Aero

Back by popular demand, the general all-things Road forum!

Moderator: robbosmans

Franklin
Posts: 400
Joined: Mon Aug 04, 2008 11:09 am

by Franklin

ergott wrote:
Calnago wrote:Surely it can't be assuming a constant 30mph whether you're on a flat grade of 0% versus a steep grade of 15%??


There is no assumption of constant speed, just constant power. That's why aero resistance is so much more of the total resistance on more flat terrain where you are going faster. I like that chart because I can relate to the power output if only for relatively brief periods (compared to a pro's threshold).

It really shows that the importance of aero really depends on the course. It doesn't have to be that steep for weight to be a bigger concern.

=> We are looking at systems, not parts. IOW: The aerodynamics of a bike and the weight of a bike is less than a very small part of those bars.
=> Weight has a downward limit (regulations)
=> Most important: Weight and aerodynamics are not mutually exclusive.

rchung
Posts: 97
Joined: Sat Jul 09, 2011 5:01 am

by rchung

davidalone wrote:ERO sports test LOTS of atheletes. in a velodrome, with real time riding conditions. so I would say they have possibly the most 'real world' data there. I do believe they use the chung method.

They do.
davidalone wrote:at 11m/s ( 40 kph- a realistic racing speed for most of us) - that works out to anything between 50-80 watts saved from changing from a road to a TT position. Now, I'm assuming you did your chung method test at similar conditions ( unless you're telling me you did the test faster, by which any means, you have strong legs, sir.)

You don't actually need to test at 11m/s. As for the watts savings, this is why I prefer to use "difference in CdA" rather than "watt savings." For weird historical reasons (blame John Cobb) results are often expressed in terms of watt savings at 30 mph. I hate that. Anyway, 30 mph = 13.33 m/s and at standard air density that means a difference of about .021 m^2 in CdA. For reference, that's about the difference Cervelo found in the wind tunnel (and we subsequently found in field tests) between their then-new P3C and old P2K TT bikes. The current Cervelo S5 frame and fork is actually pretty close to the now-old P3C frame and fork, while the ancient P2K frame and fork was close to the old Cervelo Soloist, which isn't quite as good as the new S3 frame and fork but is better than some round tube aluminum or steel frames. So although I've never actually tested a big round tube Cannondale CAAD frame and fork, if you were comparing something like a current S5 frame and fork to something like that, yeah, a difference of .02 m^2 is plausible.

As for the Reynolds numbers, our empirical findings are that when doing field tests at speeds up to at least the current hour record speed, the CdA appears to be constant.

by Weenie


Visit starbike.com Online Retailer for HighEnd cycling components
Great Prices ✓    Broad Selection ✓    Worldwide Delivery ✓

www.starbike.com



tranzformer
Posts: 846
Joined: Mon Jan 21, 2008 10:36 pm

by tranzformer

Thanks rchung, 53x12, tapeworm, BeeSeeBee for posting all the data. It seems fairly clear what the empirical and objective data shows. I like that.


rchung wrote:
ScuderiaDouroux wrote:Saving watts in a wind tunnel with half of a dummy on top is not the same as an actual rider in real world circumstances. I don't care how you slice it.


Field tests on real roads under real conditions with entire riders (though especially when I'm the rider you might think of it counting as half a dummy) give the same results as wind tunnels. Whether you make the measurements in the field or in wind tunnels, the estimates for CdA that come out make the correct predictions of speed for power (or power for speed).



rchung, is there a link to a useable spread sheet for the Chung method test where one could just plug in the required data and it would spit out a CdA number?

User avatar
ergott
Posts: 2870
Joined: Tue Apr 18, 2006 3:03 am
Location: Islip, NY
Contact:

by ergott

Franklin wrote:=> Weight has a downward limit (regulations)


I'm not a pro so I don't have a downward limit. Right now my bike is under that. I know I can get a UCI limit full aero bike, but don't have that kind of money to burn since again, I'm not a pro. If I were, they'd give me one.

rchung
Posts: 97
Joined: Sat Jul 09, 2011 5:01 am

by rchung

I'm terrible at spreadsheets. I use a series of custom functions I wrote for myself in R which are completely opaque to everyone else on the planet. There are a couple of spreadsheets floating around on the internet but the simplest thing to do is to use Golden Cheetah's Aerolab module. This page gives a little bit of discussion and an example.

