Wednesday, January 20, 2010

Wind farm wakes

I'd like to highlight a stunning photograph and an interesting paper.

The image shows clouds forming in the wakes of the front row of wind turbines of Horns Rev wind farm.
The paper, Wake effects at Horns Rev and their influence on energy production, by Mechali, Barthelmie, Frandsen, Jensen, and Rethore, describes measurements of the effects of these wakes on wind power production.

The message of the paper is interesting. The downstream wind turbines lose 20% or 30% of their power, and sometimes even more, relative to the front row. The spacing of the turbines is 7 diameters.


RPA said...

It makes for an interesting operations research problem: given N turbines of a particular radius, a patch of land and a distribution over wind direction, what is the optimal placement? You would presumably want to minimise the expected shadowing under your distribution. It would seem that a good solution would look like some quasi-random low-discrepancy tiling and not like a regular grid.

DanH said...

There's some recent modelling work on the optimal tiling for vertical-axis turbines (Whittlesey and Dabiri, 2009, Bull. Am. Phys. Soc. 54(19):BAPS.2009.DFD.LV.9). Don't know about horizontal-axis turbines, though.

AlexC said...

That graph seems somewhat badly labelled. I think it's charting relative power, not "relative power drop" which is what it claims to be. That would be 0, 0.28, 0.3, 0.34 - in other words, it'd be one minus the actual figures. Unless "power drop" is some industry lingo for "power output", I suppose.

Fabulous Photo Gifts said...

I'm sure this is going to seem a very 'low-brow' question, none the less, the Internet is all about sharing and discovery so here goes.

The picture in itself (i won't pretend to understand he maths) suggests the cloud 'mass' is following the direction of the wind - no great surprise if the turbines follow the winds direction (i'm assuming they do).

That being the case, the staggered layout / spacing of subsequent rows is at fault and a auditorium style arrangement of staggered seating would surely be better? Is there any data / has it even be considered on mounting the turbines at slightly different heights?

Layman's questions to be sure but a voyage of discovery none the less.

It's not hard to imaging such a picture scary the living daylights out of folk and putting them off supporting wind generated power.... "we'll all be enveloped in cloud"....."arghhhhhh".


(previous resident of Gt Chishill nr Cambridge).

pawb1 said...

Whatever pattern of turbines is used, for different wind directions, there will be wind shadow effects.
Large wind farms must suffer from the wake effect. Not only does it result in reduced power from downstream turbines, but the increased turbulence results in increased fatigue loading on downstream turbines. It also causes increased noise from downstream turbines, which is a major concern for onshore turbines.

Michael Tobis said...

Spectacular photo, thanks.

The large wind farms I've driven through in West Texas struck me as having a disordered placement pattern, despite having an essentially flat and homogeneous landscape. Perhaps this is intended to minimize the effect you note and illustrate.

Lord Blagger said...

Imagine a wind farm producing all that polution!

Just like all those pictures of cooling towers show CO2 production (and not steam)


David MacKay FRS said...

Yes, the graph is badly labelled, and I agree with AlexC's interpretation of it, as my original comment (I hope) made clear. It's amazing how often people mislabel graphs.

Wardski said...

Wow, impressive clouds.

I've always been convinced that any power source (be it fossil fuels, solar, geothermal or wind) that is deployed on a relatively large scale will have some non-negligible impact on the earth's energy balance and thus potentially its climate.

We've recently discovered the side effect of burning our fossil fuels (who could have predicted 100 years ago that burning petroleum would rise the earth's CO2 level?).

Concerning wind power, I was pondering about the potential environmental impact of large scale deployments. But I hadn't expected clouds to be formed to such a great extend. I've read somewhere that the DROP in clouds after the 9/11 attacks (because of the grounded airplanes) was quite spectacular. It was a unique opportunity to measure the impact of aviation industry on the amounts of clouds in the sky. It isn't to hard to imagine that all this extra soft white stuff might have a large impact on our climate as well.

