EV battery management systems (BMS) and high-voltage battery packs: Engineering Safety, Failure Analysis and Validation
Introduction and application problem
An EV pack combines high-energy cells, sensing, contactors, cooling and embedded control. A BMS may detect an abnormal voltage or temperature, yet the hazard can persist if the measurement is wrong, the fault is diagnosed too late, a contactor is welded, or the commanded shutdown is not physically achieved. Battery safety is therefore a system argument linking vehicle hazards to layered protection, defined reactions and test evidence—not simply a threshold table or a cell-monitor data sheet.
- Overvoltage, undervoltage, overcurrent and excessive temperature can contribute to cell damage, venting, fire or loss of propulsion; the path depends on chemistry, pack construction and operating conditions.
- Isolation faults can create electric-shock risk during driving, charging, crash response or service.
- Plausible-but-wrong measurements may be more dangerous than obvious open circuits because the BMS may continue permitting charge or discharge.
- Welded contactors, pre-charge faults and failed power switches can defeat an isolation command.
- Battery protection interfaces with charging, inverter torque, thermal management, crash response, service procedures and emergency response.
- Define item boundary, chemistry, pack topology, voltage, charge paths, cooling, service modes and interfaces before safety classification.
- Perform vehicle-level HARA; use FMEA/FMEDA, fault trees and dependent-failure analysis for sensing, switching, thermal and software paths.
- Derive measurable safety requirements for detection, plausibility, reaction time, current interruption, warnings, torque limitation and safe service state.
- Design layered diagnostics: cross-channel plausibility, watchdogs, communication supervision, contactor feedback, isolation monitoring and self-tests.
- Verify faults at component, ECU, pack and vehicle levels, maintaining traceability from hazard through requirement to test.
Application-specific safety and security challenges
- Cell and pack variation: Chemistry, ageing, imbalance and thermal gradients alter the operating envelope; validate justified worst-case corners.
- Fast charging and regeneration: Current direction changes rapidly; define mode-specific limits, sensor bandwidth, filtering delay and disagreement response.
- Thermal runaway precursors: One sensor may miss a local hot spot; evaluate placement, cooling-loss response, propagation assumptions and vehicle warning/egress.
- Crash, service and residual energy: Coordinate crash signals, isolation, discharge and service procedures; a software indication alone is not proof of de-energisation.
Application architecture and operating context
Boundary includes cell monitors/tap harnesses, current and pack-voltage sensing, temperature sensors, BMS controllers, contactors/pre-charge, fuses, insulation monitoring, thermal actuators, charger, inverter, vehicle controller and diagnostics. Define which element owns each safety action and prove that the physical switching path can achieve it. Cover startup, charging, drive, regeneration, crash, sleep/wake, maintenance and emergency modes. Record assumptions for cell specifications, cooling, isolation monitor, communications and contactor lifetime.
Applicable standards and boundaries
- ISO 6469 series: electric-propulsion vehicle safety, including relevant electric-shock/energy hazards; select applicable parts.
- UNECE R100 may apply to electric powertrain and REESS type approval in relevant markets; verify current series and vehicle category.
- ISO 26262: malfunctioning E/E behaviour and safety lifecycle; ASIL comes from vehicle-level HARA, not chemistry or product name.
- ISO/SAE 21434 and UN R155: cybersecurity risk management where applicable, not substitutes for functional safety.
- UN R156 and ISO 24089: update management/engineering where relevant.
- Cell, pack, transport and regional battery requirements must be selected for the product and market; they do not automatically establish vehicle functional-safety compliance.
Applicability note: this is a screening set, not a declaration that every standard applies to every product. Confirm the exact item, vehicle category, market, contractual requirements and current edition/amendment status. A draft standard must not be presented as a published requirement.
Application safety lifecycle
- Item definition and context: define intended use, interfaces, modes, assumptions, users and external dependencies.
- Hazard analysis and safety goals: identify hazardous vehicle-level behaviour and derive goals through the applicable process; do not invent an ASIL.
- Functional and technical safety concepts: specify required behaviour, fault reactions, safe/degraded states and interface assumptions.
- Architecture and implementation: allocate requirements to hardware, software, mechanical/electrical elements and operations; analyse independence and common causes.
- Safety analysis and integration: maintain traceable FMEA/FMEDA, fault trees and dependent-failure analysis where appropriate, with assumptions and data sources.
- Verification and validation: combine reviews, analysis, unit/integration tests, HIL, fault injection, environmental testing and vehicle/application validation as relevant.
- Production, service and change: control versions, calibration, supplier changes, repair, software updates, field feedback and change-impact analysis.
