How fast can offshore wind be deployed? What are infrastructure requirements?
Guest post by Andy Dawson
One anti-nuclear argument that’s frequently made is that nuclear is slow to deploy – that renewables can make inroads into carbon production rather faster than can building new nuclear stations. I was recently provoked into taking a look at this, in the context of the UK’s 2020 and 2030 CO2 and Renewables targets, and out plans for nuclear new-build.
Our average capacity factor for onshore wind over recent years has consistently been around the 26% mark.
The first thing for US readers to understand is that space is at a (relative) premium here in the UK – and hence the options to deploy new onshore wind-farms (the dominant renewables option for us) is limited. We added approximately 500MW of onshore wind in 2009-10 (the last Renewables Obligation reporting period). To all accounts this is likely to slow as resistance to the siting of windfarms increases, and in fact, new accreditations of onshore wind have fallen in each of the last three years (700MW in 2007-8, 600 in 2008-9).
On that basis, we’d be well advised to assume at most adding an extra 2500 or so MW by 2020.
Our average capacity factor for onshore wind over recent years has consistently been around the 26% mark. Applying that, we can anticipate average production equivalent to about 650MW.
Let’s assume that EDF get just one reactor up and running in that time (their schedule would have three, so let’s be pessimistic). And, let’s assume it hits the same average capacity factor that Sizewell B has through life, to date – about 88%.
That one reactor gives me 1425 MW average equivalent generation. More than double the amount estimated from onshore wind.
So, rather more progress towards the 2020 goals from the “too little, too late” nuclear option than from wind deployment.
Now, the next thing I’m sure you’ll say is “what about offshore wind”. Well, that’s hardly flying out there – but we’ll ignore that problem too, and simply concentrate on logistics.
We’ll make an easy assumption – we’ll go for a floating system (much faster to deploy than sea-bed mounted systems, especially if we’re assuming moving far offshore). And move far-offshore we will have to, since near-shore potential is ultimately limited.
We’ll assume 5MW units. We’ll assume that the ambition is deploy the same average equivalent generation rate as I’d get from an EPR – 1425 MW. We’ll also assume we’ll beat the current capacity performance for offshore wind in the UK (same source as above) of 27%, by hitting the dizzy heights of 30%.
We’ll need to deploy 950 units.
OK, it’s now 2011. Let’s assume we’re a year in the design, survey etc. for these various schemes, but after that, can push the button at the end of 2012, giving me eight years until the end of 2020.
I need, on average to build and emplace 2.3 5MW units per week. Between friends, let’s call it one per two working days.
Let’s assume a VERY conservative 13 weeks to build the floating platform for each turbine (that’s rather better, btw, than was achieved on much simpler liberty ships in WW2). That’s 65 working days. Allow 5 days at each end for getting the completed unit out and floated, and then the dry-dock cleaned up for the next.
Roughly, I need about 35 working dry docks large enough to build the equivalent of a vessel of several thousand tonnes.
Problem is, there are probably only a couple of hundred such dry-docks worldwide. And about half a dozen in the UK.
So, let’s assume we can build an extra 30 or so in perhaps two years. But, if we need two years, that ups the build rate to about three per week. We’d actually need 40 extra dry docks. Never mind….let’s assume that’s all OK.
But, there’s a problem. I’ve worked offshore, in the north Sea. It’s a b******d of a place for about eight months of the year. Basically, you do the absolute minimum for eight months of the year – don’t even think about trying to emplace a structure other than between March and September, or anything that needs more than about a three-four day period of good weather.
So, we need the ability to manage rather more than two units a week – I need about to emplace 950 units over about six years (having built the dry docks) in about 100 weeks – 700 days. About 1 per day.
Putting an anchored platform for an oil platform in place takes about six weeks at an utter minimum. Lifting the pylon and nacelle for a wind turbine on land needs a week or so, and erecting the blades about the same.
So, about enough anchoring capability to emplace (950*6/100) units at a time – I make it 57. So far as I’m aware, the entire North Sea oil sector supports one or two. We also need to provide floating cranage to life 100 metre length blades onto a 120 metre pylon – about 20 such units. I thinks there’s one or two that service the North Sea oil business.
