Repository: github.com/ImperialCollegeLondon/StrataTrapper-models
endurance/
Folder with Endurance CCS site sector models in PFLOTRAN-OGS input formateast-mey/
East Mey CCS site modelsDockerfile
Docker image with clean installation of PFLOTRAN-OGS-1.8Endurance CCS site sector model in PFLOTRAN-OGS’s input deck format.
In endurance/
directory:
one-
/four-
stands for the number of injection wells in the model.
-base
stands for reference viscous-limit upscaling with zero capillary pressure.
-upsc
stands for capillary-limit upscaling conducted by StrataTrapper toolkit.
In detail, specifications can be found in preprint [2].
The East Mey models in this repository are based on a model developed by the ACT consortium (Accelerating CCS Technologies). Our work is based on that model, as opposed to later versions of the East Mey model that might exist outside of the public domain.
East Mey CCS site model in PFLOTRAN-OGS’s input deck format.
In east-mey/
directory:
1-
/2-
stands for two particular well placements and respective injection scenarios.-base
stands for base coarse dynamic model (directly based on ACT’s developments).-base-pc
stands for the -base
model with added capillary pressure model.-upsc
stands for capillary-limit upscaling conducted by StrataTrapper toolkit.To run the supplied models, PFLOTRAN-OGS-1.8 installed on Ubuntu 22.04 is required. A dedicated Docker image can be used instead, which is covered in the next section.
For a model folder <model_folder>
containing .in
files:
chc[xyz].data
from chc[xyz].zip
located at <model_folder>/include/
.Go to a model directory. Using the command line from repository root:
cd <model_folder>
Start a simulation using the run.sh
shell script
bash ./run.sh <model_name> [number_of_processes]
<model_name>
is an input file name with .in
omitted[number_of_processes]
is optional. It is responsible for MPI-based parallelization and equals 4
by default.Examples:
cd endurance
# runs with 8 parallel processes
bash ./run.sh one-well-base 8
# runs with 4 parallel processes
bash ./run.sh four-wells-upsc
Currently, memory demand per process is significantly higher for upscaled models.
Thus, it’s preferable to use fewer processes for -upsc
models, expecting slightly higher computation times.
Visualize simulation results using ResInsight or equivalents.
Viscous-limit upscaling | Upscaled with StrataTrapper |
---|---|
To avoid manual installation of PFLOTRAN-OGS, we provide a Docker image described by the Dockerfile
.
Once Docker is installed and Docker engine is running:
Build the image using command from docker-build.txt
docker build --progress=plain --platform linux/amd64 -t pflotran-ogs .
Run a container from the built image using command from docker-run.txt
docker run -it pflotran-ogs