Slowman wrote:
Jim@EROsports wrote:
Tunnels and velodrome have both been good predictors of time, so have proven results. Both, though, can also be poor predictors of time depending on how well the tests are completed. In the tunnel, many fall into the trap of testing a position that just isn't sustainable in the real world. On the velodrome, it can be the opposite, athletes tend to keep their head higher than they would outdoors looking for the next corner. This is where field testing comes in, no? Aero sticks will fill this void (if one wants to refer to it as a void), and we've shown the results are quite good predictors of time in the past few months with the one we're using. You need to include yaw, though, and I don't think there's one available to the general public that measures yaw...yet. It's good to be me. ;-)
i guess i'm a pretty bad explainer of my query. i'm comfortable with wind tunnels. i'm comfortable with field testing. i'm comfortable that there's agreement among the models. i'm not casting doubt on the validity of wind tunnels. but if i could *wheel* this back to the point i made: a tunnel using a rolling road *might* teach us something we aren't getting from what we now have. no well funded tunnels were made for bikes. they're made for subsonic airplanes, or 200mph cars, and those vehicles don't shuck and jive the way we do on bicycles. we've been hacking these tunnels with our own contrivances that give us pretty good guesses. i'm not complaining. i'd just like to see a bike & rider mimicking their movements on the road in a tunnel. wouldn't you?
I would think the movement of the bike and rider is going to be both very specific to individual riders and inconsistent even for an individual depending on effort level, fatigue, terrain & road conditions, weather conditions, perception of danger, etc. Surely that would make test data from a free riding athlete in a tunnel pretty impossible to usefully analyse?
I hadn't even considered the idea of using a tunnel with a rolling road in this way, and I don't think it's what I would do.
The purpose of a rolling road is normally to eliminate the tunnel boundary layer to more accurately model movement of an object through still air. In normal testing, the idea is the bike stays still and the air flows around it to simulate the relative movement of the two. However out on the road, in still air, the road and air are not moving relative to each other so there is no wind velocity gradient (ground boundary layer). The bottom of the wheel has the same impinging airflow as the top of the rider's head. That's not the case when it's windy and it's not the case in a tunnel without a rolling road. Different wind tunnels have different boundary layer characteristics. For most testing you want a uniform airflow, but you also don't mount your test piece on the floor of the tunnel. It's mounted in the middle of the tunnel where the airflow interacts least with the surfaces of the tunnel. For testing of entire ground based vehicles, it's necessary to test in proximity to a plane representing a road. You may still want uniform airflow to represent movement through still air, or you may want a gradient to represent the ground boundary layer. A rolling road allows you to model the relative movement of an object and a surface (e.g. a bike moving on a road surface). There are various set-ups and devices sometimes used to manipulate the tunnel's BL to produce or eliminate the BL immediately upstream of the test section.
I've never seen any of this mentioned in bicycle related tunnel test discussions. That may be because it's not considered, or because there's a conventional setup considered to be the standard and assumed for all cycling tunnel testing. Or I just missed it (I admit I don't usually spend much time on tunnel testing discussions).
TLDR?
Okay, what I'm getting at is this:
I'd expect the bike to be mounted rigidly on a rolling road for the purpose of more accurate airflow modelling.
I don't see a convincing reason to try using it like a cycling treadmill to achieve more accurate modelling of rider movement.
How do you measure drag for a bike moving freely on a rolling road anyway?
You couldn't directly measure it, so what do you gain over field testing?