SPE11B#
SPE11B is a two-dimensional field-scale case. The configurations compare Cartesian and corner-point grids, a convective-dissolution model, and a fine 1 m reference simulation.
Configurations#
r1_Cart_10mUniform 10 m Cartesian grid with 1 m cells at the left and right boundaries.
r2_cp_10mishCorner-point grid with an approximate cell size of 10 m and 1 m boundary cells.
r3_cp_10mish_convectiveApproximate 10 m corner-point grid using the convective-dissolution model for facies 2 and 5.
r4_Cart_1mUniform 1 m Cartesian grid.
Run the cases#
Run these commands from the benchmark/spe11b directory:
pyopmspe11 -i r1_Cart_10m.toml -o r1_Cart_10m -m all -g all -r 840,1,120 -t 5 -w 0.1
pyopmspe11 -i r2_cp_10mish.toml -o r2_cp_10mish -m all -g all -r 840,1,120 -t 5 -w 0.1
pyopmspe11 -i r3_cp_10mish_convective.toml -o r3_cp_10mish_convective -m all -g all -r 840,1,120 -t 5 -w 0.1
pyopmspe11 -i r4_Cart_1m.toml -o r4_Cart_1m -m all -g all -r 840,1,120 -t 5 -w 0.1
See pyopmspe11 -m, pyopmspe11 -g,
pyopmspe11 -r, pyopmspe11 -t, and
pyopmspe11 -w.
Key observations#
For quantities reported in box A, the convective model in r3 compares well
with the fine-scale r4 simulation while running approximately 500 times
faster.
The convective implementation continues to be improved for regions where dissolved CO2 accumulates. See Mykkeltvedt et al. (2025) for the model description.
Results#
Performance data#
Sparse data#
Spatial maps#
Continue#
See Configuration reference for configuration-variable definitions.
Follow the Tutorial for a guided introductory workflow.
Use Visualizing benchmark results with plopm to reproduce the simulation-grid maps.
See Convergence for the SPE11B grid-refinement study.