[EGNIR]

I am trying to make flap motion for wavemaker. I have Z axis as my flap and X axis for forward motion. I tried to make a code for flap motion although the when tried mesh preview, the whole flap zone is moving forward. I want the bottom coordinate to be stationary and upper coordinate to move from - X to X axis.

#include "udf.h"
#include "dynamesh_tools.h"

static real period = 1.0;
static real amplitude = 0.1;
static real pi = 3.14159265358979323846;
static real depth = 0.5;
static real s = 0.1; /* define the stroke */


DEFINE_CG_MOTION(oscillate,dt,vel,omega,time,dtime )
{
/* define the variables */
Thread *t;
face_t f;
real dv, ang_freq;
real x[ND_ND];
real z;
/* get the thread pointer for which this motion is defined */
t = DT_THREAD(dt);
ang_freq = 2 * pi / period;


/* loop over each face in the zone to create an array of data */
begin_f_loop(f,t)
{
/* set x-component of velocity */
F_CENTROID(x,f,t);
z = x[2];
vel[0] = ((z+depth)/depth)* s * cos(ang_freq * time)/4 ;
}
end_f_loop(f,t)

Message("time = %f, x_vel = %f\n", time, vel[0]);
}

[vinerm]

The macro used, as its name implies, is meant for moving CG of a complete zone. For flapping, I suppose you wish to move different nodes by different displacement values. Therefore, you should use DEFINE_GRID_MOTION. If the flap is rigid, then you may not require deforming mesh; moving mesh might do the job.

[EGNIR]

Originally Posted by vinerm

The macro used, as its name implies, is meant for moving CG of a complete zone. For flapping, I suppose you wish to move different nodes by different displacement values. Therefore, you should use DEFINE_GRID_MOTION. If the flap is rigid, then you may not require deforming mesh; moving mesh might do the job.

I tried making it with nodes although i used CG motion ,

here is the code,

#include "udf.h"
#include "dynamesh_tools.h"
DEFINE_CG_MOTION(oscillate,dt,vel,omega,time,dtime )
{
Thread *tf = DT_THREAD (dt);
face_t f;
Node *z2;
real NV_VEC (velocity), NV_VEC (axis);
real NV_VEC (origin), NV_VEC (rvec);
real velc;
real freq;
real amp;
real w;
int n;
real s;
real depth;
real z1;
/* set deforming flag on adjacent cell zone */

SET_DEFORMING_THREAD_FLAG (THREAD_T0 (tf));
freq=1;
w=2*3.141592654*freq;
s= 0.2 ;
depth =0.5;
z1= 1;
velc = ((z1+depth)/depth)*s* cos(w* time)/4;
NV_S (velocity, =, 0.0);
NV_D (axis, =, 0.0,1,1);
NV_D (origin, =, 0.0, 0.0, 0.0);/*center of gravity for my flap*/

begin_f_loop (f, tf)

{
f_node_loop (f, tf, n)

{
z2 = F_NODE (f, tf, n);

if (NODE_Z (z2) < 1.1 && NODE_POS_NEED_UPDATE (z2))
{
velocity[0] = NODE_Z (z2)*velc;

NV_V_VS (rvec, =, NODE_COORD (z2), +, velocity, *, dtime);
NV_V (NODE_COORD (z2), =, rvec);
}
}
}
end_f_loop (f, tf);
Message("time = %f, x_vel = %f\n", time, velocity[0]);
}

although it is moving like a flap the nodes on the edge are fixed and the rest are moving, I want the whole zone to move. I have attached a flap motion picture for reference.

[vinerm]

I am not sure if I completely understand your requirement but I hope the following is helpful

1. CG_MOTION is used for the rigid body motion where the motion is specified by the user and the nodes or cells of the rigid zone maintain their relative position. Fluent allows this macro to be used if rigid body is selected under the Type in Dynamic Mesh Zone definition. A flap may be governed by this if all its boundaries maintain their shape and size.

2. GRID_MOTION is used for the zone that may have relative motion between its nodes or cells. A flap that deflects like a cantilever beam with deflection being a nonlinear function of distance from fixed edge is governed by this macro. If this is the requirement, you should NOT use CG_MOTION and switch to GRID_MOTION.

[EGNIR]

thanks my requirement was first. I have a rigid body for flap. Now I have made the flap movement possible also my wave elevation are not regular. They are getting damped with time. 2m waveprobe.PNG

here is the code for flap motion.

#include "udf.h"
#include "dynamesh_tools.h"

real pi, freq, phase, w, AVel, A1, s,depth,z1,y,x[ND_ND],velc,A2;

DEFINE_CG_MOTION(wallmotion, dt, vel, omega, time, dtime)
{
Thread *t;
face_t f;
NV_S(vel, =, 0.0);
NV_S(omega, =, 0.0);

/* Defined Constants*/
pi = 3.141592654;
freq =1.458;
s= 0.2 ;
depth =0.5;

/*Define Wall Motion via. Angular Velocity */
t = DT_THREAD(dt);
w=2.0*pi*freq; /*angular freq in radians*/
begin_f_loop(f,t)
{
F_CENTROID(x,f,t);
y = x[2];
velc = ((y+depth)/depth)*s*w*cos(w* time)/4;
AVel = velc;
omega[1] = -AVel; /*Assign AVel to omegaz */
}
end_f_loop(f,t)
Message("time = %f, x_vel = %f\n", time, omega[1]);
}

[vinerm]

Few important points

1. x[2] implies z-coordinate. You can call it y but within a case, it will address z-coordinate

2. omega[1] is velocity about y-axis.

3. Rigid body implies no relative motion across space. But by using omega[1] inside the face loop implies each face has its own omega. That will distort the geometry leading to a conflict of the body being flexible.

[EGNIR]

yes I wanted z coordinate as that is my flap length. In my case the Z coordinate shows the height of the wave tank, X shows the lenght and y shows the width. That's why I choose omega [1] so that my flap moves accordingly.

According to your 3 point, can it be the reason for my irregular wave??

I tried omega[1] outside the face loop, it gives negative error volume. I tried the time step according to the 1/2 of the lenght scale. So is it because my omega[1] was not inside the face loop??

[vinerm]

With omega inside face loop, every node on the boundary will be rotated by a different value of omega. This may cause drastic mesh deformation very close to the flap and negative volume may result. Do note that negative volume may result due to any motion if either the mesh is not of good quality initially or the setup for deformation is wrong.