Skip to content
Tutorials Center
My learning

Topics

Lesson 37 of 49 ยท Computational Aerodynamics

Chapter 8 overview: computational aerodynamics

A map of computational aerodynamics: panel methods, grids and finite differences, turbulence modelling, and how to trust a simulation, with a cost estimator.

Wind tunnels and flight tests still decide the final answer in aerodynamics, but nearly every design begins on a computer. Computational aerodynamics replaces experiments, at least partly, with numerical solutions of the equations of fluid motion. A modern airliner wing is shaped by thousands of simulations before a model ever enters a tunnel. This chapter shows how those simulations work, where they are strong, and where they quietly mislead.

A structured mesh of layered cells wrapped around an airfoil section, shrinking toward the wall, next to four workflow steps: geometry, mesh, solve and check.
A typical simulation: prepare the geometry, mesh the air around it, iterate the equations, then check that the answer can be trusted.

What you will learn

  • The range of methods, from cheap potential-flow panel codes to full Navier-Stokes solvers.
  • The workflow of a simulation and where errors enter.
  • How grids, finite differences and turbulence models work in outline.
  • Why verification and validation are as important as the solver.

The chapter in slides

A ladder of fidelity

Potential flow Fast, no viscosity
Euler Shocks, no viscosity
RANS Viscous and modelled
LES and DNS Resolve the eddies
1 / 4

What each lesson covers

  1. Panel methods: potential flow, singularities, the Kutta condition and a working source-panel code.
  2. Grids and finite differences: how derivatives become arithmetic, what accuracy and stability mean, and why schemes smear or wiggle.
  3. Turbulence modelling and validation: Reynolds averaging, RANS, LES, DNS, wall spacing, grid convergence and credibility.
  4. Chapter 8 review.

Key terms

CFD (computational fluid dynamics): numerical solution of the equations of fluid flow.

Mesh (grid): the set of cells or points that discretizes the flow domain.

Convergence: the iterative solution approaching a steady answer; not the same as grid convergence.

Verification and validation: checking that the equations are solved correctly, and that they represent reality.

The governing equations in one glance

Everything in this chapter approximates the same laws: conservation of mass, momentum and energy. For a viscous compressible gas they are the Navier-Stokes equations. Drop viscosity and heat conduction and you get the Euler equations, which still capture shocks. If, in addition, the flow is irrotational and incompressible, a single linear equation remains, Laplace's equation โˆ‡ยฒฯ† = 0 for a velocity potential ฯ†, which is cheap enough for a calculator and is the basis of panel methods.

Try it

How expensive would it be to simulate everything, even the smallest eddies, in a boundary layer? Direct numerical simulation needs a number of cells that grows roughly with Re^(9/4) for a whole flow. The table below is an order-of-magnitude illustration only, not a prediction for a specific code.

Python The cost of resolving everything

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Which equations include viscosity and heat conduction?

    Navier-Stokes is the full model; Euler drops viscosity and conduction.
  2. 2. What does verification check?

    Verification concerns solving the equations right; validation concerns solving the right equations.
  3. 3. Why is DNS not used for full aircraft?

    Resolving all scales is far too expensive at flight Reynolds numbers.
  4. 4. Which is the cheapest of these methods?

    Panel methods solve a small linear system for the body surface only.

Summary

  • Computational aerodynamics ranges from panel methods to DNS, trading cost for fidelity.
  • The workflow is geometry, mesh, solve, check; errors enter at every step.
  • Verification and validation decide whether results can be trusted.

Next: Panel methods