Fire alarm system

A fire alarm system detects or receives an indication of fire, processes that information and produces the warning or control signals required by its design. The SFPE Handbook describes these functions as detection, processing and signalling. The system can warn occupants and initiate other protective actions, such as smoke control or door closing.

Its performance depends on the complete sequence from fire development to the intended response. Smoke or heat must reach an automatic detector before it can operate. Processing and verification may introduce further delay, followed by notification or intervention. A detector's activation time is therefore only one part of the system's response time.

Engineering assessment starts with the safety objective and the fire scenarios against which it will be checked. Detector selection, location, warning arrangements and human response are then considered together. Testing and management support those assumptions, while analysis identifies delays, uncertain inputs and shared failures that could prevent the intended action.

Detection, processing and signalling

Schifiliti, Meacham and Custer divide a fire alarm system into three functional elements: detection, processing and signalling. Detection senses a fire characteristic or receives an initiating input. Processing evaluates that input according to the system's configured functions. Signalling then communicates the warning or command required by the design.

An automatic initiating device responds to a physical fire signature: a Smoke detector responds to particles, while a Heat detector responds to thermal conditions. A Manual call point allows a person who recognises a fire to initiate the alarm. These inputs depend on different mechanisms and need not operate at the same time.

The outputs can include an Evacuation alarm system and commands to other protective systems. The SFPE Handbook gives smoke control, lift recall, door closing and signalling to the fire service as examples. Which outputs operate depends on the intended function of the particular installation. An initiating signal and a completed protective action are therefore different events in the analysis.

This functional description does not prescribe a particular control-panel arrangement. The handbook treats panel selection and detailed auxiliary control design as outside its detection calculation chapter, and its method instead follows what each function needs to achieve and when. That distinction allows detector performance to be assessed without assuming that every alarm installation has identical hardware.

Fire alarm components serve a shared safety goal. Based on Schifiliti, Meacham, and Custer (2002), pp. 4-1–4-3.

Defining what the system must achieve

The system objective determines what counts as sufficiently early detection. For life safety, warning must leave enough time for the intended occupant response before conditions become untenable. Other objectives can concern damage to equipment, business interruption or environmental consequences. The SFPE Handbook begins its design process by identifying these goals.

A performance objective becomes more useful when expressed through measurable conditions. Examples include limiting the fire's heat release rate or preventing damage beyond a specified item. Heat release rate describes how quickly the fire releases energy. The selected objective determines the latest acceptable point for detection and the subsequent response.

The fire may continue growing after the detector operates. Processing, warning, staff response or suppression initiation can consume additional time. Schifiliti and colleagues therefore distinguish the limiting design fire from the smaller fire that must trigger detection. The difference depends on fire growth during the remaining response period.

Several fire scenarios are considered because fuels, growth rates and locations can vary; the handbook describes testing a candidate design against alternative scenarios after establishing its detector type and spacing. A system that satisfies one scenario may fail its objective in another. This examination connects the detection arrangement to the hazards in the actual occupancy.

Fire signatures and detector response

A fire signature is a measurable effect of combustion that can be sensed by a detector. Smoke particles, elevated temperature and emitted radiation are different signatures, and their strength and transport depend on the fuel and the fire's development. Selecting a detector therefore involves identifying the signature expected before the limiting fire condition occurs.

Heat detection depends on gases transferring energy to a sensing element, and the gas temperature and velocity at the device affect its rate of heating. Smoke detection depends on particle transport, entry into the sensing chamber and the sensor's response to those particles. Neither process is represented completely by a device's nominal alarm threshold alone.

Ceiling height, beams and other obstructions can change the route taken by fire products, and a higher plume can entrain more air before reaching the ceiling, reducing temperature and concentration. Air currents are particularly relevant to smoke transport during weak early fires. These effects explain why detector location is part of performance assessment.

Product testing supplies evidence about response under defined conditions; the handbook distinguishes this evidence from a calculation of the installed system's response to a particular fire. Test fuels, airflow and geometry may differ from those in the building. A stated installation spacing therefore needs to be interpreted alongside the conditions and objective of the assessment.

