Learning Forum

[Dimitrios Lamprakis]

Hi, I am trying to simulate a 2D turbine cascade (I generate the geometry and the passage domain using ANSYS Design Modeller) and obtain flow deviation and losses. I want however to obtain the fully mixed-out flow.

Instead of having a very large downstream domain to enable mixing of the flow, I was thinking of either doing that in FLUENT, using the mixed-out mixing plane or in CFX using stage mixing plane - the stage average velocity option as the constant total pressure to my understanding is not introducing total pressure losses. To do that I am creating a small domain downstream the blade passage to create an interface between the 2 and impose a mixing plane (see Fig.).

I understand that in both cases the domain has to be 3D (to enable mixing planes in FLUENT and CFX doesnt run in 2D), but in FLUENT the mixed-out option doesnt appear(doing that through the command window - legacy options). In CFX, I cannot find in the theory guide how the total pressure and constant velocity mixing planes are implemented (e.g. free vortex, forced vortex etc.) to decide whether or not the formulation would be appropriate to model mixed-out flow. I am also planning of doing something similar to a 2D radial impeller.

My questions are: Can I enable mixed-out plane in fluent? What is the formulation that cfx mixing planes (cont velocity) uses? Can I run in purely in 2D in Fluent? Is there a better way to do this? Would also that be appropriate for radial impellers?

 

[CFD_Friend]

Hi Dimitrios,

CFX's formulation for mixing planes involves circumferentially averaging the fluxes in bands and transmitting the average fluxes to the downstream component. This is described in the CFX Best Practices Guide for Turbomachinery, where the Stage (Mixing-Plane) model is discussed. 

11.4.2. Stage (Mixing-Plane) (ansys.com)

[Dimitrios Lamprakis]

Hi, thanks for the quick reply. I have read that, although the formulation unlike FLUENT is not given. The formulation sounds like the mixed-out approach of Amecke where the flow is split up into notional series of thin radial layers of 2D flow (see Analysis of Averaging Methods for Nonuniform Total Pressure Fields, ASME journal of turbomachinery, page 4 right column).

Are you aware if thats the mixing plane formulation used? I assume the constant velocity option will introduce total pressure losses due to mixing out of the flow?

Do you know if in Fluent I can use a mixing plane in 2D simulations and why the mixed-out plane is not showing on the gui when accessing it through the command window or if there is any other way to activate it?

[Dimitrios Lamprakis]