Fire & hazard modeling
Model fire spread or smoke and report tenability results.
20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.
- Typical applications
- FDS · Smokeview · FARSITE
- Class GUI-need label
HY· Hybrid
Task proposals · 20
Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.
R10.04.01.02-T01 Run a supplied FDS case and report peak HRR and detector activation
Original instruction
input/office_fire.fds is a ready FDS 6 case of an office fire with two smoke detectors. Run it, then from the output CSVs report the peak heat release rate (kW) and its time, and the activation time of each detector, in output/results.json.
- Platforms
lin· Linuxwin· Windowsmac· macOS- Applications
- FDS
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Grader reruns the same case (fixed seed, same FDS build): peak HRR within 1% and time within one output interval; detector activation times within 1 s of the reference _devc.csv; output files must be from a run of the unmodified input (CHID and checksum).
Proposed partial credit
40% peak HRR and time; 60% detector times.
R10.04.01.02-T02 Build a single-room FDS model and find the time visibility falls below 10 m
Original instruction
From input/room_spec.pdf (6 x 4 x 2.7 m room, one open door 0.9 x 2.1 m, gypsum walls, polyurethane fire with medium t-squared growth to 1 MW), write output/room.fds with a 0.1 m mesh, visibility and temperature devices at 1.8 m in 4 marked locations, run it, and report the first time any device reads visibility < 10 m or temperature > 60 C in output/tenability.json.
- Platforms
lin· Linuxmac· macOS- Applications
- FDS
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Parsed input: geometry within 0.1 m, vent and material correct, RAMP reproduces t-squared growth (alpha within 5%), devices at specified points; reported tenability time within 10% of the reference model's.
Proposed partial credit
35% geometry/vent/materials; 20% fire ramp; 15% devices; 30% tenability time.
R10.04.01.02-T03 Render visibility slices in Smokeview and find when the corridor becomes untenable
Original instruction
input/corridor_case/ holds completed FDS output for a hotel corridor fire. In Smokeview load the visibility slice at z = 1.8 m, set the colorbar to 0-30 m with 10 m as the split, export PNG frames at 60, 120, 180 and 240 s to output/, and record the time at which the far exit door first falls below 10 m visibility in output/answer.json.
- Platforms
win· Windowsmac· macOSlin· Linux- Applications
- Smokeview
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
click· Point & clicknav· Pan/zoom/scrollvis· Visual judgment- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
PNG frames vs reference renders SSIM >= 0.9 (same view, colorbar bounds checked from saved .ini); untenable time within 10 s of gold derived from slice data at the door cell. Smokeview's -script/.ssf scripting is disabled and Python FDS readers (fdsreader) are absent from the VM, so frames and settings must come from the Smokeview GUI.
Proposed partial credit
45% four frames correct; 15% colorbar; 40% untenable time.
R10.04.01.02-T04 Trace a floor plan in PyroSim and export an FDS model
Original instruction
In PyroSim, import input/level2_plan.pdf as a background image, calibrate it to scale, draw walls, doors and the corridor for the two offices and corridor shown (wall height 3 m), place a 1.5 MW fire in Office B and 6 visibility devices at the marked points, set a 0.2 m mesh covering the floor, and export output/level2.fds plus the project output/level2.psm.
- Platforms
win· Windows- Applications
- PyroSim
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
path· Continuous pathgeom· Precise geometryclick· Point & click- Work product
WP04· 3D/spatial model- Conditions
licensed· Licensed pro app- Proposed difficulty
- 4 / 5
Proposed verifier
Exported OBST/HOLE volumes vs gold geometry IoU >= 0.85; door openings within 0.1 m; fire and devices within 0.2 m of marked points; mesh bounds cover plan. .psm must exist and the .fds header must carry PyroSim's generator line; a hand-written .fds that differs from the grader's re-export of the .psm in PyroSim is detected and fails.
Proposed partial credit
45% walls and doors; 20% fire; 20% devices; 15% mesh.
R10.04.01.02-T05 Run a mesh-sensitivity study and report convergence
Original instruction
Using input/compartment.fds, run three cases at cell sizes giving D*/dx of about 4, 8 and 16 (compute D* from the peak HRR), and report the upper-layer temperature at 300 s from the THCP tree for each, the percent change between successive refinements, and which resolution you would use (<= 5% change) in output/mesh_study.json.
- Platforms
lin· Linux- Applications
- FDS · Python
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Three .fds files with dx within 10% of required values; reported temperatures within 2% of grader reruns; percent changes consistent; chosen resolution equals the gold choice.
Proposed partial credit
30% correct cell sizes; 40% reported temperatures; 30% convergence conclusion.
