What does ARC testing actually measure and how is it used in pack design?

Accelerating Rate Calorimetry (ARC) measures the self-heating rate of a cell as a function of temperature in an adiabatic environment — eliminating heat loss to determine the cell's intrinsic heat generation profile. The output is a dT/dt vs T curve that defines onset temperature, self-heating threshold (typically 0.02°C/min), heat-wait-search protocol parameters, and total heat of reaction. These parameters feed directly into thermal abuse models and are used to set BMS protection thresholds with actual kinetic basis rather than empirical approximation.

What is the Frank-Kamenetskii parameter and why does it matter for pack design?

The Frank-Kamenetskii (FK) parameter (δ) is a dimensionless number that determines whether a reactive exothermic material will self-ignite for a given geometry and boundary conditions. δ = (Q·A·E·ρ·L²)/(k·R·T²) where L is the characteristic half-length. When δ exceeds a critical value (~0.88 for a slab), the heat generation rate exceeds the maximum possible heat removal rate — thermal runaway is inevitable regardless of cooling. For pack designers, this defines the maximum allowable cell group size and minimum required thermal conductivity for a given chemistry and ambient temperature.

What are the AIS-156 Phase 2 specific requirements for thermal runaway propagation?

AIS-156 Phase 2 (aligned with UN ECE R100 Rev 3) requires that a single-cell thermal runaway event must not cause fire or explosion outside the battery system for at least 5 minutes after thermal runaway onset in any cell. The standard specifies the nail penetration trigger procedure, the telltale definitions (fire, explosion, case rupture), and requires a thermal runaway warning signal to occupants at least 5 minutes before the event meets telltale criteria. Testing must be performed on representative production packs, not development samples.

How do you build a credible thermal propagation model for a new pack design?

A credible propagation model requires: (1) cell-level ARC data for the specific cell lot being used, (2) measured thermal conductivity and heat capacity of all module components (cell, busbar, adhesive, housing), (3) CFD or FEA model validated against a single-cell thermal runaway test with instrumented adjacent cells, (4) propagation barrier material data (onset temperature, expansion ratio, thermal resistance post-expansion). The model should predict cell N+1 onset time within ±20% of measured test data before being trusted for design decisions.

What is the most common root cause pattern in field thermal runaway events in Indian commercial EVs?

Based on publicly reported field events and investigation data from Indian commercial EV incidents, the dominant pattern is: (1) mechanical damage from road hazards or minor collisions creating slow-developing internal short circuits that the BMS does not detect because voltage divergence thresholds are set too wide; (2) degraded cells in high-cycle applications (e-buses, 3-wheelers) operating at elevated ambient with inadequate BTMS cooling, where cumulative lithium plating provides the internal short nucleus. Manufacturing defect-triggered events are less common in well-sourced cell lots but have occurred in early-generation packs.

Thermal Runaway — Kinetics, ARC Testing, Propagation Modelling, and Pack Certification

The electrochemistry is settled. The stages are understood. This article is about the engineering rigour needed to design a pack that survives a single-cell event, passes AIS-156 Phase 2 certification, and gives you defensible analysis when something fails in the field.

What This Article Assumes
ARC Testing — What It Actually Measures
Arrhenius Kinetics and the Self-Heating Model
Frank-Kamenetskii Criticality — The Designer's Tool
Pack-Level Propagation — Modelling Architecture
CFD vs FEA — Which Approach For What Question
Propagation Barrier Material Selection
Vent Path Design for Prismatic Cells
AIS-156 Phase 2 Compliance Architecture
UN ECE R100 Rev 3 vs AIS-156 Phase 2 — Divergences
BMS Firmware Architecture for Thermal Event Detection
HF Toxicology and Pack Venting Strategy
Field Failure Investigation Framework
Lithium Plating as a Propagation Nucleus
Key Design Decisions and Their Thermal Safety Impact

ℹ

This is the Master level of the EVPulse Thermal Runaway series. It assumes you have read the Expert article and have working experience with pack design, BMS architecture, thermal management systems, and certification testing. It is not a reference document — it is an engineering decision framework.

What This Article Assumes

You know what ARC stands for. You've specified cells, not just selected them. You've reviewed at least one thermal runaway incident report. You understand that the five stages are a simplification of a continuous kinetic process and that the real question is never "will it run away" but "at what timescale and with what consequence."

