Electrical Fire Monitoring System

What an Electrical Fire Monitoring System Is

An electrical fire monitoring system is a distributed instrument whose job is to detect the electrical conditions that precede an electrical fire and to alarm on them, while leaving the circuit energized. It is built from two layers: field detectors installed inside distribution boards, cabinets, and at cable junctions, and a central monitoring panel (the controller) that collects detector signals, logs events, and drives local and remote indication. The defining characteristic, and the source of most confusion among first-time buyers, is that the system warns rather than trips. A residual current circuit breaker exists to disconnect power and protect life; an electrical fire monitoring detector exists to flag a slowly developing insulation fault long before it would ignite, without de-energizing the load.

The problem it solves is specific. A large share of building fires originate in the fixed electrical installation, from three slow mechanisms: insulation that degrades and leaks current to earth, bolted or clamped connections that loosen and overheat, and conductors that develop low-level series or parallel arcs. None of these draws enough fault current to operate an overcurrent breaker. Insulation leakage of a few hundred milliamps, a terminal running at 90 degrees Celsius, or an intermittent series arc all sit comfortably below the trip curve of the protecting device, yet each dissipates enough heat to char insulation and eventually ignite surrounding material. The monitoring system is engineered precisely to catch the band between normal operation and the overcurrent trip threshold.

The regulatory history runs on two tracks. In China the concept was formalized by the GB 14287 series, Electrical fire monitoring system, whose detector parts were revised in 2014 (Parts 2, 3, and 4) and extended in 2025 with Part 5 for pyrolysis particle sensing. Installation and design rules were folded into GB 50116-2013, the code for design of automatic fire alarm systems, which sets the residual current alarm band and dictates where detectors sit in the distribution hierarchy. Internationally, the same fire-precursor concern crystallized around arcing: IEC 62606 was first issued in 2013 and amended in 2017 and 2022, partly derived from the older North American UL 1699 arc fault circuit interrupter standard, and IEC 62606 devices were progressively written into national wiring rules such as the IEC 60364 family and BS 7671.

In scale terms, an electrical fire monitoring system spans from a single distribution board in a small commercial unit to thousands of monitored points across a metro line, a data center, or a high-rise residential tower. A typical mid-size project monitors residual current and temperature on every outgoing feeder of the main and sub distribution boards, which can mean several hundred channels reporting over a serial bus to one or two panels. The system is therefore as much a data-acquisition and communications problem as it is a sensing problem, and modern installations increasingly stream detector data to a building management system or cloud platform for trend analysis.

Four engineering attributes determine the quality of an electrical fire monitoring system: detection coverage (which fault mechanisms it senses), setpoint correctness (alarm thresholds matched to the real standing leakage and thermal profile of each circuit), nuisance-alarm immunity (the ability to ignore normal leakage, inrush, and operational arcing), and communications reliability (whether every channel reports without bus collisions or dropouts). A system that nuisance-alarms is quickly ignored by operators, which defeats its entire purpose, so discrimination quality matters as much as raw sensitivity.

Detector Types and System Architecture

By sensed quantity, electrical fire monitoring detectors fall into four families plus a combined category, each mapped to a part of GB 14287. Choosing which families a project needs is the first selection decision, because each family catches a different fault mechanism and no single detector covers all of them. The table below summarizes the detector families, the standard part that governs each, and what each one senses.

Residual current detectors are the workhorse of the GB system. A zero-sequence current transformer encircles all live conductors of a circuit (line and neutral together) so that, in a healthy circuit, the vector sum of currents is zero. Any leakage to earth, through degraded insulation or a moisture path, unbalances the sum and induces a residual current the detector measures. The detector does not trip; it compares the measured leakage against a configured alarm setpoint and signals the panel. Because every real feeder carries some natural standing leakage from cable capacitance and connected equipment, the alarm band is set well above that baseline, which is why GB 50116-2013 places it at 300 to 500 mA rather than the 30 mA used for shock protection.

