CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   Wind Tunnel Testing (

Vinnende April 12, 2012 02:54

Wind Tunnel Testing
Hello all,

Firstly yes I know this isn't technically CFD but I'd really appreciate some input from you intelligent fluids folk.

I'm in my final year of engineering so I'm required to complete a thesis as part of that. My topic is to investigate the wind loading on a typical conveyor belt + truss arrangement, because there's a lot of uncertainty in industry about drag coefficients etc. I've attached a model of the steel structure below (simplified by removing members parallel to flow), but there's also lots of other services and the conveyor belt itself not pictured. It's 18m long.

Anyway, my plan is to construct a scale model ~(1:10) and do a variety of wind tunnel tests for different arrangements/alignments/etc, with the major output being drag force (assuming the other forces and moments are negligible). The problem is, my supervisor doesn't think the whole project is theoretically complex enough, and I guess he's right - at the end of the day all I'm doing is measure drag on a model.

So my question is, how can I make this project more complex? Right now the only real calculations involve some dimensional analysis/similitude (which I'll try to make as elegant and clever as possible). What else can I do/investigate? Essentially I want to show/prove that I've done my research and know the theory behind the results, so I'd love to include some sort of theoretical analysis, but obviously the complicated geometry makes anything I do by hand irrelevant. I was thinking perhaps include flow visualization in some wind tunnel tests and then discuss the boundary layers/flow separation/turbulence onset, but then what's the point? It's not like the dynamic behaviour of the wind will adversely affect its structural integrity.

Essentially I just want to prove I know my stuff, whilst keeping it relevant to the topic. I'd love to do some sort of CFD analysis, but I fear the geometry is far too complex and I also don't know much about CFD :p. But if performing a simulation over a small section could provide further insight (and not take forever) I'd love to have a crack!

Any help/comments are much appreciated :)

Nickul April 12, 2012 17:11

As part of the project how about you do a feasibility study into the appropriatness of a CFD model, then its a case of how deep is the rabbit hole.

At first glance the mesh looks pretty complex, but then your realise that most of the geometry is repeated, so meshing up may not be that painful.
What about the wind loads? transient? and then where do you get the data for such a loading?
You mention bounday layers, for the geometry you've shown there is unlikely to be much BL growth, I would hazard a guess thats its all down to vortex shedding and the downstream interaction of the structure.

If you do want to go with CFD then start with a much simplified model and get some results then build upon that, there's no point in going mad with mesh to not finish it in time. You haven't done much CFD? If you've got access to good tools, then I would advise it, with a time constraint (I'm assuming you have to complete by summer).
Start in 2d as well, how about a bluff shape with another downstream bluff shape and look at the interactions?, baby steps.

just my 2 pennies.

Vinnende April 12, 2012 22:28

Thanks Nickul! A CFD feasibility study could definitely add some content :) The problem is that picture I've posted is really only half the story (or less). Once you add the piping, electrical cables, handrails, conveyor belt, idlers and return idlers the geometry becomes a real nightmare. Taking a 2D section might still be feasible for something extra to do, and I've done some basic things in FLUENT before, but I don't know much about the turbulence modelling side. Do you know of any good introductory resources that give a good overview of turbulence modelling?

And I was talking more about a general discussion of boundary layers, what they are, how they form, when they separate, why they separate, what that does to the flow, etc. just to add some content.

Thanks again

agd April 13, 2012 13:42

Without even getting into CFD, have you addressed the issues of translating subscale model results to full-scale geometry? One thing that I have observed is that most people tend to view wind-tunnel testing like they do CFD - it's a push-button exercise that can't be that tough. But wind tunnel testing, and correctly applying the results of wind tunnel tests, is as much an art as properly using CFD. For example, is the flow laminar, turbulent, or transitional over your full-scale geometry? How will you match that in your testing? How will you account for the inevitable sacrifices to modeling fidelity that arise in reducing a full-size geometry to a 1/10th scale model? Depending on how the model is installed in the tunnel, how will you account for any sting or strut or mounting interference?

i mean no disrespect to your instructor, but are you both certain that a wind tunnel test that generates good, well-documented and scaled data, is really such a trivial exercise? For all I know you go to a school where wind tunnel testing is a way of life, and the points I have raised really are considered trivial. If so then ignore my comments. But having worked around wind tunnel testers for many years, I don't know of any of them that would consider these issues simple. Just some thoughts.

Vinnende April 13, 2012 15:15

Not trivial, just not enough to warrant a decent mark. It's a year-long thesis that is used to award graduation honours, not a highschool project.

All times are GMT -4. The time now is 20:34.