prometheus-equilibrium 0.1.1

Combustion equilibrium solver for Python

Prometheus

Prometheus is an open-source Python package for chemical equilibrium and rocket-combustion analysis. It parses multiple thermo-data formats, builds a unified species database, and solves equilibrium problems such as adiabatic chamber combustion (HP) and isentropic nozzle expansion (SP).

Project Status

Pre-1.0 (v0.1.0), active development. Core solver infrastructure and all thermo-data parsers are complete; the recommended solver is GordonMcBrideSolver. Expect breaking changes to the API and data formats!

TO-IMPLEMENT Roadmap

Nice-to-have items for upcoming releases, grouped by impact area.

Core Solver and Data

• Add structured non-convergence diagnostics to EquilibriumSolution (failure reason, residuals, element-balance error, last step norm).
• Implement PEPSolver._tp_equilibrium and add regression coverage for TP/HP/SP behavior.
• Add optional numerical-stability fallback modes for difficult edge cases (adaptive damping / tighter line search controls).
• Add lightweight profiling hooks to report per-iteration timing and major matrix-solve costs.
• Investigate if key species from the species database can also be used as propellant ingredients e.g. methane, oxygen

GUI

• Implement full report export from the GUI (TXT/CSV/JSON) including sweep metadata and final plots.
• Add progress + cancel support for long sweep runs in PerformanceWorker.
• Persist user session settings (units, selected species databases, sweep mode, and recent inputs).

CLI and Tooling

• Add a non-interactive solve CLI for HP/TP/SP runs with machine-readable output (--json / --csv).
• Add benchmark baseline comparison mode in tests/benchmark.py with threshold-based pass/fail output for CI.
• Add strict non-interactive build options in thermo tooling (--no-prompt, --fail-on-missing-label) for reproducible automation.

Features

Thermodynamic databases

• NASA-7 and NASA-9 (including CEA-derived polynomials)
• JANAF tables
• TERRA / Shomate-equivalent (translated from binary)
• AFCESIC ionic and condensed species (translated and calibrated)

Equilibrium problem types: TP, HP, SP, TV, UV, SV

Solvers

Solver	Role	Notes
GordonMcBrideSolver	Production (default)	Fastest in current implementation; matches NASA CEA / RocketCEA algorithm
MajorSpeciesSolver	Alternative	S×S Newton, quadratic convergence
PEPSolver	Reference	Linear convergence, not recommended, currently unstable solve properties

Rocket performance: frozen and shifting isentropic nozzle expansion with c*, Isp (vacuum and sea-level), and area ratio.

Installation

Install from PyPI (recommended)

python -m pip install prometheus-equilibrium
python -m pip install "prometheus-equilibrium[gui]"

Install from source (development workflow)

This repository uses uv.

uv sync --group dev
pip install -e .
pip install -e ".[gui]"

Quick Start

Note: SpeciesDatabase() now resolves package-relative thermo data paths automatically. You can still override any source path in the constructor.

from prometheus_equilibrium.equilibrium import (
    SpeciesDatabase,
    EquilibriumProblem,
    ProblemType,
    GordonMcBrideSolver,
)

db = SpeciesDatabase()
db.load(include_nasa9=True, include_nasa7=True, include_terra=True)

h2 = db.find("H2", phase="G")
o2 = db.find("O2", phase="G")
products = db.get_species({"H", "O"}, max_atoms=6)

T_react = 298.15
H0 = sum(n * sp.enthalpy(T_react) for sp, n in {h2: 2.0, o2: 1.0}.items())

problem = EquilibriumProblem(
    reactants={h2: 2.0, o2: 1.0},
    products=products,
    problem_type=ProblemType.HP,
    constraint1=H0,
    constraint2=30e5,   # 30 bar
    t_init=3500.0,
)

solution = GordonMcBrideSolver().solve(problem)
print(solution.summary())

Or access the GUI:

prometheus-gui

If you are running from source without installing scripts, use:

uv run prometheus-gui

For a longer walkthrough, open demonstration.ipynb.

