November 10, 2014
FLO Cycling - What I've Learned from My Time in a Wind Tunnel - Part 2
This article is a follow up to last week's article FLO Cycling - What I've Learned from My Time in a Wind Tunnel - Part 1. I suggest reading the previous article before you read this one, especially if you want a bit of background on wind tunnels. Here we go with the rest of the lessons I've learned.
Lesson Three: Should We Test Products Alone?
This lesson really depends on what you are testing. When testing wheels, you are faced with the question of testing the wheel by itself, with a bike, or with a bike and rider. There are literally endless possibilities and each is unique. Someone with shaved skinny legs will affect the air flow around the back wheel in a different way than someone with thick hairy legs. Is there a water bottle on the bike? If so, what shape does it have? Is it a TT frame or a road frame and who makes it? The list goes on and on.
Different people have different opinions but personally I like to single out wheels by themselves. I feel I can get a better understanding of what is going on with the wheel. That’s not to say that testing additional features is bad. To be honest, it makes a lot of sense, but I feel because of the number of options, specific testing is best done with individuals and their gear. I don’t think it’s fair to add a frame and rider and then assume that it works for everyone. I think it gives us an idea, but I don’t think it’s the answer.
You do have to take things into consideration when you test items individually. A wheel is part of a system. So without the system the results will be different. Remember the discussion about what the air sees. The wheel on the back of the bike sees the air differently than the wheel on the front. Additionally, when the air gets to the back wheel it has already passed the front wheel, the bike and the rider. Air that was laminar (traveling in a smooth flow before it interacts with anything, or clean air) can now be turbulent (air that is not traveling in smooth flow pattern, or dirty air). Clean air and dirty air modify the drag of an object in a large way.
This matters because reducing drag is about saving time. If the drag is reduced more when a wheel is tested by itself than when a wheel is tested on a bike, the calculated time savings are exaggerated. I don’t think this is deceitful if it’s clearly spelled out. We openly admit to calculating time savings by testing wheels by themselves but will also tell you that you have to test the whole system (you included) to get an accurate answer on time savings.
Again, for me it comes down to clearly stating what you are doing. If you make it clear, the reader understands what is going on.
Lesson Four: Attachment And How You Sweep
First we have to discuss attachment. Watch the video below for an explanation.
Now that you understand attachment, let’s talk about how it works in practice. The longer the air stays attached to a wheel as you sweep it through yaw angles, the more you will reduce the drag. Attachment is a good thing. If I start a wheel at zero degrees yaw, I can assume that the air is attached on both sides of the rim until it hits the spokes. As the yaw angle increases, we eventually get to the point where the flow detaches. Let’s assume that the flow detached at 15.3 degrees. It would make sense that if we moved the wheel back to 15.2 degrees the air would reattach. However, it doesn’t. It’s a fact that once the air detaches, it take longer to reattach when going back in the opposite direction.
So why does this matter? Well, how do I report my results? Do I start the test at zero degrees and sweep it through different yaw angles until the air detaches? Or, do I start the test with the flow detached and sweep it until the air attaches? The first option would give me a better result than the second. The best answer is to sweep in both directions and take an average. In the real world, we experience a number of different situations and the average seems like a fair approach.
I learned this after talking to some people much smarter than I am. Swaying the results can make a big difference so make sure you ask the question if you don’t see it spelled out. Our most recent results do not include an average because we didn’t run full sweeps. I wish I had read this article before I made my last trip =]. The next time we will take an average.
Lesson Five: Small Things Can Make A Really Big Difference
The last time we visited the A2 Wind Tunnel we tested a number of different tires. While testing we noticed that the same tire put on in the reverse direction can make a big difference in the drag. Why? I’m not quite sure. I can see a couple of grams but this was way more than couple. Up to 20 grams for our tests and the guy running the test had seen far worse.
Tire pressure is another factor that makes a difference. We have not extensively studied tire pressures but we plan to in the future. I’ve heard through the grapevine that 90psi is the magic number. In the wind tunnel I am sure it’s pretty accurate but what happens on the road with different rider weights may be a different story. We are working on a new testing protocol which, I will talk about later.
If I had to take a guess at why small things like tire direction and psi make such a big difference I would say it's the tires shape. Increasing or decreasing tire pressure changes the shape of the tire and each tire leaves the mold looking a little different than the next. This may explain the difference in drag from one side of a tire to the next.
Make sure when you are in the wind tunnel to look for the little things. It may save you from scratching your head. Then again it may have you scratching your head even more.
Lesson Six: What’s The Equation To Solve For A Fast Wheel?
Science can be broken down into two categories. The first allows you to develop a mathematical formula so that you are able to predict results and see changes before you perform the experiment. The other category does not allow you develop a formula. The only way to get more info on a topic is to perform more experiments. A part of aerodynamics falls into this category. There are equations in aerodynamics, but there isn’t one that allows you to solve for a shape with the least amount of drag in various conditions. Therefore you have to test each shape you design to see how it performs.
Lesson Seven: Black Magic
Since there is no equation that solves for the fastest wheel, and small things make a big difference, a lot of what we do in a wind tunnel is guess and check work. That’s not to say that general rules don’t apply. People with experience can predict behaviors with a high degree of accuracy. There are, however, things that seem like black magic.
The Continental GP 4000 S II tire is a good example. It’s a very fast tire when tested but understanding exactly why is very difficult. Modeling the surface of a tire in CFD is very difficult so testing plays a big role in tire design. Sources have told me that even the engineers at Continental aren’t exactly sure why the tire is so fast. I don’t want to say they got lucky because a lot goes into designing a product. However, I do believe you can have a very fortunate educated guess.
Spokes in wheels are another example. There have been a number of studies on spokes but they are very hard to completely understand. The air leaving the spokes is much different than the air entering the spokes and understanding how that works is nearly impossible. Speed changes, yaw angle changes, dirty air, it all adds up.
There are people working on these difficult areas like Matthew N. Godo at Intelligent Light. He’s doing amazing work. We interviewed him a while back if you are interested in reading about him and his work.
That’s A Wrap
Wind tunnels are amazing places and we will continue to use them when we design products. I do, however feel that real world testing will change as technology evolves. We are currently developing a new testing protocol that we believe will make testing in the real world very accurate and more accessible to the general public. Since wind tunnel time is so expensive ($600-$900/hour), the general public does not readily have access to them. In order to understand more about aerodynamics, I think it is important to collect as much data as possible and our new testing protocol plans to do that. We will have more on this shortly.
Thanks for reading my thoughts. If you have any questions about them, please let me know.