How we will be sure about fluent.

Is fluent properly validated for the models and options it offers, if so, then where we can find these geomery , case and data files to give refernce in our work

regards Sijal Ahmed Memon

It is a generally adopted approach among developers of commercial CFD codes to use methods that have been already published and validated in the open literature. Their implementation is certainly tested and you can see various benchmarks and successful applications on their pages. But case and data files (at least generally) are not available. You may give it a try for a concrete case through technical support though.

However, you can never be sure about any commercial CFD code. It is simply out of question for the developers to test all admissible combinations of all options in such a code. In short: Bugs are omnipresent. That's a valid Murphy's law. And the same impplies unfortunately also that Fluent implementation of a validated model will not necessarilly behave exactly as one would expect. Of course not only due to bugs, but more often due to users deploying models and methods in combinations not envisaged by the developer of the model/method.

Hope this helps,

George

does it imply that we cannot use it for basic research

Of course you can, I just wanted to give a general warning, as I have several recent experiences with bugs and unexpected behaviour of some solutions in Fluent.

then up to where we can believe in fluent solution.

if fluent has bugs then why people are using it.even many big compienies using it

There is simply no cut and dry answer to this. If you are using a commercial code such as Fluent for fundamental research, there is no alternative but to carry out experiments to validate the numerical solution. Personally i find it very tough to blindly accept what a commercial code spits out as a solution. But when validated, the amount of information that can be obtained is alot more than from experiment, simply as one cannot measure each and every parameter

do u want to say that even the researcher who make their own code for their particular case cannot validate thier code for other parameters expect than thier own area of interest

Ag George, had mentioned earlier, the implementation of the numerical scheme that solves the Navier Stokes or the Euler equations need not be the same if donw by two different ppl. Modifications to models to enable quicker solutions are startard in commercial codes.

Ny numerical code requires validation regardless of what the area of interest is. My advice would be to find an experimental case similar to the experiment that you are trying to model, and then run the code to try and match the results, whis will atleast give you confidence.

wat u suggest i any one want to simulate the turbine rotor

my field of experiance is not turbomachinery, let me say that right upfront. I am mostly into high speed gas dyamics, shockwaves, and jet flows.

I find it hard to imagine that there are no validated cases for turbomachinery, as fluent sems to sell the code as a turbomechinery solver. i would advice you to try and get a hold of your local agent who should be able to provide you with literature from Fluent regarding this.

thanks i will try to contact fluent

regards Sijal Ahmed Memon

As a CFD engineer, problems are an everyday occurance. If there was a code that was "perfect" then there wouldn't be Fluent, Star-CD, or CFX, and numerous other codes. It's a matter of application and verification of CFD results to experimental results that clarify validity. For example, one viscous model may work better than another. You just need to do your research in what codes are best for turbomachinery. Though my area of research is combustion analysis and Fluent is really good at this. From what I have seen from presentations from Fluent that Fluent is good for turbomachinery applications. You just need to sit down set up a case from a experiment and make that your validation and then move forward.

this is the problem , in turbomachinery i have no test case, as turbomachinery is the area of interest for 1.giant companies: for commercial products 2. the big goverments : for defence purpose.

i am sure that no one from both is going to share and single thing from there expermental setup

so wat to do?

See http://www.fluentusers.com/fluent6/d...d/main_pre.htm

FLUENT 6.0 Validations:

1. Flow in a Rotating Cavity

2. Natural Convection in an Annulus

3. Laminar Flow Around a Circular Cylinder

4. Flow in a 90 Planar Tee-Junction

5. Flows in Driven Cavities

6. Periodic Flow in a Wavy Channel

7. Heat Transfer in a Pipe Expansion

8. Propane Jet in a Coaxial Air Flow

9. Non-Premixed Hydrogen/Air Flame

10. 300 kW BERL Combustor

11. Flow through an Engine Inlet Valve

12. Turbulent Flow in a Transition Duct

13. Solid Body Rotation with Central Air Injection

14. Transonic Flow Over a RAE 2822 Airfoil

15. Mid-Span Flow Over a Goldman Stator Blade

16. Compressible Turbulent Mixing Layer

17. Scramjet Outflow

18. Turbulent Bubbly Flows

19. Adiabatic Compression and Expansion Inside an Idealized 2D In-Cylinder Engine

Frank

thanks a lot

regards sijal ahmed memon

But these all files are inluded in fluent cd for validation one needs the expermental data

If you look in the solution files http://www.fluentusers.com/fluent6/d.../solutions.htm you'll find experimental data.

Frank

If you can get hold of it, the AGARD Adivsory Report No. 275 "Test Cases for Computation of Internal Flows in Aero Engine Components" July 1990, contains a wealth of experimental 2D and 3D test cases for cascades, compressors and turbines (single row and single stage). There is enough information here to validate FLUENT (and any other code) for both steady state single row calculations and unsteady single stage calculations.

can u send me copy of this report , i will gratful to u bye

I could, but as this is a very comprehensive report, it is several hundred pages long (including the annexes). You would be better contacting the AGARD and ordering a copy.