Key Considerations for Selecting an Arc-Flash Relay
Arc-Flash Relays are an effective defense against dangerous Arc-Flash events, and the decision to include such a relay in a design is an easy one. Less easy, however, is selecting the optimal relay for an application.
According to OSHA, industrial arc-flash events cause about 80% of electrically related accidents and fatalities among qualified electrical workers. Even if personnel injuries are avoided, arc flash can destroy equipment, resulting in costly replacement and downtime. In response, many designers are adding arc-flash relays to electrical systems. These devices greatly mitigate the effects of an arc flash by detecting a developing incident and sending a trip signal to a breaker to disconnect the current that feeds it. Arc- flash relays are complex devices; an understanding of the technical details of their operation and features is essential. This white paper covers key points of arc-flash relay technology so that specifying engineers, OEM designers, and end users can make an informed selection decision.
NFPA 70E goes into great detail on procedures to avoid electrical shock and arc-flash events by opening and locking out circuit breakers before working on electrical equipment. When work on a live system is required, this standard spells out approach distances, use of personal protection equipment and apparel, and other precautions.
Arc-flash relays (Figure 1) are an important part of an arc-flash mitigation strategy and are often installed in electrical cabinets. These compact devices are designed to detect a developing arc flash extremely quickly and send a trip signal to a circuit breaker, which significantly reduces the total clearing time and subsequent damage. This is accomplished by providing an output that directly activates an electrical system circuit breaker to cut off current flow to the arcing fault.
Arc-flash relays can help companies comply with the NEC, which in some cases requires workers to adjust the circuit protection device to zero-delay mode when working inside an arc flash boundary. The code states that workers don’t need to take this extra step if an arc-flash relay is protecting the cabinet.
The fastest arc-flash relays available on the market today will detect a developing arc flash and send a trip signal to a breaker in less than 1 ms. The breaker will typically take an additional 35 –50 ms to open, depending on the type of breaker and how well it is maintained. Because an arc flash can draw a fraction of bolted-fault current, especially in the early stages, circuit breakers alone cannot be relied upon to distinguish between the arcing current and a typical inrush current. That’s why installing an arc-flash relay to detect those developing incidents rapidly reduces the total clearing time and the amount of energy released through an arcing fault greatly. In turn, there is less damage to equipment, as well as fewer and less severe injuries to nearby personnel. Generally, this minor damage is limited to the fault point where the arc originates, and avoids the more widespread and severe damage that occurs in a full-blown arc flash.
Using a system circuit breaker to clear the fault also avoids the need to purchase and install a separate switching device. This allows a more compact arc-flash relay design that can fit in an existing switchgear cabinet.
Key Selection Criteria
It is important to evaluate all the design aspects of the arc-flash relays being considered for this important function. The most important selection criteria are:
1. Reaction time
2. Trip reliability
3. Ease of installation
5. Sensor design
6. Avoidance of nuisance tripping
7. Scalability and flexibility
1. Reaction Time
When evaluating an arc-flash relay’s reaction time, keep in mind the timing of events that typically occur during an arcing fault. In the early stages of an arc flash, the arcing current is too low to trip a circuit breaker, which is sized to tolerate temporary rises in current caused by transients, such as motor inrush current. For example, on a 50 kA bolted fault between 480 Vac and ground, as the current increases, cable insulation will catch fire in about 50 ms; within 100 ms, the copper conductor begins to vaporize.
In addition to the intense light, the arc flash generates extreme heat and an explosive high-pressure wave.
Because light is the earliest detectable indication that an arc-flash is occurring, arc-flash relays use optical light sensors to detect the arc that is forming. The output of the light sensor is hard-wired to the arc-flash relay, which trips a circuit that interrupts the energy supply in the arc. The speed with which this occurs depends on the arc-flash relay design. If you examine the specifications for several arc-flash relays on the market, you will find the fastest possible trip times range from less than one millisecond to about nine milliseconds.
These reaction times are principally a function of the arc-flash relay’s light sensor input sampling scheme and the design of its trip output circuit. An important aspect of the sensor sampling scheme is avoidance of nuisance tripping (more on light sensing and nuisance trips later). For example, consider the sampling design of the Littelfuse PGR-8800 Arc-Flash Relay. During normal operation of these devices, light sensor inputs (up to six allowed per relay, up to 24 per system) are sampled every 125 microseconds (i.e., a sample frequency of 8 kHz). Thus, the device will make a positive arc-flash detection if light intensity higher than the trip level is captured at a sensor when the unit samples its related input. The unit counts the number of consecutive samples above the trip level and activates the output when a sufficient number of samples has been counted.
