Velocity decrease over a boundar (COMSOL)
 

Hi all,

First ill let you know that I'm very new to CFD; I've only been playing around for a few days.

Anyway, what I'm trying to achieve is the modelling of a bunch of 'turbines' in a fluid flow to see the down stream effects. For my purposes, all these turbines have to do is resist or slow the flow. I've had mixed success with COMSOL getting this working through a number of methods including: Increasing the viscosity at these 'turbines', implementing a resistive volume force (in the NS eqs), and finally, trying to implement a boundary condition at these turbines to slow the velocity predictably.

As I said, I have only had mixed success. The first method, increasing the viscosity, worked pretty well, but it was hard to choose a viscosity that made things correspond to the real turbines, and also there was no velocity dependence of the force in this case...

For this reason the second seemed a better option, since I could use the standard drag equation to give me what I wanted, but the whole using a volume force didn't seem quite right, at least when I calculated the drag force and then divided by the volume of a 'turbine', the area dropped out and the force became negligible…

Finally the simplest option, just reducing the velocity with a boundary condition, while almost the perfect concept (I know how much these turbines should slow the field: 59% (Betz) at best), I can't get it working properly.

I guess this is a bit of an open question, but does anyone have any advice the can offer, maybe (most probably) I'm doing something wrong, or maybe some one with experience can recommend another method or idea?

Thanks in advance,

Tim

Re: Velocity decrease over a boundar (COMSOL)
 

COMSOL is a finite element solver for the Navier Stokes equations, all FEM Solvers use the concept of artificial viscosity to stabilize the solution, ask your support engineer to give you guide lines on how to use that technique (it is a matter of experience, there is no rule of thumb you can rely upon) or if you have a deep pocket, change to a FV solver Be patient and Good Luck