pifira

Pressure-based Inverse Fire-heat-flux and Radiative-Area analysis

A data-agnostic Python toolkit for estimating fire heat input and effective flame-contact area for pressurized LPG tanks under localized (non-engulfing) flame exposure.

pifira combines the closed-vessel pressure-rise method (heat input from a measured pressure history) with a radiative (T?? weighted angular-temperature inversion (effective fire-contact area from surface temperatures), and provides a dimensionless framework that unifies localized and full-engulfment exposure across tank sizes, fill levels and orientations.

No experimental data are bundled with this package. All routines operate on user-supplied arrays. See Data sources for how to obtain the public validation datasets.

Why

Most published fire-engulfment work assumes uniform, full engulfment of horizontal tanks. Small vertical LPG tanks under localized flame exposure are under-studied, and standards (API 521, KGS) do not cover heat-input estimation for this case. pifira addresses this by inverting a measured pressure rise for heat input and a measured angular temperature distribution for the effective contact area ??without assuming full engulfment.

Installation

pip install pifira

Dependencies: numpy, scipy, pandas, CoolProp. Optional plotting extras:

pip install "pifira[plot]"

Quick start

import numpy as np
import pifira

# 1. Define the tank (volume, fluid, fill fraction)
tank = pifira.TankSpec(
    volume=0.619,
    fluid="HEOS::Propane[0.95]&n-Butane[0.05]",
    fill=0.70,
)

# 2. Heat input from a pressure history (bring your own arrays)
#    t [s], P [bar gauge], closed-vessel window before PRV opens
Q_W = pifira.heat_input(P1_bar=10.83, P2_bar=16.30, dt=544.5, tank=tank)
print(Q_W / 1000, "kW")

# 3. Effective contact area from angular wall temperatures
#    flame centre at 0 deg; off-flame points at +/-90, 180
res = pifira.effective_arc(
    angles_deg=[0, 90, 180, -90],
    T_wall_C=[271, 49, 39, 46],   # period-averaged wall temps
    T_amb_C=35.0,
)
arc_deg = res["arc_deg"]          # area-preserving equivalent width

h_eff = pifira.effective_height(flame_height_m=0.767, liquid_height_m=0.885)
A_eff = pifira.effective_area(diameter_m=0.75, arc_deg=arc_deg, height_m=h_eff)

# 4. Heat flux
print(pifira.heat_flux(Q_W, A_eff), "kW/m^2")

A full time-varying workflow (moving-window heat input, per-instant effective arc, heat-input-weighted representative value) is in examples/analyze_localized.py.

Public API

Method summary

1. Closed-vessel pressure-rise. Before pressure-relief venting, incident heat converts entirely into the contents' internal-energy rise: Q = dU/dt = d/dt(m_L u_L + m_V u_V). Mass and volume are fixed; the liquid/vapour split is re-flashed at each pressure (CoolProp).

2. Radiative (T?? angular inversion. Radiative flux scales with T_wall????T_amb?? (Stefan-Boltzmann). Angular wall temperatures are fit with a Gaussian; the area-preserving equivalent width ??2?)·?` gives the effective arc.

3. Axial intersection. Effective height = min(flame contact height, wetted liquid height). Vapour-space wall is excluded.

4. Representative value. Heat-input-weighted mean Σ(q·Q)/ΣQ, since fire safety is governed by high-heat-input instants.

5. Dimensionless framework. ? (coverage), q* (normalized flux), ? (pressurization), Ω (orientation) unify localized/full and vertical/horizontal.

Data sources

pifira ships no data. The public datasets used to validate the method in the associated study are available from their original publications:

• Moodie, K., Billinge, K., Cutler, D.P. (1988). Fire Engulfment of LPG Storage Tanks. IChemE Symposium Series No. 93, pp. 87??06. Institution of Chemical Engineers. ??pressure history (Fig. 5) and initial conditions / heat fluxes (Tables 1??). Obtain from IChemE or a library.

• Birk, A.M., Poirier, D., Davison, C. (2006). On the thermal rupture of 1.9 m³ propane pressure vessels with defects in their thermal protection system. Journal of Loss Prevention in the Process Industries 19(6), 582??97. https://doi.org/10.1016/j.jlp.2006.02.002 ??pressure (Fig. 8) and lading-temperature stratification (Fig. 10) for the 25% partial-engulfment test. Obtain from the publisher.

To reproduce the validation, digitize the relevant figures from the original papers and pass the values to the pifira API. Digitized values are not redistributed here.

Experimental data from the authors' own fire tests are not included and are reported separately.

Citing

If you use pifira, please cite it via CITATION.cff and the archived release DOI (Zenodo). See the repository release page for the DOI.

License

MIT ??see LICENSE.

Author

Woo-gui-yeon Lee, Korea Gas Safety Corporation (KGS), AI Safety Research Team (ORCID 0009-0008-8976-5363).