Monday, Feb. 18, 11:29

Boundaries are important.

Case in point is Jim Brasseur’s Cyber Wind Facility at Penn State, which is using Blacklight — as well as other XSEDE machines — to figure out why electricity-generating wind turbines lose so much power.

You’ll find out more about Brasseur’s work in future PSC communications; at the moment they’re running the “1.0” version of the simulation, so the results are not yet in. But in a larger sense, it’s no mystery. The wind is always changing its direction and speed, and the turbines can’t keep up. Worse, the ensuing turbulence whips the rotors to and fro, stressing the structures and lowering their working lifetime. In an industry that measures durability in decades — because it takes about 20 years for one of these things to start making money — that’s a big deal.

Here’s the neat bit. In order to be accurate, Brasseur and his group have to model a boundary layer of air on the surface of the wind rotors — those huge propellers you see from the Pennsylvania Turnpike — that is only 5.2 micrometers thick.

Brasseur pointed out that it’s half the width of a red blood cell. Maybe more interesting, it is about one 20-millionth the length of the rotors.

I have a copy of a dissertation — Jennifer Lynn Page’s,* from Georgia Tech — that’s also been on my mind, again because of the boundaries. Page studies crabs’ ability to follow a scent plume in turbulent water streams to its source. When scent — airborne or waterborne — encounters turbulence, it gets strung out into delicate filaments. These are thin but intense bits of scent that get farther apart as you move away from the source.

Crabs, you see, have scent receptors on their antennules and their legs. Page’s data filled us in on why.

The antennules sense the filaments as they wash past. Together with tactile sensors that tell the crab the direction of the current, it tells the crab how to move upstream, toward the source of the scent.

But like the wind washing over Brasseur’s rotors, water currents don’t maintain a constant direction. As the current changes, it puts kinks — dog legs, you could say — in the scent plume. If you just follow the filaments into the current, you’re likely to shoot past one of these turns and lose the plume.

Which is where the scent detectors in the legs come in. Close to the seabed, there’s a boundary layer that mixes very thoroughly, dispersing the filaments and creating a uniform but dilute layer of scent. Because of this, there’s an abrupt drop in scent when you move past the scent plume’s envelope into the clean water beyond. The crabs use their legs to tell them when they’re starting to pop out of the plume — and need to turn back into it.

Which brings me to Friday night (and almost the end of this admittedly long post) — I couldn’t stay late to finish my blog entry, because we had a nighttime dog training.^

My little farm collie, Rosie, and I pulled a task that was a search of a new construction site — about 20 acres, no structures yet, but a lot of empty space, some dirt mounds, and piles of construction materials. At the east side, the search area was bounded by a new road that started running south, then curved to the west.

Well, we’re searching this site in the dark for more than an hour — way longer than it should have taken. Rosie starts whining, because she’s getting frustrated with re-crossing the same ground. I can’t figure out what the problem is.

Finally, I cave — radio to my practice subject for a hint.

“I’m in the big metal tube at the top of the hill.”

Big metal tube? What big metal tube?

The hill lay at the southeast corner of my area — downwind of us. No wonder Rosie wasn’t getting anything.

Climbing up the hill, I see — on the far side of the road that was supposed to be my boundary — an array of corrugated metal culverts.

Irritated, and getting ready to give my practice subject a lecture on navigation, I walk Rosie to the far side. Once she’s downwind she makes short work of it, finding my subject in seconds.

But I’m a little distracted, by the road on the other side of the culverts. It ran parallel to the one I’d thought was our boundary, and less than 100 feet away from it — a tiny sliver, compared with the size of our area. But big enough to put egg on my face.@

Boundaries are important.


*Kudos to Georgia Tech for making their PhD dissertations web-downloadable.

^I am a volunteer search-and-rescue dog handler. So now we know that.

@Never work with dogs or kids. If everything goes well, they get the credit. If it doesn’t, you look like an idiot.

This entry was posted in Bioinformatics, HPC Research, Outreach, People, XSEDE and tagged , , , , , , , , , , , , . Bookmark the permalink.

Comments are closed.

People. Science. Collaboration.

PSC's Bi-annual Publication (select issue to download PDFs)

PSC Spring2014c

PSC2013 covers web    Projects2012

Subscriptions: You can receive PSC news releases via e-mail. Send a blank e-mail to

Media Contacts

Media Contact(s):

Ken Chiacchia
Pittsburgh Supercomputing Center
Shandra Williams
Pittsburgh Supercomputing Center

Subscriptions: You can receive PSC news releases via e-mail. Send a blank e-mail to

PSC Logo Download: PSC's logo is available for use in print, e-media and presentation application. Various formats are available here.

Use of PSC materials: To request permission to use PSC materials, please complete this form.

News Archive: 2012201120102009