ceaToFoam

From a NASA CEA chamber solution to a ready-to-run OpenFOAM thrust-chamber case

_{Mach number (upper half) and temperature (lower half) in an RP-1/LOX thrust chamber at 500 psia and O/F 2.4, sized for 2250 lbf at sea level with the frozen mixture and computed with the k-ω SST model.}

ceaToFoam takes the propellants, chamber pressure, mixture ratio and design thrust of a liquid-rocket engine and writes an axisymmetric rhoCentralFoam case for the flow through its conical thrust chamber. The chamber composition computed by CEA is frozen and evaluated with one set of thermodynamic and transport data, and this one description of the gas both sizes the nozzle and defines the thermophysical model of the simulation. The design point and the CFD therefore describe the same gas, so the difference between them is the multidimensional and viscous loss of the thrust chamber, not an artefact of the handoff.

Documentation: https://mcf2-lab.github.io/ceaToFoam/

How it works

flowchart LR
    A["Operating point<br/>propellants, O/F, pc,<br/>thrust, ambient"] --> B["NASA CEA<br/>(CEA_Wrap)"]
    B --> C["Frozen chamber mixture<br/>(Cantera)"]
    C --> D["Design point<br/>throat, exit, c*"]
    C --> F["janaf and Sutherland<br/>fits"]
    D --> E["Conical contour<br/>wedge mesh"]
    E --> G["rhoCentralFoam case"]
    F --> G
    G --> H["ceatofoam post<br/>thrust, mass flow,<br/>exit state"]

Loading

Verified and validated

Check	Result
Frozen sizing vs. CEA's frozen solution	exit temperature, Mach number and Isp within 0.1 %
Simulated vs. design exit pressure (frozen sizing, inviscid)	within 2.03 % for RP-1, CH₄ and H₂ with LOX at 500 and 1500 psia
Simulated vs. design mass flow rate (same cases)	within 0.54 %
Wall pressure in conical nozzles, Back et al. (1965)	RMS deviation of p/pt at most 0.018
Discharge coefficient, Back and Cuffel (1971)	within the measurement uncertainty for 5 of 7 nozzles
Base mesh vs. finest mesh	thrust, mass flow, exit pressure and Isp within 0.27 %

Sizing the same nozzles with CEA's equilibrium isentropic exponent instead leaves the frozen-flow simulation 16 to 21 % below the design exit pressure, while the mass flow rate stays within 0.7 % of the design value. Details are on the verification and validation page.

Quick start

git clone https://github.com/MCF2-Lab/ceaToFoam.git && cd ceaToFoam
pip install .                     # or: pip install '.[plot]'

ceatofoam design examples/rp1-lox-500psi.yaml            # print the design point
ceatofoam case   examples/rp1-lox-500psi.yaml cases/rp1  # write the case
cd cases/rp1 && ./Allrun                                 # blockMesh ... rhoCentralFoam
ceatofoam post   .                                       # thrust, mass flow -> results.json

Requirements: Python 3.9 or newer, an OpenFOAM.com release (tested with v2606) and the CEA executable. CEA_Wrap ships the executable for Windows only; on Linux and macOS compile it once:

ASSETS=$(python3 -c "import CEA_Wrap, os; print(os.path.join(os.path.dirname(CEA_Wrap.__file__), 'assets'))")
mkdir -p ~/.local/bin && cd "$ASSETS"
gfortran -O2 -std=legacy -o ~/.local/bin/Dan_CEA2 dan_cea2.f
export CEA_EXE_LOCATION=~/.local/bin/Dan_CEA2    # add to ~/.bashrc

Commands

Command	What it does
ceatofoam design INPUT	Print the CEA results, the frozen mixture and the design point
ceatofoam case INPUT CASE_DIR	Write the OpenFOAM case and its record ceatofoam.json
ceatofoam post CASE_DIR	Average the last quarter of the run and write results.json
ceatofoam sweep --fuel F --oxidizer O --pressure P	Run CEA over a range of O/F
ceatofoam interactive CASE_DIR	Ask for the inputs on the console, then write the case

Every case records its input, the CEA results, the frozen mixture, the fit errors, the design point, the contour, the mesh and the versions of every component, including the source revision of ceaToFoam.

Documentation

Page	Contents
Installation	requirements, CEA on Linux and macOS, tests
Usage	a complete walkthrough, from the design point to results.json
Input reference	every option of the YAML input
The OpenFOAM case	domain, mesh, boundary conditions, thermophysics, function objects
Method	frozen mixture, sizing models, geometry, solver, performance evaluation
Verification and validation	grid convergence, conical-nozzle measurements, handoff and viscous losses

The documentation source is in docs/; mkdocs serve shows it locally (pip install -r docs/requirements.txt).

Tests

python3 -m unittest discover -s tests

Tests that need the CEA executable or OpenFOAM are skipped when these are not available.

Limitations

The composition is frozen at the chamber (no finite-rate or shifting chemistry in the nozzle), the gas is ideal, the walls are adiabatic and the contour is conical. Injectors, combustion and wall cooling are not modelled.

Citing

If you use ceaToFoam, please cite it; GitHub's "Cite this repository" button reads CITATION.cff. Every release is archived on Zenodo: 10.5281/zenodo.23637114 cites all versions, and 10.5281/zenodo.23637115 cites version 2.0.0. The method, its verification and its validation are described in L. Cerda and E. Madadi-Kandjani, Consistent thermochemistry for frozen-flow simulations of liquid-rocket thrust chambers: sizing, verification and validation; the full reference will be added once it is published.

License and authors

Copyright (C) 2025-2026 Luke Cerda and Ehsan Madadi-Kandjani.

ceaToFoam is free software, distributed under the GNU General Public License v3.0 (GPL-3.0-only), without any warranty.

Luke Cerda and Ehsan Madadi-Kandjani, Department of Mechanical and Aerospace Engineering, California State University Long Beach.