Hughes on Degradation

In December 2012 the Renewable Energy Foundation published a study by Dr. Gordon Hughes [backup link] that detailed the degradation of capacity factors in Denmark and the UK over time.  As it happens, I’ve been posting on this issue ever since John Harrison first brought it to my attention over a year ago.  John and I and others have run our own numbers, obtaining results that are fairly consistent with Hughes’ results.  First I’ll summarize his report and then discuss the similarities.

Hughes’ report runs a total of 52 pages, of which the first 22 are the body and the remainder are appendices.  The body’s format is 45 points, each of which provides a paragraph on his various findings.  I’m not (thankfully!) going to go through all 45 as the body is fairly well written and is accessible to a general audience.  In summary he went through the same sort of process as John and I did in a more rigorous manner.  He took the published output of wind projects in the UK and Denmark, normalized them with wind speeds and graphed the declines over time.  His sources of data look to be more complete than what John and I had to work with, so his numbers are likely more accurate.

He calculated the capacity factor declines for 3 different groups of turbines:

  1. UK Onshore: the CF went from 24% to 15% over 10 years and further decreased to 11% over the next 5 years.  This averages 0.9% decline per year.
  2. Denmark Onshore: the CF went from 22% to 18% over 15 years, a decline of 0.26% per year.
  3. Denmark Offshore: the CF went from 39% to 15% over 10 years, a decline of 2.4% per year.

These declines were statistically significant; there is little chance the slopes were the wind industry’s hoped-for zero.

To me the more important of Hughes’ 45 points are:

  • Large wind projects in the UK perform less well (and degrade faster) than small wind projects.
  • More recent projects in the UK perform less well (start at a lower point) than older projects.
  • Given the declines of the projects in the UK, a repowering would seem to be necessary at about the 15-year mark.
  • Given the declines of the projects in the UK, getting to their renewable energy goals will be much more difficult and expensive than current plans.
  • Denmark’s declines of onshore turbines may be due to better maintenance, or smaller projects, or better citing.  Or some combination.
  • Denmark’s large declines of their offshore turbines may be due to a harsher environment or that they are earlier in the development cycle.
  • The UK’s plans for offshore turbines might be in trouble, given the large declines of Denmark’s offshore turbines.

While of course one can quibble with his numbers and conclusions, Hughes’ work appears to be fundamentally sound.  The fact that his work is consistent with what I’ve posted here previously certainly helps.

The issuance of this study was duly reported in the press.  It got enough notice that EWEA felt obliged to respond.  I’ll have to note that EWEA’s response is pretty typical for wind proponents – lots of ad hominems but no numbers of their own and no refutations of his methods.  Science at its finest.

How do his numbers compare with what I’ve posted?  Here’s a listing (all in annual CF percentage declines):

UPDATE – Bach has posted a response to the Hughes study, in which he discusses how his results can differ from Hughes.  It gives you a good sense of the issues all of us encounter when trying to dig out the truth.  In addition, if you want a non-glossy view of Denmark’s electrical system his entire site is worthwhile.


7 thoughts on “Hughes on Degradation”

  1. You would reason, given that we here in North America are a “car culture”, that we would understand the depreciating efficiency character of power generation technology. Whilst the engineering of the newer “permanent” magnet generators of Industrial Wind Turbines is a technological marvel from my perspective, this is not a testament to their efficacy as efficient and long term power generation. The expected inevitable maintenance issues of gear driven turbine generators aside, even the ongoing efficiency of “permanent magnet” generators is in fact not permanent. Over time the flux density increases and eventually results in core saturation, resulting in diminishing output, meaning that they must be replaced. Keep in mind that we are talking about mechanical mechanisms that are composed of for example, 1.5 tons of rare earth minerals (neodymium and dysprosium) for a 2.5 MgW turbine, that will require great expenditure of fossil fuel energy sources to create. As well, the mineral elements in these can be recycled, but again, only with significant expenditure of fossil fuel energy. Interestingly, Vestas, the German owned manufacturer of IWTs, at one time devoted a page on their web site, to advocate the greener nature of their “copper wound” generators over the “polluting” and “energy intensive” aspects of their permanent magnet counterparts. I’ve noticed that as their “bottom line” began to suffer from the wind industry travails, they have largely switched to incorporating permanent magnet generators… whatever happened to corporate virtue?… gone the way of the DoDo I suspect. The absurdity of this is that without significant continued inputs of oil, coal and gas, resulting in alarming amounts of GHG pollution and other environmental degradation, these inefficient and obsolescent Industrial Wind Turbine technologies would not exist or could continue to exist.

