Example Notebooks¶

Fifteen Jupyter notebooks in examples/ demonstrate the full SCPN-Fusion-Core stack, from equilibrium solves to closed-loop control with uncertainty quantification. Pre-built HTML exports are published to GitHub Pages on every push to main.

Golden Base¶

Notebook	Description
`neuro_symbolic_control_demo_v2`	Golden Base v2 – formal proofs, closed-loop SNN control, DIII-D shot replay, and deterministic reproducibility. The single most comprehensive demonstration of the neuro-symbolic pipeline.
`neuro_symbolic_control_demo`	Legacy v1 frozen base (preserved for reproducibility of published results).

Core Tutorials (01–10)¶

Notebook	Description
`01_compact_reactor_search`	MVR-0.96 compact reactor optimizer: multi-objective sweep over \(R\), \(A\), \(\kappa\), \(B_t\), \(I_p\) subject to Greenwald, Troyon, and Lawson constraints.
`02_neuro_symbolic_compiler`	Petri net to SNN compilation pipeline: define places, transitions, and contracts; compile to LIF neurons; execute in closed loop.
`03_grad_shafranov_equilibrium`	Free-boundary Grad-Shafranov solver tutorial with Picard iteration, coil geometry, and SPARC GEQDSK validation.
`04_divertor_and_neutronics`	Divertor heat flux calculation, TBR estimation, and blanket neutronics with port-coverage and streaming correction factors.
`05_validation_against_experiments`	Cross-validation against 8 SPARC GEQDSK equilibria and the ITPA H-mode confinement database (53 shots, 24 machines).
`06_inverse_and_transport_benchmarks`	Inverse equilibrium reconstruction and QLKNN-10D neural transport surrogate benchmarks.
`07_multi_ion_transport`	D/T/He-ash multi-species transport evolution: fuel depletion, He-ash accumulation and pumping, independent Te/Ti, coronal tungsten radiation.
`08_mhd_stability`	Five-criterion MHD stability suite: Mercier interchange, ballooning (\(s\)-\(\alpha\) diagram), Kruskal-Shafranov, Troyon beta limit, and NTM seeding (modified Rutherford equation).
`09_coil_optimization`	Free-boundary coil current optimization via Tikhonov-regularised least-squares with self-consistent GS-coil outer iteration.
`10_uncertainty_quantification`	Full-chain Monte Carlo UQ: IPB98(y,2) scaling-law exponents, transport model coefficients, and equilibrium boundary perturbation propagated through to \(Q\), \(P_{\mathrm{fus}}\), and \(\beta_N\) confidence envelopes. ITER vs SPARC comparison.

Supplementary Demos¶

Notebook	Description
`Q10_closed_loop_demo`	ITER-like Q=10 closed-loop simulation with PID ramp-up and H-infinity flat-top control, including a tearing-mode disruption scenario with controller recovery.
`platinum_standard_demo_v1`	Vertical integration demo: NMPC-JAX, Rutherford MHD island evolution, quantitative SOC (calibrated MJ), Bosch-Hale D-T reactivity.
`02_neuro_symbolic_compiler_secondary`	Extended compiler walkthrough with alternative Petri net topologies.

Running Notebooks¶

All notebooks can be run locally after installing the package:

pip install -e ".[dev,ml]"
jupyter lab examples/

Or launch directly in the cloud: