm for myself that we were really conservative in our 2.4:1 EROI for many different reasons and factors. But I will not publish more data. I will go back now to my organic garden and wish you all the best for what I suspect may be a grim future.

Antonio Gramsci: “I’m a pessimist because of intelligence, but an optimist because of will.”


2015-4-11 Ted Trainer responds to criticism of Prieto & Hall

Trainer is the author of “Renewable Energy Cannot Sustain a Consumer Society” 

It is very disappointing that so much confusion and acrimony surrounds the crucial issue of boundaries, and that they seem not to be moving to a resolution as quickly as they should be. There are of course big interests at stake, with the conventional high EROI assumption suiting the industry and the theorists who put out such claims. At the very least Prieto and Hall should be commended for getting the whole messy issue of boundaries and components, and appropriate energy cost assumptions for the various components, on the agenda. Sadly the disputation over this issue illustrates the way scientists are not immune from prejudiced and nasty behavior, (a considerable amount of which my efforts to analyze renewables has evoked.) When large scale research funding is at stake there can be strong incentive for competitors to reinforce perspectives that suit them.

As I see it, the goal should not be a single EROI figure for PV, because much depends on the situation and conditions. We need values for modules operating at the average site in Spain with its level of radiation and losses, and we need figures for the various components in the system, such as energy used to produce modules in the factory, energy used to produce the factory, energy lost in inversion, in typical inefficiency due to dust, poor alignment…, and in transmission, energy embodied in inverter replacement, energy used to get workers to the factory, energy used for Operations and Management at the solar farm, energy “retrieved” when the modules are recycled. A fairly thorough provision of these elements would enable anyone to work out the EROI for a particular plant at a particular location, and most importantly the EROI assuming a given set of boundary assumptions. Graham Palmer has just begun a PhD at Melbourne U intended to sort all this out.

I strongly object to Raugei’s comments to you re peer review. I have little respect for the entire peer review edifice, due to my unsatisfactory experience in trying to get critical analyses published. Very often I have found the comments of reviewers to range between nit picky imposition of the way they would have expressed things or gone about the job, through reasoning that I see as at least challengeable and at times dead wrong, to rejection on utterly idiotic grounds … such as being told that my recent 20 page detailed critique of the 2014 IPCC report on renewables was “not scientific”, after waiting seven months for review. (That phrase constituted the full case given for rejection.) On another occasion, where it took over a year to get through the difficulties, I was presented with a seven page essay disagreeing with elements in my case. If that reviewer wanted to express a different view he should have done it somewhere else, not try to insist that I say what he would have said. I have another case where a 50 word review from probably the most prestigious individual in the field said the paper was good, but the paper was rejected because a second even shorter review was unfavourable. The reasons were so unintelligible that I had to ask what they meant. It eventuated that the editor said he didn’t think it was the kind of paper his journal published … after I had waited seven months.

I see the process as far too prone to the whims, prejudices and in fact arrogance of reviewers and editors. They should get out of the way and let people say what they have found or think, and focus only on things like pointing out mistakes or pointing to overlooked evidence or assumptions, or logical errors. Their role should be to help get ideas and analyses out to others, and to block only as a last resort. Too often I have found that reviewers think their role is to make authors conform to their preferred style and they assume the right to condemn work that doesn’t proceed as they would have. I have written reviews in which I say I think the argument is wrong and the procedure not satisfactory but I think the paper should be published, because I could be mistaken and the paper does present a case that it is important for us to think about.

Ultimately what matters is not whether some guru approves of your analysis, what matters is whether the case is sound/convincing/persuasive/well supported, and that judgment should be up to readers, and the quality of the work should be established over time as others in the field comment on it. My main concern here is what must be the large amount of time and good work that doesn’t get published because of the whims of some guru. I would assume that most of us have had papers rejected by one set of reviewers but regarded highly by those from another journal.

So I see any attempt to block publication of controversial, and even flimsy/challengeable cases, on grounds to do with “peer review” as very annoying. I have no interest in whether or not it was peer reviewed; what matters is whether or not the case it argues is sound, or valuable, or ought to be heard. (Theses that are dead wrong can turn out to be valuable contributions, by helping subsequent discussion to clarify an issue.) Whether or not it was peer reviewed has nothing to do with whether or not it is correct, or a valuable contribution, and, Alice, should certainly not be regarded as “a valid criticism”.

In my view Raugei raises some important problems, such as the effect on the Pietro and Hall conclusions had by the Spanish subsidy system, but it’s appropriate to now sort these, not to regard them as reasons why the gook should be rejected.

The most important issue he raises is in claiming that the energy input to PV production should be reduced to one-third, on he grounds that it is electricity and PV produces electricity. As I see it this simply depends on whether the electricity used to produce the modules is coming from PV (or wind or CSP) generating systems … and at present it isn’t. In a world where all electricity came from PV farms it would make sense to put the value of the electricity input into the denominator of an EROI, but in the presenter world the energy going into production is (mostly) coal.




This entry was posted in Charles A. S. Hall, EROEI Energy Returned on Energy Invested, Pedro Prieto, Photovoltaic Solar, Solar EROI and tagged , , , . Bookmark the permalink.

Comments are closed.