CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Siemens (
-   -   Accuracy degree for an airflow simulation (

Paris Hdez February 27, 2006 04:03

Accuracy degree for an airflow simulation
hi! I am trying to simulate the flow in a ship's hold. At one area of this hold some measurements of wind velocity where taken. When comparing the air speed at this same points in my simulation with those measured the differences are of the order of 30%. Should I content myself with these results? Which grade of accuracy it is possible to reach with CFD's? The dimensions of the model are really big (180x20x7m) this does not allow me to make cells so small as I would like to, though I think the actual dimensions of the cell (0.5 the biggest ones and 0.125 the smallest ones) should be enough to get decent results. Please help me, I've allready tried it all and I am really going mad. Thank you very much, Paris

kris February 28, 2006 02:31

Re: Accuracy degree for an airflow simulation
you need to reveal your information of STAR-Cd setting for your CFD. for example using a turbulent model, properties, or Boundary condition...

In your case, 30% error is big.

in CFD, it is important to implement appropriate turbulent model.

Paris Hdez March 9, 2006 09:21

Re: Accuracy degree for an airflow simulation
hi! I used COMET for my calculations. I have tried it with all the possible turbulent models that the program offers, but mainly with k-e, but also k-w and I have just once tried it with RSM but somehow it didn't converge.The results were still the same, in some cases even worse.

The thing is that the measurements were taken over a period of 180 s, and values were taken every 0.1 s. The velocity which has been taken was the average. However I'm not sure if I could rely on such results as the values vary much the difference between the greatest and the smallest values are really big. That's why I have also tried to run a transient calculation, this one showed very small variations in time from the steady state results. Thus the flow simulation being apparently not so turbulent like it is on the real case.

All times are GMT -4. The time now is 21:03.