Propellant Database

Prometheus ships a TOML propellant database covering hundreds of ingredients (AP, HTPB, aluminium, MMH, UDMH, N₂O₄, kerosene, and many more) that has been directly converted from PROPEP. This database is in active development and may change in the future as propellants are added / removed / consolidated. PropellantDatabase.mix() returns a PropellantMixture with the element set, reactant moles, and total H₀ pre-computed so you don't need to calculate them manually. Pass the mixture directly to PerformanceSolver.solve_from_mixture() to get paired frozen and shifting results in one call:

from prometheus_equilibrium.equilibrium import SpeciesDatabase, PerformanceSolver
from prometheus_equilibrium.propellants import PropellantDatabase

db = SpeciesDatabase()
db.load(include_nasa9=True, include_nasa7=True, include_terra=True)

prop_db = PropellantDatabase(
    "prometheus_equilibrium/propellants/propellants.toml"
)
prop_db.load()

# AP/Al/HTPB composite solid propellant at O/F ≈ 2.74 (68/18/14 by mass)
mixture = prop_db.mix([
    ("AMMONIUM_PERCHLORATE",    0.68),
    ("ALUMINUM_PURE_CRYSTALINE", 0.18),
    ("HTPB_R_45HT",              0.14),
])

result = PerformanceSolver().solve_from_mixture(
    mixture, db,
    p_chamber=70e5,    # 70 bar
    area_ratio=10.0,
)

print(f"T_chamber = {result.shifting.chamber.temperature:.0f} K")
print(f"c*        = {result.shifting.cstar:.1f} m/s")
print(f"Isp (vac, shifting) = {result.shifting.isp_vac:.1f} m/s")
print(f"Isp (vac, frozen)   = {result.frozen.isp_vac:.1f} m/s")

The mix() call accepts amounts in any consistent mass unit — only the ratios matter. For liquid bipropellants simply name the two ingredients:

mixture = prop_db.mix([("HYDRAZINE", 1.0), ("NITROGEN_TETROXIDE_LIQ", 1.33)])

Use prop_db.ingredient_ids and prop_db.formulation_ids to browse what is available. Named formulations stored in the TOML can be loaded with prop_db.expand("formulation_id") instead of mix().

Documentation

Online: https://prometheus-equilibrium.readthedocs.io/en/latest/

Build the Sphinx docs locally:

uv run --group docs sphinx-build -b html docs docs/_build/html

The docs include:

• API reference — every public class and function with docstrings
• Thermodynamic database guide — polynomial formats and reference-state calibration across NASA-9, TERRA, and AFCESIC
• Solver comparison — algorithm analysis and implementation notes for MajorSpeciesSolver, GordonMcBrideSolver, and PEP

Development

uv sync --group dev            # Install dev extras (pytest, black, rocketcea …)
uv run pytest                  # Run the test suite
uv run python tests/benchmark.py  # Benchmark 170 cases vs RocketCEA
uv run black prometheus_equilibrium tests
uv run isort prometheus_equilibrium tests
uv run prometheus-build-all-thermo   # Re-generate all thermo databases
uv run prometheus-build-legacy all   # Re-generate TERRA + AFCESIC from binaries

Package Layout

prometheus_equilibrium/
  core/constants.py              element masses, R, P° = 1×10⁵ Pa
  equilibrium/
    species.py                   Chemical, Species hierarchy, SpeciesDatabase
    mixture.py                   Mixture (mutable moles array + all properties)
    element_matrix.py            ElementMatrix, basis selection
    problem.py                   EquilibriumProblem, ProblemType
    solution.py                  EquilibriumSolution + derived properties
    solver.py                    MajorSpeciesSolver, GordonMcBrideSolver, PEPSolver
    performance.py               Frozen / shifting nozzle expansion
  propellants/loader.py          PropellantDatabase, TOML-based definitions
  thermo_data/                   Compiled JSON/CSV databases + raw source files
  tools/                         CLI build scripts (prometheus-build-*)
  gui/                           PySide6 desktop interface

Validation

Prometheus has been developed by cross-referencing multiple existing solver implementations:

• NASA CEA
• PyProPEP / cpropep
• pep-for-zos

The tests/benchmark.py script runs 170 cases (H₂/O₂, CH₄/O₂, N₂H₄/N₂O₄, NH₃/O₂ across 11 O/F ratios × 5 pressures) and reports temperature, molar mass, γ, Cₚ, and speed-of-sound error against RocketCEA.

Contributing

• Open an issue first when possible to discuss the change.
• Add or update tests for any behaviour change.
• Keep docstrings in Google style (Args:, Returns:, etc.).
• Install the pre-commit hooks after cloning so that black and isort run automatically on every commit:

uv sync --group dev
uv run pre-commit install

License

GPL-3.0-or-later. See LICENSE.txt.

Acknowledgements

• Bonnie J. McBride and Sanford Gordon (NASA CEA)
• Boris Trusov (TERRA thermodynamic database)
• Alexander Burcat (BURCAT thermochemical database)