I've never encountered any scientific studies on this though (i.e. unexpected INDIRECT environmental impacts of alternative energies like wind turbines and solar panels). If anybody has any pointers, I would be very interested to hear about it.

Paul Robins said...

@Fabulous Photo Gifts

Wind companies I'm sure take considerable care over the arrangment of turbines to maximise output and thus revenue.

Most sites will have a prevailing wind (google wind rose) that they can use to set up sites that are slightly asymmetrical to take advantage of this. As shown in the paper MacKay links to, Horns Rev is slightly slanted, as presumably the wind is stronger in the north/south direction (correct me if I'm wrong, I haven't read all the paper yet).

Smaller wind farms have much less of a problem with wind shadows because they usually have much less depth, so providing there is a decent prevailing wind, they can be spaced closer together perpendicular to the wind direction (e.g Royd Moor).


I'm sure the increased vibration/wear and tear etc. aren't exactly welcome, especially offshore with the difficulty of maintence. But I doubt it's something that is seriously detrimental to wind power and can't be overcome by engineering solutions, or at worst tolerated and incorporated into the bottom line.


Indeed it is worth wondering about the side effects of any technology, especially as wind farms become larger and more common. It's hard to say whether cloud formation from wind farms is a problem yet as it depends how frequently the situation occurs (possibly not a lot given it's the first time any of us here have seen this), how long it lasts for, the height they form at etc etc. But I'd suggest the bigger concern is mining the materials needed to build them (or indeed any other form of power generation) given our increasing demand for energy.

Andrew Smith-Gibbs said...

These scale of wake losses are only when the wind direction is precisely aligned with the grid on which the turbines are laid out (±2°), and when the wind speeds are within the steepest range on the wind-speed power-output curve. At much above 11m/s, there are very low wake losses, because the power curve is horizontal there.

Real-world wake losses, averaged over a wind farm over the year, tend to be more of the order of 5% or so.

Whether Whittlesey & Dabiri's work will give us better layouts than the rectangular grid, we'll see. It's intuitively appealing that a fish shoal optimises wake losses; but, of course, the shoal reorientates itself entirely, when currents change. It's much harder to change the entire layout of your wind farm each time the wind shifts direction a bit!

Paul Robins: with energy payback times on wind turbines being of the order of 3-9 months (including all processes from mining raw materials to installation), then there isn't a scaling problem with increasing the amount of wind turbines. And with Britain being the Saudi of renewables, we really should be going for it.

Andrew Smith-Gibbs said...

And by the way, I assume that both Ward and DJCM are joking, and that you both realise that's not a photo, right? It's a computer visualisation of the turbine wakes.

Unknown said...

"It's a computer visualisation of the turbine wakes."
Really? I quick search didn't locate any evidence in support of this claim.

Andrew Smith-Gibbs said...

There are two opposing claims, aren't there Dave? One is that on this bright, dry day with good visibility, wind turbines created clouds in their wake. The other claim is that the image is a visualisation of the wakes. Did you find any evidence for the first claim, either?

Does it strike you as odd that this phenomenon has never been documented before, except in relation to this one image?

Now, I could be wrong on this. But as far as I can tell, the image was created by the Risoe research lab in Denmark. It's used in their presentations on modelling the wake of wind turbines.

And here's the illustration of why the precise wind direction matters so much: a deviation of just a few degrees sharply reduces the wake losses (taken from Flow and wakes in large wind farms in complex
terrain and offshore
, Barthelmie et al)

Wardski said...

your statement that these are not clouds but just a computer generated visualization certainly isn't clear from the 'picture' or accompanying text. It looks pretty real to me (for what that is worth; Photoshop is a powerful tool)

Unknown said...

Andrew, it is quite a serious accusation to say that this photo has been tampered with. I have had a good look at and find no evidence that this isn't a photo. Also, of course clouds can be formed on a bright, clear day, especially where there is a change in pressure/air velocity as would be found near a turbine. You seem to have an axe to grind about everything. What's your problem?

