Simulating a 2D turbine as an actuator disk
 

I'm trying to simulate a 2D axisymmetric wind turbine duct with a pressure drop along the turbine using the actuator disk boundary condition. However, running the analysis immediately gives an error that SU2 has diverged. Changing the boundary condition to a wall allows the solver to run smoothly.

The mesh was made in Pointwise by making a connector to represent the turbine and setting one side to disk_in and the other side to disk_out. I've tried many different combinations of the settings for the actuator disk boundary condition, but none seem to work.

Could someone further explain the actuator disk boundary condition? What do the rotation angles mean? Is this boundary condition usable for a 2D compressible mesh? Do root and tip radius mean the inner and outer radius of the disk? Are there other ways to model momentum sinks?

Hi,
I've been working with the actuator disk feature of the last few months (I apologize if this reply comes late). I've found that you cannot use a 2d approach, it must be 3d. When using 2d, the actuator disk doesn't "turn on", flow just goes through as if the disk were not present.

If the solution diverges, it is likely a mesh/configuration file issue unrelated to the actuator disk setting. If you look at the configuration template, there are a number of actuator disk commands, but some important ones are 'ACTDISK_DOUBLE_SURFACE'. If you've already split the surface (sounds like you have), then set this to 'YES'. Otherwise specify 'NO' and use the ACTDISK_SU2_DEF command (set to 'YES') and SU2 will create a back side. There is an example configuration file in TestCases/rans/propeller, but it does not mention these options. I think SU2 automatically assumes the disk needs to be split.

I haven't had to mess with rotation angles and other features, those may not be the problem.

Good luck,
Nat

Hi all,



do you know if now is it possible to use a 2D actuator disk condition over a line? If this is not possible, do you know from where it is convenient to start in the code to implement it?



Thanks,


Andrea

anyone, is it still not possible to use pressure jump boundary condition on line (for 2D cases)