- Alexis Balayre
- Deng Chuan
- Chung Yueh
- Julien Godfroy
- Lyderic Faure
- Majuran Chandrakumar
- Marcell Gyorei
This project focuses on the innovative integration of machine learning with computational fluid dynamics (CFD) to address the limitations of traditional turbulence models. By incorporating Physically Informed Neural Networks (PINNs) and employing the Sparse Identification of Nonlinear Dynamical Systems (PySINDy) approach, we aim to improve the accuracy and efficiency of turbulent flow simulations. This interdisciplinary effort aims to harness the power of high-performance computing to revolutionise the predictive modelling of fluid dynamics.
- Software_repo: This directory contains the core codebase developed for the software aspect of our project. It includes the machine learning models and the necessary infrastructure to integrate these models with traditional CFD workflows.
- CFD_repo: The code pertinent to the computational fluid dynamics component of our research is stored here. It encompasses simulations, traditional turbulence models, and the interface for integrating with our machine learning enhancements.
- PINNs_repo: This section houses the codebase for the Physics-Informed Neural Networks (PINNs) model. It provides a detailed overview of the model architecture, data handling, training, evaluation, and visualisation components.
- PySINDy_Models_repo: Early exploratory work involving PySINDy models is documented in this folder. It contains Jupyter notebooks that elucidate our journey in understanding and applying PySINDy models to the project's objectives. This section provides valuable insights into the developmental phase of our research
For ease of understanding and navigation, each directory is accompanied by a detailed README.md file. These guides are meticulously crafted to provide a roadmap through the different components of the project, ensuring that researchers, collaborators and interested parties can explore and benefit from our work in an efficient manner.
- Initial Setup: Familiarise yourself with the root directory's README files before diving into the specific sections. This will give you an overview of the project structure and where to find relevant pieces of code and documentation.
- Environment Preparation: Each repository section includes instructions for setting up your environment, ensuring compatibility, and running the codebase effectively.
- Explore and Experiment: With the environment set up, you're encouraged to explore the code, execute simulations, and visualise results through the Streamlit application in Software_repo. This hands-on approach will enhance your understanding of the practical applications of our research.