High-freq oscillations in residuals
 

Hello there,

I am wondering if someone could help me. I am running a transient simulation and encountering oscillations, particularly in the continuity and energy residuals within a single timestep, when other timesteps are fine. After a few timesteps, these oscillations disappear and return to relative normalcy.

Advanced solution control options are initially left to their defaults (F-Cycle), but I've also variously tried V-Cycle multigrid with and without BCGSTAB stabilization on Flow and Energy.

I've also gone into the text-interface options /solve/set/amg-options/amg-gpgpu-options and played with using different solvers (FGMRES PBICGSTAB AMG), pre-/post-sweeps and smoothers (ILU GS JACOBI).

Thanks in advance!

http://i193.photobucket.com/albums/z...psqxsarzvc.png

The second to last timestep showed the beginning of the oscillations with no changes from the start of the simulation and was interrupted. There is nothing physically going on that should cause the oscillations. The Energy under-relaxation factor was reduced from 1 to 0.75 and Density from 1 to 0.9 in the last shown timestep.

FYI: Resolved by combination of remeshing and stopping the transient simulation before the timestep of observed oscillations/flatlining and reducing underrelaxation factors. Using hybrid initialisation with proper patching also helped overall convergence greatly (0.5X number of iterations to complete entire transient simulation run of 512 timesteps).