As part of a sensor sampling scheme, a programmable time elay may be used to filter the inputs. This will establish the number of samples needed to trip the relay and thereby filter out, for example, photo flashes that could cause unintentional trips. Typically, a programmable time delay filter can be set between zero (instantaneous light detection) and one or two seconds for special application conditions.
After an arc-flash relay’s input time delay, it takes a certain amount of time for its electronic output to turn on. This time is a function of the type of output relay used. Solid-state outputs (for example, insulated gate bipolar transistors (IGBTs)) are much faster than electromechanical relays and can operate within 200 microseconds.
In the case of the Littelfuse PGR-8800 and AF0500 Arc-Flash Relays, their default delay of 500 microseconds includes three consecutive samples above the trip threshold and the 200 microsecond turn-on time of their IGBT output. The IGBT turn-on time and the sample interval of 125 microseconds correspond to a minimum trip time (with no time delay filtering) of less than 0.5 ms. The overall sampling time is proportional to the number of sensors used (up to a maximum of six in this case), so the reaction time specification of that arc-flash relay is listed as <1 ms. For the fastest response time, the programmable delay should be set to the minimum value consistent with avoiding nuisance tripping.
2. Trip Reliability
Next to reaction time, reliable tripping is the most important characteristic of an arc-flash relay because this ensures mitigation of an arcing fault. Two aspects of reliability should be considered: trip redundancy and system-health monitoring.
Redundant Tripping. Few arc-flash relays offer a redundant tripping feature, which is analogous to using both seat belts and air bags in an automobile. It has both primary and secondary trip path logic. The primary path is controlled by the internal microprocessor and its embedded software, and works by activating the coil of the primary trip relay.
The redundant path typically uses a discrete solid-state device that does not go through the microprocessor. Any failure in the primary (microprocessor) path will cause the unit to switch automatically to its redundant path, which activates a shunt-trip relay without delay when a sensor input is above the light detection threshold. A solid-state redundant path is not influenced by programmable settings such as time delay. Some relays provide an option for under-voltage or shunt trips; the redundant path usually operates in shunt mode, so an under-voltage coil will trip immediately if the microprocessor fails.
One often-overlooked advantage of a solid-state trip path over a microprocessor-based circuit is the reaction time when the relay is first powered up. It is always best if the relay can remain powered at all times, which is why some relays include options for charging and operating from a battery backup, but consider the situation where the plant (including the arc-flash relay) is shut down for maintenance. Wiring mistakes, tools left in hazardous locations, and the regular stresses of powering up all contribute to the risk of an arc flash on power up. A microprocessor can require 200 ms or longer before it is able to start scanning the optical sensors. However, a solid-state trip path can detect an arc and send a trip signal in as little as 2 ms.
The significance of such redundant types of design is higher reliability because there is no single point of failure that will incapacitate the system. In addition, there are fail-safe features that alert operators when, for example, the microprocessor fails.
Health monitoring. It is often said, “A chain is only as good as its weakest link.” Health monitoring ensures the system is in good operating condition and it should extend from the light sensors to the output of the arc-flash relay trip circuitry.
Most arc-flash relays have some degree of internal health monitoring, but designs can vary considerably. In the case of the PGR-8800 and AF0500 relays, their built-in health monitoring starts right on the sensors. A signal is sent from the relay to the light sensors, where a test light is detected by the sensor and sent back to the relay. In the case of a fiber-optic sensor, this also verifies the entire length of the fiber is not pinched or broken.
Although many arc-flash relays indicate sensor status on the face of the relay, keep in mind that the relay will not be visible to a worker in a different compartment. Most arc-flash relays do not provide on-sensor indication of the scanning and health status of the system. The PGR-8800 and AF0500 relays are exceptions. Their point sensors have an LED indicator that shows if light detection is functioning properly. On-sensor health indication can be critical in preventing maintenance work on equipment where protection is inactive because the worker can immediately see that the sensor LED is off. It also has the added benefit of providing rapid fault location.
Following the path of a trip signal from the sensor, internal monitoring must also include the primary and, where applicable, the redundant trip circuit. The next link in the chain is the trip output, in this case, an IGBT switch, which is connected to the breaker trip coil. Low voltage across the IGBT indicates a wiring fault or an error in the trip coil, and a high voltage is a sign of an error in the IGBT switch, both of which are also reported and logged. The IGBT is also thermally protected against overloads, and will turn off if it overheats. However, the thermal protection has a 100 ms delay before acting, meaning that even a dangerously overheated coil will attempt to signal a trip before resuming thermal protection. In this way, complete sensor-to-breaker system-health monitoring is achieved.
