Hydro-mechanical simulation of CO2 Injection into a faulted aquifer

Summary

This dataset contains the results of geomechanical simulations conducted on a faulted aquifer under conditions of CO2 injection. The primary focus of the simulations is the pressure evolution within the rock matrix and along the fault, as well as the associated changes in the mechanical state, including rock deformation and fault slip. Additionally, the simulations explore the sensitivity of fault stability under varying orientations of far-field stress.

The dataset includes raw data in VTK format, as well as an illustrative Jupyter notebook that provides a comprehensive explanation of the problem's geometry, boundary and initial conditions, and an interpretation of the observed physical phenomena. The Jupyter notebook is designed to be run both online and locally.

These simulations were performed using an open-source FEM-based geomechanical simulator. Detailed instructions for running the notebook, along with a link to the geomechanical simulator, are provided in the description below.

Contributions

Emil Gallyamov did contribute to the production of the dataset and its visualisation.

Ismaël Gomes did contribute to the development of the visualisation interface through Jupyter Notebooks and Streamlit.

Guillaume Anciaux did contribute to the development of the visualisation interface and data curation.

Data collection: period and details

From 20 April, 2024 to 30 August, 2024, the .pvt and .npy files were generated, curated, and visualisation routines were developed.

Funding sources

ENAC Interdisciplinary Cluster Grant project OSGEOCGS.

Notebook demonstration

Online

An interactive notebook showcasing visualisations of the dataset is available on RenkuLab.

Running locally

Alternatively, you can launch the notebook on your computer. Download the dataset, install dependencies, and launch Jupyter notebook:

pip install -r requirements_freeze.txt

jupyter notebook

Then, open notebooks/DataVisualisation.ipynb.

Reproducing the dataset

To recreate the results found in this dataset, install the solver and go through the example at examples/injection_fault.

Data structure and information

The repository has the following structure:

data: simulation results

reservoir_vs_time: bulk, solid, cohesive and fault fields for the duration of the simulation

paraview: vtk files

*.npy: numpy arrays to store numerical data at specified locations

fault_vs_angle: only cohesive fields for 180 degrees of rotation

cohesive_0*.vtu: vtu files containing absolute values of the fields

cohesive_init_0*.vtu: all the fields in these files are frozen to the initial (at rest) system state

cohesive_parsed.pvd

cohesive_init_parsed.pvd: postprocessing of Ismaël is substracting one set of data from the other one and plots the difference - increase of slip

notebooks: data visualisation through Jupyter Notebooks

images: images used for illustration in notebooks (e.g. schema of stress rotation, model geometry, etc.)

stress_rotation: contains images with scheme of rotation for angles from 0 to 180

DataVisualisation.ipynb: main notebook with visualized DataVisualisation

library: all the scripts needed to visualize DataVisualisation