Skip to content

Runaway confirmation: the (a OR b) AND c rule

What is this: the test for "did thermal runaway actually happen" during a propagation verification test. From Appendix C clause C.5.3.7.

The rule, verbatim

"C.5.3.7 Thermal Runaway Triggering Conditions: a) The triggering object experiences a voltage drop, and the drop exceeds 25 % of the initial voltage; b) The temperature at the monitoring point reaches the manufacturer's specified maximum operating temperature; c) The temperature rise rate at the monitoring point, dT/dt, is ≥ 1 °C/s, and it lasts for more than 3 seconds. When a) and c) or b) and c) occur, thermal runaway is considered to have happened."

The decision logic: (a OR b) AND c.

a — voltage drop > 25 % b — T reaches max op T c — dT/dt ≥ 1 °C/s for > 3 s TR confirmed?
no no no no
yes no no no
no yes no no
no no yes no
yes no yes yes
no yes yes yes
yes yes yes yes

Condition c) (the dT/dt ramp) must always be present. Either of conditions a) or b) — or both — must accompany it.

Engineering interpretation (non-normative)

The decision logic in clause C.5.3.7 is (a OR b) AND c. Diagram below.

flowchart TD
    M([Continuous monitoring]) --> AB
    AB{a OR b}
    AB --> A[a) Voltage drop > 25 % of initial]
    AB --> B[b) Temperature at monitor reaches mfr's max operating T]
    M --> C[c) dT/dt ≥ 1 °C/s, sustained > 3 s]
    A --> AND{AND}
    B --> AND
    C --> AND
    AND -->|both true| TR[Thermal runaway confirmed]
    AND -->|either side false| Cont([Continue monitoring])

The triggering object must satisfy a rate condition (c) and at least one of the electrical/thermal conditions (a or b). Confirmation requires both halves; either condition alone is not enough.

Why two electrical / thermal conditions plus one rate condition

  • a) Voltage drop > 25 % is the electrical signature of the cell having lost a substantial fraction of its capacity, typically because the cathode and anode are now in direct contact through a failed separator.
  • b) Temperature reaches max operating T is a fallback for cases where the voltage drop is not clean (e.g. trigger method shorts the cell briefly without immediately dropping its terminal voltage).
  • c) dT/dt ≥ 1 °C/s for > 3 s is the thermal signature of self-sustaining exothermic reactions inside the cell — the runaway itself. This is the discriminator that distinguishes TR from heating-driven temperature rise (which could in principle reach max op T without the cell actually running away).

The > 3 s duration matters because momentary spikes in dT/dt can come from the trigger device (a needle making contact, a heater pulse) and are not TR. Sustained ≥ 1 °C/s for more than three seconds is hard to fake with a trigger device alone — it requires the cell's own chemistry to be releasing energy. See monitoring and sensors for why the sensor must sit on the side opposite the heater for this to work.

The 25 % voltage drop — measured against what

"...exceeds 25 % of the initial voltage."

"Initial voltage" is the cell's voltage at the start of the trigger application. For a cell at high SOC (≥ 90 % or ≥ 95 % per C.5.2 b), terminal voltage will be roughly at the manufacturer's upper cutoff. A 25 % drop from, say, 4.2 V is 1.05 V — i.e. terminal voltage falls below ~3.15 V. That's a clear separator-failure signature for a Li-ion cell.

Stop conditions for the trigger

Each trigger method specifies its own stop condition tied back to C.5.3.7:

  • Needle: continue until TR occurs or depth reaches 90 % of cell dimension (C.5.3.3 f). See needle penetration.
  • External heating: stop when TR occurs or monitor T reaches 300 °C (C.5.3.4). See external heating.
  • Internal heating plate: stop when TR occurs or monitor T reaches 300 °C (C.5.3.5.2). See internal heating plate.

What if no method triggers TR?

"If the thermal runaway triggering method used does not trigger thermal runaway in the battery cell, to ensure the safety of the occupants and property, it must be demonstrated that none of the three recommended methods mentioned above will cause thermal runaway." (C.5.3.7)

This is the fail-safe in the standard. A cell chemistry that genuinely cannot be driven into TR by any of the three recommended methods is, by definition, not a propagation risk in this test. But the burden of proof shifts to the manufacturer: they must show all three methods fail, not just the one they tried.

Engineering note (non-normative): The "none of the three" demonstration is much harder than just running the test with one method that succeeds. If you suspect your cells will not run away under your preferred trigger, plan for a second and possibly third trigger campaign. In practice, OEMs reach for the internal heating plate when the others fail because it deposits energy directly inside the jellyroll, making TR difficult to avoid for any standard Li-ion chemistry at high SOC.

Engineering note (non-normative): The (a OR b) AND c rule is what the verification report in C.4 must show was met. The OEM's BMS alarm thresholds (Report 1, C.2) are a separate matter — and should generally be tighter than this rule, because the BMS needs to issue the warning before TR is fully under way, not at the moment TR is confirmed.

Cross-references

Source: GB 38031-2025, Appendix C section C.5.3.7 (PDF p. 38).