[asking]

Hi,

I am running a simulation of a vibrating cylinder subjected to a constant flow rate. The model works fine and the cylinder moves as expected.

I saved the simulation every 100 time-steps and I have problems when I resume it from the last saved time-step in a different computer. Basically, the scaled residuals increased drastically and most of the parameters that I am saving also increase significantly.

From time step 14000 to 14001
Lift: from -0.02 to 0.45
Yplus: from 0.86 to 2.29
Drag: from 0.07 to 9.88

Due to this increment, eventually my model fails due to negative cell volume problem.

I resume the simulation from Fluent going to File - import - case and data, and selecting the last saved time-step. Then, I copy the UDF with its corresponding folder "libgen" and just run the simulations.

Any suggestions would be really appreciated.

Regards

[asking]

I have narrow it down to the UDF. It seems that my problem is almost the same as the one in this thread:

problem restarting dynamic mesh cases

Specifically, the solution seems to be in this response:

Originally Posted by `e`

Not quite. Consider the restoring force from earlier:



If you only want a single DOF such that the restoring force of the spring acts purely in the y-direction, then the force on the cylinder could be calculated with:

force = -k*(DT_CG(dt)[1] - x_equilibrium[1]);

where the DT_CG macro is the centre of gravity vector of the dynamic thread zone and x_equilibrium is the vector where the spring is at equilibrium (defined within the UDF). The UDF manual has examples and details for further clarification and reference.

Note: remember that indexing in C starts from zero instead of one, so DT_CG(dt)[1] corresponds to the second spatial coordinate (y for Cartesian).

But I am not sure how to implement it in my UDF, here is my code that solves the equation of motion

for (i=0;i<3;i++)
{
cg[i]=DT_CG(dt)[i];
vcg[i] = DT_VEL_CG(dt)[i];
}

/*Numerical methods*/
/*Numark-beta*/
real beta = 0.25;
real gamma = 0.5;
real term0 = (1/(beta*dtime*dtime))*(mass+ad_mass) + (gamma/(beta*dtime))*c;
real term1 = (1/(beta*dtime))*(mass+ad_mass) + ((gamma/beta)-1)*c;
real term2 = ((1/(2*beta))-1)*(mass+ad_mass) + dtime*((gamma/(2*beta))-1)*c;

real Keffx = k + term0;
real Reffx = fx*(water_depth/length) + term0*cg[0] + term1*vx + term2*ax;
real Keffy = k + term0;
real Reffy = fy*(water_depth/length) + term0*cg[1] + term1*vy + term2*ay;
Message("Velocity: %f \n",vy);

dx = Reffx/Keffx - cg[0];
dy = Reffy/Keffy - cg[1];

x += dx;
y += dy;

real vprev_x = vx;
vx = (gamma/(beta*dtime))*dx + (1-(gamma/beta))*vx + dtime*(1-(gamma/(2*beta)))*ax;
ax = (1/(beta*dtime*dtime))*dx - (1/(beta*dtime))*vprev_x - ((1/(2*beta))-1)*ax;

real vprev_y = vy;
vy = (gamma/(beta*dtime))*dy + (1-(gamma/beta))*vy + dtime*(1-(gamma/(2*beta)))*ay;
ay = (1/(beta*dtime*dtime))*dy - (1/(beta*dtime))*vprev_y - ((1/(2*beta))-1)*ay;

It seems that when I resume my calculations, some values go back to zero and not the last values they had in the last saved time step.