AMPS I · Week 13 — Advection–Diffusion: Péclet Number · Galerkin Oscillations · Upwinding · SUPG

File What When
W13_lecture.html (.qmd source) Lecture slides — Parts A (theory), B (SUPG in DOLFINx 0.11), C (measurements), D (LLM, HW12) in class
W13_lab.ipynb (Colab) Lab: the Galerkin Péclet threshold in 1D, SUPG and nodal exactness, errors away from boundary layers in 2D, an interior layer (skew advection), a \tau that survives \mathbf b=0, LLM demo in class (≈ 80 min), finish at home
W13_HW12.ipynb (Colab) HW12 — 1D difference equation and optimal diffusion (theory); Péclet sweep with Galerkin, upwind and SUPG on a new problem; designing \tau; LLM debugging of an SUPG script due before Week 14
../dolfinx_0.11_api_reference.md DOLFINx 0.11 API reference sheet — paste it into your LLM always

The session starts with a 10-minute quiz on the Week 12 lecture (closed-AI).

All cells run in seconds to a minute. Measure errors away from the layers as well as on the whole domain, and always print the range of u_h.

Before submitting: Runtime → Restart session and run all, fill in the AI-usage log, download the .ipynb. An error inside a %%fenicsx cell does not stop Run all, and nan in a matrix may only show up as a negative KSP reason — read every output.

Week 14 starts with a 10-minute quiz on this week’s lecture (closed-AI).