Another reliability feature on some arc-flash relays is the ability to power the unit with either AC or DC sources. This allows backup battery operation in the event of failure on an independent AC system that powers the unit.
3. Ease of Installation
What makes an arc-flash relay easy to install? First, look for a relay that does not require PC configuration. The Littelfuse AF0500 Arc-Flash Relay has a plug-and-play, simple, and flexible design. It does not require PC configuration, which simplifies installation and gives users immediate confidence the relay is set up correctly. (Although PC configuration for both the AF0500 and PGR-8800 is optional, it will be needed to configure certain advanced behaviors or multiple power sources.)
Second, look for a relay whose wiring ports are clearly marked, allowing contract electricians to see quickly how to install the device. By the way, the point sensors used with the PGR-8800, AF0500 and AF0100 have a blinking light to indicate the health of the sensor, so workers can see immediately if installation was successful.
Third, look for a relay whose inputs accept both point sensors and fiber-optic sensors. The AF0500, PGR-8800 and AF0100 arc-flash relays allow for a flexible mix of point sensors and fiber optic sensors using the same inputs. The installer can modify the sensor configuration and adapt at installation if necessary, and easily repurpose the relay for a different cabinet in the future. Some relays require users to purchase a separate relay or to specify sensor types in advance. This complicates installation and increases cost.
If an arc flash relay is easy to install, then it becomes easier to justify its use in smaller electrical panels. Beyond feeder switchgear cabinets, arc-flash protection can be extended to transformers, power converters, and motor control centers.
A few arc-flash relays, including the PGR-8800 and the AF05000, have software that provides event logging, which is useful for tracking trends in the system’s performance. To make troubleshooting easier, this software should record the specific sensor that initiated the fault in the data records. Both the PGR-8800 and AF0500 relays can log up to 1000 events, and the PGR-8800 supports data analysis tools such as graphs, etc.
Various data communication interfaces are included on arc-flash relays, which can be used to configure the units. For example, some arc-flash relays have a USB interface that makes setup easy; even complex scenarios with multiple modules, current sensors, and customized trip levels take only minutes to configure. (See Figure 3) Events can be stored in a log file, which can also be accessed remotely via the USB interface to generate reports and graphs of historical data.
Power system analysis software companies are including arc-flash relays in their component libraries. This allows users to perform “what-if” analysis for a variety of relay configurations, circuit variables, and fault conditions.
5. Sensor Design
Most arc-flash relay installations utilize multiple fixed-point light sensors near vertical and horizontal bus bars where arcing faults are apt to occur in feeder switchgear cabinets. Sufficient numbers of sensors should be installed to cover all accessible areas, even if policy is to work only on de-energized systems. At least one sensor should have visibility to an arc fault if a person blocks another sensor’s field of view. Light sensors may also be installed in other electrical cabinets and on panels that are subject to routine maintenance and repairs, such as those associated with motor control centers (see Figure 4).
In addition to fixed-point sensor inputs, some arc-flash relay manufacturers supply sensors with fiber-optic strands, which have a 360° field of view for detecting light. This allows more flexible positioning of the light sensing locations because the fiber-optic strands can be looped throughout an enclosure or panel to cover challenging component layouts. Fiber strands ranging from 26 to 65 feet are supplied by various arc-flash relay manufacturers; some also offer interconnection hardware for even longer lengths.
However, long “open” fiber strands designed for light reception over their entire length should be used with caution. Depending on the location of an arc flash relative to the far end of such a strand, it’s possible that light arriving at the detector end may not be bright enough to cause a relay trip due to attenuation along the length of the fiber.
Some manufacturers avoid this problem by using interconnecting hardware with hard-wired outputs from the interconnection and detector points back to the arc-flash relay. For example, the Littelfuse PGR-8800 allows up to six light sensor inputs per relay, and up to four relays can be interconnected to act as a single unit. This means up to 24 light sensor inputs can be used to monitor an electrical equipment configuration.
6. Avoidance of Nuisance Tripping
Typically, arc-flash relays use light sensors with fixed detection thresholds set somewhere in the range of 8,000 to 10,000 lux light intensity. In most installations, this will avoid nuisance tripping because an arcing fault produces light intensity exceeding the ambient light level inside a closed switchgear cabinet (See Table 1).
Typical Light Levels
Typical Light Levels
Illuminance (lux) - Source
320 - 500 - Typical office lighting
1,000 - Overcast day or typical TV studio lighting
10,000 - 25,000 - Full indirect daylight
32,000 - 130,000 - Direct sunlight
1,000 - Overcast day or typical TV studio lighting
10,000 - 25,000 - Full indirect daylight
32,000 - 130,000 - Direct sunlight
Nevertheless, it can be useful to raise the setpoint to minimize nuisance tripping. For example, opening a switchgear cabinet to bright ambient light, a camera flash, or a worker welding nearby could set off the arc-flash relay inadvertently, causing costly downtime.
