Syllabus / Part III — Partial Differential Equations
Module 10
Numerical Methods for PDEs
From the 5-point stencil to multigrid: finite differences, stability analysis, advection schemes, elliptic solvers, finite volumes, spectral and finite element methods.
Builds on: Module 2 — Multivariable & Vector Calculus, Module 9 — Classical PDEs: Heat, Waves & Potentials, Module 4 — Numerical Methods for ODEs — The Theory, Module 8 — Boundary Value Problems & Fourier Series, Module 1 — Linear Algebra, Reloaded
Grids, stencils, deriving FD formulas from Taylor series, truncation error, the discrete Laplacian, and grid convergence studies.
🌊 Every grid-based fluid solver (including Stam's) is built from these stencils.
FTCS (forward-time centered-space), method of lines, the stability limit Δt ≤ Δx²/(2α), and watching instability explode.
🌊 Why Stam calls naive explicit diffusion 'unstable' — and why he switches to an implicit solve.
Fourier modes on a grid, amplification factors, stability conditions, and the Lax equivalence theorem.
🌊 Von Neumann analysis was invented for fluid simulations at Los Alamos.
Backward Euler in time (BTCS), Crank–Nicolson, tridiagonal systems and the Thomas algorithm, ADI for 2D.
🌊 Implicit viscosity is standard in fluid solvers for large time steps.
Leapfrog for waves, the Courant–Friedrichs–Lewy condition, numerical dispersion, and absorbing boundaries.
🌊 The CFL number dictates the time step in essentially every explicit CFD code.
Why centered differences fail for advection; upwind, Lax–Friedrichs, Lax–Wendroff; numerical diffusion vs dispersion; flux limiters.
🌊 Numerical diffusion is why simulated smoke looks blurry — the motivation for higher-order advection.
Stam's key idea: trace particles backward and interpolate. Unconditional stability, interpolation choices, MacCormack/BFECC corrections.
🌊 This is the 'advect' step of Stable Fluids, built from scratch.
Solving Poisson's equation on a grid: the matrix structure, Jacobi, Gauss–Seidel, red-black ordering, SOR, convergence rates.
🌊 Stam's 'lin_solve' routine is Gauss–Seidel; the pressure solve is the bottleneck of every incompressible solver.
Why Jacobi kills high frequencies but not low ones; coarse-grid correction; V-cycles; near-optimal O(N) Poisson solves.
🌊 Production fluid solvers use multigrid or multigrid-preconditioned CG for pressure.
Cell averages and fluxes; conservative discretization; Godunov's method and Riemann problems; finite volumes on unstructured meshes.
🌊 OpenFOAM, ANSYS Fluent and most industrial CFD codes are finite-volume codes.
Collocated vs staggered grids, checkerboard pressure, MAC layout, ghost cells, no-slip/free-slip/inflow/outflow boundaries, obstacles.
🌊 The MAC grid (Harlow & Welch, 1965) is the standard layout for incompressible solvers — and what you'll use in the wind tunnel.
Representing solutions in Fourier modes; spectral differentiation; pseudo-spectral methods; aliasing and dealiasing.
🌊 Stam's 'FFT fluids' and turbulence research codes are spectral.
Weak forms, basis 'hat' functions, assembling a stiffness matrix, 1D and 2D triangular elements, and when to choose FEM.
🌊 Structural analysis of an airframe (and some CFD) uses finite elements.