CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   ERCOFTAC testcase, pipe flow by Anwer (

Olof Daunius April 26, 1999 05:00

ERCOFTAC testcase, pipe flow by Anwer
Hello everyone,

I wanted to benchmark our codes on a pipe flow testcase and found just what I wanted in the ERCOFTAC database: "Swirling and non-swirling flows in curved pipe -- experiments by Anwer"

However, I have a hard time deciphering the measurement data. Firstly, there is no way of knowing the xyz coordinate system orientation. Secondly, the inflow velocity profiles have large components of velocity in all directions eventhough it is supposed to be fully developed straight pipe flow..

To summarize, I'm pretty confused. Has anyone tried this case and managed to understand the measurement data, please post!

If you know where to contact Dr Anwer, I would be interested too.

/Olof Daunius Volvo Car Corp., Engine Division.

reyman April 27, 1999 16:11

Re: ERCOFTAC testcase, pipe flow by Anwer
Olof, I too have attempted validation from the benchmarked codes archived in the ERCOFTAC database. The procedure I followed: (1) collected the experimental data from the archive and organized it (2) found all published papers associated with the data (3) built a virtual model (hand draw the flow scenario with portions of the data as labels for various locations within the experimental flow field) use the info from the archive and from experimental papers to help guide the recreation of the original experiment (4) built a mesh (5) set boundary conditions (6) run model (7) adjust mesh (8) run again, compare results, adjust mesh, fine tune upwinding, adjust turbulence parameters, adjust mesh, etc, you know the rest. My advice is to pick a very, very simple benchmarked case that has several published papers associated with it and keep all the data very well organized then do your best to use this information to rebuild the thinking process of the original experimenters. Ideally you will either have success or know early enough when to quit and look for another benchmark. cheers, reyman

All times are GMT -4. The time now is 00:25.