Every thermal runaway failure investigation eventually arrives at the same conclusion: the design team understood the general risk, but did not have quantitative values for their specific cells in their specific configuration at their specific operating conditions. They were designing with the wrong numbers, or with no numbers.

This article is about having the right numbers.

ARC Testing — What It Actually Measures

Accelerating Rate Calorimetry operates on one principle: remove the boundary condition. In a standard calorimeter, the cell loses heat to the environment. In an ARC, the environment temperature tracks the cell temperature — perfect adiabatic conditions — so every joule of heat generated by the cell stays in the cell.

The standard protocol is Heat-Wait-Search (HWS):

Heat step

Raise cell temperature by a defined increment (typically 5–10°C) using the ARC heater

Wait step

Hold at that temperature for a soak period (typically 15–30 minutes) to allow thermal equilibration

Search step

Monitor self-heating rate (dT/dt). If dT/dt < threshold (typically 0.02°C/min), the cell is considered thermally stable at this temperature — proceed to next heat step

Exothermic detection

If dT/dt ≥ 0.02°C/min, the ARC switches to tracking mode — following the cell temperature as it self-heats. This is the exothermic onset temperature (T_onset)

Tracking

The calorimeter tracks until runaway or until cell temperature stabilises

The output is a self-heating rate curve — dT/dt as a function of temperature in adiabatic conditions. This curve contains everything needed to characterise thermal runaway kinetics:

T_onset (exothermic onset)

Temperature at which dT/dt exceeds threshold — typically 70–120°C for NMC, 130–160°C for LFP under HWS protocol

T_max

Peak adiabatic temperature during runaway

ΔT_ad (adiabatic temperature rise)

T_max − T_initial — total heat available; 500–1000°C for NMC, 200–400°C for LFP

Heat of reaction

Calculated from ΔT_ad and cell heat capacity (J/g)

Activation energy (Ea)

Extracted from Arrhenius fit to the dT/dt curve

Pre-exponential factor (A)

Arrhenius parameter from curve fitting

⚠

ARC data is cell-lot specific. A different batch of cells with the same part number from the same manufacturer can have T_onset variations of ±15°C and heat of reaction variations of ±20%. Always ARC-test the actual lot being used in production, not a qualification sample from 18 months prior. The BMS protection thresholds derived from ARC data should include a safety margin that accounts for this lot-to-lot variability.

Important limitations of ARC data:

ARC measures a fully charged, pristine cell in isolation. Production cells at end-of-life, partially discharged, with lithium plating, or in mechanical contact with thermal interface material will behave differently. Aged cells typically show lower T_onset due to SEI thickening and accumulated anode degradation products. ARC on fresh cells gives you the best-case baseline — actual pack design should assume degraded cell behaviour.

Arrhenius Kinetics and the Self-Heating Model

The heat generation rate from a thermally abused cell follows Arrhenius kinetics to a first approximation:

Arrhenius Heat Generation Rate

\dot{q} = Q \cdot A \cdot \exp\left(-\frac{E_a}{RT}\right) \cdot f(\alpha)

Where:

Q = total heat of reaction (J/g)
A = pre-exponential frequency factor (s⁻¹)
Eₐ = activation energy (J/mol)
R = universal gas constant (8.314 J/mol·K)
T = absolute temperature (K)
f(α) = conversion function representing reactant depletion (α = extent of reaction)

The energy balance for a single cell in a pack environment:

Cell Energy Balance

m c_p \frac{dT}{dt} = \dot{q}_{gen} - \dot{q}_{cool}

Where q̇_cool = UA(T − T_ambient) for a cell with overall heat transfer coefficient U and surface area A.

Extracting Arrhenius parameters from ARC data

From the ARC self-heating curve, the relationship between dT/dt and T in the pre-runaway region follows:

ln(dT/dt) = ln(Q·A·f(α)) − Eₐ/RT

For the initial stages where f(α) ≈ 1 (reactant not yet significantly depleted), plotting ln(dT/dt) vs 1/T gives a straight line with slope −Eₐ/R. This is the standard method for extracting activation energy from ARC data.

For multi-reaction models (separate Arrhenius parameters for SEI decomposition, anode-electrolyte reaction, cathode decomposition), the curve is decomposed into regions and each reaction is fitted separately. This is standard practice for inputs to pack-level thermal models. Published Arrhenius parameters for common cell chemistries exist in the literature but should always be validated against your own ARC data before being used in safety-critical design calculations.