Figure 5. Phase current transformers (CTs) can be used to help avoid nuisance trips; an Arc-Flash relay trip output requires both a light sensor input and rapidly rising phase current.
Some arc-flash relays also use a phase-current sensing scheme to help avoid nuisance tripping. One way of handling this is to use current transformer inputs to the arc-flash relay’s microprocessor (see Figure 5); three CTs are used to measure each of the three phase currents in the system. If the microprocessor logic receives an input from a light sensor, it checks for a rapidly rising input from the current transformers. If present, it then sends an output signal to the disconnect device.
However, using RMS current sensing is not an optimal way to implement this feature. A better approach, used by the PGR-8800 relay, is to use instantaneous current values to detect rapidly rising current that is indicative of an arcing fault. This avoids unnecessary delays in tripping the relay when an arc flash occurs. With this method, the microprocessor looks for the numerically largest of the three phase currents at any given sample, and compares that to the limits programmed by the user. In addition, delays should be available in the configuration for over-current trips to describe how long the maximum value must stay above the limit for the unit to trip the output.
To ensure no negative impact to the protection provided by the arc-flash relay, the total trip reaction time should be minimally affected by use of CT-current verification. In the case of the PGR-8800 relay, for example, the total trip time is still less than 1 millisecond, subject to the user-selectable time delay setting.
7. Scalability and flexibility
Several arc-flash relay designs allow the interconnection of multiple devices, such as multiple relays with several sensors each. A unique feature of using such a network is the ability of a downstream arc-flash event to trip the upstream circuit breaker. This can be very useful where the upstream substation is feeding downstream motor control centers (MCCs). Working on live equipment is very common on MCCs, raising the odds of an arc-flash event there.
However, the MCC may be a main-lug-only type or the main circuit breaker in the MCC may not allow it to be tripped remotely. That means there is no method to remove power if an arc-flash event is detected. An arc-flash relay installed in the MCC can detect the arc flash and tell the upstream relay or a networked relay to trip the main circuit breaker feeding the MCC.
Upstream breaker trip functionality is built into the AF0500 relay. If the local circuit breaker fails to open, then another command is sent after a short delay to the upstream breaker to maintain arc flash protection. This is easily accomplished because the AF0500 relay has two IGBT outputs. The same functionality is available with the PGR-8800 relay, however it has only one IGBT output, so upstream breaker tripping is accomplish through the relay’s logical TRIP output instead.
Relays (Figure 6) can be linked in a manner similar to zone- selection interlocking protection devices to provide back-up protection. For example, an arc-flash event in the MCC will cause the arc-flash relay to send a trip command to the main circuit breaker in the MCC. If it does not trip, the arc-flash relay will send a signal to the linked arc-flash relay upstream to trip the upstream feeder circuit breaker.
Some relays, such as the PGR-8800 relay, accomplish zone tripping is by networking multiple relays. The AF0500 relay, in contrast, has built-in zone trip capability. One AF0500 relay can monitor two separate zones, with two sensors assigned to one zone and two sensors assigned to another zone. The relay can trip different breakers for each zone, or trip the same breaker, or trip a local breaker and upstream breaker for a single zone.
Zone tipping may be used to minimize an outage. For example, an arc-flash event may trip only the affected portion of a cabinet, rather than shutting down the whole cabinet. Or zone tripping may be used to open a circuit breaker on the incoming feeder and simultaneously open a tie breaker so that power is not routed from a second feeder.
Zone tripping. The AF0500 can trip two separate zones.
Upstream breaker tripping. If the local circuit breaker fails to open, another trip command is sent after a short delay to an upstream breaker to clear the fault.
Tie-breaker tripping. In case of an arc in one section of the switchboard, theAF0500 can trip both the incoming feeder and the tie breaker simultaneously.
Arc-flash relays are an important adjunct to mitigation devices built into switchgear, such as ducts to channel the explosive forces of an arcing fault away from personnel. Relative to the equipment and personnel they protect, these relays are a very cost-effective insurance policy, capable of decreasing the risk of arc flash significantly. This in turn can prevent costly downtime, equipment replacement, and lawsuit expense. Arc-flash relays effectively and efficiently minimize these problems and provide additional protection where traditional OCP devices fall short. Still, when selecting an arc-flash relay, be sure to examine the functions of the device carefully, and ask the manufacturer to explain how and what it can do in your specific application.