How bad is wind for aero field testing with the Chung Method? Here’s some data:
http://andyfroncioni.com/2010/04/the-chung-method-in-windy-conditions/
AndyF
http://andyfroncioni.com
I just downloaded GC with Aerolab this morning and I’m looking forward to experimenting to see what I can accomplish with it. Just want to say thanks to all of you (you, RChung, TomA, Andy C, AFM, etc.) for sharing your hard work. The wind tunnel might not be dead, but for those of us who can’t justify the travel and cost or just like to experiment, this is an amazing option.
I just downloaded GC with Aerolab this morning and I’m looking forward to experimenting to see what I can accomplish with it.
Thanks for trying it out. If you need any help, don’t be shy to contact me.
The wind tunnel might not be dead, but for those of us who can’t justify the travel and cost or just like to experiment, this is an amazing option.
Curiosity is one of human nature’s strongest driving forces, isn’t it? We’re just trying to see how far we can take all this stuff. It’s certainly fun!
How bad is wind for aero field testing with the Chung Method? Here’s some data:
http://andyfroncioni.com/2010/04/the-chung-method-in-windy-conditions/
Those aren’t data, those are theoretical calculations. As I told you in that email, you could show the same thing with a mathematical proof.
Hence, the question still remains open: is it possible to estimate CdA with high precision under windy conditions when you have on-bike measurements of wind speed and direction? To answer that, you need some actual data - have you considered contacting Mark Cote to see if he would share the data from the “control” rider they tested at Lowes Motor Speedway? They calibrated their probe (which appears to be comparable to the one you are building) in the A2 wind tunnel, so presumably they know the rider’s actual CdA when the device is attached to the bike. Moreover, the Lowes track is not completely flat, so differentiating between CdA and Crr should be easier.
BTW, I hope you realize how low 0.5 m/s is in terms of a ground level wind speed…even 1 m/s is barely sensible if you’re just standing there with your face to the wind. Or to put it another way:
Wind is thine enemy.
Those aren’t data, those are theoretical calculations. As I told you in that email, you could show the same thing with a mathematical proof.
I haven’t seen a mathematical proof and I guess don’t immediately see the answer in this way. My strength is numerical analysis and I’m trying to leverage that to gain an understanding of the problem. I don’t think data precludes numerical experiments, but I won’t quarrel with you about the term. It is what it is, and my “data”, if you will, should definitely be taken at face value.
Andy, you’ll have to believe that I’m trying my very best that I can with the tools that I have. As one of my employers once told me, “Everyone tries their best. Not everyone’s best is the same, though.”
Hence, the question still remains open: is it possible to estimate CdA with high precision under windy conditions when you have on-bike measurements of wind speed and direction? To answer that, you need some actual data - have you considered contacting Mark Cote to see if he would share the data from the “control” rider they tested at Lowes Motor Speedway?
Yes. Unfortunately, it isn’t Mark’s data to share. They answered the question they were trying to answer. We have a different question here, and it will have to wait for Chung-On-A-Stick.
They calibrated their probe (which appears to be comparable to the one you are building) in the A2 wind tunnel, so presumably they know the rider’s actual CdA when the device is attached to the bike. Moreover, the Lowes track is not completely flat, so differentiating between CdA and Crr should be easier.
It is comparable. But I now have a family of probes that will be tested. In the absence of other sources of data, I’ll have to make my own.
AndyF
http://andyfroncioni.com
Those aren’t data, those are theoretical calculations. As I told you in that email, you could show the same thing with a mathematical proof.
I haven’t seen a mathematical proof and I guess don’t immediately see the answer in this way. My strength is numerical analysis and I’m trying to leverage that to gain an understanding of the problem. I don’t think data precludes numerical experiments, but I won’t quarrel with you about the term. It is what it is, and my “data”, if you will, should definitely be taken at face value.
I don’t mean to seem quarrelsome, but think about what you have done: you have used the equation of motion for a cyclist to generate your test file, then used that same equation of motion to extract two of the constants that you used to generate that test file in the first place. While that approach may lend insight into how, e.g., unaccounted for wind may bias the results, it doesn’t answer the question you’re trying to answer.
have you considered contacting Mark Cote to see if he would share the data from the “control” rider they tested at Lowes Motor Speedway?
Yes. Unfortunately, it isn’t Mark’s data to share.
Ah, too bad. Well, how about this, then: add some random flucuations in wind speed and direction when generating your original test file, then see if you can estimate CdA.
While you’re at it, you should try adding some random error in quantifying wind speed and direction using your probe, as well…note how the probe used by Mark et al. provided noisier data at 5-10 deg of yaw than at 0 deg of yaw or at higher angles, either because it couldn’t entirely differentiate the wind direction, or because it was physically moving itself (tough to have something out on a thin stalk not bob around a little bit when hit by the wind).
How bad is wind for aero field testing with the Chung Method? Here’s some data:
http://andyfroncioni.com/...in-windy-conditions/
AndyF
http://andyfroncioni.com
How about the simple answer? It doesn’t work well. As someone “blessed” with flat but windy conditions the method never works to the degree I need it to. Typical reported winds in my area are 10-20 mph, so you have to figure 4-8 mph at the ground. Fortunately at certain times of the year the wind is pretty much one direction +/- 10-20 degrees. As long as you do loops, it’s just better to get deltas for equipment rather than nail an absolute value.
I don’t mean to seem quarrelsome, but think about what you have done: you have used the equation of motion for a cyclist to generate your test file, then used that same equation of motion to extract two of the constants that you used to generate that test file in the first place. While that approach may lend insight into how, e.g., unaccounted for wind may bias the results, it doesn’t answer the question you’re trying to answer.
No quarrel taken, Andy. Think of what I’ve done. I’m using a windless Chung method to analyse a … umm… “windfull” ride file. That’s not a trivial test. At least not for me.
Ah, too bad. Well, how about this, then: add some random flucuations in wind speed and direction when generating your original test file, then see if you can estimate CdA.
Yup. I’ll try to add successive layers of realism to the mix. Generating a ride file with known properties is a real benefit. Eventually, I’ll try to sniff out the entire CdA(yaw) curve.
While you’re at it, you should try adding some random error in quantifying wind speed and direction using your probe, as well…note how the probe used by Mark et al. provided noisier data at 5-10 deg of yaw than at 0 deg of yaw or at higher angles, either because it couldn’t entirely differentiate the wind direction, or because it was physically moving itself (tough to have something out on a thin stalk not bob around a little bit when hit by the wind).
The next versions of the probe will have a small 3-axis accelerometer chip on it, and will subtract out vibration. I hope I can get enough frequency resolution for this. The risk is all part of the fun, I guess.
What can I say – I’m unemployed right now and this is the best challenge I can think of. 
How about the simple answer? It doesn’t work well.
That’s almost what I did. I tried to answer “how much doesn’t it work well?”, instead.
As someone “blessed” with flat but windy conditions the method never works to the degree I need it to. Typical reported winds in my area are 10-20 mph, so you have to figure 4-8 mph at the ground. Fortunately at certain times of the year the wind is pretty much one direction +/- 10-20 degrees. As long as you do loops, it’s just better to get deltas for equipment rather than nail an absolute value.
As soon as I find out how to quantify the error in a estimating deltas, I will. Just thinking about a good test that I can measure.
Robert Chung has a moment-by-moment delta-based method that I’d like to be able to try. Perhaps something like:
- assume CdA of run1 is perfectly well known.
- determine CdA2-CdA1 from the second run.