R10.04.01.02-T06 Compare ASET and RSET for an assembly hall using FDS and Pathfinder
Original instruction
Assembly hall in input/hall_dwg/ (DWG) with 420 occupants. Build an FDS smoke model of the design fire in input/design_fire.pdf and a Pathfinder egress model from the same DWG. Report ASET (first of visibility < 10 m, FED >= 0.3 or T > 60 C at 2 m along exit paths), RSET (detection + pre-movement + movement), and the safety margin in output/aset_rset.json with supporting slice images.
- Platforms
win· Windows- Applications
- FDS · Smokeview · Pathfinder · PyroSim
- Interfaces
desk· Desktop GUIterm· Terminal- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
HY· Hybrid- Interaction primitives
geom· Precise geometrycode· Code & commandsvis· Visual judgmentmulti· Cross-app- Work product
WP07· Simulation result- Conditions
licensed· Licensed pro appmultiapp· Multi-applong· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Geometry IoU >= 0.85 vs gold in both models; Pathfinder occupant count and exit widths exact; RSET within 10% and ASET within 10% of reference models; margin sign correct; images present.
Proposed partial credit
25% smoke model; 25% egress model; 35% ASET and RSET values; 15% margin and images.
R10.04.01.02-T07 Compute surface fire spread and flame length in BehavePlus
Original instruction
In BehavePlus (Windows), run the SURFACE module for fuel model GR2 with 1-h/10-h/100-h dead fuel moisture 6/7/8%, live herbaceous 60%, 20-ft wind 15 mi/h with wind adjustment factor 0.4, slope 30%, wind and spread upslope. Save the run worksheet and report rate of spread (ch/h) and flame length (ft) in output/behave.json.
- Platforms
win· Windows- Applications
- BehavePlus
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G3· GUI-only- Interaction primitives
type· Text & formsclick· Point & click- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 1 / 5
Proposed verifier
Rate of spread and flame length within 2% of the reference BehavePlus run; saved worksheet (.bpw) present with the same inputs. Python fire-behavior libraries are not installed on the VM, so values must come from the BehavePlus run recorded in the saved worksheet.
Proposed partial credit
50% rate of spread; 30% flame length; 20% worksheet inputs.
R10.04.01.02-T08 Run FlamMap fire behavior and report area above a flame-length threshold
Original instruction
In FlamMap, load input/ridge.lcp, set weather and wind from input/conditions.txt (gridded wind off, fuel moisture file input/fms.fms), run Basic Fire Behavior, export the flame length raster to output/fl.tif, and report the percent of burnable area with flame length > 2.4 m in output/summary.json.
- Platforms
win· Windows- Applications
- FlamMap
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
click· Point & clicktype· Text & formsnav· Pan/zoom/scroll- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
fl.tif vs reference raster: cell agreement >= 98% (same grid); reported percentage within 1 point; run inputs read from the saved project match the spec. The TestFlamMap command-line build, GDAL and Python are not installed on the VM, so the run and export must be done in the FlamMap GUI.
Proposed partial credit
50% raster agreement; 30% percentage; 20% inputs.
R10.04.01.02-T09 Simulate a 24-hour fire spread with FARSITE and export perimeters
Original instruction
In FlamMap's FARSITE module, load input/canyon.lcp, ignite at the point in input/ignition.shp, use the weather and wind streams in input/wx.wxs for a 24-hour simulation with 30-minute perimeter output and a fuel moisture conditioning period, and export perimeters to output/perimeters.shp plus the final burned area in output/area.json.
- Platforms
win· Windows- Applications
- FlamMap (FARSITE)
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
click· Point & clicktype· Text & formsnav· Pan/zoom/scroll- Work product
WP07· Simulation result- Conditions
tutorial· Follow docs/tutorial- Proposed difficulty
- 4 / 5
Proposed verifier
Final perimeter IoU >= 0.85 with the reference run; burned area within 10%; 48 perimeter timesteps present; conditioning start date matches spec (read from the saved run inputs). The TestFARSITE command-line build, GDAL and Python are not installed on the VM, so the simulation must be run in FlamMap's FARSITE module.
Proposed partial credit
40% final perimeter IoU; 25% area; 20% timesteps; 15% conditioning settings.
R10.04.01.02-T10 Estimate time to flashover with a CFAST two-zone model
Original instruction
For input/storage_room.pdf (room, door, materials, fire curve), build a CFAST input output/storage.in, run CFAST, and report the time when the upper-layer temperature first reaches 600 C and the peak upper-layer depth before flashover in output/cfast.json.
- Platforms
win· Windowslin· Linux- Applications
- CFAST
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Input parsed: compartment, vent and fire definitions within 2% of spec; grader rerun gives the same results; reported times within 5 s of reference.
