Multiscale Simulation of Electron Hydrodynamics in Graphene

Abstract

This repository contains a multiscale computational framework designed to simulate electron hydrodynamics in strained graphene. The project bridges three distinct domains of physics and computer science:

1. Quantum Mechanics: Tight-binding Density Functional Theory (TB-DFT) via Fireball.
2. Computational Statistics: Machine Learning surrogate models (MLP) trained via PyTorch.
3. Fluid Dynamics: Macroscopic continuous flows via FEniCS / Lattice Boltzmann Method.

Repository Structure

Based on standard separation of concerns (ISO/IEC 25010):

• project-graphene/configs/: YAML configuration files separating hyperparameters from code logic.
• project-graphene/pipeline_fireball/: Scripts for local quantum geometry perturbation and TB-DFT eigenvalue extraction.
• project-graphene/pipeline_ml/: Object-Oriented PyTorch implementation of the MLP mapping function $f_{\text{ML}}: \mathcal{X} \to \mathcal{Y}$.
• project-graphene/pipeline_fenics/: Finite Element Method solvers for the electronic Navier-Stokes (Gurzhi) equations.
• workshops/: Archival codes and exercises from the IFD course.

Installation

Ensure you have Python 3.12+ installed.

# Clone the repository
git clone https://github.com/YOUR_GITHUB_USERNAME/puw-ifd-workshop.git
cd puw-ifd-workshop

# Create and activate virtual environment
python -m venv venv
source venv/Scripts/activate  # On Windows

# Install dependencies (requirements.txt to be added)
pip install -r requirements.txt