Frank-Kamenetskii Criticality — The Designer's Tool

The Frank-Kamenetskii (FK) analysis answers a specific question: for a given cell geometry and boundary conditions, is spontaneous ignition inevitable?

The FK parameter δ:

Frank-Kamenetskii Parameter

\delta = \frac{Q \cdot A \cdot E_a \cdot \rho \cdot L^2}{k \cdot R \cdot T_a^2} \cdot \exp\left(-\frac{E_a}{R T_a}\right)

Where:

L = characteristic half-length of the cell (for a cylinder, L = radius)
k = thermal conductivity of the material (W/m·K)
ρ = density (kg/m³)
T_a = ambient temperature (K)

Critical δ values:

Infinite slab: δ_crit = 0.878
Infinite cylinder: δ_crit = 2.00
Sphere: δ_crit = 3.32

When δ > δ_crit, steady-state heat removal is impossible for any boundary temperature — runaway is thermodynamically guaranteed. When δ < δ_crit, there exists a stable steady-state.

🔑

The FK analysis tells pack designers the maximum permissible cell group dimensions for a given chemistry, ambient temperature, and thermal interface material conductivity. This is not the same as the BMS protection threshold — it's the physical geometry limit below which runaway cannot self-initiate even without active cooling. Designing module geometries where δ < δ_crit at maximum design ambient is a meaningful structural safety argument.

Practical implications for Indian EV design at 45°C ambient:

At T_a = 318 K (45°C), the exponential term in the FK parameter is significantly larger than at 25°C, which substantially increases δ for any given geometry. Pack designs that are FK-safe at 25°C ambient may not be FK-safe at Indian summer operating conditions. This is a calculation that should be run explicitly, not assumed.

Pack-Level Propagation — Modelling Architecture

A production-credible propagation model requires four validated input data sets:

1. Cell Thermal Characterisation

ARC-derived Arrhenius parameters for the specific cell lot
Heat capacity (Cp) vs temperature measured by DSC
Thermal conductivity (axial and radial for cylindrical; through-plane and in-plane for prismatic/pouch)
Gas generation volume and pressure profile vs temperature

2. Module Geometry and Material Properties

Thermal conductivity of busbar material (typically aluminium or copper, well-characterised)
Thermal conductivity and heat capacity of cell-to-cell adhesive/TIM (highly variable, must be measured)
Module housing material thermal properties
Cell-to-cell gap dimension (critical for radiation heat transfer calculation)

3. Propagation Barrier Characterisation

Onset temperature of intumescent expansion
Expansion ratio vs temperature
Thermal resistance of expanded material (post-expansion conductivity)
Mechanical integrity at temperature (does it stay in place under gas pressure)

4. Vent Path Characterisation

Vent opening pressure for specific cell design
Gas composition and volume at Stage 2, 3, 4 temperatures
Flow path resistance from cell vent to pack exit

The validation loop — why one-way modelling is not enough

A model built from literature values and datasheet numbers will give you a number. That number may be wrong by a factor of 2–5 for propagation timescales. The validation loop is:

1. Build model with best available inputs
2. Instrument a representative module (thermocouples on each cell face, one cell instrumented internally if possible)
3. Trigger single-cell thermal runaway via nail penetration in the instrumented module
4. Compare predicted cell N+1 onset time and temperature profile against measured data
5. Update model inputs based on discrepancy — typically the thermal conductivity of TIM materials and the gas flow path resistance are the highest-uncertainty inputs
6. Re-validate

A model that predicts cell N+1 onset within ±20% of measured data across three independent tests is credible for certification submissions. A model that has only been run forward without validation data is not.

CFD vs FEA — Which Approach For What Question

CFD (Computational Fluid Dynamics)	FEA (Finite Element Analysis)
Best for: Gas flow and convective heat transfer from venting	Best for: Conductive heat transfer through solid pack structure
Key output: Vent gas temperature and velocity at adjacent cells	Key output: Temperature distribution in cells and module housing
Typical tool: ANSYS Fluent, OpenFOAM	Typical tool: ANSYS Mechanical, COMSOL
Computational cost: High — fluid domain is large	Computational cost: Moderate — solid domain only
Critical input: Vent gas volume and composition vs time	Critical input: TIM and adhesive thermal conductivity
Limitation: Cannot easily model solid-phase conduction through cell	Limitation: Cannot model convective gas flow within module

For a complete pack-level propagation analysis, both are needed. The standard approach is to use FEA for the conductive propagation path (cell to cell through structure) and CFD for the convective propagation path (hot gas flow from venting cell). In practice, many engineering teams use FEA only and apply conservative estimates for convective contribution — this is acceptable for certification purposes if the conservatism is explicitly documented.

