geoinr-faults 0.1.0

Fault-aware implicit neural representation toolkit for geologic modeling.

GeoINR-faults

GeoINR-faults is an extension of the GeoINR framework, designed to explore and extend the capabilities of implicit neural representations for structural geology.

In addition to fault modeling, GeoINR-faults re-implements a substantial portion of the core functionality of the original GeoINR framework, providing a unified implementation for both standard GeoINR workflows and fault-related extensions.

Relationship to GeoINR

GeoINR-faults is based on the methodology introduced in the original GeoINR project:

Hillier, M., Wellmann, F., de Kemp, E., Schetselaar, E., Brodaric, B., & Bédard, K. (2023). GeoINR 1.0: an implicit neural representation network for three-dimensional geological modelling. Geoscientific Model Development Discussions, 2023, 1-40. https://doi.org/10.5194/gmd-16-6987-2023

This repository is developed as a research extension focusing on fault modeling.

Note:

• This is not the official GeoINR repository
• The implementation in this repository is developed independently

Repository Structure

• main branch: new implementation for fault modeling
• GeoINR_original branch: original GeoINR codebase (forked)
• NN_fault branch: related fault modeling works in INRs (forked)

Installation

We provide the latest release version of GeoINR-faults via PyPi package services. We highly recommend using PyPi,

$ pip install geoinr_faults

The dependencies are: numpy, pandas, scipy, sklearn, torch, matplotlib, and pyvista.

Documentation

After installation, you can either check the notebook tutorials or go to the documentation site for further information.

• Notebook tutorials
• Documentation

License

This repository includes components from the original GeoINR project. All original GeoINR copyright and permission notices are retained in accordance with the MIT License. Additional code in this repository is subject to separate copyright notices.

Citation

If you use this work, please cite:

• Hillier, M., Wellmann, F., de Kemp, E., Schetselaar, E., Brodaric, B., & Bédard, K. (2023). GeoINR 1.0: an implicit neural representation network for three-dimensional geological modelling. Geoscientific Model Development Discussions, 2023, 1-40. https://doi.org/10.5194/gmd-16-6987-2023
• Gao, K., & Wellmann, F. (2025). Fault representation in structural modelling with implicit neural representations. Computers & Geosciences, 199, 105911. https://doi.org/10.1016/j.cageo.2025.105911
• This repository (to be updated)