Bridgeman
Posts: 742
Joined: Wed Jan 12, 2005 6:04 am
Location: USA
Contact:

by Bridgeman

rchung wrote:I'm terrible at spreadsheets. I use a series of custom functions I wrote for myself in R which are completely opaque to everyone else on the planet. There are a couple of spreadsheets floating around on the internet but the simplest thing to do is to use Golden Cheetah's Aerolab module. This page gives a little bit of discussion and an example.


Thank you rchung.

climbandpunishment
Posts: 62
Joined: Fri Mar 20, 2015 4:40 am

by climbandpunishment

Rick wrote:Reynolds Number for 40 kph in 20°C Air for a characteristic length of .5 meters (the bicycle overall ?) is about 3.7 X 10^5


Please correct if I'm mistaken, but from what I understand, I think it's probably most appropriate to use the horizontal length of the bike section of interest for the Reynolds number (though wheels and rotating parts might be different and I'm no expert on those). So the appropriate characteristic length would then be the horizontal length of that individual part of the bike. For bike tubes and speeds where you'd start to worry about aerodynamics, 3.7 X 10^5 is probably close to the upper limit of the Re #s you'd expect, with a lower limit of closer to roughly 5 X 10^4, depending of course on exact tube size and speed.

rchung wrote:As for the Reynolds numbers, our empirical findings are that when doing field tests at speeds up to at least the current hour record speed, the CdA appears to be constant.


Thank you for sharing! It's good to have that confirmed, given how expensive good testing can be. That seems to agree with the above, unless I'm really off the mark. Out of curiosity, was that testing with round(-ish) tube frames or aerodynamically shaped tube frames? Or both?

rchung
Posts: 97
Joined: Sat Jul 09, 2011 5:01 am

by rchung

Almost all of the testing I've done or seen done is with frames that people would actually think about racing on, so they've tended to be at least nominally shaped. A few years ago I tested my round-tubed steel commuter bike with handlebar bag as a lark but, as you can guess, the speeds were lower and I wasn't really interested in getting a very careful estimate. I've occasionally thought about testing someone in a business suit on a Dutch bicycle but I don't have a Dutch bike and I'm starting to know fewer and fewer guys with business suits.

climbandpunishment
Posts: 62
Joined: Fri Mar 20, 2015 4:40 am

by climbandpunishment

Thank you rchung! The Dutch test sounds fun to watch. :)

User avatar
Rick
Posts: 2034
Joined: Sat Aug 29, 2009 4:30 pm

by Rick

climbandpunishment wrote:
Rick wrote:Reynolds Number for 40 kph in 20°C Air for a characteristic length of .5 meters (the bicycle overall ?) is about 3.7 X 10^5

Please correct if I'm mistaken, but from what I understand, I think it's probably most appropriate to use the horizontal length of the bike section of interest for the Reynolds number (though wheels and rotating parts might be different and I'm no expert on those). So the appropriate characteristic length would then be the horizontal length of that individual part of the bike. For bike tubes and speeds where you'd start to worry about aerodynamics, 3.7 X 10^5 is probably close to the upper limit of the Re #s you'd expect, with a lower limit of closer to roughly 5 X 10^4, depending of course on exact tube size and speed.

Both "characteristic lengths" are used in aerodynamic calculations, but for slightly different situations.
If you are measuring surface drag or boundary layer parameters, then the Re is typically length ALONG the unit being tested.

But for overall drag of a body, the Re is usually based on the "cross sectional area"; and since area is length squared, it is usually the diameter of a circle of equal X-sectional area. That is why it is tough to get too precise on complex shaped bodies. That is why emprical measurements still rule in areodynamics. The experimenter could really choose any characteristic length athat is reasonably associated with the body, so long as he clearly states what he is using.

rchung wrote:As for the Reynolds numbers, our empirical findings are that when doing field tests at speeds up to at least the current hour record speed, the CdA appears to be constant.


So that would indicate that the "complex body" is in the "turbulent flow regime", like in the flat section of the "flat disk" in the chart. That makes sense and it is a good thing to keep in mind.

