CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)

 thecfduser December 3, 2009 12:06

Hi all
i have some little questions:
1-can URNANS models predict unsteadiness in flows? like vortex schedding???
2-The average solution for URANS flows is it the same than the solution of steady RANS ???
thanks and regards

 sfalsharif December 9, 2009 19:28

Quote:
 Originally Posted by thecfduser (Post 238615) Hi all i have some little questions: 1-can URNANS models predict unsteadiness in flows? like vortex schedding??? 2-The average solution for URANS flows is it the same than the solution of steady RANS ??? thanks and regards

1. Yes they can, but not with complete confidence. You need to do some validation against experimental or perhaps higher accuracy, trusted simulations to make sure your results are acceptable.
2. Not necessarily. It depends on the problem. There might be non-linear interactions between 'organised' unsteadiness and the random fluctuations that alter the time-mean solution. In such cases, you need to be wary of URANS results.

 thecfduser December 10, 2009 14:42

Thank u very much, but i didt understand too much the second point
In fact, what u are saying is true: URANS will not give a trusted result, but this will be the same for steady RANS.
My question is if we do the same simulation with steady RANS, than with Unsteady Rans and we take the average of the URANS solution, will the 2 solutions be the same????
I know that the mean solution of non-linear equations is not necessarely a solution of the equation , but i am not 100% sure....
Thank u

 sfalsharif December 11, 2009 22:12

Quote:
 Originally Posted by thecfduser (Post 239520) My question is if we do the same simulation with steady RANS, than with Unsteady Rans and we take the average of the URANS solution, will the 2 solutions be the same???? Thank u
Not in general. In fact, if the problem is intrinsically unsteady and you try to solve it with a steady RANS solver, you'll most likely have difficulty getting converged results.
But even if you do get a converged solution using a steady RANS solver, it may not be the same as the time-mean of the solution obtained with an URANS solver. If you have a quasi-steady rate of variation, where the large-scale variations of flow conditions over time are slow enough such that at each instant in time the turbulence is able to re-adjust to the new conditions as if it were at a new equilibrium state, then you might expect the to get agreement between the time-mean and the steady solution. But if not, i.e. if the rate of variation is too fast for the turbulence to adjust to new conditions, then you probably will not have such an agreement.
There can be other complicating factors, such as periodic laminarisation and re-transition in pulsatile pipe flow (happens under certain conditions), for example, which might throw-off the turbulence models in a cyclical way (models are less than perfect in predicting transition). In which case, the time-mean solution you get will be different from the corresponding steady solution without forced pulsations. This is to name one example.

 thecfduser December 12, 2009 17:12

OK thank u very much for ure answer

 All times are GMT -4. The time now is 12:18.