CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Oscillatory cylinder in a static flow medium (http://www.cfd-online.com/Forums/fluent/95257-oscillatory-cylinder-static-flow-medium.html)

teopetre December 11, 2011 16:44

Oscillatory cylinder in a static flow medium
 
As you can read in the title, I have a box inside of whom I put a cylinder which is moving vertically according to a harmonic function, written in a simple udf as follows:

#include "udf.h"
/*y(t)=a*sin(2*pi*f*t) */
DEFINE_CG_MOTION(oscilation,dt,vel,omega,time,dtim e)
{
Thread *t;
face_t *f;
NV_S(vel,=,0.0);
NV_S(omega,=,0.0);
if(!Data_Valid_P())
return;
t=DT_THREAD(dt);
vel[1] = 10*cos(10*time);
}

For a 2D model, I've tried both smoothing+remeshing and layering, but still the same error, from the first time step of motion, the existence of negative volume, preceded of course by the warning of left handedness(when I use an axisymmetric model). returning to our 2d model(not axi!), the cylinder wall was set as rigid body, and the 4 boundaries of my domain were set as deforming(for smoothing and remeshing) and for layering those 4 were set as static. Moreover, Fluent don't allow me to access the info mesh scale for the boundaries of the box, giving me the overflow error.
Has anyone idea what's wrong and what should be to do?
Besides I've tried both uniform and non-uniform grids, with both an inviscid model and a k-epsilon model.

Many thanks in advance!

teopetre December 16, 2011 18:36

Case resolved
 
It was just a matter of time step choice. Another question would be, how realistic would be a fluid-structure interaction analysis with Fluent, in my case, if I put a velocity bc on the left boundary of the box(square) and the cylinder moves with the harmonic law. Had anyone studied something like this?
Sincerely yours,
Petruţ.

karthickeyan December 20, 2011 01:31

reply
 
hi friend
u can use layering for hex mesh and remeshing for tet mesh it depends on time step u given it will show the error. if time step size is greater than ur cell size it will show negative volume error .send me cas file i will work on it

Quote:

Originally Posted by teopetre (Post 335453)
As you can read in the title, I have a box inside of whom I put a cylinder which is moving vertically according to a harmonic function, written in a simple udf as follows:

#include "udf.h"
/*y(t)=a*sin(2*pi*f*t) */
DEFINE_CG_MOTION(oscilation,dt,vel,omega,time,dtim e)
{
Thread *t;
face_t *f;
NV_S(vel,=,0.0);
NV_S(omega,=,0.0);
if(!Data_Valid_P())
return;
t=DT_THREAD(dt);
vel[1] = 10*cos(10*time);
}

For a 2D model, I've tried both smoothing+remeshing and layering, but still the same error, from the first time step of motion, the existence of negative volume, preceded of course by the warning of left handedness(when I use an axisymmetric model). returning to our 2d model(not axi!), the cylinder wall was set as rigid body, and the 4 boundaries of my domain were set as deforming(for smoothing and remeshing) and for layering those 4 were set as static. Moreover, Fluent don't allow me to access the info mesh scale for the boundaries of the box, giving me the overflow error.
Has anyone idea what's wrong and what should be to do?
Besides I've tried both uniform and non-uniform grids, with both an inviscid model and a k-epsilon model.

Many thanks in advance!



All times are GMT -4. The time now is 14:37.