So when different values are obtained for different rider positions are evaluated, are you measuring the actual change in cross-sectional frontal area, or assuming that is constant and the Cd value is what is changing....or are both being tracked ?
Just sort of a rhetorical question. I am not really hung up on the technicalities of the answer; but it seems like some people like to probe the intimate mathematical details. :)

rchung
Posts: 97
Joined: Sat Jul 09, 2011 5:01 am

by rchung

Usually we just look at CdA as if it were one quantity, not the separate Cd and A. One of the test protocols is, within one test run (i.e., for fixed position and equipment), for the rider to vary speed from slow up to about as fast as they can go while still holding position. If the CdA were speed-dependent then we'd see a mismatch between the predicted speed for power (or power for speed or VE) at low vs. high speeds. At least over the ranges of speeds we've been seeing, we don't see that mismatch so our empirical result is that we can treat the CdA as if it were constant. To me I'm most happy that not only do we have a way to estimate the CdA but also a diagnostic that tells us when something is different.

climbandpunishment
Posts: 62
Joined: Fri Mar 20, 2015 4:40 am

by climbandpunishment

Rick wrote:But for overall drag of a body, the Re is usually based on the "cross sectional area"; and since area is length squared, it is usually the diameter of a circle of equal X-sectional area. That is why it is tough to get too precise on complex shaped bodies. That is why emprical measurements still rule in areodynamics. The experimenter could really choose any characteristic length athat is reasonably associated with the body, so long as he clearly states what he is using. . . . So that would indicate that the "complex body" is in the "turbulent flow regime", like in the flat section of the "flat disk" in the chart. That makes sense and it is a good thing to keep in mind.


Thanks, Rick. I'm aware of the potential choices an experimenter can make regarding defining characteristic lengths. I believe in the chart you posted the typical definition of the characteristic length (i.e., the actual length, or diameter of the sphere in that case) was what was used. So it's probably a mistake to conclude based on that chart that a bike is primarily in the turbulent regime. Drawing conclusions about the flow about a complex body using a chart for a specific shape, especially considering you're using two different Reynolds number characteristic length definitions, is very tenuous at best. Many parts of a bike can have largely laminar flow, and many are also primarily turbulent in nature.

User avatar
ergott
Posts: 2870
Joined: Tue Apr 18, 2006 3:03 am
Location: Islip, NY
Contact:

by ergott

Um, am I wrong or do you have to consider the angle of the tube? I keep hearing talk of sphere and circle. The downtube is the predominate tube on the frame and even when it's round its cross section as far as the wind is concerned is an oval.

climbandpunishment
Posts: 62
Joined: Fri Mar 20, 2015 4:40 am

by climbandpunishment

Yes, that's correct and a great point to keep in mind! A circular tube shape will of course "look" like an oval as far as the air is concerned if it's angled like a downtube would normally be, for example. Rick was just helpfully pointing out that it's possible that coefficient of drag could potentially change abruptly at certain speeds, which is better documented for a sphere/circle, but will also be the case for ovals and many other more bike-appropriate shapes too. It sounds like, based on rchung's testing, this isn't something that we have to worry about too much for a bike.

by Weenie


Visit starbike.com Online Retailer for HighEnd cycling components
Great Prices ✓    Broad Selection ✓    Worldwide Delivery ✓

www.starbike.com



SuperDave
in the industry
Posts: 534
Joined: Sun Jun 15, 2008 4:57 am
Location: San Juan Bautista, CA
Contact:

by SuperDave

airwise wrote:
53x12 wrote:Jonathan Caron, professional triathlete who did a "blind" test of the S3 vs. R3 for comfort and his inability to distinguish between the two based on comfort. http://forum.slowtwitch.com/gforum.cgi? ... ASC;mh=25;

Hilarious.
That's what he calls "rough pavement"?
He should try riding even the main roads in the South of England. That looks like a carpet in comparison.


The riders' butt, hands and feet are all connected to the bicycle through a series of "springs" or elements that introduce flex and what we perceive as "comfort"
In mechanics, two or more springs are said to be in series when they are connected end-to-end so as to act as a single spring, this is how a bicycle works. The tire, wheel, frame, seatpost (butt), stem, bar (hands), BB and crankset (feet) act as a single spring with deflection measured from the origin of the stress or force applied. The amount or deformation of the system is the sum of the strains of the individual springs. The compliance of a spring is the reciprocal of its physical constant "k". The spring constant of a wheel and tire at 100psi is so far less than that of a modern carbon frame that when adding the contribution of total system compliance of the frame to the equation you get beyond the range of what is humanly detectable.

This isn't good for marketing "comfortable" frames however and does not relate to hinged bicycles like the Slingshot, Castellano designs, and Trek's Iso.

-SD

Post Reply