Propagation Barrier Material Selection

Intumescent barriers between cells or between modules are the most effective passive propagation mitigation available. Selection criteria:

Expansion onset temperature

Should be ≤ T_onset of cathode decomposition for the cell chemistry (≤170°C for NMC, ≤260°C for LFP)

Expansion ratio

Minimum 3:1 volumetric expansion to fill cell-to-cell gap reliably. 5:1 or greater preferred.

Post-expansion thermal resistance

≥ 0.01 m²·K/W to meaningfully slow conductive propagation

Gas permeability (post-expansion)

Should be low enough to deflect vent gas away from adjacent cells

Mechanical retention

Must maintain position against pack vent gas pressure (typically 0.5–3 bar inside module during runaway)

Commercially available materials in the Indian market that meet these criteria include silica-based intumescent sheets, mica-vermiculite composites, and ceramic fibre boards with intumescent additives. Pure aerogel blankets offer excellent thermal resistance but limited gas barrier performance — they are more appropriate as module-level insulators than cell-level propagation barriers.

⚠

Foam-based materials (polyurethane, polyethylene) should never be used as propagation barriers regardless of their initial thermal resistance values. They melt and combust at Stage 2 temperatures, removing the barrier at exactly the moment it is needed and adding fuel to the event.

Vent Path Design for Prismatic Cells

The vent path is the most neglected thermal safety element in most commercial pack designs. The engineering requirement is simple: vent gas from a cell in runaway must be directed away from adjacent cells and towards a safe exit path from the pack enclosure.

In practice, this requires:

1. Vent channel dimensioning. The cross-sectional area of the vent channel must accommodate peak vent gas flow without creating back-pressure that could delay vent opening or cause the cell to rupture through a non-designed path. Gas flow rate during Stage 3–4 venting for a 100 Ah prismatic cell is approximately 0.5–2 L/s. Size the channel accordingly.

2. Direction control. The cell pressure relief vent must face the vent channel, not the adjacent cell. This sounds obvious — it is routinely violated in pack designs where the vent orientation was determined by electrical layout rather than thermal safety layout.

3. Pack exit sizing. The pack enclosure exit vent (burst disc or pressure relief valve) must be sized for the peak mass flow of all cells in the largest module segment that could realistically fail simultaneously.

4. Exit direction. The pack vent exit must direct gas away from vehicle occupants, fuel lines, and high-voltage cables. For floor-mounted packs, the standard direction is downward and rearward. This must be validated against the vehicle underfloor structure — not just designed on paper.

AIS-156 Phase 2 Compliance Architecture

AIS-156 Phase 2, effective for new type approvals from October 2023, is India's primary regulatory framework for EV battery safety. It aligns substantially with UN ECE R100 Rev 3 but has specific Indian amendments.

AIS-156 Phase 2 Clause 6.1.7 — Thermal Runaway Warning Signal

The BMS must issue an advance warning to vehicle occupants at least 5 minutes before the battery pack meets any of the following telltale conditions: fire, explosion, or physical case rupture.

The warning signal must be:
- A visual indicator inside the passenger cabin
- Distinct from standard battery temperature warnings
- Not suppressible by the driver

The 5-minute requirement means the BMS must detect thermal runaway precursors — not the event itself. Gas detection or dT/dt monitoring meeting this window is required. Absolute temperature threshold monitoring alone typically provides 0–2 minutes at best.

AIS-156 Phase 2 Clause 6.1.8 — Propagation Test

Test procedure: Single-cell thermal runaway is triggered by nail penetration (nail diameter 3–8 mm, penetration speed 25 mm/s, penetration depth through the full cell thickness) on the cell identified as having the highest propagation risk in the pack geometry.

Pass criteria: No fire or explosion external to the battery system for at least 5 minutes after thermal runaway onset in the nail-penetrated cell.

