Examples

Water injection

In this example we consider the configuration file ‘h2o.toml’ available in the examples folder (link to the file), where the co2store model is used and only water is injected in a radial grid.

If the generated files are to be saved in a folder called ‘h2o_radial’, then this is achieved by the following command:

pyopmnearwell -i h2o.toml -o h2o_radial

We change the grid to a 3D catesian grid:

grid = "cartesian" #Grid type: cake, radial, core, cartesian2d, coord2d, tensor2d, cartesian, cpg3d, coord3d, or tensor3d
adim = 100 #Grid cake/radial: theta [degrees]; core: input/output pipe length [m]; cartesian2d, coord2d, tensor2d: width[m]

and increase the injection rate 6 times (this to be comparable since the radial grid has an angle of 60).

inj = [[1e-1,1e-1,5e-2,0,360000]]

We run again the configuration file and save it in a different folder:

pyopmnearwell -i h2o.toml -o h2o_cartesian

To compare the results in the radial and cartesian folder, then it is enough to write in the terminal:

pyopmnearwell -c compare

The following figure compares the pressure profile for both simulations (pressure_1D_single_layer_xnormal.png inside the compare folder):

CO2 cyclic injection

In this example we consider the configuration file described in the configuration file section, which is available in the examples folder as ‘co2.toml’.

If the generated files are to be saved in a folder called ‘co2’, then this is achieved by the following command:

pyopmnearwell -i co2.toml -o co2

The execution time was c.a. 20 seconds and the following is an animation using ResInsight to visualize the gas saturation:

Simulation results of the gas saturation.

The following are some of the plots created by the pyopmnearwell executable:

Permeability (top), CO2 injection schedule (middle), and saturation values over time on the cells along the well location at three different locations (bottom).

CCUS (machine learning)

See this folder for an example of how to use pyopmnearwell to generate data for different input parameters (e.g., injection rates) and read the data (e.g., production volumes). An additional example can be found in the data_generation folder. These examples could be used as a starting point for the ones interested in ML.

Publications

For the simulation results published in this paper about the impact of intermittency on salt precipitation during CO2 injection, see/run these configuration files.

For a study where pyopmnearwell is used to generated a machine-learned near-well model, click here.