README.md 3.49 KB
Newer Older
ylan's avatar
ylan committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76
## 127602 (non-touching) pebbles in an annulus cylinder (Last updated: Jan. 27, 2021)
ann127602 N, Version 3.2, E39,240,190

- New features (after Aug, 2020):
  - fixed edge insert
  - mesh smoother+optimizer in Nek5000
  - tol adjustment in edge collapse

- Notes
  - DEM data from Yiqi (StarCCM++?) 
  - smaller edge collapse tol depends on geom + few local locations
  - Geometries, pebble locations are stored at `ann127k.dat`

- Geometry and BC in re2 file:
  - linear mesh, no curved sides
  - Sphere distribution: free-free (selected from some range in Z, free = non-flat)
  - Sphere radius = 1.0 (Wall)
  - Cylinder (inner) radius = 24.256 (W01, convert to Wall in production run)
  - Cylinder (outer) radius = 87.895 (W02, convert to Wall in production run)
  - Zbot of the cylinder = -27.010 (W03, convert to Inflow in production run)
  - Ztop of the cylinder = -29.031 (W04, convert to Outflow in production run)
      ```
        xyz min    -87.895      -87.895      -27.010
        xyz max     87.895       87.895       29.031

        mesh metrics: (linear mesh)
        GLL grid spacing min/max    : 6.77E-04 3.52E-01
        scaled Jacobian  min/max/avg: 1.70E-02 9.99E-01 3.89E-01
        aspect ratio     min/max/avg: 1.05E+00 9.23E+01 1.58E+01
      ```

- Mesh file (`mshmesh_retry2.fld+X`)
  - lx1=3 GLL points projected onto curved sides
  - no issues for interpolation onto lx1=8 and mid-multigrid levels at lx1=4
  - no issues for lx1=2 corase grid

- Production run
  - Use re2 for boundary conditions, but use restart file for the curved mesh
  - Time is NOT set to 0 when reading the mesh file, physical time = simulation time - 20.0
  - We use dimless scale to run Re10000 with vz=1.0
  - NekRS version (old)
    ```
      commit 636fe6281c1db7e45ef0a223e16745bd0fb95a16
      Author: Malachi <malachi2@illinois.edu>
      Date:   Mon Nov 2 09:08:19 2020 -0600
  
      Time all reduce and copy to device in ellipticResidualProjection (#168)
    ```

  - Runs (check `dat_logfiles`):
    ```
     - run1 
         step= 1     t= 2.00000100e+01  dt=1.0e-05  C= 0.07 
         step= 1170  t= 2.00117000e+01  dt=1.0e-05  C= 0.07  U: 2  V: 2  W: 2  P: 15  eTime= 7.32e-01, 8.82609e+02 s
     - run2 
         step= 1     t= 2.00100500e+01  dt=5.0e-05  C= 0.34 
         step= 6590  t= 2.03395000e+01  dt=5.0e-05  C= 0.35  U: 2  V: 2  W: 2  P: 12  eTime= 7.06e-01, 4.36998e+03 s
     - run3 
         step= 1     t= 2.03101000e+01  dt=1.0e-04  C= 0.72 
         step= 3960  t= 2.07060000e+01  dt=1.0e-04  C= 0.84  U: 2  V: 2  W: 2  P: 45  eTime= 2.04e+00, 5.60561e+03 s
     - run4 
         step= 1     t= 2.06103000e+01  dt=3.0e-04  C= 2.59 
         step= 4090  t= 2.18370000e+01  dt=3.0e-04  C= 3.66  U: 3  V: 3  W: 3  P: 120  eTime= 4.87e+00, 1.33436e+04 s
     - run5 start using Pr-proj=F
         step= 1     t= 2.18102000e+01  dt=2.0e-04  C= 1.91
         step= 5740  t= 2.29580000e+01  dt=2.0e-04  C= 1.92  U: 2  V: 2  W: 3  P: 56  eTime= 2.29e+00, 1.31354e+04 s
     - run6 
         step= 1     t= 2.28102000e+01  dt=2.0e-04  C= 1.89 
         step= 5580  t= 2.39260000e+01  dt=2.0e-04  C= 2.84  U: 3  V: 2  W: 3  P: 54  eTime= 2.45e+00, 1.31585e+04 s
     - run7 
         step= 1     t= 2.38101000e+01  dt=1.0e-04  C= 1.21 
         step= 6430  t= 2.44530000e+01  dt=1.0e-04  C= 0.95  U: 2  V: 2  W: 2  P: 43  eTime= 2.07e+00, 1.30440e+04 s
     - run8 
         step= 1     t= 2.44102000e+01  dt=2.0e-04  C= 2.10 
         step= 5090  t= 2.54280000e+01  dt=2.0e-04  C= 1.89  U: 2  V: 2  W: 3  P: 55  eTime= 2.76e+00, 1.31923e+04 s
    ```