cfd

KO | EN

gitlite — search

#python #hacktoberfest #react #ai #typescript #golang #go #machine-learning #deep-learning #docker #macos #windows #cli #awesome-list

(★ 19)

CFD from scratch in Python — finite differences to turbulence, Navier-Stokes to k-ω SST, validated against Ghia et al.

cfd Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# CFD from Scratch — Foundations to Turbulence

A self-contained, notebook-based course in Computational Fluid Dynamics. Every concept is built from first principles in Python — no black-box solvers, no pre-packaged CFD libraries. You write the physics, the numerics, and the code yourself.

The course runs from the very first equation (what *is* a finite difference?) all the way to turbulence models used in industrial CFD codes like OpenFOAM and Fluent.

---

## About this repo

I built this while learning CFD myself — these notebooks are the notes and exercises I wrote as I worked through each concept from scratch. I'm still actively learning and adding to it.

I'm sharing it publicly so other self-learners can follow the same path without having to piece together resources from a dozen different places.

The entire course was built with the help of a [Claude Code](https://claude.ai/code) AI agent acting as an interactive CFD tutor — asking me questions before explaining, making me implement each concept before moving on, and catching my mistakes along the way. The Socratic + Feynman approach it follows is baked into how each notebook is structured.

**What the AI agent made vs. what I made:**

- `curriculum/` — curriculum notebooks and teaching notes written by the Claude Code agent. These contain theory, derivations, and worked examples.
- `exercise/` — **written entirely by me**. Every solver, every function, every line of code in this folder is my own implementation, written during the lesson after learning the concept. The agent only asked the questions; I wrote the code.

Think of it this way: the agent is the textbook and the professor, `exercise/` is my homework.

---

## Who this is for

- Engineers or scientists who want to understand what CFD solvers actually do inside
- Grad students preparing for research in fluid mechanics or numerical methods
- Anyone who has taken a fluids course but never implemented anything from scratch

**You need:** Python + NumPy comfort, basic calculus, some exposure to PDEs (the course builds intuition as it goes).

---

## What you will build

By the end of this course you will have implemented, from scratch:

| What | Key idea learned |
|------|-----------------|
| Finite difference schemes (1st–4th order) | Truncation error, order of accuracy |
| 1D linear advection (FTBS, FTCS, FTFS) | CFL condition, upwinding, numerical diffusion |
| 1D diffusion (explicit, implicit, Crank-Nicolson) | Stability vs accuracy, tridiagonal solvers |
| 1D Burgers' equation (inviscid + viscous) | Shock formation, characteristics, entropy condition |
| 2D scalar transport | Multi-dimensional upwinding, combined stability |
| Finite Volume Method | Flux balance, exact mass conservation |
| 2D Navier-Stokes solver | Pressure-velocity coupling, projection method |
| SIMPLE algorithm | Pressure correction, under-relaxation |
| Lid-driven cavity flow | Ghia et al. validation, iterative solvers |
| Flow over a cylinder | Drag, lift, vortex shedding |
| Boundary conditions | Dirichlet, Neumann, ghost cells |
| Turbulence: mixing-length model | RANS, Reynolds decomposition, Boussinesq hypothesis |
| Turbulence: k-ε and k-ω SST | Transport equations, blending functions, wall treatment |
| Mesh generation | Structured/unstructured/hybrid, y⁺ estimation, geometric clustering |

---

## Curriculum

---

### Module 1 — Foundations

**Reference notebooks** (written by the AI tutor):

| Notebook | Topic |
| -------- | ----- |
| [curriculum/module_01_foundations/01_what_is_cfd.ipynb](curriculum/module_01_foundations/01_what_is_cfd.ipynb) | What is CFD — discretization, grids, cost vs accuracy |
| [curriculum/module_01_foundations/02_python_for_cfd.ipynb](curriculum/module_01_foundations/02_python_for_cfd.ipynb) | Python and NumPy for CFD |
| [curriculum/module_01_foundations/03_conservation_laws.ipynb](curriculum/module_01_foundations/03_conservation_laws.ipynb) | Conservation laws — mass, momentum, energy |
| [curriculum/module_01_foundations/04_continuity_equation.ipynb](curriculum/module_01_foundations/04_continuity_equation.ipynb) | Continuity equation, incompressible flow |
| [curriculum/module_01_foundations/05_finite_difference_method.ipynb](curriculum/module_01_foundations/05_finite_difference_method.ipynb) | Finite difference schemes, order of accuracy, convergence |
| [curriculum/module_01_foundations/06_1d_linear_advection.ipynb](curriculum/module_01_foundations/06_1d_linear_advection.ipynb) | 1D linear advection — FTBS, CFL, upwinding, periodic BCs |
| [curriculum/module_01_foundations/07_1d_diffusion_equation.ipynb](curriculum/module_01_foundations/07_1d_diffusion_equation.ipynb) | 1D diffusion — explicit FTCS, implicit BTCS, Crank-Nicolson |
| [curriculum/module_01_foundations/08_1d_burgers_equation.ipynb](curriculum/module_01_foundations/08_1d_burgers_equation.ipynb) | Burgers' equation — shock formation, viscous vs inviscid |

