Skip to content

Some changes in project files for 3D toy model

wenqing requested to merge wenqing/dgr:project_improve into 2D_pyBCs_smk
  1. added project files for TH simulation,

  2. some data correction in 3D/simTHM_init.prj and 3D/simTHM_galaciation.prj, e.g. geometry sizes in function parameters, the Dirichlet BC functions from my TH tests, time stepping, and etc. The stabilization is applied for the init simulation as well.

  3. updated 3D/simH_1.prj, and changed its name to 3D/simH_galaciation.prj. Changed simH_1 to sim_H in Snakemake, which means the process name for snakemake becomes as simH for the hydraulic simulation, e.g.:

snakemake -F --profile profiles/default results/processes/3D/Ton-Nord/simH/glaciation.prj

Issue: For THM simulation, even with coarse spatial discretization, the computation takes long time. Therefore the parallel computing is required. However I found a bug in ogs in the restart computation with parallel computing: the integration point stress and strain data cannot be read after running RemoveGhostData for multiple vtu files, and they are not considered in XDMF output. Therefore the restarting computation with parallel computation for M related processes only works without initial stress. The bug will be fixed. So far only THM/init can be run with parallel computing. With a spatial resolution of 150 m and 20 partitions, the parallel computation of THM/init on eve cluster took 2735.11 s: THM0

Edited by wenqing

Merge request reports