Repository to analyze Wellbore flow using Neural Networks

Repository Structure

The repository is organized as follows:

wann/
├── conferences/
├── docs/
├── firedrake/
├── geo/
├── input/
├── mathematica/
├── pytorch/
├── pyscripts/
├── src/
├── targets/
└── training/

Directory Details

• conferences/: Contains materials and presentations for conferences related to this work.
• docs/: Documentation files, including guides and references for using the repository.
• firedrake/: Firedrake code to run nonlinear-1d wellbore analysis.
• geo/: Gmsh files used to generate the mesh used in the C++ code for coupled reservoir-wellbore analysis.
• input/: Input files (.json) and configurations for running the C++ simulations and and generate the training data to be used in the Neural Networks algorithm.
• mathematica/: Contains Mathematica notebooks and scripts for symbolic computations and data analysis related to wellbore flow studies.
• pytorch/: PyTorch-specific implementations and utilities for neural network training.
• pyscripts/: Python scripts for preprocessing, postprocessing, and automating tasks related to data generation and analysis.
• src/: Core C++ source code for the repository.
• targets/: Where the executables with the main functions are located.
• training/: Output directory for the training data generated by the C++ code.

Contributions and suggestions for improvement are encouraged!