Quite an expansion, you’ll agree. And that’s just to deliver average generation capacity, in the same time it takes to build a single 1600 MW EPR. One reactor between 2011 and 2020.
So, what’s the actual history of deployment of offshore wind? Well, we’ve yet to emplace a single far-offshore unit. And the record even on nearshore, seabed anchored systems is chequered, to say the least. In 2003, about 50MW. In 2004, much the same. In 2005, about 180MW. In 2006, none. In 2007 about 180MW, and in 2008 back to 90. In 2009-10, we hit about 320 MW. We seem to have nothing due for completion in 2011, then about 2GW is due in 2012, and nothing in 2013. Meaning, it’ll taken us about ten years to install capacity sufficient to give an average ouput equivalent to about half an EPR.
For comparison, France alone built and commissioned about 2 ½ reactors per year through the 70s and 80s. Or, 18 times the rate implicit in the calculations above. The current plans of our three potential nuclear consortia would commission just under 16,000MW from construction of the first unit starting in 2013 to the last starting up in 2025.
I really don’t thing a renewable enthusiast should be making an argument about the slowness of deployment………..
Rod, small request – in the first paragraph, “proiked” is a typo for “provoked”. I like to think my points are good, but I’m a lousy typist.
Done.
It is worth noting that, in the UK, wind power has the longest preparation/licensing period of any technology. DECC’s 2010 analysis shows that offshore wind needs 5-7 years preparation before contruction starts, plus 1.5-3 years of construction. (p84-86)
http://www.decc.gov.uk/assets/decc/statistics/projections/71-uk-electricity-generation-costs-update-.pdf
Basically anything that is not in the pipeline by now is unlikely to be generating much before the end of the decade.
Andy writes:
“We’ll also assume we’ll beat the current capacity performance for offshore wind in the UK (same source as above) of 27%, by hitting the dizzy heights of 30%.”
Some time ago L. M. Miller, F. Gans and A. Kleidon published and article (http://www.earth-syst-dynam.net/2/1/2011/esd-2-1-2011.pdf) there they estimated maximum global land surface wind power extractability. They showed that there are not that many places in the world there the onshore extractability at the par with the north of the UK. Based on information in the article (although limited) I did some back of the envelop calculations of the capacity factor we should expect and came to approximately 26%, which is what we have here in the UK on average at the moment. It may be physically impossible to reach capacity factors higher than the currently observed figures.
Here is my quick estimation.
Area of the the largest wind farm (http://en.wikipedia.org/wiki/Scout_Moor_Wind_Farm) is 545 Hа or 5450000 м^2. From fig6b in the article wind dissipation for the north part of the UK is something around 2.5….3 W/m2. In another word this is how much power can be extracted from the boundary layer. So 5450000 м^2 x 3 W/м^2 = 16.4 МWт.
From the farm specification:
Turbines 26
Installed capacity 65 MW
Capacity factor 27%
Annual generation 154 GW·h , which roughly corresponds to 17.6MWt.
Figures are very close. Similar estimations for other farms produce approximately similar result.
For example to provide 10% of the UK annual electricity capacity from wind farms an area of 1440 km^2 will be required ( [(378.5 TW·h/365/24)*0.1]/3 Вт/м^2 = 1440 км^2). On average a single turbine requires 400×400 m^2. Therefore about 9000 turbines will be needed. At the moment there are about 3000 in the UK. Granted, off-shore extractability may be higher, but still the staggering number makes me wonder.
@Andy,
Thank you for writing this article and for looking at what it takes to actually do what one says that they will do.
Have you ever looked at or know of anyone that has attempted to fully compare the reliability of wind and that of nuclear in a levelized cost analysis? I have not been able to come up with a way or find any way of incorporating the entropy into a levelized cost model. I think that until we do that we will not compare apples to apples.
I ran some numbers from the Bonneville Power Association 2007-2011 for their wind generation. They had an expected capacity factor of 29.8% and an entropy of 9.5.
