Mitigating arc flash hazards in medium-voltage switchgear

Engineers should be aware of design alternatives that can reduce arc flash hazards in medium-voltage systems.

By Syed M. Peeran, PhD, PE, CDM Smith, Cambridge, Mass. December 10, 2013

The term “arc,” which literally means part of a circle, is attributed to Humphrey Davis, an English scientist. In 1802, Davis demonstrated that electric current can flow between two carbon rods separated in air by a short distance in the form of a band of ionized air that looks like an upward bow. In fact, electrical science started with the study of the electric arc. Soon, a number of inventions came forth, such as arc lamps, arc furnaces, spark plugs, arc welders, and others. Today, the electric arc is again a subject of great interest and study because of the hazards it creates in electrical distribution systems due to its intense heat, which can destroy equipment and cause severe or fatal injuries to unprotected personnel who are unfortunate to be in close proximity to it. 

In all electrical equipment, a serious hazard exists for operating personnel due to possible arcing between energized parts and between energized parts and grounded metal enclosures. Hazardous arcing can take place in electrical equipment because of one or more of the following:

  • Accidentally dropping metal tools in energized parts
  • Incorrect alignment of contacts in draw-out circuit breakers
  • Loose connections can cause overheating and minor arcing, which can escalate to an arcing fault
  • Rodents and vermin in switchgear enclosures
  • Defective cable and bus insulation. 

The arc behaves like a flexible conductor and consists of ionized air at very high temperature, in the order of 35,000 F—more than three times hotter than the surface of the sun. It can burn holes in copper bus bars. It can vaporize copper, which when condensed on other parts can cause secondary faults. It can cause pressure buildup and/or an explosion in enclosed equipment. It can cause severe burns and can ignite clothing. 

OSHA and the National Fire Protection Association (NFPA) have adopted specific requirements with regard to the arc flash hazard. OSHA requires that all equipment be marked with a label that indicates the arc flash boundary, the incident energy in the arc, the safe working distance, and the category of clothing and other protective equipment to be used by personnel. Article 110.16, which states that the equipment be clearly and visibly labeled to warn personnel of the potential arc-flash hazard, was introduced in NFPA 70: National Electrical Code in 2002. In 2004, NFPA 70E: Standard for Electrical Safety in the Workplace required that shock and arc-flash hazard analyses be completed to determine the level of personal protective equipment required in each location. 

Incident energy, working distance, and hazard risk category

Incident energy is the measure of the severity of the hazard to workers. This quantity is defined as the energy density in calories/cm2 or Joules/cm2 to which the worker’s face or body is exposed in an arc flash event at the working distance. The working distance is the typical distance between a potential source of the arc in the equipment and the face or body of the person performing the work on the equipment. The value of the incident energy determines the type of mandatory protective clothing to be worn by the worker. Typical working distances defined by IEEE Std. 1584 include:

  • 15kV switchgear: 36 in.
  • 5kV switchgear: 36 in.
  • Low-voltage switchgear: 24 in.
  • Low-voltage motor control centers and panel boards: 18 in.
  • Cables: 18 in. 

Arc flash hazard is quantified by a number called the hazard risk category (HRC). According to NFPA 70E, the relationship between the HRC, the available incident energy, and the type of protective equipment is listed in Table 1. 

Arc flash equations, solution

In 1982, Ralph H. Lee published a paper in the “IEEE Transactions on Industry Applications” on calculation of the incident energy in open air arcs, such as in outdoor substations. This paper triggered renewed interest in the arc flash phenomenon. In 2002, the IEEE Industry Applications Society published IEEE Standard 1584: IEEE Guide for Performing Arc Flash Hazard Calculations and issued subsequent amendments in 2004 and 2011 as 1584a and 1584b. The equations in this standard are empirically derived using statistical analyses and curve-fitting algorithms on a huge collection of experimental data (see “Calculating arcing faults”). The equations can be used for systems from 208 V to 15 kV, 50 to 60 Hz, available short-circuit current from 700 A to 106,000 A, and for arcing distances from 0.5 in. to 6.0 in. 

