Problem with convection term in linearized Navier-Stokes equations
 

Hi guys,

i'm trying to write a solver for Linearized Navier- Stokes equations.
I modified icoFoam solver, splitting the convection term in two parts:

fvm::div(phi_zero, U) + fvc::div(phi,U_zero).

U_zero is the base flow, phi_zero is defined as phi_zero = linearInterpolate(U_zero) & mesh.Sf().

I tried to run a simulation around a cylinder and the result seems correct.
Now, to prove that what i have written is correct, i should run two separate simulations:

1) The first one with fvc::div(phi,U_zero) = 0;
2) The second one with fvm::div(phi_zero, U) = 0;

In both cases i should re-obtain a solution equal to the base flow U_zero.
The first simulation works well but the second one doesn't.
It seems that the fvc term without the fvm one causes some problems.

Could anyone please help me?

Regards,

Andrea

hi, Andrea!

Have you solved this problem? I'm facing a similar question now.

hi kkpal,

unfortunately I haven't solved this problem yet. Actually I'm not working on it anymore. It was the topic of my Master's Thesis, but me and my supervisor decided to change it because it would have required too much time to solve it (the graduate session was too near and my project too big).
I hope my threads can help you.
Let me know if you come up with a solution.

Best Regards,

Andrea

Hi, Andrea!
I found several papers concerning this topic, especially this one:
http://www.sciencedirect.com/science...70963813001879

Currently I'm trying to running some validation cases according this method.

have you succeeded in implimenting Linearized Navier- Stokes equations in openFOAM?