Processing decisions and response delays

Processing can introduce a delay between an initial detector signal and the required output. Alarm verification is one example discussed in the SFPE Handbook, and it is intended to address nuisance signals, but its delay remains part of the response sequence. A timing calculation that omits it can overstate the time available for intervention.

The International Fire Engineering Guidelines (IFEG, 2005) describe coincident detection, which requires two or more detectors to operate before an alarm is recognised. The inputs may come from similar detectors at different locations or from different detector types. Such logic changes the condition required to initiate the output. It must therefore be represented when estimating activation rather than treating the first detector as sufficient.

Unwanted fire alarms and genuine fire response need to be considered together. Schifiliti and colleagues describe signal-processing algorithms intended to reduce nuisance alarms and identify fire signatures, and their operation can vary between products and may not be fully published. A simple threshold calculation may consequently omit part of the actual decision process.

A warning or control output is useful only in relation to its intended recipient or equipment; the same initiating event may lead to several different actions in the building. The functional description therefore follows each output to the action relevant to the safety objective. This prevents a control-panel indication from being treated as if evacuation or suppression were already complete.

Worked example of the response time budget

Example 20 in the SFPE detection chapter considers a hotel ballroom and an intended staff response to a chair fire. It assigns 10 s to the system delay between detector activation and alarm initiation. Alarm verification ranges from 15 s to 60 s in that example. Staff take a further 60 s from the indicated alarm to arrival.

The post-detection subtotal is found by adding these successive delays. Using the longer verification time gives 130 s. Using the shorter verification time gives 85 s. Both calculations exclude the time taken for smoke to reach the detector, which the source identifies separately.

$$t_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}=t_{\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{m}}+t_{\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{f}\mathrm{y}}+t_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{f}\mathrm{f}}$$

Symbol	Description	Unit
$t_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}}$	Subtotal after detector activation, excluding smoke transport	s
$t_{\mathrm{s}\mathrm{y}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{m}}$	System delay used in the example	s
$t_{\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{f}\mathrm{y}}$	Alarm verification delay	s
$t_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{f}\mathrm{f}}$	Time from alarm indication to staff arrival	s

$$t_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t},\mathrm{l}\mathrm{o}\mathrm{n}\mathrm{g}}=10+60+60=130\mathrm{s}$$ $$t_{\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t},\mathrm{s}\mathrm{h}\mathrm{o}\mathrm{r}\mathrm{t}}=10+15+60=85\mathrm{s}$$

These are source-specific example inputs, not standard delays for fire alarm systems. The difference illustrates why a processing feature can change the time available for the intended intervention. The complete assessment must still account for fire development and transport to the detector. It must also justify which verification and staff-response assumptions apply to the scenario.

A response-time budget separates each delay. Based on Ministry of Business, Innovation and Employment (2025), pp. 26–30.

Connecting the alarm to occupant response

Notification is the point at which information reaches the occupants, not the moment at which an automatic detector first senses a fire. New Zealand's C/VM2, second edition (2025), separates detection time from notification time, and it also distinguishes pre-travel activity after notification from movement towards safety. These periods perform different roles in its required safe egress time calculation.

For standard evacuation strategies, C/VM2 assigns 30 s to notification. Where management investigates an activation before notifying occupants, the method requires the extended notification period to be considered. Its pre-travel activity values also vary with occupant group, location and warning type. These are provisions of the New Zealand method rather than universal properties of alarm equipment.

A Voice alarm system can supply information that a tone alone does not convey. Proulx describes the influence of informative messages, staff instructions and familiarity on the decision to start moving. Audibility and intelligibility both affect whether a spoken warning can serve its purpose. The alarm's electrical operation and the occupant's response therefore require separate evidence.

Occupants may also receive direct fire cues or warnings from other people. IFEG requires the cue used in an evacuation analysis to be identified and its selection justified. Different groups in different enclosures can receive different information. Treating every occupant as responding immediately to the first panel signal would bypass this part of the assessment.

