Fire Science Show

Modeling the consequences of a fire is what most of us do day to day. Modeling how a fire actually grows across real materials, under changing heat flux and ventilation, is the part we tend to avoid, hiding behind a prescribed design fire. That gap is exactly what we tackle with Jason Floyd and Jonathan Hodges from UL Research Institutes, the developers of S-PYRO, a new engineering pyrolysis model already included in Fire Dynamics Simulator (FDS).

We walk through the origin story in railcar fire growth modeling, where “traveling fire” assumptions and fixed HRR curves can fall apart once the compartment becomes ventilation-limited. S-PYRO takes a more practical route to predictive fire modeling: it uses cone calorimeter data (multiple heat flux levels, and optionally multiple thicknesses) and scales burning rate and time so mass consumption stays consistent while the exposure conditions come from your CFD simulation. That means the local incident heat flux in FDS drives which cone response the surface should follow, instead of guessing the fuel output up front.

The hardest detail is also the most important: flame feedback. A burning sample sees extra heat flux from its own flame, and the cone heater flux can be partially absorbed by soot and hot gases. Jason and Jonathan explain how they built a “digital twin” cone using high-resolution FDS runs to generate lookup tables for flame heat flux and cone absorption across different fuel properties, plus how validation against large public cone databases reveals where the self-similarity assumption holds and where it breaks (dripping polymers, delamination, collapse, long-duration char oxidation, and complex assemblies).

If you’ve ever wanted fire growth modeling that’s more realistic than a scripted curve but more approachable than full kinetics, this is the roadmap. Subscribe for more fire science, share this episode with a colleague who uses FDS, and leave a review if it helps. What material or scenario would you try S-PYRO on first?
Resources:
Jack Bono Award winning paper "An Approach for Flux and Thickness Scaling of Cone Calorimeter Data for Predicting the Pyrolysis of Materials" (improved SPyro)
An engineering model for the pyrolysis of materials
FDS github with all the resources
Cover image from a paper: J. Hodges et al., MODELING THE FIRE GROWTH IN A RAILCAR, Interflam 2025
----
The Fire Science Show is produced by the Fire Science Media in collaboration with OFR Consultants. Thank you to the podcast sponsor for their continuous support towards our mission.

ASET/RSET is the fire safety engineering idea that sounds almost too clean to argue with: the building must give you more time to escape than you need to get out. But once you try to prove it in a real project, the simplicity could and and the hard questions start. Where does “time zero” actually begin: ignition, detection, or alarm? What does it mean for tenability to fail in a place where nobody is anymore? And are we designing safer buildings, or just winning a seconds-on-a-spreadsheet contest?

I break down the Available Safe Egress Time and Required Safe Egress Time concept from both sides: the human evacuation model and the fire model. On the RSET side, we talk about detection and alarm strategy, pre-movement time distributions, and why congestion, stair capacity, and occupant behavior can dominate the result. On the ASET side, we get into CFD fire modeling, tenability criteria like visibility, heat, and toxicity, and the huge responsibility that comes with choosing a design fire that is both plausible and fail-safe. I also explain why fire location can completely change outcomes even when the heat release rate curve is identical.

We also tackle the famous criticism of ASET/RSET, where it truly breaks down (especially residential scenarios), and why I still find it useful in complex spaces like public venues and tunnels. To move beyond one “magic number,” I share practical alternatives like scenario-by-scenario comparisons, probabilistic thinking, ASET/RSET maps, and agent-based tenability approaches, plus a look at digital twin possibilities for smarter evacuation guidance.

If this helped you rethink how you use ASET/RSET in performance-based fire engineering, subscribe, share the episode with a colleague, and leave a review so more engineers can find it. What’s the toughest ASET/RSET argument you’ve had to defend?
-----
Three mentioned links in the episode:
Rino's pre evacuation time distributions: https://www.sciencedirect.com/science/article/pii/S0379711218302212
Vyto's paper here: https://onlinelibrary.wiley.com/doi/abs/10.1002/fam.1025
Tenability survey paper here: https://www.scopus.com/pages/publications/105046523596?origin=resultslist
----
The Fire Science Show is produced by the Fire Science Media in collaboration with OFR Consultants. Thank you to the podcast sponsor for their continuous support towards our mission.

In this episode I take you into practicality of introducing wind into fire safety engineering considerations, and focus on it's implementation as a part of fire oriented CFD analyses. In here, we dig into why wind is tied to so many catastrophic fires, why firefighter safety is often shaped by sudden wind shifts, and why designing for calm air can quietly turn fire safety into a coin toss.

I walk through what wind actually is from a fire engineering perspective: an atmospheric, turbulent phenomenon full of vortices and gusts, not a tidy uniform flow. We talk about the trap of modeling only the average wind speed, why gusts matter so much for smoke movement and vent performance, and how the wind profile changes with height and terrain roughness. From there, I connect the dots to computational wind engineering and the messy reality of cities, where nearby buildings, parks, towers, and “urban canyons” can completely reshape the local wind field right at your façade openings.

