PR 4/7: Pyomo multi-period DAE model

[bernalde]

Summary

Full trajectory optimization model. Discretizes the drying process into multiple time periods, encoding the physics (vapor pressure, product resistance, energy balance, mass transfer) as Pyomo constraints. Supports both finite-difference and collocation discretization. Log-transform for vapor pressure to avoid numerical issues with exponentials.

PR 4 of 7 in the Pyomo integration series.

Changes

• lyopronto/pyomo_models/model.py — Full multi-period DAE model
• tests/test_pyomo_models/test_model_multi_period.py — Multi-period tests (741 lines)
• tests/test_pyomo_models/test_model_advanced.py — Advanced model tests
• tests/test_pyomo_models/test_model_validation.py — Validation against scipy
• tests/test_pyomo_models/test_physics_equations.py — Physics equation tests

Key Physics Encoded as Constraints

Psub = 2.698e10 * exp(-6144.96 / (Tsub + 273.15))     # Vapor pressure (log-transformed)
Rp = R0 + A1 * Lck / (1 + A2 * Lck)                   # Product resistance
dmdt = Ap / Rp * (Psub - Pch)                          # Sublimation rate
Kv * Av * (Tsh - Tbot) = dmdt * dHs                    # Energy balance

PR Chain

#	PR	Status
0	Sync upstream (#5)	Merged
1	CI/CD for Pyomo (#6)	Open
2	Pyomo dependencies (#7)	Open
3	Utils & single-step (#8)	Open
4	Multi-period model (this PR)
5	Optimizer functions	Pending
6	Benchmarks	Pending
7	Docs & examples	Pending

Testing

~2075 lines of new tests covering model construction, discretization, physics constraints, and validation against scipy baseline.

[bernalde]

I found several blocking issues in the core model formulation and in the new test coverage. The main concerns are that the product-temperature constraint is applied to the wrong variable/direction, scaled solves return scaled rather than physical units, warmstart initializes dmdt 10x too high, the cake-length ODE does not match the existing scipy drying-rate update, and one new Pyomo test fails when networkx is not installed.

[bernalde]

Review summary:

1. Blocking issues: the default scaled create_multi_period_model() result does not expose the documented unscaled model attributes and cannot be passed to warmstart_from_scipy_trajectory; dmdt[0] is unconstrained but is returned as part of the solution trajectory.
2. Nonblocking issues: none.
3. Questions: none.
4. Tests run and outcomes: PYTHONPATH=/tmp/LyoPRONTO_pr9 pytest tests/test_pyomo_models -m "pyomo and not slow" -n 0 --tb=short --maxfail=10 passed with 67 passed, 9 skipped, 14 deselected; RUN_SLOW_TESTS=1 PYTHONPATH=/tmp/LyoPRONTO_pr9 pytest tests/test_pyomo_models/test_model_multi_period.py::TestModelOptimization::test_optimization_runs tests/test_pyomo_models/test_model_validation.py::TestOptimizationComparison::test_optimization_improves_over_scipy tests/test_pyomo_models/test_model_validation.py::TestOptimizationComparison::test_optimized_solution_satisfies_constraints -n 0 --tb=short --maxfail=3 passed with 3 passed; ruff check lyopronto/pyomo_models/model.py tests/test_pyomo_models/test_model_advanced.py tests/test_pyomo_models/test_model_multi_period.py tests/test_pyomo_models/test_model_validation.py tests/test_pyomo_models/test_physics_equations.py passed; git diff --check origin/pr/pyomo-utils-singlestep...HEAD passed.
5. Merge as-is: no. I would not merge this until the blocking issues above are addressed.

[bernalde]

Propagating review feedback from the lower Pyomo PRs where the same issue still applies.