So in wind tunnel tests usually you find decreasing drag at higher yaw angles until the point at which you stall. But doesn’t the forward/backward vector (the direction the rider is facing) decrease in magnitude as yaw angle goes up, if wind speed is constant? So if you’re trying to hold 25 mph, it’s not categorically less drag if you have a side wind, yeah? Because then the total wind flow going over you (vector sum) would be higher (wind vector plus velocity vector, right?) Only if your airfoil had sail-like qualities would that be the case? Just curious.

Also, along those lines, I’m wondering why they keep the wind speed constant across yaw angles, instead of forward vector constant (i.e. increasing the wind speed as you increase yaw angle, so modeling wind drag as if you’re holding x velocity with increasing side winds)

So if you’re trying to hold 25 mph, it’s not categorically less drag if you have a side wind, yeah? Because then the total wind flow going over you (vector sum) would be higher (wind vector plus velocity vector, right?) Only if your airfoil had sail-like qualities would that be the case? Just curious.

That’s my understanding. With that said, a lot of modern frames and wheels are pretty fantastic at yaw (speed concept, firecrest, etc)

This is typical in Engineering, trying to solve the problem correctly/completely (technically, analytically) is usually too difficult, burdensome, or - typically - is too costly, so assumptions are made to simplify the problem. Assumptions noted in the report, everyone is happy (although some of the white papers bike companies put out, really take creative license calling their departments Engineering departments). If the “correct” or perhaps a more-applicable answer is desired, a simulation could be performed, accounting for the assumption made and then adapting to the specific problem.

So, yes, technically wind tunnel testing is flawed as you’ve described. But what’s the detriment if every frame and every wheel is tested in the same purportedly fallacious way? In the same case, if the tunnel speed is 30mph at 15° degrees, and we know Martin/Cancellara can run 30mph plus the additional velocity from the crosswind, we can calculate that he and the bike are seeing (let’s say) 35mph at 15°. We know drag is proportional to V^2, so we can directly calculate drag in Cancellara’s run. But at the end of the day, it’s who crosses the line first - I can be way better at math than everyone else, but if they trained harder, they’re going to win.

If you think wind tunnel testing is scandalous, probably should give composite (material) analysis a wide berth.

How is the airflow vector varied in CFD modeling?

If what you are asking is whether you can be faster with a slight crosswind than no wind at all, the answer is a definitive yes, you can, if your bike+position has enough decreasing drag with yaw. Total ‘wind flow’ going over would still be more but it is not retarding your motion as much, at certain wind speeds/yaw angles.

So in wind tunnel tests usually you find decreasing drag at higher yaw angles until the point at which you stall. But doesn’t the forward/backward vector (the direction the rider is facing) decrease in magnitude as yaw angle goes up, if wind speed is constant? So if you’re trying to hold 25 mph, it’s not categorically less drag if you have a side wind, yeah? Because then the total wind flow going over you (vector sum) would be higher (wind vector plus velocity vector, right?) Only if your airfoil had sail-like qualities would that be the case? Just curious.

What you want to know is the CdA at each yaw angle. Varying the speed within cycling speeds tends not to change the CdA much, so it wouldn’t matter much, and would make testing slower, and more expensive, and would likely be able to tease out fewer small differences.

In the end, with CdA at each yaw angle in hand, you can accurately simulate how you will perform at given power on a given course.

Also, along those lines, I’m wondering why they keep the wind speed constant across yaw angles, instead of forward vector constant (i.e. increasing the wind speed as you increase yaw angle, so modeling wind drag as if you’re holding x velocity with increasing side winds)

Not whether or not you can be, but whether or not you are, on average. I assume that with the cutting edge modern equipment you STers buy nowadays you probably are.

Not whether or not you can be, but whether or not you are, on average. I assume that with the cutting edge modern equipment you STers buy nowadays you probably are.

Yes with a good position and equipment it is pretty common. If you have a guess at how your position’s drag decreases with yaw, (if you look like zabriskie, use his data) you can simulate scenarios with bestbikesplit.com or analyticcycling.com to see what wind speed ranges and directions will cause it.

Hi BoyWithACoin,

As Jack has mentioned, it really happens.

Although the wind speed in the tunnel is not varied with yaw, the measured drag can be corrected for the effect you’re talking about. We call this the “cosine squared correction.”

It is left as an exercise for the reader to derive the correction.

Cheers,

Assuming the drag coefficient doesn’t change with these slight changes in wind speed (which is a reasonable simplification to make), all you have to do to correct for this effect is to take the measured drag force and divide it by the square of the cosine of the yaw angle, since since drag force is a function of velocity squared. And as has been said already, this calculation is a lot cheaper and easier than constantly changing wind speed in the tunnel.

Very nice!