Most of the time a lithium-ion battery is a quiet, well-behaved store of energy, but under the wrong conditions a single cell can begin releasing its stored energy faster than it can shed the heat, and once that happens the cell can drive itself to destruction. That failure mode is thermal runaway. In a battery energy storage system it is the safety event that operators most need to catch early, because a runaway in one cell can spread to its neighbours and cascade through a whole module. This guide explains what thermal runaway is, what triggers it, the early warning signs an operator can watch for, and the detection and suppression measures that codes such as NFPA 855 address.
Battery thermal runaway in one line: Battery thermal runaway is a self-sustaining chain reaction inside a cell in which internal heat drives exothermic chemical reactions that generate still more heat, so the cell temperature climbs uncontrollably even after the original trigger is removed. As it runs away the cell vents flammable gas, can catch fire, and can heat adjacent cells enough to make them run away too, cascading through a module. Because it feeds on itself, the priority is to detect it early through rising temperature and off-gas signatures and to contain propagation before it spreads.
Thermal runaway is defined by its self-sustaining nature. Something first heats a cell beyond a threshold, and past that point the chemistry inside the cell begins to break down in ways that themselves release heat. That released heat pushes the temperature higher, which accelerates the reactions further, which releases still more heat. The loop feeds itself, so the cell continues to heat even if whatever caused the initial rise is gone. This positive feedback is what separates thermal runaway from ordinary overheating, where removing the heat source lets a component cool back down.
As the cell climbs through this cascade it passes through recognisable stages. Early on the internal chemistry begins to decompose and the cell starts to swell and generate gas. The safety vent then opens and the cell releases a mixture of hot, flammable off-gas, an event usually called cell venting. If the temperature keeps rising the released material can ignite, and the cell can rupture. Each stage happens faster than the last, which is why the window to act is widest at the very beginning, before venting and fire.
The energy that drives all of this was already stored in the cell; runaway simply releases it in an uncontrolled way. That is why a fully charged cell tends to release more energy in a runaway than a partly charged one, and why the gases and heat produced are treated as a hazard in their own right. Understanding runaway as an internal, self-feeding release of stored energy, rather than as a fire that arrives from outside, is the key to reading the warning signs and to appreciating why containment matters as much as extinguishing.
Several conditions can push a cell past its runaway threshold. Mechanical damage that creates an internal short circuit, electrical abuse such as overcharging or over-discharging, a manufacturing defect that seeds an internal fault, and external heating from a nearby fire or a failed neighbour are the common triggers. Whatever the specific cause, they share a result: enough localised heat inside the cell to start the self-feeding reactions. Once one cell is in runaway, the heat it produces becomes the trigger for the cells packed around it, and this cell-to-cell spread is called propagation. Preventing or slowing propagation is central to keeping a single-cell event from becoming a module or rack event.
The value of watching for early signs is that the first stages of runaway announce themselves before there is any fire. Cell or module temperature rising abnormally fast, a cell voltage that collapses or behaves erratically, and physical swelling are all indicators. The most useful early signal in many systems is the off-gas that a venting cell releases, because gas detection can pick up the characteristic vented species before temperature alone would trigger an alarm. An off-gas detector positioned in the battery enclosure can therefore give operators the earliest possible warning that a cell has begun to vent.
Reading these signs is a symptom-to-cause exercise. A symptom of a sharply climbing cell temperature together with a dropping voltage points to a cell entering runaway, whose likely causes include an internal short from damage or defect or sustained overcharge. A symptom of a gas-detection alarm with little temperature change yet points to early venting, whose likely cause is a cell that has just opened its safety vent. The diagnostic steps follow the same logic: confirm which cell or module the alarm maps to, check whether charge or discharge was in progress and stop it, verify the reading against a second instrument or an adjacent sensor to rule out a faulty transducer, and treat any genuine off-gas or fast temperature rise as an incipient runaway that calls for the system's protective response rather than a slow investigation.
Because runaway is fast and self-sustaining, a battery energy storage system is built to detect and respond to it rather than merely record it after the fact. Standards for stationary storage, notably NFPA 855, drive the presence of measures such as gas detection, temperature monitoring, ventilation or deflagration management to handle vented gases, and separation and spacing intended to limit propagation between units. The battery management system provides the electrical picture, watching cell voltages and temperatures and cutting off charge or discharge when a cell drifts outside safe bounds, while dedicated gas and heat detection provides the safety-event picture. Together they aim to catch a runaway at the venting stage and to keep it from cascading.
This is where cloud SCADA and remote monitoring earn their place, especially for storage assets that sit at unmanned or lightly staffed sites, including battery systems paired with oil and gas facilities or remote power. A cloud SCADA platform such as Merobix can gather cell and module temperatures, off-gas detector states, and battery management system alarms and surface them centrally, so that a fast temperature climb or a gas alarm raises an immediate notification to on-call staff no matter where the site is. Trending these values over time also helps, because a cell that is slowly warming or drifting relative to its neighbours can be flagged as a maintenance concern long before it reaches a runaway threshold.
It is worth being clear about what remote monitoring does and does not do. The physical protection - venting, spacing, suppression, and the battery management system's own cut-offs - is what actually contains a runaway, and those act locally and automatically because the timescales are too short to wait for a human. Cloud monitoring adds the earliest possible human awareness, an auditable record of temperature and gas trends, and the ability to distinguish a real event from a nuisance alarm by looking at multiple signals together. Treated that way, gas-detection and temperature alarms surfaced through SCADA become the operator's early-warning layer on top of the automatic safety systems, not a replacement for them.
The earliest signs are usually an abnormally fast rise in cell or module temperature, a cell voltage that drops or behaves erratically, and physical swelling of the cell. In many systems the off-gas released when a cell vents is detectable even sooner, so a gas-detection alarm in the battery enclosure can give the first warning before temperature alone would. Catching any of these early gives the widest window to stop charge or discharge and let the protective systems act.
Once a cell is fully in runaway it is self-sustaining and continues to heat even without the original trigger, so the realistic goal shifts from stopping that cell to containing it and preventing propagation to neighbouring cells. Removing electrical abuse, cooling, and the enclosure's suppression and separation measures aim to limit spread and manage vented gas rather than to reverse the reaction in the affected cell. This is why early detection at the venting stage, before neighbours are involved, matters so much.
NFPA 855 is the standard for the installation of stationary energy storage systems, and it drives the safety measures meant to detect and contain events like thermal runaway. In practice that includes provisions around gas detection, ventilation or deflagration handling for vented gases, and spacing and separation intended to limit propagation between units. It shapes how a battery energy storage system is arranged and instrumented so that a single-cell runaway is caught and kept from cascading.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.