Thermal runaway onset is defined as the point at which the nail-penetrated cell temperature exceeds the maximum temperature of any other cell by more than 50°C and is rising at ≥1°C/s.

The test must be performed on a production-representative pack at 100% SOC.

Required Documentation for AIS-156 Type Approval

- Thermal runaway warning system description and detection logic
- Propagation test report from an accredited test laboratory (ICAT, ARAI, or NATRAX)
- BMS architecture document describing detection thresholds and their derivation
- Cell-level thermal characterisation data (may be submitted as confidential)
- Pack-level thermal model and validation data (may be submitted as confidential)
- Vent path analysis with gas flow calculations
- Post-test inspection report documenting the extent of damage and verifying no external fire

UN ECE R100 Rev 3 vs AIS-156 Phase 2 — Divergences

UN ECE R100 Rev 3	AIS-156 Phase 2
Thermal runaway warning: 5 minutes minimum lead time	5 minutes minimum lead time — aligned
Propagation test trigger: Nail penetration OR heating	Nail penetration only
Test SOC: 100%	100% — aligned
Propagation pass criteria: No external fire/explosion for 5 min	No external fire/explosion for 5 min — aligned
Scope: All EV categories	M and N category vehicles; separate AIS for L category
Test agency: Notified Body (European)	ICAT / ARAI / NATRAX
Certification validity: Type approval	Type approval with annual surveillance audit

The most significant practical divergence is in the trigger method. UN ECE R100 allows either nail penetration or external heating as the trigger — testing teams sometimes prefer heating because it's more controllable and reproducible. AIS-156 Phase 2 mandates nail penetration only, which is a more severe and variable test. Packs that pass R100 using the heating trigger may not pass AIS-156 Phase 2 using nail penetration — a detail with significant commercial consequences for imported vehicle OEMs entering the Indian market.

BMS Firmware Architecture for Thermal Event Detection

A BMS that can reliably provide 5 minutes of thermal runaway warning requires a fundamentally different detection architecture than a standard temperature threshold monitoring system:

Tier 1 — Steady-State Anomaly Detection (hours to days lead time)

Voltage divergence monitoring at rest: Flag any cell whose open-circuit voltage drops more than 30 mV below pack average during a defined rest period (minimum 2 hours at zero current). This detects slow internal short circuits.

Requires: Accurate OCV measurement, reliable current zero detection, cell-level voltage measurement resolution of ≤5 mV.

Tier 2 — Dynamic Anomaly Detection (minutes lead time)

dT/dt monitoring with load compensation: Calculate predicted cell temperature rate of rise from I²R model (using measured current and impedance). Flag cells where measured dT/dt exceeds predicted dT/dt by a defined threshold (typically 0.5–1.0°C/min above prediction) for more than 60 seconds.

Requires: Cell-level or group-level temperature sensors with ≤2°C absolute accuracy, accurate impedance map, current measurement with ≤0.5% error.

Tier 3 — Precursor Chemical Detection (2–8 minutes lead time)

CO₂ or VOC sensor inside pack enclosure: Flag any sensor reading above defined threshold (typically 1000 ppm CO₂ above ambient baseline). This detects Stage 2 gas venting directly.

Requires: Gas sensor in pack enclosure with CAN output to BMS, sealed enclosure to prevent dilution from external atmosphere, periodic baseline recalibration.

Tier 4 — Runaway Confirmation (immediate response)

Absolute temperature threshold: Flag any cell above defined maximum (chemistry-specific — typically 70°C for NMC, 80°C for LFP in an Indian ambient context). This is the final confirmation, not the primary warning.

🔑

AIS-156's 5-minute warning requirement cannot be reliably met by Tier 4 alone. A BMS firmware architecture that achieves compliance will have Tier 2 and ideally Tier 3 detection active. Tier 3 (gas detection) is the lowest-cost path to reliable 5-minute warning compliance for production packs.

HF Toxicology and Pack Venting Strategy

Hydrogen fluoride generation during thermal runaway is the acute human health risk that is most frequently under-engineered in pack design. The design decisions that affect HF exposure:

HF IDLH (Immediately Dangerous to Life and Health)

30 ppm (NIOSH)

HF threshold for acute symptoms

3 ppm

HF release per 100 Ah prismatic cell (full runaway)

15–30 mg HF

HF release per 96S1P NMC pack (worst case)

1,440–2,880 mg HF

Lethal concentration (rat, 1hr exposure LC50)

966 ppm — extrapolation to human risk is complex

For a pack in a passenger vehicle cabin volume of approximately 3 m³, 96S1P full-pack HF release would produce concentrations far in excess of IDLH limits if the venting path is directed into the cabin.

