Wind tunnel testing

[Shadow Catcher]

One of the concerns many of us have is wind resistance and I have been playing with the idea of a small home made wind tunnel. I mean after all the Wright brothers did it (I have seen theirs).
The one half of the equation was coming up with a scale model of TV and trailer and that idea has been supplied by the foamy builders.
This would give us the empirical data to see what does and does not work in drag reduction (the bubble thingy comes to mind).
I have a suck your house inside out fan that can be used (we use it to suck air out of the up stairs).
Build a couple of plywood box's , the goesinta and goesouta and a poly carbonate middle box test chamber and use a cheap air filter to break up the air pattern into the test chamber.
Still in the thinking about it stage.
Tom

[Kharn]

It might be easier to find someone with the right software to simulate it. Scaling down models for wind tunnel testing isn't just 12":1"

[Shadow Catcher]

Using the software approach only works with a very regular surface i.e. a wing from what I understand, looking at the information I could find. The auto manufacturers seem to do quite a bit of wind tunnel testing some involving a rolling road and wheels http://www.youtube.com/watch?v=eszhVxE_9-8

[Mightydog]

Why?

[gww25]

One of the guys over at the Junk Rig Forum built a small scale testing tunnel. I'll try to dig up the link but it was years ago that I saw it.

[angib]

Very, very, very difficult - or maybe a bit more difficult than that! The Wright Brothers had it easy as any knowledge was more than they had before, plus they were only working in one medium - air.

The problems are scale effects and scale speed. It is almost impossible to make a model that actually matches the real thing - even things like the gaps between door and body panels make a difference. A foam model would have to be surfaced somehow as otherwise the skin friction would be utterly different.

Real road vehicles do not sit with non-rotating wheels on a road that has a 70mph wind flowing over it - so you have to think about rotating model wheels and a moving floor to the test chamber (so that the road and the air are travelling along together).

And then you have to decide at what scale speed you are going to run and that means that you only get to model one component of drag correctly and you have to let the others go hang. And then you need to calibrate the wind tunnel to work out how to convert the test results to full scale results.

You were going to build the test chamber at least ten times the cross-sectional area of the vehicle, weren't you, so that the air could actually flow around it and not just stop?

The one thing that you could do is to compare two similar models at a scale speed that matched pressure drag, since that is the largest drag component. If one model had a lower drag than the other, then that might mean the full-size version would have lower drag too - or it might mean that you changed something between the two models that made the result useless - there would be no way of knowing. But it couldn't harm to try.

Anyone who wanted to learn about road vehicle aerodynamics can't do better than read The Leading Edge. This book is about solar race vehicles but it is the best explanation of road vehicle aerodynamics that doesn't need a masters degree just to read it.

Just don't expect anything like "this design detail works better than this design detail" as that depends entirely on the circumstances.

For overall road vehicle aerodynamics, including a page or two on trailer drag, the world-wide reference book is The Aerodynamics of Road Vehicles, now published by SAE. But it is a dry read with lots of equations.

It probably is cheaper and easier to use CFD (Computational Fluid Dynamics) software to model the airflow over a tow vehicle and trailer - I guess that might cost less than $100,000 even including the genius you need to work the software, whereas a decent wind tunnel is going to be in the millions.....

[Shadow Catcher]

Mightydog
Each of the things potentially disrupting air flow has to be factored in to the program. This includes the shape of the tow vehicle the shape of the trailer (both fairly easy relatively speaking, but unique) also the space between the TV and the trailer and what is in that space i.e. tongue box.
Compass Rose was built with a platform for the largest tongue box I could find which is 17 cubic feet. The space between the trailer and the TV is one of the most important as I understand.
Try modeling this,


Angib
I realize this is sub optimal as far as accuracy and comes under the heading of gross exploration but my feeling is that it might be useful. I think this come under the heading of for absolute accuracy yes you have to figure in things like wheels road etc. but in reality they are things I can not easily change and so come under the heading of "oh well" would a bubble or wind deflector make any difference. I am going for cheep!