For any electrical equipment, there are two significant parameters that determine the incident energy and, therefore, the type of protective clothing to be used. These parameters are the arcing fault current “Ia” and the duration of the arc “t”. The arcing fault current Ia is less than the bolted fault current (Ibf) because of the voltage drop across the arc or because of the arc resistance. For a given arc length, the arc voltage drop is almost constant for a wide range of current. Consequently, the arc exhibits negative incremental resistance. The term “bolted” signifies a fault through zero resistance, as when the 3-phase wires are stripped, lugged, and bolted together. 

Figure 1 simplifies the relationship between the arcing fault current and the arc voltage drop. The drawing shows why the arcing fault current Ia is considerably less than the bolted fault current Ibf in low-voltage equipment, while it is about 90% of Ibf in medium-voltage and high-voltage equipment. This is because the arc voltage drop, which is approximately 200 V for a 2-in. arc, is a significant part of the circuit voltage in 480 V equipment, while it is less than 10% of the circuit voltage in 4.16 kV and 13.8kV equipment. 

The arc voltage drop depends on several factors including the clearances in different classes of equipment. The relation between Ia and Ibf and the relation between the incident energy E and Ia and t are given in Section 5 of IEEE 1584. These equations are programmed into the arc flash evaluation programs of most distribution system analysis software. These programs require that a short-circuit study be performed first to determine Ibf at the equipment in question. 

Duration of the arc

The duration of the arcing fault has a direct impact on the available incident energy. Arcing faults, like all other faults, must be detected and cleared by the first upstream circuit protective device. Therefore, the total arcing time is the total clearing time of the device, which, in the case of circuit breakers, equals the sum of the relay or sensor time and the breaker operating time. Relay or sensor time depends on the setting of the relay and the fault current. Typical circuit breaker operating times are listed in Table 2.

Mitigating hazards in medium-voltage equipment

There are many reasons why mitigation of arc flash hazards is of greater concern in medium-voltage equipment. First, medium-voltage switchgear occupies a higher hierarchical position in most radial distribution systems. Consequently, medium-voltage protective devices must be set to operate with a greater time delay to allow the low-voltage downstream devices to operate first in the event of a fault. Second, medium-voltage circuit breakers take more time to clear a fault than do low-voltage circuit breakers. In addition, the arcing fault current is very nearly equal to the bolted fault current. The increased arcing time and the higher arcing fault current contribute to greater incident energy and HRC. Because of the higher hierarchical position, de-energizing the medium-voltage switchgear for maintenance work is often not an option because it would shut down a significant portion of a facility. Therefore, one must look seriously at various methods of reducing the HRC. 

Design alternatives that can reduce arc flash hazards in medium-voltage systems include:

  • Use of smaller and higher impedance transformers
  • Bus differential and transformer differential protection
  • Current limiting fuses
  • Maintenance switch
  • Arc flash detecting relays
  • Arc-resistant switchgear
  • Crowbar methods
  • Remote operator panels.

The engineer must evaluate each option and select one or more most appropriate for a given system.

Smaller and higher impedance transformers: Most distribution systems are radial. Instead of specifying one large-capacity, medium-voltage transformer to feed the plant, two or more small-capacity, higher-impedance transformers can be used to supply individual areas of the plant. The idea is to reduce the available bolted fault current and the arcing fault current. Reducing the arcing fault current does not necessarily increase the fault clearing time. Relays can be set to minimize the fault clearing time. For example, a 3,000 kVA, 13.8 kV/4.16 kV transformer with typically 6% reactance would be a source of 6,940 A of short-circuit current at the 4.16 kV switchgear, while a 1,500 kVA transformer with 8% reactance can supply only 2,603 A of short-circuit current. The incident energy in the event of an arcing fault would be reduced by 62%. However, the capital cost and the space requirements for two 1,500 kVA transformers would be more than those for the 3,000 kVA transformer. In addition, higher transformer impedance would cause a greater steady-state voltage drop and a greater transient voltage dip during motor starting. These drawbacks must be evaluated and weighed against the advantage of reduced arc flash incident energy. 