Pack venting strategy implications:

Vent exits must discharge to external atmosphere, not into the cabin or footwell. For floor-mounted packs, the standard approach (downward discharge) satisfies this. For side-mounted or rear-compartment packs, specific vent routing is required.

The 5-minute warning window mandated by AIS-156 is in part a concession to HF toxicology — it is intended to give occupants time to exit the vehicle before HF concentrations in the cabin reach dangerous levels. A BMS that provides only 1–2 minutes of warning provides insufficient time for occupants to exit and move a safe distance from the vehicle.

Field Failure Investigation Framework

When a thermal runaway event occurs in a deployed vehicle, the investigation must establish: trigger mechanism, propagation pathway, BMS detection performance, and contributing design or operational factors. The framework:

Secure the evidence

The pack and BMS flash memory are the primary evidence. Secure BMS data download before any disassembly. Photograph the pack in situ. Do not clean or disassemble until the investigation team has reviewed the intact scene.

BMS data extraction

Extract the full event log including cell-level voltage, temperature (all channels), current, contactor state, fault flags, and timestamps. The 30–60 seconds before the first fault flag are the most valuable data. Reconstruct the timeline from first anomaly to contactor open.

Post-event cell mapping

Identify the origin cell from physical evidence (centre of maximum damage, aluminium liquefaction patterns — aluminium melts at 660°C, copper at 1085°C — provides temperature calibration of the event). Map propagation pathway through the module.

Trigger classification

Determine trigger class: thermal abuse (external heat source pattern), electrical abuse (BMS log shows overcurrent or overvoltage preceding event), or mechanical (post-event inspection shows internal short at specific location). For mechanical triggers, inspect for external damage or manufacturing defect evidence at the failure origin.

Root cause analysis

Apply 5-why or fault tree analysis to identify the root cause. Common findings: BMS voltage divergence threshold too wide to detect slow internal short; thermal sensor not positioned to detect Stage 1 in origin cell location; propagation barrier installed incorrectly (gap in coverage); vent channel blocked by module assembly deformation.

The 'healed short' phenomenon in field failures

A recurring pattern in slow-developing field failures is the healed short — a metallic particle contamination or separator defect that creates a brief internal short circuit, appears to resolve (voltage recovers), but leaves thermal damage that seeds a subsequent faster event.

BMS logs for these events typically show: one or more brief voltage dips in a specific cell, weeks or months before the final event; above-average self-discharge rate in the affected cell during subsequent storage periods; and a final event that progresses from Stage 1 to Stage 3 faster than expected from a fresh internal short.

Detecting healed shorts requires the Tier 1 voltage divergence monitoring described above, with logging that retains anomalies for post-event review. Many production BMS implementations discard voltage anomalies below the protection threshold — meaning the warning signs are generated but not retained.

Lithium Plating as a Propagation Nucleus

In high-cycle commercial EV applications (e-buses, 3-wheelers at 2–4 cycles per day), lithium plating on the graphite anode is the dominant degradation mechanism leading to thermal runaway risk.

Plated lithium is metallic lithium — not intercalated lithium — sitting on the anode surface. It is: highly reactive (much more reactive than LiC₆), electrically conductive (capable of creating dendrite short circuits through the separator), and exothermic when it reacts with electrolyte at temperatures lower than Stage 2 onset for healthy cells.

Plating onset conditions:

Charging at temperatures below 10°C
Charging above the cell's rate capability at high SOC (above 80%)
Degraded graphite with reduced intercalation capacity

⚠

An e-bus operating in Northern India in winter, being opportunity-charged at high current during short layovers at high SOC, with cells that are 18 months into a 5-year design life — is accumulating lithium plating every cycle. The BMS voltage curves may look normal (plated lithium shows some similar electrochemical signatures to intercalated lithium). The risk is building invisibly. Post-charge rest period temperature monitoring (plated lithium stripping is exothermic) is the most practical detection method without EIS.

