The mission desk
Describe a crew cabin. We'll rank real tests by similarity and show how much each test reports.
Set up a scenario
Kofi · A close match is evidence, not a safety prediction.
Describe a cabin, and the atlas ranks every NASA test by how closely it matches. It shows the evidence and how sure it is — it does not predict whether a fire will start or spread.
Loading Mission Lab…
Start from a mission
ISS-like reference atmosphere; BASS/BASS-II tested a range of oxygen levels near sea-level pressure
The atlas holds the most evidence here: 56 BASS-II tests ran near sea-level pressure at 14 to 21 % oxygen.
Drag a marker. Each tick underneath is one NASA test record at its recorded value, and the dashed window is the tolerance the Evidence Ladder uses.
44 records within ±1.5 points · BASS and BASS-II, Saffire, LUCI
57 records within ±10 kPa · BASS and BASS-II, Saffire, LUCI
76 records ran in orbit
78 records of any material
44 records within ±50 % of your airflow · BASS and BASS-II
All conditions together20 records match every condition
BASS and BASS-IISaffireTolerance window Each tick is one NASA test record at its recorded value.
20 NASA test records match every condition you set.
This is a statement about evidence, not a prediction: MicroFire does not estimate whether a fire would start or spread.
BASS and BASS-IISaffireLUCIFLEXACME
Finding relevance ranks scientific observations and context. Experiment relevance ranks physical conditions. Neither is a safety rating. How findings are ranked
Matches: airflow
Why #1?
Ahead of #2 because the two tie on every ranking key, so only the finding ID orders them.
Why this is relevant
NASA · Reported observation
“The whole combustion process took about 40 s.”
Limitation
Curated related records are not independent replications; the finding may summarize them together.
Matches: airflow
Why #2?
#1 is ahead because the two tie on every ranking key, so only the finding ID orders them.
Why this is relevant
NASA · Reported observation
“Compared with the 10-cm/s case, the 5-cm/s case has a shorter flame and spreads more slowly.”
Limitation
Curated related records are not independent replications; the finding may summarize them together.
Matches: airflow
Why #3?
#2 is ahead because the two tie on every ranking key, so only the finding ID orders them.
Why this is relevant
NASA · Reported observation
“Because the flow speed is lower, it took more than 60 s to burn the whole sample.”
Limitation
Curated related records are not independent replications; the finding may summarize them together.
Fixed evidence-type ordering; weights vary ±25% across 200 seeded runs. This measures sensitivity to MicroFire’s ranking choices, not scientific confidence. Showing 3 of 14 topic/material matches.
Evidence Ladder
Climbing up means stronger evidence. Tests from 3 experiment families are compared, each kept in its own physical regime.
“A principle objective was to determine the limiting oxygen index (LOI), the oxygen percentage below which combustion would not occur.”
FLEX burned methanol and heptane droplets (2 to 5 mm), a different physical regime: it can explain mechanisms such as oxygen limits, not how a solid material behaves. FLEX droplet results, PDF p. 8
“Observed LOI values for these fuels appear to be slightly lower than initially expected.”
FLEX burned methanol and heptane droplets (2 to 5 mm), a different physical regime: it can explain mechanisms such as oxygen limits, not how a solid material behaves. FLEX droplet results, PDF p. 8
“The droplets had initial diameters between 2 and 5 mm, and they were burned in oxidizing atmospheres at ambient temperature and various pressures between 0.70 and 3.05 atm, over ranges of dilution of oxygen with nitrogen, with carbon dioxide, and with mixtures thereof.”
FLEX burned methanol and heptane droplets (2 to 5 mm), a different physical regime: it can explain mechanisms such as oxygen limits, not how a solid material behaves. FLEX droplet results, PDF p. 8
Tolerances: oxygen ±1.5 points, pressure ±10 kPa, airflow ±50 %. A missing value never counts as a match. How the ladder works
Tolerance check. Across 1,000 variations of these tolerances (oxygen ±1–2 points, pressure ±5–15 kPa, airflow ±30–70 %), the direct rung always holds evidence (5–20 records). BASS-II test GMT216-T25 stays the closest record in 51 %.
Among them, NASA recorded: 19 end state not stated, 12 went out, 4 kept burning, 4 did not ignite. This is a count of past tests near your conditions, not a probability that a fire will behave this way.
Dashed cyan lines mark your oxygen and airflow. Oxygen above 21.5 % lies off the top of the chart: no test went there.
Mission Relevance, Evidence Confidence and Ranking Robustness are three separate project heuristics, not NASA ratings. Robustness reranks every test 1,000 times with each weight varied ×0.75–1.25, the pressure scale 5–15 kPa and the similar-material credit 0.25–0.75. See the formula
“At the lowest speeds examined (less than 1 cm/s) all the flames tended to become dim blue and very stable. However, heat loss at these very low convective rates is small so the flames can burn for a long time.”
Low-visibility combustion is relevant to the detection question under weak-flow conditions. This observation does not establish detection performance for the selected habitat.
20 NASA test records match every condition you set, so no single condition is missing. Their limits (sample size, duration, platform) still apply; open each record to read them.
Source trail: NASA documents and open data