Bus differential, transformer differential protection: Differential protection is a means of clearing the fault inside the zone of protection without intentional delay and without interfering with the overcurrent protective device coordination. The zone of protection is defined by the location of the current transformers (see Figure 2). Another common instance where differential protection would considerably reduce the arc flash hazard is shown in Figure 3A. Transformer primary protection is provided by a fuse. The fuse is chosen to provide adequate protection to the transformer and to permit the magnetizing inrush current. A fault at the line side of the secondary main breaker must be cleared by the primary fuse only. Often the HRC for the line side fault in this situation is excessive. If the fuse is replaced by a circuit breaker and differential protection is provided, the line side fault would be cleared without delay and the HRC can be brought down considerably (see Figure 3B). 

Current limiting fuses: Current limiting fuses have the capability to clear faults within a half cycle (less than 0.0083 sec) in addition to limiting the let-through current. Current limiting action of the fuse results from the melting of the silver filaments inside a sand filling inside the fuse, thus creating multiple arcs inside. Great reduction in the available incident energy is possible because of the fast clearing of the fault. However, this is possible only when the fault current lies in the current-limiting range of the fuse characteristic. For example, in a 15 kV 300 A current limiting fuse, the current limiting action takes place for fault current in excess of 6,000 A. The benefit of current limiting fuses can be realized only if the available short-circuit current is in excess of 6,000 A. One must also recognize that it is difficult to coordinate current limiting fuses with downstream protective devices. 

Maintenance mode on solid state protective relays: A maintenance switch is now available in most medium-voltage circuit breakers as a means of temporarily adjusting the settings of the solid state protective device during scheduled maintenance such that arcing faults are cleared without delay, while still maintaining the desired settings for coordination with downstream protective devices. Figure 4 shows the application and the benefit of a maintenance switch in 4.16 kV switchgear. Figure 4 A shows the single-line diagram of the switchgear. Figure 4 B shows the time-current curves of the main and the feeder breaker relays. The calculated arcing fault current is 8.44 kA for a bus fault. The fault is cleared by the main breaker in 1.303 sec (including the breaker time), the incident energy is 12 cal/cm2, and the HRC level is 3. 

When the maintenance switch is engaged, the main breaker relay’s instantaneous setting is reduced from 80 (16,000 A) to 30 (6,000 A), below the expected arcing fault current. The arcing fault will now be cleared in 0.015 sec, the incident energy is reduced to 1.2 cal/cm2, and the HRC level is reduced from 3 to 1. 

While using the maintenance switch, plant supervisors must enforce an error-proof method of ensuring that the maintenance switch is disengaged after the scheduled maintenance work is completed. Otherwise, there will be nuisance tripping of the main breaker. 

Arc flash protective relays: Light emitted by the arc can be used to detect an arcing fault instead of current sensing. This is the principle of operation of arc flash protective relays now being marketed by some companies in the U.S. The result is the same as that of the maintenance switch except that no human action is necessary. Arcing inside the switchgear enclosure is detected by either a photoelectric receptor or a length of fiber-optic cable. The input is given to a single-function or a multifunction electronic protective relay, which can trigger instantaneous tripping of the breaker. This method is independent of the magnitude of the arcing fault current and can detect arcing in the early stage of its development. One company claims that the detection takes place in 1.0 msec. These relays have not gained wide acceptance yet, but they surely present a better way of detecting arcing and immediate tripping than current sensing. 