Key Design Decisions and Their Thermal Safety Impact

✅ Pros

LFP chemistry: 100°C higher onset, 5× lower heat of reaction, slower propagation — the single highest-impact safety decision
Intumescent barriers between all cell groups: 3–10× propagation delay at ₹200–500/module cost premium
Gas sensors (CO₂/VOC) in pack enclosure: reliable 2–8 minute warning lead time at ₹400–600/pack
Cell-level voltage monitoring with tight rest-period divergence thresholds: detects slow internal shorts hours to days in advance
ARC testing on each cell lot: accurate BMS threshold derivation, eliminates arbitrary safety margins
Directed vent path design: prevents HF from entering cabin, reduces propagation via hot gas convection

❌ Cons

NMC chemistry for range: accepted risk, requires compensating measures (tighter BMS thresholds, mandatory barriers, gas sensing)
Cell-to-pack (CTP) architecture: eliminates module housings that provide propagation delays — requires rigorous barrier integration at pack level
High energy density targets: drive thinner barriers, tighter cell spacing, smaller vent channels — all reduce propagation resistance
Cost pressure on sensor count: fewer temperature sensors means larger detection dead zones in Stage 1

Battery Technology Timeline

1999

First ARC studies on lithium-ion thermal abuse published — Dahn group, Dalhousie University

2006

Sony laptop recall (6 million cells) — first large-scale field thermal runaway event; establishes need for pack-level protection standards

2008

UN Model Regulations for lithium battery transport codified — precursor to UN38.3

2013

Boeing 787 Dreamliner battery incidents — first high-profile aviation thermal runaway; leads to mandatory electrochemical impedance monitoring requirements

2018

UN ECE R100 Revision 2 — first international vehicle regulation with specific propagation requirements

2020

BYD Blade Battery nail penetration demonstration — LFP cell-to-pack architecture without fire; resets industry expectations for LFP safety

2021

AIS-156 Phase 1 notified in India — first Indian EV battery safety standard

2023

AIS-156 Phase 2 effective for new type approvals — adds 5-minute warning and propagation test requirements

2024

UN ECE R100 Revision 3 alignment with Global Technical Regulation on EV Safety — harmonises propagation test across markets

Resources and References

ℹ

All references verified as of May 2025. Master-tier references include primary research, standards, ARC methodology papers, and Indian regulatory documents. DOIs provided for all journal articles.

Standards and Regulations

AIS-156 Phase 2 (2023) — Automotive Industry Standard for Electric Power Train Vehicles, Phase 2. BIS / MoRTH. Mandatory for new type approvals from October 2023. https://morth.nic.in
UN ECE R100 Revision 3 (2021) — Uniform provisions for EV approval: thermal runaway warning + propagation requirements. https://unece.org
UN GTR No. 20 (2021) — Global Technical Regulation on the Electric Vehicle Safety — harmonised basis for R100 Rev 3 and AIS-156 Phase 2. https://unece.org/transport/documents/2022/01/standards/gtr-20-evs
IEC 62660-2 — Reliability and abuse testing for lithium-ion cells in electric road vehicles.
ISO 12405-4:2018 — Lithium-ion traction battery packs: performance testing specification.
SAE J2464 — Electric and Hybrid Electric Vehicle Rechargeable Energy Storage System (RESS) Safety and Abuse Testing.

ARC and Thermal Characterisation

Dahn, J. R., Fuller, E. W., Obrovac, M., & von Sacken, U. (1994). Thermal stability of Li_xCoO2, Li_xNiO2, and λ-MnO2 and consequences for the safety of Li-ion cells. Solid State Ionics, 69(3–4), 265–270. DOI: 10.1016/0167-2738(94)90415-490415-4) — foundational ARC characterisation paper.
Richard, M. N., & Dahn, J. R. (1999). Accelerating rate calorimetry study on the thermal stability of lithium intercalated graphite in electrolyte. Journal of the Electrochemical Society, 146(6), 2068–2077. DOI: 10.1149/1.1391893
Roth, E. P., & Doughty, D. H. (2004). Thermal abuse performance of high-power 18650 Li-ion cells. Journal of Power Sources, 128(2), 308–318. DOI: 10.1016/j.jpowsour.2003.09.068