Reliability and shared dependencies

Several initiating routes can support the same protective action. IFEG gives a shopping-centre example involving detection by people, smoke detectors or sprinklers. Smoke control can likewise receive automatic or manual activation signals. These alternatives provide useful redundancy only when the complete functional paths are considered.

A redundancy study looks for single points of failure within apparently separate arrangements. If several paths depend on one component, failure of that component can remove their shared function. IFEG therefore asks what backup remains after an individual failure. Counting initiating devices alone does not establish the independence of the resulting warning or control actions.

The same assessment distinguishes failure from reduced performance. A delayed response can matter even when the equipment eventually operates. IFEG describes sensitivity studies that vary important inputs and uncertainty studies that examine confidence in the outcome. These methods support an explicit treatment of the timing and reliability assumptions used in the design.

The analysis also considers whether failure would be recognised by those responsible for the building. In IFEG's example, the objective includes avoiding unrecognised equipment failure and retaining a backup route, and this connects technical reliability to the arrangements for operating and managing the building. An alternative activation path cannot help if its availability is only assumed.

Redundant paths can share a single point of failure. Based on Australian Building Codes Board et al. (2005), p. 1.2-24.

Testing, management and the limits of models

Fire safety management supports the response credited to the installed system. Staff who receive an indication need procedures consistent with the assumed action and timing. Proulx links training to the particular building and its evacuation arrangements. The modelled benefit of staff intervention therefore depends on more than the presence of an alarm output.

A specialised example appears in the 2010 supplement to NUREG/CR-6850 for United States nuclear facilities, and its very early warning detection assessment depends on acceptance testing, maintenance, calibration and documented settings. It distinguishes an incipient alert from a later fire alarm. These conditions belong to that nuclear cabinet application and are not presented as general building rules.

The same report requires reductions in sensitivity to be assessed for their effect on early detection. It also relates the credited benefit to personnel locating the source and taking the intended action. This example demonstrates how an engineering claim can depend on both equipment and operating procedures. Changing either can change the performance assumed in the assessment.

A verification method can prescribe analysis inputs that differ from the characteristics of the installed product. In New Zealand, C/VM2 (2025) requires its tabulated detector values to be used regardless of make or model. Its heat-detector representation uses an activation temperature of 57 °C and a response time index of 30 m^1/2s^1/2. The response time index characterises the thermal lag of the sensing element.

The same table assigns an optical density of 0.097 m⁻¹ at alarm for its smoke-detector representation. Optical density describes attenuation of light by smoke along a specified distance. Projected-beam detectors are treated separately using the beam length and total obscuration setting. These distinctions concern the calculation required by that method, so they do not establish universal product sensitivities.

A performance report therefore distinguishes prescribed analysis values, measured device characteristics and assumptions about system operation. Each answers a different question about the proposed arrangement. The handbook's detector calculations need suitable fire and sensor inputs, whereas the New Zealand method specifies some of those inputs directly. Reporting their origin makes the calculation reproducible and exposes where a different method could produce a different result.

Detector models also have limits independent of maintenance. Schifiliti and colleagues identify uncertainty in smoke properties, transport and manufacturer response algorithms. Simplified response estimates need to be interpreted within those limits. Where the required evidence is unavailable, a precise calculated time does not by itself establish precise system performance.

See also

• Reliability of fire protection systems
• Flame detector
• Aspirating smoke detection
• Fire curtain

References

• Australian Building Codes Board, National Research Council of Canada, International Code Council, & Department of Building and Housing New Zealand. (2005). International fire engineering guidelines. Australian Building Codes Board.
• Ministry of Business, Innovation and Employment. (2025). C protection from fire: Verification method C/VM2 (2nd ed.).
• Proulx, G. (2002). Movement of people: The evacuation timing. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-342–3-366). National Fire Protection Association.
• Schifiliti, R. P., Meacham, B. J., & Custer, R. L. P. (2002). Design of detection systems. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., pp. 4-1–4-43). National Fire Protection Association.
• U.S. Nuclear Regulatory Commission, & Electric Power Research Institute. (2010). Fire probabilistic risk assessment methods enhancements (NUREG/CR-6850, Supplement 1).