CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   CFX and gravity direction (https://www.cfd-online.com/Forums/cfx/199920-cfx-gravity-direction.html)

dibloff March 19, 2018 10:59

CFX and gravity direction
 
Hello Analysts

I’m running a complex model in CFX with multiphase, rotating domain and gravity. the domain is initialized with air, and then coolant is being introduced through the coolant inlets. The CFX solver will not run in steady state mode if the gravity vector is perpendicular to the axis of rotation. the ANSYS guys confirmed that this is a solver limitation. I was told it’s the same in fluent too. I want to run a steady state analysis compared to transient, because my fluid time scale is ~4 ms, while my thermal time scale would be in 10’s of seconds.
to simplify the analysis I ventured into using a negative pressure boundary @ the outlet, to “lure” the fluid there, but I’m not sure what should be the pressure I would have to use, since it’s 2 phase flow. the volume does not have a breather hole, or other opening on it. so if I apply a negative pressure boundary at outlet it’ll create a relative vacuum and it’ll facilitate the oil flow from the oil inlet(s).
I’ve done a transient analysis to see what the (stabilized) pressure would be @ the outlet (over a 1 sec run time) but it did not converge. I’ve done numerous small scale studies to investigate this but at the end of the day I’m not sure I can believe the numbers.
I’m wondering if anyone has a better idea on how to tackle this problem.
Any suggestion is appreciated,
thanks.

A.

Gert-Jan March 19, 2018 11:18

My fall back scenario is to extent the domain with a knock-out drum where oil and gas separate. And then create a top outlet for gas and bottom outlet for oil.

Opaque March 20, 2018 10:55

You may have misunderstood the support answer, but it is not possible to run a steady state simulation with a rotating domain and the gravity direction normal to the axis of rotation.

It is not a limitation of ANSYS Fluent, ANSYS CFX or any other code. It is not mathematically possible to solver such a problem because it is incorrectly posed.

In the rotating frame the gravity vector is moving opposite to the rotation direction; therefore, the problem is always transient for the observer in the rotating frame.

Hope the above helps,


All times are GMT -4. The time now is 06:36.