**My exercises** (written by me):

- [exercise/1d_advection.ipynb](exercise/1d_advection.ipynb) — 1D linear advection, FTBS, CFL
- [exercise/1d_diffusion.ipynb](exercise/1d_diffusion.ipynb) — explicit and implicit diffusion schemes
- [exercise/1d_diffusion_with_ghost_cells.ipynb](exercise/1d_diffusion_with_ghost_cells.ipynb) — ghost cell boundary conditions
- [exercise/1d_inviscid_burgers_equation.ipynb](exercise/1d_inviscid_burgers_equation.ipynb) — inviscid Burgers, shock formation
- [exercise/1d_viscous_burgers_equation.ipynb](exercise/1d_viscous_burgers_equation.ipynb) — viscous Burgers, shock + diffusion

---

### Module 2 — Core CFD Methods

**Reference notebooks** (written by the AI tutor):

| Notebook | Topic |
| -------- | ----- |
| [curriculum/module_02_core_methods/01_2d_scalar_transport.ipynb](curriculum/module_02_core_methods/01_2d_scalar_transport.ipynb) | 2D advection-diffusion, multi-dimensional stability |
| [curriculum/module_02_core_methods/02_finite_volume_method.ipynb](curriculum/module_02_core_methods/02_finite_volume_method.ipynb) | Finite Volume Method — flux balance, mass conservation |
| [curriculum/module_02_core_methods/03_navier_stokes_equations.ipynb](curriculum/module_02_core_methods/03_navier_stokes_equations.ipynb) | Navier-Stokes equations — derivation, non-dimensionalization |
| [curriculum/module_02_core_methods/04_pressure_velocity_coupling.ipynb](curriculum/module_02_core_methods/04_pressure_velocity_coupling.ipynb) | Pressure-velocity coupling, projection method |
| [curriculum/module_02_core_methods/05_simple_algorithm.ipynb](curriculum/module_02_core_methods/05_simple_algorithm.ipynb) | SIMPLE algorithm — pressure correction, under-relaxation |
| [curriculum/module_02_core_methods/06_lid_driven_cavity.ipynb](curriculum/module_02_core_methods/06_lid_driven_cavity.ipynb) | Lid-driven cavity — theory and Ghia benchmark |
| [curriculum/module_02_core_methods/07_flow_over_cylinder.ipynb](curriculum/module_02_core_methods/07_flow_over_cylinder.ipynb) | Flow over a cylinder — drag, lift, vortex shedding |
| [curriculum/module_02_core_methods/08_boundary_conditions.ipynb](curriculum/module_02_core_methods/08_boundary_conditions.ipynb) | Dirichlet, Neumann, ghost cells |

**My exercises** (written by me):

- [exercise/2d_scalar_transport.ipynb](exercise/2d_scalar_transport.ipynb) — 2D advection-diffusion, multi-dimensional stability
- [exercise/fvm_1d_scalar.ipynb](exercise/fvm_1d_scalar.ipynb) — FVM flux balance, exact mass conservation
- [exercise/lid_driven_cavity.ipynb](exercise/lid_driven_cavity.ipynb) — full SIMPLE solver, Ghia et al. validated
- [exercise/cylinder_flow.ipynb](exercise/cylinder_flow.ipynb) — flow over a cylinder, drag and lift
- [exercise/flow_over_vertical_plate.ipynb](exercise/flow_over_vertical_plate.ipynb) — flow over a vertical plate
- [exercise/airfoil_flow.ipynb](exercise/airfoil_flow.ipynb) — airfoil flow simulation *(generated by the AI tutor as a demonstration; not independently written by me)*
- [exercise/airfoil_aoa_sweep_colab.ipynb](exercise/airfoil_aoa_sweep_colab.ipynb) — angle-of-attack sweep, Colab-ready *(generated by the AI tutor as a demonstration; not independently written by me)*

