コンテンツにスキップ

forest-spheric-dyn-mixed

英語 | 日本語

« Home · « 自由表面

SPHERIC Test 2 -- 3D dam break against an obstacle box (Kleefsman et al. 2005, MARIN experiment; J. Comput. Phys. 206:363) on the DYNAMIC interface-following mixed-resolution free-surface path -- the Stage-4 free-surface AMR stepper (freesurface_dynamic_mixed.hpp) applied to the violent SPHERIC run-up. The single-phase free-surface LBM (free_surface.hpp; Koerner/Thuerey VOF + MRT + Smagorinsky LES + free-slip tank walls) runs at COARSE (full tank) + FINE (a 2x bottom slab) simultaneously; the fine slab DYNAMICALLY FOLLOWS the interface (mixedfs_dynamic_regrid) and the coarse/fine seam is FLUX-MATCHED each coarse step (Berger-Colella reflux), so total fill-mass is conserved even when the collapsing reservoir / far-wall run-up pushes the seam into fluid.

概要: SPHERIC Test 2 を動的混合解像度 AMR で解く。受圧面の空間平均を取ることで単点スパイクを除いた圧力を評価できる。

Geometry

forest-spheric-dyn-mixed geometry

Fields

この計算が最後の出力ステップで書き出した場をそのまま描いたもの。3D ケースは中央断面。

forest-spheric-dyn-mixed fill

Plots

この計算が書き出した QoI の時系列・分布。参照値がある場合は水平線で重ねている。

forest-spheric-dyn-mixed impact-pressure

Build & run

cmake -S native -B native/build && cmake --build native/build -j
./native/build/orheo-forest-spheric-dyn-mixed

このページの図を作った計算のコスト(Miyabi-G Grace CPU (72 cores, -stdpar=multicore)): 239.3 s wall。

Validation

Tier T2 — 単点比較
QoI P1-P4, H1-H4
Reference Kleefsman et al. (2005) / MARIN
Gate — ゲートなし

Source & run details

View source

Item Value
Target orheo-forest-spheric-dyn-mixed
Source native/benchmarks/forest_spheric_dynamic_mixed.cpp
Lattice —
GPU twin yes (orheo-forest-spheric-dyn-mixed-gpu)
Modules free_surface.hpp, freesurface_dynamic_mixed.hpp, freesurface_mixed_step.hpp
Gallery run 2026-07-27 @ d7395df2
Args 20 3500 1e-5 1e-3 0.16 1.0 15 2 40 2 1 .
Wall time 239.32 s
CPU time 5893.33 s
Device Miyabi-G Grace CPU (72 cores, -stdpar=multicore)

Computation log

この計算の実際の出力(先頭と末尾)。

SPHERIC Test 2 DYNAMIC-MIXED (free-surface AMR): coarse 64x20x20 (dx=0.0500 m), fine slab 128x31x40 (dx=0.0250 m, +ghost), cby0=15 (0.750 m) DYNAMIC [min 4], dt_c=2.258e-04 s, glat=1.0e-05, nu=0.0010, cs=0.16, MRT, free-slip tank walls (no-slip box), drop_isolated=2, band=2 buffer=2 regrid/40, ratio_cf=0.5030 ratio_fc=1.9881
box coarse cells x[13,16] y[0,3] z[5,14]; reservoir coarse x[39,64) y[0,11); m0=4.036000e+04
  [phase1] phase1_init.csv  step 0  t=0.000 s  front_x=2.013 m
step      0/3500  t=0.000 s  front_x=2.013 m  P1=0 Pa  cby=15 (0.750 m)  M=4.036000e+04 (dM/M0=0.00e+00)  seamMis=0.00e+00 drift=0.00e+00
step    175/3500  t=0.040 s  front_x=1.988 m  P1=0 Pa  cby=15 (0.750 m)  M=4.035657e+04 (dM/M0=-8.51e-05)  seamMis=9.40e-07 drift=9.40e-07
step    350/3500  t=0.079 s  front_x=1.963 m  P1=0 Pa  cby=15 (0.750 m)  M=4.034865e+04 (dM/M0=-2.81e-04)  seamMis=3.94e-06 drift=3.94e-06
step    525/3500  t=0.119 s  front_x=1.913 m  P1=0 Pa  cby=15 (0.750 m)  M=4.034297e+04 (dM/M0=-4.22e-04)  seamMis=7.52e-07 drift=7.52e-07
step    700/3500  t=0.158 s  front_x=1.812 m  P1=0 Pa  cby=15 (0.750 m)  M=4.036178e+04 (dM/M0=4.42e-05)  seamMis=3.58e-06 drift=3.58e-06
step    875/3500  t=0.198 s  front_x=1.738 m  P1=0 Pa  cby=15 (0.750 m)  M=4.037342e+04 (dM/M0=3.32e-04)  seamMis=4.63e-06 drift=4.63e-06
step   1050/3500  t=0.237 s  front_x=1.638 m  P1=0 Pa  cby=15 (0.750 m)  M=4.039021e+04 (dM/M0=7.48e-04)  seamMis=5.76e-06 drift=5.76e-06
step   1225/3500  t=0.277 s  front_x=1.513 m  P1=0 Pa  cby=15 (0.750 m)  M=4.039014e+04 (dM/M0=7.47e-04)  seamMis=2.77e-06 drift=2.77e-06
step   1400/3500  t=0.316 s  front_x=1.413 m  P1=0 Pa  cby=15 (0.750 m)  M=4.039912e+04 (dM/M0=9.69e-04)  seamMis=1.02e-06 drift=1.02e-06
  [phase2] phase2_break.csv  step 1419  t=0.320 s  front_x=1.388 m
step   1575/3500  t=0.356 s  front_x=1.288 m  P1=0 Pa  cby=15 (0.750 m)  M=4.040481e+04 (dM/M0=1.11e-03)  seamMis=1.07e-06 drift=1.07e-06
...

step   2975/3500  t=0.672 s  front_x=0.288 m  P1=4234 Pa  cby=15 (0.750 m)  M=4.048119e+04 (dM/M0=3.00e-03)  seamMis=3.10e-06 drift=3.10e-06
step   3150/3500  t=0.711 s  front_x=0.163 m  P1=3811 Pa  cby=15 (0.750 m)  M=4.048032e+04 (dM/M0=2.98e-03)  seamMis=2.02e-06 drift=2.02e-06
step   3325/3500  t=0.751 s  front_x=0.062 m  P1=3621 Pa  cby=15 (0.750 m)  M=4.046533e+04 (dM/M0=2.61e-03)  seamMis=7.70e-08 drift=7.70e-08
  [phase4] phase4_wall.csv  step 3359  t=0.758 s  front_x=0.013 m
step   3500/3500  t=0.790 s  front_x=0.013 m  P1=3887 Pa  cby=15 (0.750 m)  M=4.047284e+04 (dM/M0=2.80e-03)  seamMis=1.03e-06 drift=1.03e-06

wrote ./sensors.csv + phase{1,2,3,4}*.csv.
Conservation: MAX per-step drift-after-reflux = 3.818e-04 (rel), MAX per-regrid drift = 0.000e+00 (rel); FINAL total mass drift dM/M0 = 2.796e-03.
Phases reached: init=1 break=1 box=1 wall=1.