  2. I’ve seen claims that the results of this study may have been negatively influenced by a focus only on the oldest turbines. So it would be nice if the analysis also showed if more recent farms also already start to show the same type of ageing in their first years or not (reading the document, it seem that Hughes in fact also includes newer farms in the analysis, but in any case doesn’t make explicit if they are variation or not in the ageing between the turbines generations).

    About point 34, the analysis of differences between Danemark and England shows they might be mostly linked to poorer maintenance of sites in England. If the operators become aware of the actual impact it has, they may try to move closer to the state of things in Danemark and improve their result. So a significant improvement might not be as improbable as the author thinks. A case in point would the realization by Spanish operator of how important the cleaning of blades was, and the development of improved and cheaper methods for it. The maintenance problem idea is quite reinforced by the fact offshore danish farms also have a poor age performance, because maintenance for them is much harder than onshore (of course there’s also tha fact that the structure stress is also higher).

    It would be interesting also to try to link ageing to the manufacturer of the turbine, and as noted just above, to the use of permanent magnet or not. Some manufacturers may well have done a better job of handling all of this.

    The discussion on page 28 appendix D is important too. The author, if I’m not mistaken, claims that the wind turbines themselves are the best wind speed measuring tool we can have, and that adequately deriving the wind conditions from their production results will always be more precise that the monthly index published by the DECC. The claim by itself does not seem outrageous at all.
    However there’s also something of a closed loop in using the same data to estimate both the actual wind speed and the time variation of the efficiency of turbines in capturing the energy of that wind, which means the end result could possibly end up being very sensible to unexpected errors or hidden mistakes in that calculation. Whilst the values of the DECC will be less precise, it seems to me they are also less at risk of being affected by unexpected biases.
    So I would actually do both calculations, and check if they are any surprising divergence between the two results.

    It would great if the REF and Dr Hughes made the raw data available to other researchers, like P-F Bach for example, so that they can dig inside it, and check if their own findings support their initial ones.

  3. Referring to the declines in capacity factor with percentages makes my eyes hurt. Going from 40% to 20% is a 50% decline, not a 20% decline.

    It would be clearer if people wrote about capacity factor in terms of hours. Going from capacity factor of .40 to .20 is obviously a 50% decline – or a decline of .20. But nobody would accidentally read that as a 20% decline and think they still have 80% of their original capacity.

    That said, it is weird that wind plant output seems to degrade linearly in absolute term. In relative terms, that means degradation accelerates fairly constantly with age. Maybe this is normal behavior for gearboxes and bearings, but it is strange for electrical equipment.

  4. Vboring – thanks for your comments. All of us (myself, Hughes, Harrison and Bach) used absolute declines and graphed them in a linear fashion (Hughes did some of both), compared with your suggestion of using relative declines and a logarithmic slope. I can’t speak for the others, but I found it is easier to calculate and graph (and present to our readers) in our form. But your point is well taken, and is probably more accurate. But having started this way we all will likely continue, if for no other reason than to maintain consistency. The lesson in either case is the same – unless properly maintained, wind turbines degrade surprisingly quickly.

  5. Looking at the data on the Danish offshore turbines (405 total), I don’t see a dramatic decline in capacity factor at all.

    Let’s start with the oldest – Vindeby:

    Max possible yearly output for a 450 kW unit is 3942000 kWh or 3.94 GWh

    Avg per-turbine power output averaged over 5 yrs and shifted by 3 is:

    Yrs Annual Avg kWh Cap Fac
    ’92 – ’96 969127 .246
    ’95 – ’99 886745 .225
    ’98 – ’02 903564 .229
    ’01 – ’05 941155 .239
    ’04 – ’08 956802 .243
    ’07 – ’11 918452 .233

    Tunø Knob

    Max possible yearly output for a 500 kW unit is 4380000 kWh or 4.38 GWh

    Yrs Avg Ann. kWh Cap Fac
    ’96 – ’99 1378113 .315
    ’98 – ’02 1407247 .321
    ’01 – ’05 1337737 .305
    ’04 – ’08 1351033 .310
    ’07 – ’11 1380879 .315

    The max possible annual output for one of the 20 2 MW turbines at Middelgrunden would be 17520000 or 17.52 GWh.

    5 yr Avg ann. nrg Cap Fac
    ’01 – ’05 4385612 .250
    ’04 – ’08 4706426 .269
    ’07 – ’11 4700099 .268

    Horns Rev 1, 82 2 MW Vestas turbines installed in 2002

    Max possible annual output per turbine – 17.52 GWh
    Due to the shorter timeline, changing to 4 yr spans with 2 yr shifts
    Last period will be only 3 yrs avg

    4 yr Annual Avg Cap Fac
    ’03 – ’06 6415317 .366
    ’05 – ’08 7852088 .448
    ’07 – ’10 7604741 .434
    ’09 – ’11 7570480 .432

Leave a Reply

Your email address will not be published. Required fields are marked *