Temperature sensing detectors attach directly to the points that overheat first: cable lugs, busbar joints, breaker terminals, and contactor contacts. A loose or oxidized connection develops contact resistance, and the heat it dissipates rises with the square of current. The detector alarms when the measured surface temperature crosses a threshold, and better units also alarm on rate of rise so that a rapidly heating joint is flagged before it reaches the absolute limit. This family is essential because connection heating produces little or no change in leakage current, so a residual current detector would never see it.

Fault arc detectors, governed by GB 14287.4 and equivalent to the IEC 62606 AFDD, target the third mechanism: arcing. A series arc forms across a broken strand or a loosening terminal; a parallel arc forms between conductors or to earth through failed insulation. Both produce a distinctive high-frequency current signature with momentary current gaps near the zero crossing, which the detector recognizes while ignoring the harmless arcing of switches and brushed motors. Pyrolysis detectors, the newest family under Part 5, sense the sub-visible particles released when insulation begins to thermally decompose, providing the earliest possible warning. Combined detectors fuse residual current and temperature, or all three, in one housing and must satisfy every applicable part of GB 14287.

Architecturally, the field detectors report to a central monitoring panel, the GB 14287.1 equipment, over a serial bus. The dominant physical layer is RS-485 two-wire multidrop running Modbus RTU, which lets one bus segment carry dozens of detectors over distances up to roughly 1,200 m; some addressable systems instead use a proprietary loop where each detector carries an address code on the same pair as its supply. The panel aggregates alarms, time-stamps and logs events to non-volatile memory, drives local audible and visual indicators, and increasingly forwards data to a building management system or cloud over Ethernet, 4G, or LoRa. Although the electrical fire monitoring loop is kept physically separate from the building's smoke and heat detection loop, its alarms are repeated in the fire control room so that one operator sees the complete fire-safety picture.

Sensing Technologies and Principles

Underneath the four detector families sit a small set of physical sensing principles. Understanding them explains why a given detector has the range, accuracy, and immunity it does, and why one principle cannot substitute for another. The table below compares the core sensing technologies, their typical measurement span, and where each is the right tool.

The zero-sequence current transformer is the heart of residual current sensing. All live conductors of the monitored circuit pass through a single toroidal core; the magnetic fields from balanced line and return currents cancel, so the secondary winding sees nothing in a healthy circuit. A leakage path to earth removes that balance and induces a small secondary current proportional to the residual. Aperture (through-line) diameters run from about 25 mm for branch circuits up to 150 mm for main feeders, sized to the cable bundle, and the same core principle serves main-circuit currents from tens of amps up to 1,000 A or more. Because the CT measures the net imbalance and not the load current itself, a single small core can monitor a very large feeder.

Contact temperature sensing uses NTC thermistors, platinum RTDs, or thermocouples clamped or bonded to the conductor at the heat-prone point. NTC thermistors are inexpensive and sensitive over the modest range that matters for cable insulation, while Pt100 and Pt1000 RTDs offer better long-term accuracy and a wider span for busbar work. The practical challenge is mounting: the sensor must be in good thermal contact with the joint yet electrically isolated from it, so detectors use insulated ring-lug or clamp mountings rated for the system voltage. Wireless temperature spots that harvest energy from the conductor's own magnetic field are increasingly used where a wired sensor cannot reach.

Infrared (contactless) sensing reads the surface temperature of a connection without touching it, which suits moving or rotating contacts and points at hazardous potential. The detector aims a thermopile at a fixed spot and infers temperature from emitted infrared. Its accuracy depends on the emissivity of the target surface, so installations calibrate against the actual busbar finish; its advantage is complete electrical isolation and the ability to watch a spot that no clamp could reach.

High-frequency current analysis is what makes arc fault detection possible. A healthy sinusoidal load current is smooth, but an arc chops the waveform, producing momentary current gaps near the zero crossing and emitting broadband noise. The AFDD samples the current at high speed and examines spectral content, typically across the 100 kHz to 1 MHz band, feeding the result to a microprocessor that must distinguish a dangerous arc from the benign arcing of a light switch or a vacuum cleaner motor. This discrimination is the entire engineering difficulty of arc detection: too sensitive and the device nuisance-trips on every appliance, too coarse and it misses the fault it exists to catch. IEC 62606 defines the masking and discrimination tests a compliant device must pass, and for high arc currents requires a trip in under 120 ms; UL 1699 requires detection within eight arcing half cycles at 75 A or greater.