Compare that with the NERC GAR 2009 with numbers from 2006-2009. I took those numbers from the total available hours and the forced outage hours in that period. Nuclear had an expected availability of 97.8% and an entropy of 0.11.
There is an assumption in my wind numbers that it is required to be put on the grid when it is available allowing the expected capacity factor to be compared directly with the expected availability of nuclear.
So the entropy (expected uncertainty) of wind is 9.5 compared to nuclear’s 0.11 almost a two order of magnitude difference! That means it takes roughly 100 more binary questions to describe wind power than nuclear. There has to be a way we can capture that in $, £, or €.
Though I’m no expert in wind generators, I do know they require regular maintenance. Anything that is placed in water, rather than land, is going to be more time consuming and expensive to maintain. These generators will need cleaning, greasing, and anti-corrosion efforts on a continual basis. The addition of new under sea power cabling is going to be very costly as well.
Wind generators also have a “Goldilocks” sweet spot of wind speed needed to operate. Too slow and they can’t produce power. Too fast and they must be shut down otherwise they will destroy themselves.
If anyone knows a good source of wind on sea vs. wind on land maintenance costs, please post.
Rod,
Very interesting post. Do you have any source (maybe one of your own previous blog posts) which does a similar analysis of the US and Canada?
I might be wrong, but I think that in the US, we have, at least theoretically, a lot more potential for building wind farms on-shore. . . the main problem being, I believe, that most of the good on-shore sites are at places distant to the point of consumption. Places like the southwestern Deserts, western Plains, and midwestern Farmlands?
Jason,
I wouldn’t be surprised if that’s a question wind pushers don’t want an answer to. Besides, in Europe and the US, reliability and maintainability are no longer desired attributes of a power system. Safety and CO2 emissions are now the metric for new electricity generation.
This is what I was thinking too. Given the corrosive nature of saltwater and salt spray, I can’t help but wonder if offshore wind farms will last anywhere near as long as claimed — even with the expensive maintenance.
….. and at what point are the maintenance costs going to be so high that they equal or exceed the value of the electricity produced? These questions make the whole venture sound highly questionable from an investment perspective.
The bigger ones also have an “inverse Goldilocks” zone wherein they become net consumers of power because the blades have to be kept turning at low speed even if there is no wind. They draw power from the grid to do this. That avoids damage from distortion of the windmill blade which results if it is kept motionless for any length of time. The small ones don’t have this problem because their arms aren’t long enough to sag and bend like the big ones.
This is a use of the term entropy that is unfamiliar to me. Can you give a link to an explanation?
@Jim,
The best I can do is to refer you to Gibbs, Shannon, and Jaynes.
It’s taken a study of the subject to come to my understanding. Entropy is the expected uncertainty of our state of knowledge about something. You will not find that in any thermodynamics book… you’ll get a discussion on measure of disorder and other such nonsense, which obfuscates the actual meaning.
A good link but incomplete is:
http://en.wikipedia.org/wiki/Entropy_(information_theory)
Gibbs used the log of a probability distribution to describe the “index of probability”. In information theory, this is the “uncertainty” of the of the probability distribution. What Gibbs. and later Shannon do is to take the expectation of that over the probability density function. This becomes Shannon’s information entropy. It also relates information entropy to thermodynamics through Boltzmann’s k.
Also at the end you’ve got “I really don’t thing” where I think what you want is “think”.
Thinking of offshore experience, those ought to be designable against, and maintenance regimes available to mitigate the effects – blades tend to be made of materials like fibreglass, for example.
What I’d expect to be the bigger problem in somewhere like the North Sea is maintenance access. It’s one thing working on an enclosed platform , entirely another trying to do maintenance work in or on the nacelle of a wind turbine several hundred feet up, in the weather that prevails between September/Occober, and May/June.
Another problem with onshore in U.S. is that the best sites, i.e. the Appalachian ridges are also the main migratory routes for birds, raptors, and bats. Likewise, developers have applied for an “incidental take” for endangered species in a corridor from Canada to the Gulf of Mexico which also happens to be a major migratory route for endangered Sand Hill Cranes as well as waterfowl.