[JuneBug]

We were shocked at how much our mpg was reduced pulling a Little Guy 5-Wide with a Mazda MPV van on fairly flat roads between Austin and Abilene.
Anything that helps improve mileage while towing is worth exploring. There was a thread awhile back that discussed little v-shaped stick-on things (used by truckers) that help reduce drag.

[angib]

My alternative suggestion is 'coast-down testing' done with full-scale models, ie, the real thing.

A coast-down test is simply driving at a certain speed, say 60mph, and then timing how long it takes to coast down to a lower speed, say 30mph. This is surprisingly accurate - within a few percent - and allows modifications to be tested pretty quickly. It needs to be repeated in both directions and averaged if the road isn't perfectly level or there is anything more than a faintest breeze blowing.

This means you get an answer the same day and don't have to measure fuel economy over 1000 miles - which is probably no more accurate.

I've only ever owned a manual transmission so I can't say how to test with an auto gearbox - I guess running in neutral for that length of time isn't a problem, though over-shifting to park wouldn't be nice....

Obviously this also requires a full-scale 'model' of each modification but that's probably easier to do in foam, duct tape and zip ties at full scale than it is to build a scale model. OK, altering trailer ride height would be tricky, but even that could be done simply if all it had to do was survive a short trip (at least if the police weren't watching....).

[eamarquardt]

My alternative suggestion is 'coast-down testing' done with full-scale models, ie, the real thing.

A coast-down test is simply driving at a certain speed, say 60mph, and then timing how long it takes to coast down to a lower speed, say 30mph. This is surprisingly accurate - within a few percent - and allows modifications to be tested pretty quickly. It needs to be repeated in both directions and averaged if the road isn't perfectly level or there is anything more than a faintest breeze blowing.

This means you get an answer the same day and don't have to measure fuel economy over 1000 miles - which is probably no more accurate.

I've only ever owned a manual transmission so I can't say how to test with an auto gearbox - I guess running in neutral for that length of time isn't a problem, though over-shifting to park wouldn't be nice....

Obviously this also requires a full-scale 'model' of each modification but that's probably easier to do in foam, duct tape and zip ties at full scale than it is to build a scale model. OK, altering trailer ride height would be tricky, but even that could be done simply if all it had to do was survive a short trip (at least if the police weren't watching....).

Clever. Remember wind drag increase with the square of your speed so testing at higher speeds will give better results.

Years ago a friend wanted to time his 0-60 elapsed time. He couldn't drive and shift and work the stopwatch so he tape recorded it with audio cues and then timed the playback to determine his elapsed time.

Clever practical solutions to theoretical issues.

Cheers,

Gus

[wa_flyfisher]

My alternative suggestion is 'coast-down testing' done with full-scale models, ie, the real thing.

A coast-down test is simply driving at a certain speed, say 60mph, and then timing how long it takes to coast down to a lower speed, say 30mph. This is surprisingly accurate - within a few percent - and allows modifications to be tested pretty quickly. It needs to be repeated in both directions and averaged if the road isn't perfectly level or there is anything more than a faintest breeze blowing.

This means you get an answer the same day and don't have to measure fuel economy over 1000 miles - which is probably no more accurate.

I've only ever owned a manual transmission so I can't say how to test with an auto gearbox - I guess running in neutral for that length of time isn't a problem, though over-shifting to park wouldn't be nice....

Obviously this also requires a full-scale 'model' of each modification but that's probably easier to do in foam, duct tape and zip ties at full scale than it is to build a scale model. OK, altering trailer ride height would be tricky, but even that could be done simply if all it had to do was survive a short trip (at least if the police weren't watching....).

The Coast Down method was used on Mythbusters as read in this article.
http://www.popularmechanics.com/science/mythbusters/projects/how-to-drive-a-backward-porsche

[Shadow Catcher]

I know of real life tests and there are a couple of methods however the idea behind the wind tunnel is to be able to test on a small scale changes you might make prior to implementation. To be able to see what subtle changes make in wind flow patterns can guide what you are doing. The bubble on the front, should it be bigger or smaller or six inches higher or...