Arc-resistant switchgear: In extreme cases, severe arcing in enclosed equipment can cause tremendous pressure buildup and may result in an explosion. The explosion will relieve the pressure buildup but will not quench or terminate the arc, which will proceed to cause thermal damage to the bus bars and enclosures until it is cleared by circuit breakers. This is the most probable scenario that resulted in several low-voltage and medium-voltage switchgear being completely gutted by internal arcing. Arc resistant switchgear is available that is structurally strong and has means of relieving the pressure buildup. The means consist of louvers and vents in the back of the enclosure, away from the operators, to exhaust the rapidly expanding air.

There are many environments in which the extra expense of the arc resistant switchgear is justifiable. In many industries, the additional cost is much lower than the cost of repair, downtime, compensations, and litigation. 

Crowbar methods: A radically different method of dealing with arcing faults is what is known as the “crowbar” method. This method is well known in Europe and is recognized as a viable method in medium-voltage switchgear by the International Electrotechnical Commission Standard 62271-200. Unfortunately, no U.S. standard has yet been developed as a guideline for the application of this method. Essentially, the crowbar method consists of high-speed detection of arcing, intentional creation of a 3-phase bolted fault, and clearing of the bolted fault by the circuit breaker. The bolted fault is created by a grounding switch. The circuit voltage is brought to zero and the arc collapses. In the U.S., this method is marketed as a high-speed grounding switch (see Figure 5). Arcing is detected by an optical sensor. An electronic relay energizes an actuator, which closes the 3-phase grounding switch, thus creating a 3-phase bolted fault. The bolted fault is sensed by the system protective relaying and the circuit breaker is tripped. Another company is marketing a scheme in which, instead of creating a bolted fault, a second arc is created inside a confined and mechanically strong drum-like enclosure. This second arc being parallel with the fault arc serves the same purpose as the bolted short-circuit. 

Remote operating panels: Safety of personnel from arc flash hazards can be ensured by providing remote operating panels from which all manual operation of the switchgear can be performed. The remote panels must be located at a safe distance from the switchgear or in a separate room. If space is available for the remote panels, the equipment itself is not expensive. All circuit breakers in the switchgear must be electrically operated. In addition, a motor-operated draw-out mechanism must be provided. All breaker control switches, auto/manual switches, indicator lamps, ammeter and voltmeter switches, meters, and an operator interface terminal can be installed in the remote operator panel (see Figure 6).

Calculating arcing faults

The following equations are used in calculating the arcing fault current: 

For system voltage under 1 kV: 

            lg(Ia) = K + 0.662 lg(Ibf) + 0.0966 V + 0.000526 G + 0.5588 V (lg Ibf) – 0.00304 G (lg Ibf

Where:

lg = the log10 (logarithm to the base 10)

Ia = arcing current, kA

K = -0.153 or open air arcs; -0.097 for arcs-in-a-box

Ibf = bolted three-phase available short-circuit current (symmetrical rms), kA

V = system voltage, kV

G = conductor gap, mm 

For system voltage greater than or equal to 1 kV:

            lg (Ia) = 0.00402 + 0.983 lg (Ibf

The incident energy E is calculated using the following equation: 

            E = 4.184 Cf En (t/0.2) ( 610x/Dx

Where:

E = incident energy, J/cm2

Cf = calculation factor

 = 1.0 for voltages above 1 kV

 = 1.5 for voltages at or below 1 kV

En = incident energy normalized

t = arcing time, sec

x = distance exponent

D = working distance, mm 

The normalized incident energy is given by the following equation: 

            lg En = k1 + k2 + 1.081 lg(Ia) + 0.0011 G 

In these equations, the values of G and the exponent x depend on the voltage and the type of equipment. For example, for 480-V switchgear, G = 32 mm and x = 1.473. For other voltages and other equipment, Table D.7.2 of IEEE Std. 1584 gives the values of G and x.

Source: IEEE Std. 1584-2002 IEEE Guide for Performing Arc-Flash Hazard Calculations


Syed M. Peeran is a senior electrical engineer at CDM Smith. He has more than 20 years of experience in the design of electrical distribution systems. For several years, he was an adjunct professor at Northeastern University, Boston, and is a member of the Consulting-Specifying Engineer editorial advisory board.