Kinetics and Modelling

Hatchard, T. D., MacNeil, D. D., Basu, A., & Dahn, J. R. (2001). Thermal model of cylindrical and prismatic lithium-ion cells. Journal of the Electrochemical Society, 148(7), A755–A761. DOI: 10.1149/1.1377592 — Arrhenius multi-reaction model foundation.
Kim, G. H., Pesaran, A., & Spotnitz, R. (2007). A three-dimensional thermal abuse model for lithium-ion cells. Journal of Power Sources, 170(2), 476–489. DOI: 10.1016/j.jpowsour.2007.04.018
Feng, X., Sun, J., Ouyang, M., He, X., Lu, L., Han, X., Fang, M., & Peng, H. (2014). Characterization of large format lithium-ion battery exposed to extremely high temperature. Journal of Power Sources, 272, 457–467. DOI: 10.1016/j.jpowsour.2014.08.092

Propagation and Gas Emission

Lamb, J., Orendorff, C. J., Steele, L. A. M., & Spangler, S. W. (2015). Failure propagation in multi-cell lithium-ion batteries. Journal of Power Sources, 283, 517–523. DOI: 10.1016/j.jpowsour.2014.10.081
Koch, S., Fill, A., & Birke, K. P. (2018). Comprehensive gas analysis on large scale automotive lithium-ion cells in thermal runaway. Journal of Power Sources, 395, 135–144. DOI: 10.1016/j.jpowsour.2018.05.080
Sturk, D., Rosell, L., Blomqvist, P., & Brahm, A. (2019). Analysis of Li-Ion Battery Gases Vented in an Inert Atmosphere Thermal Runaway Test. Batteries, 5(3), 61. DOI: 10.3390/batteries5030061

HF Toxicology

Babrauskas, V., Lucas, D., Eisenberg, D., Singla, V., Dedeo, M., & Blum, A. (2014). Flame retardants in building insulation: a case for re-evaluating building codes. Building Research & Information, 40(6). — Contains HF emissions data from polymer fires, applicable to electrolyte combustion.
NIOSH (2019). Hydrogen Fluoride: Immediately Dangerous to Life or Health Concentrations. National Institute for Occupational Safety and Health. https://www.cdc.gov/niosh/idlh/7664393.html

Frank-Kamenetskii Theory

Frank-Kamenetskii, D. A. (1955). Diffusion and Heat Transfer in Chemical Kinetics. Princeton University Press. — Original criticality theory text.
Bowes, P. C. (1984). Self-Heating: Evaluating and Controlling the Hazards. Elsevier. — Applied FK analysis for industrial materials, extensively cited in battery thermal safety literature.

Indian Market and Regulatory Context

ARAI (2022). Testing Protocol for AIS-156 Phase 2 Compliance. Automotive Research Association of India, Pune. https://www.araiindia.com
ICAT (2023). EV Battery Safety Certification Guide for Indian Market Entry. International Centre for Automotive Technology, Manesar. https://www.icat.in
NATRAX (2023). National Automotive Test Tracks: EV Battery Safety Test Facilities Overview. NATRiP, Pithampur. https://www.natrip.in

Thermal Runaway — Kinetics, ARC Testing, Propagation Modelling, and Pack Certification

Thermal Runaway — Kinetics, ARC Testing, Propagation Modelling, and Pack Certification

Table of Contents

What This Article Assumes

ARC Testing — What It Actually Measures

Arrhenius Kinetics and the Self-Heating Model

Frank-Kamenetskii Criticality — The Designer's Tool

Pack-Level Propagation — Modelling Architecture

CFD vs FEA — Which Approach For What Question

Propagation Barrier Material Selection

Vent Path Design for Prismatic Cells

AIS-156 Phase 2 Compliance Architecture

UN ECE R100 Rev 3 vs AIS-156 Phase 2 — Divergences

BMS Firmware Architecture for Thermal Event Detection

HF Toxicology and Pack Venting Strategy

Field Failure Investigation Framework

Lithium Plating as a Propagation Nucleus

Key Design Decisions and Their Thermal Safety Impact

Battery Technology Timeline

Resources and References

Standards and Regulations

ARC and Thermal Characterisation

Kinetics and Modelling

Propagation and Gas Emission

HF Toxicology

Frank-Kamenetskii Theory

Indian Market and Regulatory Context

Further Reading — EVPulse Series

Sai Chaitanya Dasari

Frequently Asked Questions

Part of the deepdive Series