frozen relaxation zones and a moving mesh
 

Hi Niels,

Thank you for all the work you've put into this project.

I am running into issues that seem to be related to using the frozen relaxation zones and a moving mesh. If I reconstruct a certain time step after execution, and decompose it to restart the parallel simulation from that time step, I get errors about not finding "cells_inlet" files on execution. For example:

This message is then repeated for the remaining processors . It seems like perhaps the position / cell info of the frozen zones is not written / read after reconstructing and decomposing?

Greg

Hi Greg,

Out of curiosity: Does it work to restart the code, when you have not been through the reconstruct/decompose step?

Anyway, the cells_<name> lists are only created for time = 0, so there is nothing to be done in terms of restarting, if you have deleted the these lists. Of course, you could take a single time step from 0 and then restart.

I have never used this relaxation shape. May I ask, why you need it?

Kind regards,

Niels

Niels,

Yes, restarting without the recon/decomp step works just fine. Looks like maybe I need to reconstruct/decompose with the -constant flag to get the cell lists to be included. I will report back if this indeed helps.

In this case, the frozen zones are used for a ship moving through waves with the grid translating.

Greg

Hi Greg,

I suppose that you still have the 0-folder in the re-decomposed case? Then follow these steps:

1. Decompose
2. Run (one time step or at least the construction) of the re-decomposed case at t = 0 s.
3. Run the case from the start time that you want to continue from.

This will create your cell-lists and you can subsequently continue the simulation.


Good to know that you find it useful. It is the exact purpose that the frozen zone was designed for (based on a kind request of another user).


Kind regards,


Niels