---

### Module 3 — Turbulence and Meshing

**Reference notebooks** (written by the AI tutor):

| Notebook | Topic |
| -------- | ----- |
| [curriculum/module_03_advanced/01_turbulence_basics.ipynb](curriculum/module_03_advanced/01_turbulence_basics.ipynb) | RANS, Reynolds decomposition, mixing-length, law of the wall |
| [curriculum/module_03_advanced/02_turbulence_models.ipynb](curriculum/module_03_advanced/02_turbulence_models.ipynb) | k-ε, k-ω SST, turbulent channel flow simulation |
| [curriculum/module_03_advanced/03_mesh_generation.ipynb](curriculum/module_03_advanced/03_mesh_generation.ipynb) | Mesh types — structured, unstructured, hybrid, y⁺ |
| [curriculum/module_03_advanced/04_higher_order_schemes.ipynb](curriculum/module_03_advanced/04_higher_order_schemes.ipynb) | Higher-order schemes — QUICK, ENO, WENO |
| [curriculum/module_03_advanced/05_multigrid_methods.ipynb](curriculum/module_03_advanced/05_multigrid_methods.ipynb) | Multigrid methods — V-cycle, smoothers |
| [curriculum/module_03_advanced/06_unsteady_flows.ipynb](curriculum/module_03_advanced/06_unsteady_flows.ipynb) | Unsteady flows — time accuracy, dual time-stepping |
| [curriculum/module_03_advanced/07_compressible_flow.ipynb](curriculum/module_03_advanced/07_compressible_flow.ipynb) | Compressible flow — shocks, Euler equations |
| [curriculum/module_03_advanced/08_pinns.ipynb](curriculum/module_03_advanced/08_pinns.ipynb) | Physics-Informed Neural Networks (PINNs) |

**My exercises** (written by me):

- [exercise/mesh_design.ipynb](exercise/mesh_design.ipynb) — y⁺ estimation, geometric clustering, tanh stretching

---

### Module 4 — Projects (coming)

| Notebook | Topic |
| -------- | ----- |
| [curriculum/module_04_projects/01_channel_flow.ipynb](curriculum/module_04_projects/01_channel_flow.ipynb) | Turbulent channel flow |
| [curriculum/module_04_projects/02_heat_exchanger.ipynb](curriculum/module_04_projects/02_heat_exchanger.ipynb) | Heat exchanger simulation |
| [curriculum/module_04_projects/03_airfoil_simulation.ipynb](curriculum/module_04_projects/03_airfoil_simulation.ipynb) | Airfoil aerodynamics |
| [curriculum/module_04_projects/04_turbulent_jet.ipynb](curriculum/module_04_projects/04_turbulent_jet.ipynb) | Turbulent jet |

---

## How to use this repo

**Recommended path:**

1. Read a reference notebook in `curriculum/` — these explain the theory.
2. Open the matching exercise notebook in `exercise/` and implement it yourself.
3. The exercise notebooks have the structure and comments; the implementation cells are yours to fill.

**Learning order:**

```
Module 1 (curriculum/module_01_foundations/) → Module 2 (curriculum/module_02_core_methods/) → Module 3 (curriculum/module_03_advanced/)
```

You can also jump to any topic if you already know the prerequisites.

---

## Setup

```bash
git clone https://github.com/harishragul/cfd.git
cd cfd
pip install numpy matplotlib scipy jupyter
jupyter notebook
```

No additional CFD libraries needed. Every solver is written in plain NumPy.

---

## Key concepts covered

**Numerical methods:** finite differences, finite volumes, explicit/implicit time integration, upwinding, Crank-Nicolson, Jacobi/Gauss-Seidel/SOR iterative solvers, Poisson equation

**Fluid mechanics:** conservation laws, incompressible Navier-Stokes, pressure-velocity coupling, projection method, SIMPLE algorithm, boundary layer theory, Reynolds decomposition

**Turbulence modeling:** mixing-length model, k-ε model, k-ω SST (Menter 1994), law of the wall, y⁺, turbulent channel flow

**Meshing:** structured/unstructured/hybrid meshes, skewness, aspect ratio, non-orthogonality, geometric and tanh clustering, y⁺ estimation workflow

---

## Validation

The lid-driven cavity solver is validated against the benchmark results of:

> Ghia, U., Ghia, K. N., & Shin, C. T. (1982). High-Re solutions for incompressible flow using the Navier-Stokes equations and a multigrid method. *Journal of Computational Physics*, 48(3), 387–411.

![Ghia validation](https://raw.githubusercontent.com/harishragul/cfd/main/results/ghia_validation.png)

u-velocity along x = 0.5 and v-velocity along y = 0.5 at Re = 100, 41×41 grid. Red dots are Ghia et al. data; blue line is the solver output from [exercise/lid_driven_cavity.ipynb](exercise/lid_driven_cavity.ipynb).

---

## References

- Patankar, S. V. (1980). *Numerical Heat Transfer and Fluid Flow*. Hemisphere.
- Menter, F. R. (1994). Two-equation eddy-viscosity turbulence models for engineering applications. *AIAA Journal*, 32(8), 1598–1605.
- Barba, L. A., & Forsyth, G. F. (2018). CFD Python: the 12 steps to Navier-Stokes equations. *Journal of Open Source Education*, 2(16), 21.
- Ferziger, J. H., Perić, M., & Street, R. L. (2020). *Computational Methods for Fluid Dynamics* (4th ed.). Springer.

---

## License

MIT — use freely, attribution appreciated.

cfd

cfd Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

cfd

cfd Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README