Hazards, failure scenarios and safety objectives
The scenarios below are illustrative engineering examples, not a project HARA, assigned ASIL, approved safety goal or quantitative risk assessment. Final classification depends on the defined vehicle item, operating situations and applicable safety process.
| Illustrative hazardous behaviour | Possible initiating fault / condition | Potential consequence | Example analysis or evidence |
| Cell voltage under-read during charging | Tap wiring or ADC/reference fault | Overcharge and possible thermal event | Plausibility diagnostics, independent bounds and injected sensor faults |
| Contactor weld after shutdown | Contact wear, driver short or sticking | Pack remains energised | Feedback/voltage monitoring and welded-contactor test |
| Cooling failure | Pump/valve fault or sensor blind spot | Temperature exceeds validated envelope | Cooling diagnostics, derating/shutdown requirements and thermal tests |
| Isolation fault | Insulation damage, ingress or harness damage | Shock or unintended current path | Isolation monitoring and environmental/service validation |
Safety requirements, architecture and implementation
Requirements must specify measured quantity, mode, diagnostic condition, allowed latency, reaction, recovery, warning, interface behaviour and evidence. Avoid “BMS shall be safe” or “open contactors immediately” without defining trigger, timing and physical capability. Treat diagnostic coverage and hardware metrics as analysis results, not marketing claims.
Verification, validation and safety evidence
Use traceability reviews; sensor open/short/drift/stuck-value injection; stale/corrupt communication; contactor/pre-charge faults; current/voltage correlation; thermal chamber and cooling-loss tests; HIL mode transitions; pack interruption and residual-energy tests; and vehicle tests for torque limitation, charging interlocks and crash response. Record cell/IC revision, software, calibration, pack build and instruments. A passed test does not prove absence of all internal cell failures.
Common pitfalls and how to detect them
| Pitfall | Why it matters | Detection method / evidence | Better engineering practice |
| Assuming contactor command proves isolation | A welded contactor can leave the pack energised | Faulted contactor feedback and residual-voltage evidence | Monitor achieved state and validate independent interruption |
| Using a single temperature threshold | Mode, chemistry, sensor lag and location change risk | Calibration rationale and thermal-corner tests | Define mode-specific limits and validate sensor placement |
| Relying only on cell-monitor diagnostics | Harness, cooling and switching faults may be outside IC assumptions | Safety-manual assumptions and pack integration review | Maintain pack-level analysis and validation |
| Omitting post-crash/service states | Residual energy and access create distinct hazards | Crash/service procedure review | Include emergency and maintenance modes |
Relationship to adjacent disciplines
Functional safety addresses hazards caused by malfunctioning behaviour of E/E systems. Reliability engineering helps characterise failures and lifetime; cybersecurity addresses malicious threats; SOTIF is relevant where intended-functionality insufficiencies create risk; quality/process assurance supports consistent development; human factors addresses user interpretation; and environmental/electrical/homologation activities address other requirements. These disciplines exchange assumptions and evidence but are not interchangeable. Select only those relevant to the application and document their interfaces.
Currency note — Reference list prepared 24 October 2026 using linked publisher/regulator pages as a starting point. Recheck edition, amendment, status, scope and jurisdiction immediately before publication. ISO catalogue pages can show standards under revision. This informational article is not a substitute for reviewing applicable normative texts.
- ☐ Technical reviewer checked engineering argument, hazard pathways, lifecycle logic and illustrative examples.
- ☐ Standards reviewer verified current edition/status, scope, terminology and regulatory applicability.
- ☐ VerveTronics verified service, qualification, tool, experience and delivery claims.
- ☐ Marketing/editorial reviewer checked keyword ownership, search intent, metadata, canonical URL and overlap.
- ☐ CTA and internal links are live; approved visuals, captions, alt text and structured data are correct.
- ☐ No invented customer case studies, measured results, certification outcomes or unsupported numeric claims.
- ☐ Any real case study/customer detail has appropriate permission and anonymisation.
- Tap-wire, ADC/reference, current-sensor or temperature-sensor faults; drift, open/short, saturation and implausible in-range values.
- Thermal gradients, cooling loss, local hot spots, cell-to-cell propagation and delayed fault recognition.
- Contactor weld, pre-charge failure, fuse coordination, residual voltage and failed interruption.
- Mode-dependent limits and timing during fast charge, regeneration, cold start, crash, sleep/wake and maintenance.
- Stale or corrupted data between cell-monitor ICs, BMS, charger, inverter and vehicle controller.
- Unauthorized calibration or firmware changes affecting limits, diagnostics or protection.
VerveTronics automotive functional safety and ISO 26262 services
- Does ISO 26262 assign an ASIL to a battery pack? ASIL is derived from vehicle-level HARA for defined hazardous events; it is not assigned solely from voltage, chemistry or a component label.
- Does a safety-related cell-monitor IC make the BMS compliant? No. Integration must meet assumptions of use and demonstrate system architecture, diagnostics, fault reaction and validation.
- Can software alone prevent thermal runaway? No. Prevention and mitigation combine cell design, sensing, electrical protection, thermal design, containment and vehicle-level response.
