Drag reduction modifications in wind tunnel: Audi A2 from Cd 0.288 to 0.204

[NeilBlanchard]

Interesting discussion - I would think the slot just ahead of the moving "roadway" belt would be supplying air to make the relative motion of the roadway and the boundary layer of air be stationary relative to each other.

With the "extra" air flow introduced at the front of the belt - the belt's motion would possible accelerated the already moving air. By the way, what affect would the texture of the belt / roadway have on things?

By vacuuming some air away from in front of the belt would seem to INCREASE the motion of the air relative to the moving belt, and thereby make the effect of the moving belt, and therefore lowering air pressure under the car - which makes it LESS "realistic"; not more.

[Vman455]

Originally Posted by NeilBlanchard

Interesting discussion - I would think the slot just ahead of the moving "roadway" belt would be supplying air to make the relative motion of the roadway and the boundary layer of air be stationary relative to each other.

With the "extra" air flow introduced at the front of the belt - the belt's motion would possible accelerated the already moving air. By the way, what affect would the texture of the belt / roadway have on things?

By vacuuming some air away from in front of the belt would seem to INCREASE the motion of the air relative to the moving belt, and thereby make the effect of the moving belt, and therefore lowering air pressure under the car - which makes it LESS "realistic"; not more.

Don't forget: at these speeds, there is a no-slip condition between the airflow and any wall or body surface; so, with a stationary floor, there's already a pretty thick boundary layer (which thickness is proportional to the root of the distance the flow has traveled along the wall) by the time you get to where the car is (and the moving ground/belt begins). In a real car moving along a road, there is no boundary layer between the air and road since they are both stationary. To replicate this in a tunnel, you have to suck that boundary layer away by sucking the air out with a suction slot, or elevate the vehicle so it's high enough that the boundary layer on the floor doesn't affect it and skew the results.

[NeilBlanchard]

To accurately measure the drag of a car, it has to be on the ground. So, raising the car is not what you want to do.

I am talking about the effect the moving belt (and its surface texture) has in the wind tunnel. It would accelerate the air in the boundary layer, in the direction of the air in the "middle" of the tunnel. And this would lower the pressure of the boundary air at the front of the belt.

By vacuuming air out of a slot at the front of the belt - this would further lower the pressure there. And it would reduce the amount of air that would be accelerated by the belt. Which is "artificially" affecting the air pressure under the front of the car.

My take is that air would need to be ADDED at the front of the belt; not removed - to offset the effect of the moving belt.

[Vman455]

Forget about the surface texture; if the belt was moving relative to the air, there would still be a no-slip condition between the flow and the belt surface, which is why a boundary layer forms in the first place.

The airspeed and belt speed are the same; there is no acceleration of air by the belt, and no boundary layer forms between the belt and flow. This is what happens to a car on the road; to accurately represent that, you must have one boundary layer under the car, between the car underbody and the airflow. A second boundary layer between the ground and flow results in a u-shaped velocity gradient.

Elevating a car in a wind tunnel is a common practice for correcting for both wheel drag when using stationary wheels and floor boundary layer when using a fixed floor.

I'm travelling right now, but when I get home next week I can post the relevant pages from some textbooks, including Hucho's Aerodynamics of Road Vehicles, that might explain better.

[aerohead]

Originally Posted by NeilBlanchard

Interesting discussion - I would think the slot just ahead of the moving "roadway" belt would be supplying air to make the relative motion of the roadway and the boundary layer of air be stationary relative to each other.

With the "extra" air flow introduced at the front of the belt - the belt's motion would possible accelerated the already moving air. By the way, what affect would the texture of the belt / roadway have on things?

By vacuuming some air away from in front of the belt would seem to INCREASE the motion of the air relative to the moving belt, and thereby make the effect of the moving belt, and therefore lowering air pressure under the car - which makes it LESS "realistic"; not more.

Your thinking is right on.It's just a matter of 'tuning'.Some full-scale tunnels are capable of on the order of 160-mph flow velocities.I don't think you can run a moving roadway at that kind of speed.If so,then there's a possibility that the airspeed in the test section will overwhelm the 'roadway',and introduce a flow condition you'd never see in the real world.Also,the flow on either side of the roadway would be stationary,compared to the moving floor,which would introduce shearing forces and vorticity if not compensated for.The tunnel technicians need to keep sharp.

