Staged H₂ Detection — Bench Trial Against Cell-Level Off-Gas
A six-week bench trial pairing the per-vehicle thermal grid with an electrochemical H₂ sensor mesh. Goal: characterise the fusion gain across staged cell-abuse profiles.
The hypothesis going in was that fusing per-vehicle thermal with deck-area H₂ would push the detectable-onset window further out than thermal alone — at the cost of a more complex algorithmic state. Six weeks of staged events, five abuse profiles, replays against three fusion algorithms.
Abuse profiles tested
- Slow thermal injection (sub-runaway, no venting).
- Nail penetration at low SoC.
- Nail penetration at high SoC (with controlled vent).
- Overcharge to thermal runaway.
- External heat to cell pack with no electrical fault.
What the fusion bought
The slow-thermal-injection case was the outlier. With no electrochemical fault, the H₂ signature never appeared; thermal carried the detection on its own. That is informative — it tells us where the H₂ layer is and is not load-bearing.
Continue the thread
H₂, CO, and CO₂ — The Gas Signature of Early-Stage Thermal Runaway
The gas-phase signature of a stressed Li-ion cell has a specific order and a specific tempo. Knowing it tells you what to sense, where, and when.
Can Gas Sensors Detect Lithium-Ion Runaway?
Yes — and earlier than thermal in some environments. The harder questions are which gases, where to mount the sensor, and why we treat it as a complementary layer at sea.