Inferensys

Glossary

Arc Flash Detection

A protection method using optical sensors or pressure detectors to identify the intense light and pressure of an internal arc fault, triggering an ultra-fast trip to minimize equipment damage.
Technical lab environment with sensor equipment and analytical workstations.
OPTICAL PRESSURE SENSING

What is Arc Flash Detection?

Arc flash detection is a protection method using optical sensors or pressure wave detectors to identify the intense light and pressure of an internal arc fault, triggering an ultra-fast trip to minimize equipment damage.

Arc flash detection is a high-speed protection technique that identifies the radiant light and pressure wave generated by an internal arcing fault within metal-clad switchgear. Unlike conventional overcurrent relays that wait for a current threshold, optical sensors detect the flash within milliseconds, sending a signal to trip the upstream circuit breaker before the arc can cause catastrophic equipment destruction or personnel injury.

The system typically combines optical point sensors or fiber-optic loops with a current-supervision element to ensure security against false trips from external light sources. When both the light intensity and an overcurrent condition are confirmed, the arc flash relay issues an instantaneous trip command, often clearing the fault in less than 2 milliseconds—dramatically faster than traditional time-graded protection schemes.

ULTRA-FAST PROTECTION

Key Features of Arc Flash Detection Systems

Arc flash detection systems combine optical sensing with current monitoring to identify dangerous internal arc faults and trigger circuit breaker trips in milliseconds, dramatically reducing incident energy and equipment damage.

01

Optical Point Sensors

Point sensors are discrete photodetectors installed in individual switchgear compartments to capture the intense light flash emitted during an arc fault.

  • Detection mechanism: Phototransistors or photodiodes tuned to the visible and ultraviolet spectrum characteristic of arcing
  • Coverage: One sensor per compartment, typically mounted on the interior wall or ceiling
  • Threshold: Trip when light intensity exceeds a preset lux level, often configurable between 10,000 and 50,000 lux
  • Supervision: Continuous self-monitoring of sensor health and lens contamination to prevent nuisance trips or blinding

Point sensors provide deterministic zone selectivity by physically limiting detection to a single compartment, ensuring only the affected breaker trips.

< 1 ms
Light Detection Time
02

Bare-Fiber Loop Sensors

Bare-fiber loop sensors use a continuous length of unjacketed optical fiber routed through multiple compartments to detect arc flash light along its entire path.

  • Principle: Light entering the exposed fiber core at any point is guided to a photodetector at the relay end
  • Advantage: A single fiber can monitor an entire switchgear lineup, reducing per-compartment sensor cost
  • Sensitivity: Detects light levels as low as 5,000 lux, with the fiber acting as a distributed collector
  • Installation: Routed through cable glands and secured with clips, avoiding sharp bends below the minimum bend radius

This approach is particularly effective for retrofitting existing switchgear where drilling individual sensor ports is impractical.

60+ m
Single Fiber Coverage
03

Current-Supervised Logic

Current supervision prevents false trips by requiring a simultaneous overcurrent condition before the relay issues a trip command.

  • Logic: Trip = Light detected AND phase current exceeds a settable threshold (typically 1.2x rated current)
  • Purpose: Discriminates between a genuine arc fault and ambient light sources such as camera flashes, sunlight through inspection windows, or maintenance lighting
  • Implementation: The relay monitors current transformer inputs in parallel with optical sensor inputs, gating the trip output
  • Speed trade-off: Adds approximately 1-2 ms to total operating time for current threshold comparison

Without current supervision, a single exposed sensor could trigger an unnecessary outage across an entire bus section.

~1.2x In
Typical Current Threshold
04

Pressure Wave Detection

Pressure wave detectors identify arc faults by sensing the rapid pressure front that propagates through air at the speed of sound following an internal arc ignition.

  • Transducer type: Piezoelectric or MEMS-based dynamic pressure sensors with millisecond response
  • Detection signature: A sharp pressure rise rate exceeding 100 kPa/s, distinct from gradual pressure changes caused by temperature or ventilation
  • Application: Medium-voltage metal-clad switchgear where optical paths may be obstructed by internal structures
  • Complementary role: Often deployed alongside optical sensors for redundant protection in critical busbar zones

Pressure detection is immune to light contamination but responds slightly slower than optical methods due to acoustic propagation delay.

~10 ms
Pressure Propagation per 3m
05

Ultra-Fast Trip Output

The defining performance metric of arc flash detection is total clearing time—the interval from arc ignition to circuit breaker contact separation.

  • Detection latency: Optical sensors detect light in under 1 ms; relay processing adds 1-2 ms
  • Output mechanism: High-speed solid-state trip contacts or IEC 61850 GOOSE messaging over a substation LAN, bypassing slower electromechanical interposing relays
  • Breaker response: Modern vacuum or SF6 circuit breakers open within 30-50 ms after receiving a trip signal
  • Result: Total clearing times of 35-60 ms, compared to 300-500 ms for conventional overcurrent-based arc protection

Reducing clearing time from 500 ms to 50 ms can lower incident energy by a factor of 10, dramatically reducing the required personal protective equipment rating.

< 2 ms
Relay Operating Time
06

Zone-Selective Interlocking

Zone-selective interlocking coordinates multiple arc flash relays to ensure only the circuit breaker closest to the fault trips, preserving service to healthy bus sections.

  • Zones: Each relay monitors a defined protection zone—incoming main, bus coupler, or individual feeder compartments
  • Blocking signal: When a downstream relay detects a fault, it sends a block signal to the upstream relay, preventing a cascading trip
  • Communication: Hardwired binary I/O or GOOSE messaging between relays within the same substation
  • Fail-safe: If the downstream breaker fails to clear the fault within a settable time, the upstream relay trips as backup

This scheme maintains selectivity without sacrificing speed, a critical requirement for process industries where unnecessary outages carry high financial penalties.

4-6
Typical Zones per Relay
ARC FLASH DETECTION

Frequently Asked Questions

Explore the critical engineering principles behind ultra-fast optical arc flash detection systems used to protect personnel and equipment from the devastating thermal and pressure effects of internal arcing faults in medium-voltage switchgear.

Arc flash detection is a protection method that uses optical sensors or pressure wave detectors to identify the intense light and pressure of an internal arc fault, triggering an ultra-fast trip to minimize equipment damage. Unlike traditional overcurrent protection, which waits for current to exceed a threshold over time, an arc flash relay detects the blinding light (typically 10,000 to 40,000 lux) emitted during an arc event using point sensors or fiber-optic loops installed inside switchgear compartments. When the optical signal coincides with a current threshold monitored by a current transformer, the system issues a trip command to the upstream circuit breaker in as little as 2-5 milliseconds, dramatically reducing the incident energy released. This dual-criterion approach—light plus current—prevents nuisance trips from ambient light sources like camera flashes or sunlight while ensuring genuine faults are cleared before the pressure wave can rupture the enclosure.

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.