Flammability Limits Analyzer answers the gas-phase safety questions behind inerting and purge design: between which fuel concentrations does a blend burn in air, how little oxygen keeps it from burning once nitrogen or carbon dioxide is added, and what the whole flammable region looks like. An optional mode maps where a mixture explodes across temperature and pressure with a kinetic mechanism. It is the gas-phase counterpart of Reaction Calorimetry & Thermal-Safety Model, which covers liquid reactions. It runs on CPU.
Results are screening estimates. They do not replace measured limits (ASTM E681) or compliance with NFPA 68, NFPA 69, or local codes.
How it works
- Pure-fuel limits. Lower and upper flammability limits in air come from a built-in table of 30 gases and vapors, copied from U.S. Bureau of Mines Bulletin 627 (Zabetakis, 1965; public domain). Values measured above 25 °C keep the temperature the bulletin gives.
- Temperature and pressure. Zabetakis's modified Burgess–Wheeler law moves each limit to the requested temperature: the lower limit falls and the upper limit rises in proportion to the temperature change divided by the heat of combustion (computed from NASA thermochemistry). Above 1 atm the upper limit follows the bulletin's natural-gas pressure correlation; the lower limit is left unchanged. The results flag where these rules are extrapolated.
- Blends. Le Chatelier's rule combines the fuel limits into the blend's limits.
- Cross-check. The calculated adiabatic flame temperature (CAFT) method of Mashuga and Crowl finds where the flame temperature from a Cantera equilibrium calculation reaches a threshold. The threshold is set so that each fuel reproduces its tabulated lower limit, so the two methods agree closely on the lower limit and the comparison mainly tests the upper limit and the temperature rule.
- Envelope and LOC. For fuel plus air plus added nitrogen or carbon dioxide, the CAFT method finds the least oxygen that still burns at each fuel fraction. That boundary is the flammability envelope, and its lowest point is the limiting oxygen concentration.
- Explosion peninsula (optional). A grid of temperatures and pressures runs 0-D constant-volume or constant-pressure reactors with a built-in or uploaded mechanism. A point explodes when the mixture heats by half its adiabatic rise (50–400 K) before the end time. With the H₂/O₂ mechanism this reproduces the classic second explosion limit.
Inputs
- Fuels: one blend of built-in species with mole fractions that sum to 1 (fractions of the fuel blend, not of the fuel–air mixture), or a CSV table of several blends with columns
mixture_id,species,mole_fraction. A table gets limits and LOC for every blend, without an envelope. - Conditions: initial temperature and pressure, and the inert (nitrogen or carbon dioxide).
- Envelope: on or off, its resolution, and optionally a fixed CAFT threshold instead of the per-fuel calibration.
- Peninsula: the mechanism (built-in H₂/O₂ or GRI-Mech 3.0, or your own Cantera YAML or Chemkin file), the fuel and oxidizer in that mechanism's species, the equivalence ratio, and the temperature and pressure ranges (pressure is spaced logarithmically).
In a workflow, a table from another step can feed mixtures, and a mechanism from RMG Mechanism Generator or Mechanism Reduction can drive the peninsula.
Outputs
| File | Contents |
|---|---|
limits.csv | One row per blend: limits by both methods and their gap, the CAFT threshold, the LOC and the fuel fraction where it occurs, and flags |
envelope.csv | The envelope boundary: least oxygen and most added inert at each fuel fraction, and the air line |
envelope.png | The fuel–oxygen–inert triangle with the air line, the flammable region, and the LOC |
peninsula.csv | Each grid point: temperature, pressure, whether it explodes, and the ignition delay |
peninsula.png | The explosion map on a logarithmic pressure axis |
summary.json | The request, the table rows used with their source page, the methods, and versions |
Reading the results
- For methane, the calculated LOC is about 2 points of oxygen below the measured value in Bulletin 627 (figure 28). That errs on the safe side for inerting, but it is a model bias, not a tolerance.
- Hydrogen and fuels with decomposition flames (acetylene, ethylene oxide, hydrazine) are flagged: their tabulated limits are not thermal limits, and the temperature rule and CAFT method fit them poorly. The job stops with an error rather than report a limit when the temperature rule no longer applies.
- A 0-D reactor has no walls, so the lowest-pressure (first) explosion limit, which comes from radicals destroyed at the vessel wall, is not modelled.
Not covered
Autoignition temperature, minimum ignition energy, detonation limits, dust explosions, envelopes richer in oxygen than air, steam as the inert, and fuels outside the built-in table.