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
## 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)
21
  - Ztop of the cylinder =  29.031 (W04, convert to Outflow in production run)
ylan's avatar
ylan committed
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
      ```
        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
    ```