Designing better wind farms
Driving through the countryside you often see huge grids of wind turbines dotted across the fields. But does it actually matter how we lay them out? Cristina Archer, from the University of Delaware, has been studying how air flows through wind farms in order to answer this question. She spoke to Chris Smith.
Chris - So first of all, what sorts of improvements could we make then by changing the way that we format where the windmills are in these aggregations of wind farm?
Cristina - Well, my research found some staggering results actually if I can make this joke. You can actually improve the performance of a wind farm by 13 to 33% if you place your turbines in a staggered arrangement. What I mean by that is, it's easier to describe if you think about yourself at a movie theatre and you want to have a good vision of the screen. And of course, you don't want to have a person sitting right in front of you. And so, the way turbines are often arranged are kind of like the seats in a theatre. They're like aligned with one another which is not very smart because if you want to look at the screen, you will like the person in front of you to not be just exactly in front of you, but slightly staggered.
Chris - Why do you think that people chose to put them in the theatre arrangement when they didn't really have to do that? Was that just our obsession with order?
Cristina - I know. Maybe it is. If you have an event in your yard, say, a wedding or something, you will see that the natural instinct of us humans is to make things lined up. It's just easier you know, so it's very understandable.
Chris - The numbers you're suggesting of between 13 and 30% improvement in the productivity of a wind farm. Those are enormous numbers! How did you do this and why did you begin looking at this?
Cristina - First of all, the ultimate goal is to try to reduce the cost of wind energy. Wind energy is really relatively cheap compared to all the other renewable energy sources. It's possibly the cheapest. But the designs of turbines are already very good - the blades are optimised, the airfoils are great, they're even beautiful at this point. So, there's not really that much you can do with the design of turbines to make them even more productive. They're already pretty much as efficient as it gets. And so, if you want to reduce the cost of wind energy, you have to act on something else. What I was interested in is, can we place them in such a way that the wakes that the turbines generate - you can think of a wake as turbulent eddies - once that wake hits the next turbine, obviously, it impacts its productivity. So, you don't want these wakes to interfere too much with the turbines that are downstream.
Chris - So, you're saying, instead of the view in the cinema being occluded by the head of the person in front, you don't want dirty air which is all messed up and full of turbulence, hitting the turbine behind. You want to minimise that effect. So, how did you actually do that? How did you try and work out what the effect was and how best to place them?
Cristina - I used a very sophisticated simulator. This is a computer code that tries to resolve the equations of motion of the atmosphere. Very fancy terms to say that I'm trying to basically predict what the wind will be in all the points, millions of them, in a wind farm. So, you have to imagine like a cube that includes the entire farm, and I'm trying to predict exactly what the wind speed, what the temperature, what the pressure will be at all these million points. I have include obviously the presence of the turbines in this volume, the fact that they rotate and the fact that the turbulence generated by a turbine can impact the following one.
Chris - You must have had a very powerful computer to do that.
Cristina - Yes, absolutely. Actually, it's not even powerful enough. I would like more power. If any of the listeners have some kind of a super duper, even more powerful system, please let me know. Right now, I'm using a high performance computer cluster with over 200 parallel processors and it's still going to take me about 2 months per simulation.
Chris - Going to have to open a few wind farms to power that, aren't you? So before we finish, could you just tell us then, have you actually tested this and validated it? Do we know that this isn't just the product of some computer time that it really applies to the real world?
Cristina - Yes. We have data from existing wind farms. So, I can definitely access those and that's in my list of things to do. The tools I'm using right now, the simulators are more useful when you're planning for a future farm. You don't have data because it's a prediction. And so, I'm trying to tune my model so that it works very well with existing farms, and so that I can use it now and simulate conditions in other locations for future times and so on.