Five assessed laboratories aligned to the course plan. This is the full lab set for SciML for PS — we deepen these rather than adding Lab 6+.
| Lab | Topic | Status |
|---|---|---|
| 1 | Reference simulation and scientific dataset | Notebook + sciml_ps.smib helper |
| 2 | Supervised surrogate versus physics residual (PINN) | Brief only (after Lab 1) |
| 3 | Inertia and damping estimation | LS + loss-surface notebook; PINN template |
| 4 | Learned dynamics generalization | Brief only |
| 5 | Stability and CCT | Protocol ready; SMIB path + optional Kundur/ANDES notebook |
Where SciML families are practiced¶
| Family | Lab |
|---|---|
| Classical reference simulation | Lab 1 |
| Supervised surrogate and PINN residual (with physics ablation) | Lab 2 |
| Classical least squares and inverse PINN | Lab 3 |
| Neural ODE / hybrid or operator learning (held-out domain) | Lab 4 |
| Stability labels and CCT | Lab 5 |
New architectures should deepen these five labs or sit under Advanced topics / optional further practice. They do not become a sixth assessed laboratory. Method map: SciML method families.
Install helpers from the repository root: pip install -e .
How to run notebooks (local = canonical; Binder = primary online; Colab = secondary): Running notebooks. Online launch is under development and needs a public GitHub repo; contributions welcome via Join and contribute.
Suggested path when building the course: get Lab 1 solid, then Lab 3, then Lab 5; fill 2 and 4 after those run.
Books, simulators, and optional external packages live under Reference — Reading list, not as a second lab menu.