Proposed partial credit
40% correct model inputs; 40% flashover time; 20% layer depth.
R10.04.01.02-T11 Model a chlorine release threat zone in ALOHA and map it in MARPLOT
Original instruction
A 1-ton chlorine tank (synthetic site, coordinates in input/site.txt) has a 2-inch puncture at the bottom. In ALOHA set location, weather (input/weather.txt), source (tank, liquid), and compute toxic threat zones for AEGL-3/2/1 at 60 minutes. Plot the zones in MARPLOT on the site map, export output/threat_zones.kml and record the downwind distances in output/aloha.json.
- Platforms
win· Windowsmac· macOS- Applications
- ALOHA · MARPLOT
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G3· GUI-only- Interaction primitives
type· Text & formsclick· Point & clickmulti· Cross-app- Work product
WP07· Simulation result- Conditions
multiapp· Multi-app- Proposed difficulty
- 4 / 5
Proposed verifier
Distances within 3% of the reference ALOHA run for the same inputs; KML polygons IoU >= 0.9 with reference; source and weather inputs read back from the saved ALOHA file. Other dispersion models and GIS libraries are not installed on the VM, so zones must be computed in ALOHA and the KML exported from MARPLOT.
Proposed partial credit
45% threat-zone distances; 35% mapped KML zones; 20% inputs.
R10.04.01.02-T12 Calibrate an FDS model to thermocouple test data
Original instruction
input/test_data.csv holds a compartment burn test with a 12-thermocouple tree and measured HRR. Starting from input/test_room.fds, adjust soot yield, the HRR ramp (follow measured HRR) and wall material properties within the bounds in input/bounds.json until the predicted thermocouple temperatures match the data. Deliver output/calibrated.fds and output/fit_report.json with RMSE per thermocouple.
- Platforms
lin· Linux- Applications
- FDS · Python
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP07· Simulation result- Conditions
long· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Grader reruns calibrated.fds: parameters within bounds; mean RMSE across thermocouples <= 25 C and every TC <= 40 C over 0-600 s; reported RMSEs match rerun within 2 C.
Proposed partial credit
15% parameters in bounds; 55% RMSE thresholds; 30% honest report.
R10.04.01.02-T13 Produce a Smokeview animation with a camera path for a client briefing
Original instruction
From input/warehouse_case/ (completed FDS run), use Smokeview to show 3D smoke plus a temperature slice at 3 m, set a camera tour that orbits from the loading dock to the office mezzanine over 20 s, and record output/briefing.mp4 at 1280x720 covering 0-600 s simulation time.
- Platforms
win· Windowsmac· macOS- Applications
- Smokeview
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
geom· Precise geometrynav· Pan/zoom/scrollvis· Visual judgment- Work product
WP05· Time-based media- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Video 1280x720; frames at 5 checkpoints compared with reference renders from the gold tour (SSIM >= 0.8); temperature slice visible with fixed bounds; saved tour in the .ini. Smokeview's -script/.ssf scripting is disabled, so the tour must be built in the Tour dialog and the movie recorded from the Smokeview GUI.
Proposed partial credit
40% camera tour matches; 30% correct quantities shown; 30% video spec.
R10.04.01.02-T14 Hand-calculate sprinkler activation and compare with an FDS device
Original instruction
Write a script that computes sprinkler activation time for a fast t-squared fire using the Alpert ceiling-jet correlations and the RTI heat-transfer equation (RTI, activation temperature, ceiling height and radial distance in input/params.json). Compare with the activation time of the sprinkler device in input/fds_result/ and report both plus the percent difference in output/sprinkler.json.
- Platforms
lin· Linuxmac· macOS- Applications
- Python · FDS
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Computed activation time within 3% of the reference implementation; FDS activation read correctly from _devc.csv; percent difference consistent.
Proposed partial credit
60% correlation-based time; 25% FDS value; 15% comparison.
R10.04.01.02-T15 Size atrium smoke exhaust so the layer stays above the tenable height
Original instruction
input/atrium.fds models a 4-level atrium with a 2.5 MW fire. Find the minimum mechanical exhaust rate (m3/s, steps of 5) that keeps the smoke layer interface (from layer-height devices) at least 3 m above the highest walking surface for 20 minutes, with makeup air velocity below 1 m/s. Report the rate, the layer height history and a Smokeview image in output/.
- Platforms
lin· Linuxwin· Windows- Applications
- FDS · Smokeview · Python
- Interfaces
term· Terminaldesk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commandsvis· Visual judgment- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Grader reruns at the reported rate and one step below: reported rate passes and the lower one fails (layer criterion and makeup velocity); history CSV matches rerun; image present.