Then we get practical about CFD. I share the domain and mesh decisions that keep results honest (including blockage and downstream wake space), how to handle wind direction, and what turbulence modeling choices mean when you’re trying to couple wind and fire. Finally, I lay out when the hassle is worth it: constrained historic roofs, atria with limited openings, exposed natural smoke ventilators, high-rise mechanical floors, complex tunnel portals, architected natural vents with uncertain discharge coefficients, and roof features where vortex shedding can damage photovoltaics.
The key resource to implement what has been said in the episode:
- Wind and Fire Coupled Modelling—Part I: Literature Review
- Wind and Fire Coupled Modelling—Part II: Good Practice Guidelines
- Fire and Smoke Modelling Chapter
----
The Fire Science Show is produced by the Fire Science Media in collaboration with OFR Consultants. Thank you to the podcast sponsor for their continuous support towards our mission.

Smoke does not just slow people down, it changes what they believe is even possible. That simple idea is why we get so excited about pyFDS-Evac, an open source workflow that connects Fire Dynamics Simulator (FDS) results directly to modern evacuation modeling with JuPedSim. When the fire environment can remove an exit from view, reduce walking speed through extinction and soot exposure, or even incapacitate an agent, you stop “post-processing evacuation” and start modeling decisions that respond to conditions.

We sit down with Mohcine Chraibi from Forschungszentrum Jülich to unpack how their team builds pedestrian dynamics software that researchers can actually extend. We talk candidly about why legacy tools like FDS-Evac faded, why commercial evacuation software can feel like a black box when you need to publish and validate new ideas, and what it takes to keep an academic codebase alive. Mohcine walks us through the modular pipeline: run FDS as normal, read outputs through FDS Reader, compute sight lines and smoke effects with FDS VisMap, and drive agents with JuPedSim while pyFDS-Evac applies fire-driven behavior changes.

Then we go deeper into routing and wayfinding. You will hear how graphs and A-star pathfinding evolve toward flow-field methods, why “familiarity” and cognitive maps matter when occupants do not know the building, and what JuPedSim 2.0 aims to deliver next, including stairs, 2.5D to 3D buildings, and switching movement models by location.

If you care about fire safety engineering, evacuation simulation, pedestrian dynamics, or visibility in smoke, subscribe, share this with a colleague, and leave a review so more researchers and practitioners can find the tools and the ideas. What would you test first with an open FDS to evacuation pipeline?
Links to the good stuff:
Jupedsim the engine: https://www.jupedsim.org/
its webapp: https://app.jupedsim.org/

Youtube-playlist for jupedsim: https://www.youtube.com/watch?v=ekbvCdy1P3w&list=PLYiH62IDwy5fdFJRiGky0Hdq88xKwjayL

pyFDS-Evac: https://pedestriandynamics.org/pyFDS-Evac/

pyFDS-evac presentation from the visbility seminar https://pedestriandynamics.org/pyFDS-Evac/talks/visibility-seminar-2026/#/1
Cover image comes from the 1st video on the JuPedSim playlist - the Galton Board Walked by Pedestrians in JuPedSim, by Pedestrian and Fire Dynamics Channel on youtube.
----
The Fire Science Show is produced by the Fire Science Media in collaboration with OFR Consultants. Thank you to the podcast sponsor for their continuous support towards our mission.

A tool doesn’t become a global workhorse because it’s clever. It becomes a workhorse when different labs can run it, calibrate it, compare it, and trust it. It is a very long road from the invention to widespread use. And I got the chance to discuss the path of how the Cone took off with Dr. Marc Janssens, who is a key figure in this story - he was even jointly co-awarded the Phillip DiNenno Prize together with dr Vyto Babrauskas, for his efforts on developing, standardizing and promoting the Cone.

Marc takes us back to the early days of reaction-to-fire testing, when heat release rate measurement was messy, methods differed by country, and oxygen consumption calorimetry still had to prove itself. He was sent on a mission to find a heat measuring device, and shares on what was available back then. 
Later in the episode we dig into the questions raised: is a 10 by 10 cm specimen too small, what does “scaling” really mean, and how do you avoid fooling yourself when you try to turn cone calorimeter curves into real-fire predictions. We also get into why the cone’s controlled heat flux and feedback control mattered so much, and how interlaboratory round robin studies turned uncertainty into precision statements that standards could stand on.

From there, the conversation opens up into what cone calorimetry enabled: faster R&D for products, smarter screening ahead of costly room corner testing, and a path to modeling. Marc shares how cone calorimeter data supports fire modeling efforts, including flame spread approaches in FDS, and why validation at larger scales still matters. The episode lands with a powerful case study: performance-based, risk-informed nuclear power plant fire safety, where cable tray fire spread scenarios depend on models fueled by cone calorimeter measurements of cable insulation.

If you are a daily user of a Cone, an engineer who depends on high quality models and data, or simply want to hear about the history of our discipline... this one is for you.
----
The Fire Science Show is produced by the Fire Science Media in collaboration with OFR Consultants. Thank you to the podcast sponsor for their continuous support towards our mission.