RudhLab

Getting started

A self-paced path into computational fluid dynamics, for students who want to join RudhLab or simply learn how simulations are built. It starts with free material and ends with GPU programming.

How to use this page

Work through the steps in order, at your own pace. Each step pairs a little reading with something to build. If you plan to apply for a JRF or PhD position, send what you build with your application: a working, checked solver says more than a list of courses.

1 · Foundations

  • CFD Python: 12 steps to Navier–Stokes ↗L. A. Barba and G. F. Forsyth, Journal of Open Source Education, 2018

    Free Jupyter notebooks that build up from 1D convection to the full Navier–Stokes equations. A good first week.

  • An Introduction to Computational Fluid Dynamics: The Finite Volume MethodH. K. Versteeg and W. Malalasekera, 2nd ed., Pearson, 2007

    The standard first textbook on finite-volume discretisation and pressure–velocity coupling.

  • Numerical Heat Transfer and Fluid Flow ↗S. V. Patankar, 1980

    Short, clear, and still the best explanation of why finite-volume schemes are built the way they are.

2 · Your first solver

The lid-driven cavity is the classic first test: fluid in a square box, driven by a moving lid. It is simple to set up and has trusted reference results, so you can tell whether your code is right.

Try: run at Re = 100 and Re = 1000 and compare the centreline velocities with Ghia et al.; halve the grid spacing and see how the error changes; then write your own version from scratch.

3 · Add heat transfer

4 · Methods we use

5 · Programming GPUs

See what these tools produce in our simulation recordings, each with method notes.

Tools

Workshop material

Next

Built something you are proud of? Send it with your application. Questions about any of this material are welcome too.