Standards, Installation, and Setpoints

Specifying an electrical fire monitoring system correctly means satisfying both a product standard, which the equipment must be certified against, and an installation code, which dictates where detectors sit and how their setpoints are chosen. Confusing the two is a frequent submittal error: a certificate proving a detector meets GB 14287.2 says nothing about whether it is installed and set per GB 50116-2013. The table below maps the governing standards by region and function.

The single most consequential installation rule is the residual current alarm setpoint. GB 50116-2013 fixes the alarm value of a residual current electrical fire monitoring detector at 300 to 500 mA. The reasoning is physical: every real feeder carries standing leakage from distributed cable capacitance and the input filters of connected equipment, often tens to a couple of hundred milliamps, so a setpoint anywhere near the 30 mA shock-protection level would alarm constantly. The engineer measures the actual standing leakage of each circuit, applies a margin of roughly two, and selects a value inside the 300 to 500 mA window. Where a single feeder's natural leakage already approaches or exceeds 500 mA, the code requires moving the detector downstream to the next-level distribution cabinet so each branch is monitored separately and carries less leakage.

Detector placement follows the fault mechanism. Residual current detectors sit at the head end of the low-voltage installation, typically on the outgoing side of the first-level distribution cabinet, because that location sees the aggregate leakage of everything downstream. Temperature detectors mount at the specific heat-prone points: cable terminations, busbar joints, breaker and contactor terminals, since connection heating is local and would be invisible to an upstream current sensor. Arc fault detectors protect the final branch circuits, especially long socket-outlet runs and aged wiring most prone to series arcing. Pyrolysis detectors are placed inside enclosures where early insulation overheating concentrates smoke-precursor particles.

Temperature setpoints are application-specific. For PVC-insulated copper conductors, a common alarm setting is in the 55 to 70 degrees Celsius band, well below the insulation's rated limit, because the insulation begins to age and embrittle long before failure. For busbar joints and switchgear terminals, which run hotter in normal service, settings of 90 to 105 degrees Celsius are used. Rate-of-rise alarming complements the absolute threshold, catching a connection that is heating abnormally fast even before it reaches its setpoint. The general environmental ratings of the detectors themselves are modest: typical operating ranges are -10 to +55 degrees Celsius ambient at up to 95 percent relative humidity non-condensing, since the equipment lives inside distribution boards rather than in the open.

For the international AFDD path, IEC 62606 sets the device parameters: rated voltages up to 440 V AC, rated frequencies of 50 Hz, 60 Hz, or 50/60 Hz, and rated currents up to 63 A, with the standard defining how the device must mask operational arcing and how fast it must trip on a real fault. Because IEC 62606 is partly derived from UL 1699, the two converge on the same goal but differ in test current and timing detail, which is why a device sold globally often carries both marks.

Key Specification Parameters

A detector datasheet lists many parameters, but only a handful drive the selection decision and the GB acceptance test. The Key Specifications comparison below contrasts the three core GB detector families and the IEC arc fault device on the parameters that matter, drawn from published manufacturer datasheets and the governing standards. Treat the numbers as representative typical values; always confirm against the specific model datasheet before purchase.

Residual current measurement range and alarm setpoint. Detectors commonly measure from about 200 mA to 1 A with a 1 mA setting resolution, and the alarm setpoint must land in the GB 50116 band of 300 to 500 mA. The CT aperture is sized to the cable bundle, with through-line diameters from roughly 25 mm up to 150 mm; the same family monitors main-circuit currents from tens of amps to 1,000 A because the CT measures imbalance, not load. The detector's own quiescent draw is small, on the order of a fraction of a milliamp standby rising to under a milliamp in alarm.

Temperature range and threshold. Contact sensors based on NTC thermistors typically span -20 to +120 degrees Celsius, RTD-based units reach +200 degrees Celsius, and infrared spots read up to +250 degrees Celsius. The configured alarm threshold, not the sensor span, is what matters in service, and as noted it runs 55 to 70 degrees Celsius for PVC-insulated conductors and 90 to 105 degrees Celsius for busbar joints. Sensor accuracy of plus or minus 1 to 2 degrees Celsius is typical and adequate, since the alarm decision is about a margin of tens of degrees, not fine resolution.