Proposed partial credit
50% minimal rate correct; 25% layer history; 15% makeup air check; 10% image.
R10.04.01.02-T16 Build a Pathfinder egress model of an office floor and report evacuation times
Original instruction
In Pathfinder, import input/office_floor.dwg, extract rooms and doors, populate occupants at the load factor in input/occupancy.pdf (by room use), assign 2 exit stairs, run SFPE mode, and report total evacuation time, the time the last occupant leaves each stair door, and peak door flow rates in output/egress.json.
- Platforms
win· Windows- Applications
- Pathfinder
- Interfaces
desk· Desktop GUI- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickgeom· Precise geometrydrag· Drag & drop- Work product
WP07· Simulation result- Conditions
licensed· Licensed pro app- Proposed difficulty
- 3 / 5
Proposed verifier
Model file: occupant count per room exact; door widths within 0.05 m; grader re-simulates the saved .pth: reported times within 2% of the re-run and of gold model results within 10%. Pathfinder's command-line simulator is blocked for the agent account, so the model must be built and run in the Pathfinder GUI.
Proposed partial credit
30% geometry and doors; 25% occupant loading; 45% reported times and flows.
R10.04.01.02-T17 Place three smoke detectors to minimize detection time in an obstructed warehouse
Original instruction
input/warehouse.fds (high-rack storage, beams) has 4 candidate fire locations. Choose positions for 3 ceiling smoke detectors (on the 1 m grid of allowed points in input/allowed_points.csv) that minimize the worst-case detection time over the 4 fires; inspect smoke paths in Smokeview to prune candidates, run the FDS cases, and report positions and per-fire activation times.
- Platforms
lin· Linuxwin· Windows- Applications
- FDS · Smokeview · Python
- Interfaces
term· Terminaldesk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commandsvis· Visual judgmentnav· Pan/zoom/scroll- Work product
WP07· Simulation result- Conditions
long· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Grader reruns the 4 fires with the reported detectors: worst-case activation time <= 1.1 x the reference optimum; points are on the allowed grid; reported times match rerun within 2 s.
Proposed partial credit
60% worst-case time vs optimum; 20% valid positions; 20% honest reported times.
R10.04.01.02-T18 Model stairwell pressurization with door openings and check pressure limits
Original instruction
input/stair_tower.fds models a 10-story pressurized stair with an HVAC supply fan. Add the door-opening schedule in input/doors.csv, tune the supply flow so the stair-to-floor pressure difference stays within the band in input/criteria.pdf with all doors closed and with the scheduled doors open, and report the flow, minimum and maximum pressure differences per floor in output/pressurization.json.
- Platforms
lin· Linux- Applications
- FDS · Python
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Grader reruns: door schedule implemented (OPEN/CLOSE by DEVC times), pressure differences per floor within the criteria band in both conditions; reported min/max within 1 Pa of rerun.
Proposed partial credit
25% door schedule; 45% criteria met in both states; 30% reported values.
R10.04.01.02-T19 Render the same corridor slices with a Smokeview script (control)
Original instruction
Control pair of T03. Write a Smokeview script (.ssf) that loads the 1.8 m visibility slice in input/corridor_case/, sets the 0-30 m colorbar, renders the same 4 frames, and run it headless; compute the untenable time from the slice data with fds2ascii and Python.
- Platforms
lin· Linux- Applications
- Smokeview · fds2ascii · Python
- Interfaces
term· Terminal- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
code· Code & commands- Work product
WP07· Simulation result- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Same frame SSIM and untenable-time checks as T03. Measures the GUI-vs-script gap.
Proposed partial credit
45% frames; 15% colorbar; 40% untenable time.
R10.04.01.02-T20 Build a FlamMap landscape file from separate rasters
Original instruction
input/rasters/ holds elevation, slope, aspect, fuel model (40 Scott-Burgan), canopy cover, stand height, canopy base height and bulk density GeoTIFFs in two different projections. Reproject them to a common UTM grid in QGIS, then create output/project.lcp in FlamMap's landscape builder with correct units for each theme, and verify it loads with the right extent.
- Platforms
win· Windows- Applications
- QGIS · FlamMap
- Interfaces
desk· Desktop GUI- GUI need · task setting
HY· Hybrid- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
click· Point & clicknav· Pan/zoom/scrollmulti· Cross-app- Work product
WP04· 3D/spatial model- Conditions
multiapp· Multi-app- Proposed difficulty
- 3 / 5
Proposed verifier
LCP header: 8 themes with correct unit codes; grid extent and resolution equal gold; cell values match the reprojected reference within 1 unit (nearest-neighbor for fuel) on >= 99% of cells.
Proposed partial credit
30% reprojection; 30% units; 40% cell agreement.
Mapped research references
Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.