code_saturne User's Forum

Hello,

I'm running LES with internal coupling of solid-fluid conjugate heat transfer. I'm using version 6.0.5. I found in the solid region, the velocity is 0 in the x and y direction, but non-zero in z-direction. A picture showing the velocity distribution over the domain (one slice of the axisymmetric domain, solid region is an inner plate at the upper part of the domain) and a line plot of v_z in the solid region (a vertical line through the solid region, it shows v_z is about 0.0238). I also found in run_solver.log, N_iter for pressure can be as high as 2000 or above. Is that reasonable? cs_user_parameters.c is also attached.

Best regards,
Xiaoxue

Hello,

This is strange, since you have activated the stokes_model->solid option, which should take care of that.

Do you have any non-zero velocity initialization in the domain ? If you have prior movement in the solid region, it may not be removed. Otherwise, do you have a very small variant of the setup you could post here ?

Best regards,

Yvan

Hello Yvan,

It was restarted from a previous run, but I have initialized the velocity in the solid region to be 0. (cs_user_initialization is attached). The non-zero is just in one direction, and velocity is zero in the x and y directions. I've also tried to run it without restarting from a previous run, the non-zero velocity in z-direction still exits but the value is much smaller (around 0.0002). I haven't tried a simpler variant of the setup yet.

Besides the non-zero velocity in the solid, is the N_iter for pressure reasonable? N_iter for pressure is so high that the simulation is quite slow. I don't have such a high N_iter in the case without internal coupling.

Xiaoxue

Hello,

Is the restart done from a different mesh (i.e. not including the solid) or from the same mesh (with solid as fluid) ?
Did you have any issues before the restart.

Otherwise, do you have a coarse variant of the mesh (ideally less than 1M cells, less than 100000 is better) and setup that you could post here so we can try to reproduce the error ?

Best regards,

Yvan