Andrew Smith-Gibbs said...

For heaven's sake, I'm making no accusations. I'm not saying the photo has been "tampered with". All I'm saying is that this image has been innocently misinterpreted. It's a visualisation of turbine wakes - p40.

An axe to grind about everything? Not at all. I just like to see the facts emerge. Call it an old-fashioned affiliation to Enlightenment values. Sometimes I get it wrong, and I'll admit it.

We're not enemies - we just disagree. If I bring a different angle, it really is not an attack. I believe that bringing different sides of the science to the table, and understanding the differences, leads to deeper knowledge. Now surely that's a worthwhile pursuit?

Unknown said...

@Andrew: It's NOT a 'visualisation' of turbine wakes. It's a PHOTOGRAPH. It does not look like any computer visualisation (simulation or measurement) of fluid flow that I have ever seen. For a start, this is a 3D image, whereas the Aeolus project is concerned with time varying lumped (1D) data. Secondly, it is not a false-colour image. If it had been a simulation (computational fluid dynamics) or 2D/3D measured data (such as using particle image velocimetry), it would be presented in false-colour, and probably not from this angle. But if this doesn't convince you, let's entertain your hypothesis a bit further. Maybe it's a photo-realistic rendering! Well that would have no relation to any of the Aeolus project objectives. Plus if I created a rendering this good (it is a really nice photo), I would shout about it, and yet mysteriously there is no mention of the words 'rendering', 'visualisation' or suchlike on the project website. You pointed to page 40 of a pdf document. This page shows the photograph with the words above it "VISION: wake measurements on and offshore". I believe this simply means 'the vision (objective) of the project is on and offshore wake measurements'.

You're just plain wrong about this.

Bill said...

Why would you think it's not a photo? Contrails aren't a rare phenomenon.

Andrew Smith-Gibbs said...

I was wrong.

My apologies to Professor MacKay FRS and to the other readers here. I've just received confirmation from Risø that this is indeed a photograph taken from an aircraft, of a rare phenomenon: one (as far as I can tell) undocumented in the scientific literature to date.

It's interesting that the wind direction is precisely aligned with the layout of the grid: perhaps that's part of what creates the phenomenon.

Jack The Lad said...

I would have thought an answer to the power drop would be to increase the height of the turbines toward to centre of the matrix. This I know would lead toward some difficult engineering problems, probably reducing the cost effectiveness of the turbines, so why not just ignore the centre section of the array and create an array with only two turbines deep offset vertically and horozontally

Andrew Smith-Gibbs said...

Jack The Lad: on the whole, it's financially not worth doing.

Losses at the worst times, when the wind is from a very specific direction at a particular speed, are 30%. But on average, the losses over a year are only about 10%, according to slide 20 of this presentation by Leo Jensen.

If you did something very complex at Horns Rev, you might be able to increase the array efficiency from 90% to 95%. It's really not worth it. The wind farms are both economic and productive with 10% array losses, so why stress over saving a few percent? It's much easier and cheaper just to build the wind farm 5% bigger, than to reduce wake losses by 5%.

Thor Russell said...

What about this study showing that you can obtain up to 10*! more power from the same land area, doesn't this change all the conclusions regarding power from renewable energy in your book?

Unknown said...

I think it's an interesting paper, but quite preliminary. To quote: "Although the current model does not explicitly allow for complicated flow
behavior, such as vortex shedding, turbulence, and three-dimensional effects...". These are all important and non-negligible effects. I think until more work on this has been done, the results are pretty inconclusive.

Andrew Smith-Gibbs said...

Thor: utility-scale VAWTs are still in their infancy, really, so even if this wake-streaming does work, it may not be scalable within the next 10 years.

However, there are three things that make huge differences to the calculation of offshore wind resource.

