Dane Sabo cebf8c167a Initial umbrella repo: thesis + FRET pipeline + plant model with first controllers

Folds three previously-separate pieces into one preliminary-example repo
for the HAHACS thesis:

- thesis/ (submodule) → gitea Thesis.git — the PhD proposal
- fret-pipeline/ — FRET requirements to AIGER controller (was
  ~/Documents/fret_processing/; prior single-commit history abandoned
  per user decision)
- plant-model/ — 10-state PKE + lumped T/H PWR model (was
  ~/Documents/PKE_Playground/; never version-controlled before)
- presentations/2026DICE/ (submodule) → gitea 2026DICE.git
- reachability/, hardware/ — empty placeholders for Thrust 3 and HIL
- docs/architecture.md — how the discrete and continuous layers compose
- claude_memory/ — session notes and scratch knowledge pattern

Plant model refactored to thesis naming (x, plant, u, ref); pke_th_rhs
now takes u as an explicit arg instead of reading rho_ext from the
params struct. First two controllers built to the contract
u = ctrl_<mode>(t, x, plant, ref): ctrl_null (baseline) and
ctrl_operation (stabilizing, proportional on T_avg). Validated under a
100% -> 80% Q_sg step: ctrl_operation reduces steady-state T_avg drift
~47% vs. the unforced plant.

Root CLAUDE.md emphasizes that CLAUDE.md files are living documents and
that any knowledge not captured before a session ends is lost forever;
claude_memory/ holds the session-level notes that haven't stabilized
enough to graduate into a CLAUDE.md.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-04-16 16:24:11 -04:00

4.9 KiB

Raw Blame History

2026-04-16 — repo restructure

Folded three previously-separate pieces into this umbrella repo. Created the initial layout, the root CLAUDE.md, README.md, docs/architecture.md, and fret-pipeline/CLAUDE.md.

What we did

Created ~/Documents/pwr-hybrid-3-demo/ as the umbrella repo (git init, main branch, not yet pushed to any remote).
Copied ~/Documents/fret_processing/ → fret-pipeline/ via rsync excluding .git, .DS_Store, .claude, __pycache__. Git history abandoned by explicit user decision — the old single-commit history did not carry forward.
Copied ~/Documents/PKE_Playground/ → plant-model/ via rsync. Never was a git repo; nothing to abandon.
Added thesis/ as a submodule pointing at ssh://git@gitea.danesabo.com:1738/danesabo/Thesis.git (also has GitHub mirror at namelessfaceless/dissertation).
Added presentations/2026DICE/ as a submodule pointing at ssh://git@gitea.danesabo.com:1738/danesabo/2026DICE.git.
Created empty placeholders: reachability/, hardware/, docs/figures/.

What we decided and why

Monorepo with thesis as a submodule, not the reverse. Code iterates daily, thesis weekly. A slow parent with fast children is painful; a fast parent pointing at a slow submodule matches the actual cadence. Also preserves the thesis's independent git history for advisor/committee use.
Copy, not move. Old ~/Documents/fret_processing/ and ~/Documents/PKE_Playground/ still exist. User will delete them later; this gives a safe rollback window.
Folder renames: fret_processing → fret-pipeline (more accurate — it's a pipeline, not a processing step), PKE_Playground → plant-model (describes function not content). Both switched to hyphens for consistency.
No git init on PKE_Playground before the move — user said to skip it and fold in clean.
Used git submodule add, not git subtree. Submodule pain acknowledged (detached HEAD, two-step commits, --recurse-submodules on clone) but accepted in exchange for preserving the thesis's independent git life. Revisit if pain outweighs benefit.

Things that surprised us / are worth remembering

The thesis's Figure 1 (Cold Shutdown → Heatup → Power Operation + SCRAM) and the fret-pipeline/ running example PWR_HYBRID_3 are the same automaton. This is the integration point — the figure should eventually be swapped from hand-drawn TikZ to \includegraphics of a generated state-machine PNG.
The plant model's state vector [n; C1..C6; T_f; T_c; T_cold] exposes exactly the quantities the FRET predicates reference (T_avg = T_c, T_hot = 2*T_c - T_cold). The two layers are already semantically aligned; nobody wired them because they were in different folders.
Q_sg(t) is the physical input, not T_cold. Cold leg temperature is a dynamic state that results from SG heat removal. This matters for the reachability/barrier-cert formulation where Q_sg ∈ [Q_min, Q_max] is the bounded disturbance.
fret_to_synth.py adds a mutual-exclusion LTL constraint because decomposing control_mode = q_X into independent booleans lets the synthesizer assert multiple modes simultaneously. This is a synthesis encoding artifact, not a missing FRET requirement. Do not add it in FRET.

Open items (for future sessions)

presentations/2026DICE/ submodule reflects gitea HEAD, not user's local state. The original ~/Documents/Writing/2026DICE/ had uncommitted changes: M DANE_SABO_ABSTRACT.pages plus three untracked PDFs (DANE_SABO.pdf, DANE_SABO_ABSTRACT.pdf, DANE_SABO_ABSTRACT_V2.pdf, DANE_SABO_ABSTRACT_V3.pdf). If user wants these in the submodule, commit from the original location, push to gitea, then from umbrella: cd presentations/2026DICE && git pull and commit the new pointer.
Old folders not yet deleted. ~/Documents/fret_processing/ and ~/Documents/PKE_Playground/ still live at their old paths. Delete when user confirms the umbrella is working.
No initial commit yet. Everything is staged but uncommitted. User wanted to review the four new docs before committing.
No remote yet. Need to create gitea.danesabo.com/danesabo/pwr-hybrid-3-demo and git remote add origin + push.
fret-pipeline/README.md still says fret_processing/ in its directory structure section. Didn't update — it's a minor inconsistency but worth fixing on the next pass.
No bridge between discrete and continuous layers yet. The whole point of the umbrella is to eventually wire them — reachability/ is where that work lands. Not started.

Paths touched

Created: /Users/danesabo/Documents/pwr-hybrid-3-demo/ (entire tree)
Untouched: /Users/danesabo/Documents/fret_processing/, /Users/danesabo/Documents/PKE_Playground/, /Users/danesabo/Documents/Writing/Thesis/, /Users/danesabo/Documents/Writing/2026DICE/ — all still present, unchanged.

4.9 KiB Raw Blame History