Electrical and communication ratings. Detectors operate on the monitored 230 / 400 V AC system or on a separate auxiliary supply, and report over RS-485 running Modbus RTU at 9600 or 19200 baud with addresses from 1 to 247; addressable-loop variants carry a code from 1 to 255. Bus length to roughly 1,200 m and a few dozen nodes per segment are normal. The AFDD differs because it is a protective device, not a reporting detector: it carries a rated current up to 63 A, a rated voltage up to 440 V AC, and trips locally, with auxiliary contacts available to signal the monitoring panel.

Environmental and timing. Because the equipment lives inside enclosures, environmental ratings are modest, typically -10 to +55 degrees Celsius and up to 95 percent relative humidity non-condensing. Residual current and temperature detectors are alarm-only and have no trip time, but they do have an alarm-confirmation delay that filters transients; arc fault devices, by contrast, carry the hard timing requirement of detecting and tripping in under 120 ms on a high arc current under IEC 62606, or within eight arcing half cycles at 75 A or greater under UL 1699.

Selection Decision Factors

Translating the preceding chapters into a specific system specification follows the ordered sequence below. As with most instrument selection, mistakes come less from a single wrong parameter than from making a downstream decision before an upstream one is settled. These steps double as an RFQ template.

1. Define the fault coverage required: decide which mechanisms the project must catch, leakage, connection heating, arcing, or pyrolysis particles, and therefore which detector families (GB 14287 parts) are needed. Most projects start with residual current plus temperature and add arc detection on vulnerable branch circuits.
2. Map the distribution hierarchy: locate residual current detectors at the head of the low-voltage system, temperature sensors at every heat-prone joint, and arc detectors on final branch circuits. Where a single feeder's natural leakage approaches 500 mA, plan to push the residual current detector to the next-level cabinet.
3. Set the alarm thresholds: measure the real standing leakage of each circuit and choose a residual current setpoint inside the 300 to 500 mA GB 50116 band; choose temperature thresholds per insulation type, 55 to 70 degrees Celsius for PVC, 90 to 105 degrees Celsius for busbar, and enable rate-of-rise where available.
4. Size the current transformers: match the CT aperture to the cable bundle (about 25 to 150 mm) and confirm the rated monitored current covers the feeder up to 1,000 A. Verify the CT class and the residual measurement resolution.
5. Confirm certification: require GB 14287 part certificates for each detector family and the panel, and for international AFDD content require IEC 62606 or UL 1699 marks. Check that the panel and detectors are tested as a compatible set, because GB acceptance testing examines them together.
6. Specify communications and integration: RS-485 Modbus RTU is the default field bus; confirm baud rate, address range, and maximum nodes per segment, and define how the panel forwards alarms to the building management system or fire control room over Ethernet, 4G, or LoRa.
7. Verify environmental fit: confirm the -10 to +55 degrees Celsius and 95 percent humidity ratings suit the enclosure environment, and check ingress protection and mounting for the field detectors.
8. Evaluate total cost of ownership: count the panel, detectors, CTs, wiring, commissioning, and the recurring cost of periodic functional testing. A system that nuisance-alarms is operationally worthless regardless of price, so weight discrimination quality heavily.

One dimension that buyers routinely underestimate is system serviceability and single-vendor compatibility: whether the panel and detectors come from one verified family, whether spare detectors and CTs are stocked locally, whether the Modbus register map is documented for BMS integration, and whether firmware can be updated in the field. For the GB 14287 market the established suppliers are Acrel (ARCM residual current and temperature detectors with matching panels) and Four-Faith (F-EFD100 residual current monitor), alongside numerous combined-detector and panel makers. For the IEC 62606 arc fault device the global brands are ABB (S-ARC1), Schneider Electric (Acti9 iARC), Eaton (AFDD+), Siemens (5SV6), and Hager, with residual current monitoring relays also offered by Schneider Vigilohm, Socomec, and Bender. Selecting a coherent, certified set from a single source or a verified compatible pairing is what makes acceptance testing and ten-year maintenance straightforward.