One: shipping doesn't need two-thirds of our seas to be free of turbines. Small boats can manoeuvre within wind farms. Large ships need lanes that might typically be 5km wide. So, British waters available for wind farms are more of the order of 85%, rather than the 33% assumed in the book. That makes the available resource over 100kWh/d, which is higher than current energy demand! (UK Final Energy Demand is 82kWh/d, using the official figures from DUKES)

Two: the further away from the coast you are, the higher the wind power. So although offshore wind farms built to date are only of the order of 2.5W/m², ones further out at sea can yield 4-6W/m² . For now, we won't see those densities, because we're nowhere near reaching the limits of available seabed, and developers will play safe with wider turbine spacings, because seabed is much cheaper than taking any risk of reduced turbine lifetime from downstream turbulence.

Three: we now have wind turbines that can go out to 700m depth rather than the 50m limit used in the book. Indeed, several of the Round 3 wind farms are in depths beyond 50m.

All in all, the electricity potential in offshore British wind is several times as big as our energy demand.

Lord Blagger said...

So if it is such a great technology, why do you need such huge subsidies?

Here's an alternative.

Don't use wind, for now. When the price of electricity rises to make wind cost effective, implemented a wind policy. No subsidy needed

Andrew Smith-Gibbs said...

Nick: All energy technologies are subsidised. Nuclear and CCS have unlimited insurance provided by the government. Fossil fuels have their greenhouse debt subsidised by everyone. Coal and uranium are also subsidised by not having to pay for the environmental impacts of mining. Economists call these: "externalities". While there are externalities, there is no efficient market.

Furthermore, as we've seen in the credit crunch of the last two years, markets can be very bad at planning for the long-term, at allocating capital, and at balancing risks.

Wind is already one of the cheapest ways to generate electricity. If we had a Free Market for energy in Britain, we'd be getting over 50% of our energy from wind already; and all the coal and nuclear plants would have closed down.

If you wanted something approaching an efficient market for energy, you'd have to remove all insurance on nuclear and CCS: none more would get built, and the cost of waste disposal on our current set of nuclear plants would skyrocket. You'd have to add in taxes to internalise the externalities: a carbon tax of £100/tonne is a start. You'd need border taxes on the import of coal and uranium to account for the CO2 and other externalities involved in their production. And you'd need a huge programme to redistribute income, else you'd have millions living in life-threatening fuel-poverty.

The UK Carbon Census Ltd said...

This is a fantastic photograph, and a great blog. Thank you D.M. for providing such an excellent resource.

Readers may also be interested in our own contribution in this area. Last year a few of us decided to get our hands dirty, and started up a new business providing residential energy advice to UK households. It's a voluntary 'carbon census', if you like, and the idea is to make good, reliable advice easy and cheap for everyone.

The exercise is generating a lot of original data and we plan to post as much of that as we can online, as the census grows.

If you're interested, please visit us at and add your email address to our subscriber list. We're starting to build a readership, and we'd welcome your comments too.

Best Regards

Sam Arie
Director, The UK Carbon Census

Unknown said...

I think the turbine wakes are not generating the clouds, but rather that there is fog near the sea surface and that this fog is being "whirled" up into the air when the turbine wake gets down to the level of the fog. It's hard to tell for sure from the photo though.

Mike said...

This theme of "shadowing" makes me wonder why there do not appear to be any turbines using a pair of contra-rotating props. Granted the leeside prop would lose 20 - 30% efficiency (going from comments on here) and a more complex gearbox would be needed. However, compared to two single-prop units, the costs would be dramatically lower.
Can someone shed light?
Mike Higton

Lord Blagger said...

Why would you need a complicated gearbox? Just sync them. After all its the way contra-rotating props work on planes.

carlfoss said...

Consistent with Matthew's comment but not necessarily arguing for it, it seems curious to me that, if the clouds represent turbulence, the volume about one turbine diameter past the blades has no clouds. One might expect that this would be the volume of greatest turbulence.

Ralph said...

what is the provenance of this photo?

Would a higher resolution image be available for use in a presentation?