[aerohead]

Originally Posted by NeilBlanchard

To accurately measure the drag of a car, it has to be on the ground. So, raising the car is not what you want to do.

I am talking about the effect the moving belt (and its surface texture) has in the wind tunnel. It would accelerate the air in the boundary layer, in the direction of the air in the "middle" of the tunnel. And this would lower the pressure of the boundary air at the front of the belt.

By vacuuming air out of a slot at the front of the belt - this would further lower the pressure there. And it would reduce the amount of air that would be accelerated by the belt. Which is "artificially" affecting the air pressure under the front of the car.

My take is that air would need to be ADDED at the front of the belt; not removed - to offset the effect of the moving belt.

*A consideration is how far away from the settling chamber,or tunnel jet,does the moving road begin.There could be hundreds of feet of tunnel floor,wall,and ceiling for boundary layer growth before the air ever gets close to the test section and moving floor.That would have to be managed.The boundary layer could be as thick as on a short railroad train.
*Since the moving floor does not span the entire width of the test section,you'd have a differential floor boundary layer thickness,depending on location.
*If the belt is moving at exactly the same velocity as the airflow,that's one thing,if not,compensation must occur.
*And the Bernoulli effects of a 'blown' jet could perturb the streamline filaments you've spent $millions to create for your 'source',demanding vigilance with air management.

[NeilBlanchard]

Originally Posted by Vman455

The airspeed and belt speed are the same; there is no acceleration of air by the belt, and no boundary layer forms between the belt and flow. This is what happens to a car on the road; to accurately represent that, you must have one boundary layer under the car, between the car underbody and the airflow. A second boundary layer between the ground and flow results in a u-shaped velocity gradient.

Elevating a car in a wind tunnel is a common practice for correcting for both wheel drag when using stationary wheels and floor boundary layer when using a fixed floor.

At the front of the belt, the boundary layer of air IS stationary - so the moving belt would accelerate it.

If there was a streamlined plate that had a moving belt, was raised up and was in contact with the wheels, then that might improve the "accuracy" without additional measures to try to correctly mimic a rolling car on a road?

[aerohead]

Originally Posted by NeilBlanchard

At the front of the belt, the boundary layer of air IS stationary - so the moving belt would accelerate it.

If there was a streamlined plate that had a moving belt, was raised up and was in contact with the wheels, then that might improve the "accuracy" without additional measures to try to correctly mimic a rolling car on a road?

I'm trying to wrap my mind around the tunnel setup.Are we talking about the rollers being above the tunnel floor,and air is striking the forward stagnation point (line) of the roller,then accelerating up and over,onto the flat portion of the 'road'?

[Vman455]

Originally Posted by NeilBlanchard

At the front of the belt, the boundary layer of air IS stationary - so the moving belt would accelerate it.

If there was a streamlined plate that had a moving belt, was raised up and was in contact with the wheels, then that might improve the "accuracy" without additional measures to try to correctly mimic a rolling car on a road?

Only the 'layer' of air in contact with the floor; above that, there is a velocity gradient and at the top of the boundary layer, the velocity is the same as free stream velocity.

I think I see what you're saying now--the transition from the stationary flow to the moving belt, where the belt speed is the same as free stream velocity (we aren't talking about testing cars at 160 mph here...). But that's precisely why a suction slot is needed to vacuum off that boundary layer; you just get rid of it to ensure you have one boundary layer forming under the car as in a real car on a real road. I suppose you could elevate the moving belt to get it above the boundary layer over the stationary floor, but as aerohead mentioned, then you run into the issue of the airflow accelerating over the shoulder you've introduced by raising the belt and roller.

[NeilBlanchard]

Right, and vacuuming it off has the effect of lowering the air pressure under the front of the car. Combining the vacuum slot and the moving belt, this could artificially give downforce on the front of the car.

I think the ideal thing would be to extend the moving belt out ahead of the car.