Battery Failure Response Planning Guide for BESS

A lithium-ion battery incident rarely begins with visible flame. In many cases, cell damage, abnormal charging, internal faults or cooling failures first produce hydrogen and electrolyte vapours, volatile organic compounds, heat and pressure changes inside an enclosure. A battery failure response planning guide gives operators a defined way to act during that early window, before a developing fault becomes thermal runaway, fire, asset loss or a prolonged outage.

For BESS owners, data centre operators, EV charging network managers and industrial facility teams, the objective is not simply to respond faster once smoke appears. It is to identify a developing battery failure early, protect people, control escalation and preserve the information needed to return the site to service safely.

Why battery failure response planning must start before smoke

Conventional fire detection remains necessary, but it is not designed to provide the earliest possible indication of lithium-ion cell failure. By the time smoke or flame is present, conditions may already be severe, access may be restricted and the available response options may be limited.

Lithium-ion batteries can enter thermal runaway after an internal short circuit, mechanical damage, manufacturing defect, overcharge, inadequate cooling, water ingress or external heating. The pathway varies by battery chemistry, state of charge, enclosure design, ventilation and fault location. That means no response plan should assume every alarm represents the same hazard or demands the same action.

An effective plan uses early-stage indicators to create time for controlled decisions. Detection of hydrogen, VOCs and electrolyte vapours can indicate abnormal battery behaviour before conventional fire systems activate. When combined with temperature, humidity and operational data, these signals allow a site to escalate its response according to credible risk rather than waiting for a visible event.

Build the plan around clear alarm states

A response plan fails when an alarm reaches a control room and nobody knows whether to investigate, isolate, evacuate or call emergency services. Define alarm states in advance, map each state to a responsible role and test the sequence under realistic conditions.

A practical model usually has four stages:

• Advisory: An early off-gas or environmental anomaly is detected. The control room records the event, verifies equipment status and begins remote review.

• Warning: Readings persist, rise or occur alongside supporting indicators such as increasing temperature. Access is restricted, the asset owner is notified and trained personnel assess the risk from a safe position.

• Critical: Multiple indicators suggest an escalating battery fault. Stop charging or discharging where the system design permits, isolate affected equipment in accordance with approved procedures and activate site emergency arrangements.

• Emergency: Smoke, fire, venting, rapid temperature rise or confirmed thermal runaway is present. Evacuate, establish exclusion zones, call 000 and transfer incident coordination to emergency services as required.

The thresholds behind these states must be based on the battery system, monitored gases, enclosure volume, ventilation strategy and manufacturer guidance. A single generic setpoint is not a substitute for engineering review. Equally, setting alarms too high to avoid nuisance notifications can remove the very early-warning margin the system was installed to provide.

Assign ownership before an event occurs

Every action needs an owner, an alternate and a defined handover point. A facility manager may own the site emergency plan, while a BESS operator controls electrical isolation, a monitoring centre receives alarms and an electrical contractor provides technical support. If these responsibilities are not documented, valuable minutes are lost finding the right person.

Your plan should identify who can acknowledge alarms, who can command shutdown, who contacts Fire and Rescue, who communicates with tenants or nearby operations, and who approves re-entry. Include after-hours arrangements. Battery faults do not wait for business hours, and an unmonitored overnight alarm can become an operational and safety failure.

Make early detection part of the control architecture

Off-gas detection is most valuable when it is integrated into an operational response, not treated as a standalone sensor. For industrial installations, outputs should be visible through the existing building management system, SCADA platform or remote monitoring environment. Relay outputs can provide local alarm or interlock functions, while Modbus RTU integration can deliver live values, fault status and event history to the wider control system.

The Evikon E2673 is designed for this early-warning role in BESS and other critical lithium battery environments, monitoring indicators including hydrogen, VOCs, electrolyte vapours, humidity and temperature changes associated with failing batteries. Its value is the decision time it can provide before smoke and flames occur. However, detection alone does not make a site safe. The response logic, escalation workflow and training determine whether that warning produces a controlled outcome.

Detector placement also matters. Gas behaviour, airflow, enclosure geometry, extraction systems and likely failure points all influence where sensors should be located. A detector placed near a strong extraction point may see diluted concentrations later than expected. One placed without considering stagnant zones may miss local accumulation. Confirm locations through a site-specific design review rather than relying on a standard layout.

Define safe isolation and access procedures

Electrical isolation must be planned with the battery integrator, electrical engineer and operating team. In some systems, opening a disconnect may stop power flow but leave significant energy stored within battery racks. In others, abrupt shutdown can affect connected critical loads or create additional operational consequences. The correct action depends on the system architecture and approved operating procedures.

Document what can be done remotely, what requires a qualified person on site and what must never be attempted during a suspected thermal runaway event. These procedures should cover charging equipment, HVAC and ventilation operation, emergency shutdown systems, inverter controls, ancillary power and access control.

Access rules are equally important. A team should not enter a container, battery room or enclosed plant area merely to confirm an alarm. Off-gassing may create a hazardous atmosphere before there is visible evidence of failure. Establish exclusion distances, remote inspection options and personal protective equipment requirements with competent safety and emergency response advisers.

Plan communications for people beyond the battery enclosure

Battery incidents can affect far more than the immediate asset. A fault in a commercial BESS may disrupt tenants, energy customers, contractors, security personnel and neighbouring operations. For remote solar farms and infrastructure sites, delayed access and long emergency service travel times add further complexity.

Prepare short, factual notification templates for internal teams, emergency services, clients and adjacent site occupants. They should state the asset location, battery type where known, current alarm condition, access route, isolation status, known hazards and the site contact. Avoid speculation. The incident controller needs accurate, current information, not assumptions about the cause.

Keep up-to-date emergency information at the site and make it accessible to responders. This may include single-line diagrams, battery layout drawings, SDS documentation, container access details, shutdown instructions and contact numbers. Review whether emergency crews can safely reach the asset, particularly where gates, remote access systems or constrained site layouts apply.

Train for decisions, not just alarms

A response plan on a server is not a response capability. Control room staff, site operators, maintenance contractors and security personnel need to understand what the alarm means, which actions are authorised and when escalation is mandatory.

Tabletop exercises are an efficient starting point. Run scenarios such as a single off-gas alert during normal operation, repeated readings overnight, a detector fault during extreme weather, or a warning alarm coinciding with rising battery temperature. Each exercise should test notification paths, remote shutdown decisions, contractor availability and emergency-service handover.

Then test the interfaces. Confirm that alarm signals reach SCADA, that relay outputs trigger the intended annunciators or interlocks, that the monitoring centre has the correct contact list, and that event data is retained. A planned test window is far preferable to discovering a configuration gap during an incident.

Treat recovery as a controlled safety phase

The end of an alarm does not automatically mean the hazard has ended. Damaged cells can retain heat, release gases intermittently or reignite after an initial event. Re-entry, inspection, battery removal and recommissioning should occur only under an approved recovery process developed with relevant technical specialists and emergency authorities where required.

Preserve alarm trends, temperature records, BMS data, CCTV footage and operator logs. These records help identify root causes and determine whether the event was contained to one module or reflects a wider system issue. They also provide a practical basis for improving setpoints, maintenance intervals, ventilation controls and future response procedures.

The strongest battery failure response plans do not promise that every incident can be prevented. They ensure that when early warning appears, the people responsible know exactly how to turn those critical minutes into safer decisions.