CFD simulations in Aircraft industry 1

[user277418]

Dear forumers

I have a general question about CFD simulations. I am a Structural Engineer in a field of Construction. Recently I see a lot of video presentations about CFD simulations for special structures like for example high-rise buildings. A main message of the presentations is CFD (digital wind tunnels) will replace real wind tunnels. Was it already happen in Aircraft engineering?

PS: asking here, because many technologies in Construction engineering were taken from Aircraft/Aerospace after many years they were used there already.
PPS: I don't design skyscrapers. Just interesting.

BR

 

[rb1957]

CFD is good on it's own for design, then you test. With enough testing you gain confidence in your CFD models (so you'd allow small changes to be CFD'd only).

I think you are looking at ppts from CFD salesmen ! Digital wind tunnels are a thing of their imagination, and not for things that Really matter. Cars may permit this as aerodynamic loads are at most an annoyance, except in race cars where downforce is critical. But structures (and aircraft) need good aero loads. Sure some can use CFD loads directly, but only if there is a body of test data showing that the CFD is predicting correctly (ie someone else did the testing).

another day in paradise, or is paradise one day closer ?

 

[GregLocock]

Actually we do have quite a lot of aero requirements for cars. Recently they did a round robin, taking some test items into every significant wind tunnel around the world that was in the reporting group. My trust in wind tunnels declined somewhat.

The less obvious aero is minimizing yaw response for crosswind stability, and minimising changes in lift at each axle, for handling. Also a lot of CFD is done on the engine bay as it is a surprisingly large contributor to drag.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Sparweb]

Recently they did a round robin, taking some test items into every significant wind tunnel around the world that was in the reporting group. My trust in wind tunnels declined somewhat.

Was this published? Would you be willing to tell me where to find/read results from these studies?

user277418,
The safety rules for aircraft are specific on the subject of testing versus analysis. In aerodynamics, the uses of analysis are limited, unless the calculations can give results that are validated by test results. Even wind tunnel tests are "model" tests that need real-world validation. There are scale issues, and the effects of the walls that don't exist in reality.
If you're being told anything else, it's just a pipe-dream.

http://www.sparweb.ca

 

[GregLocock]

I'll contact my aero guy but I am pretty sure it won't be published. One of the obvious issues other than size is the way that roads have been treated.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[user277418]

In aerodynamics, the uses of analysis are limited, unless the calculations can give results that are validated by test results

In Construction, in general, it should be the same.

Even wind tunnel tests are "model" tests that need real-world validation

That is one thing the salesmen talk about. Each real wind tunnel have to be validated too. Hence why not to validate a digital wind tunnel. It is much cheaper.
And also a 'real-world validation'. What do you mean? Physical tests?


What all you guys think about this? Is not an advertising >>> Link

 

[Sparweb]

Hence why not to validate a digital wind tunnel. It is much cheaper.

Not necessarily. CFD software - the kind that gives you repeatable credible results (not freeware on the internet) is very expensive. You can buy time at a wind tunnel for a comparable price.
If you have one project, you don't build a wind tunnel, you rent time in one. There are many around North America, Europe, Australia and goodness knows there are probably many elsewhere in the world, too. A lab's wind tunnel will be calibrated by the laboratory owner before you arrive. If you contact an adept model-builder, you can start testing in weeks.

I've done some CFD models. They take a while to build, too, and your first runs are usually slow glitchy garbage that refuses to converge. Almost every time it seems like I promised the boss to be done on Friday, and though I got the model to work on Thursday, the GIGO is so bad that it takes until Sunday to debug, and the final run doesn't finish until Monday. After all that scramble I won't yet have anything to show validation against. So I've dropped 30k for a pack of CD-ROMS, sat in a 5k training course for 2 weeks, and been stuck in my chair for a month. All to get a calculation that doesn't prove much without some kind of real-world test anyway.

And also a 'real-world validation'. What do you mean? Physical tests?

Yes. There has to be a test that demonstrates that the conditions, assumptions, and predictions of the model (tunnel or CFD) has a trustworthy relationship with reality. Without the test, any error in your model could cause you to be vastly under-predicting a safety condition or a load that leads to failure.

http://www.sparweb.ca

 

[GregLocock]

OpenFoam is an industrial strength, free, CFD program. I know of three companies that use it.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?