Electrical Safety: Arc Flash Hazard Analysis (part 1)

1. Introduction to Arc Flash Hazards

For many years, it had been assumed that electrical shock was the greatest hazard presented by electrical work. However, more recently, it has been recognized that electric arc flash and blast effects produce greater severity of injuries and more hospital admissions than electrical shock (Lee, 1982). The problem is more severe than that indicated by the data because fatalities are not reflected in hospital admissions data. An arc flash hazard is a dangerous condition associated with the release of energy caused by an electric arc. Generally, an electric arc is formed when two conductor materials (e.g., copper) separated by a gap are shorted together, whether from inadvertent contact or insulation failure, causing passage of substantial electric currents through the gap between the conductor materials. The arc that develops is not only composed of current flowing in air, but also in vaporized conductor material, usually copper or aluminum. The temperature of the arc can reach 35,000 K, which is several times hotter than the surface of the sun. The only higher temperatures that can be produced on earth are from extremely large lasers. The heat produced by the arc can cause severe, often fatal, burns over a large portion of the body and ignition of clothing and other materials. The blast caused by the rapid expansion of copper from the heat can cause hearing loss, equipment doors to blow off their hinges, and people to be blown many feet across a room. The shrapnel associated with arc blast, consisting of small copper particles, can also cause severe injuries and death.

Since these very high temperatures vaporize the conductor materials, extremely explosive pressures are created as a result of metal vaporization and rapid heating of the air in the gap by electric current passing through it. The vaporization of metal and heating of the surrounding air results in a very rapid blast due to the buildup of high pressures. The blast causes molten metal and equipment parts to violently spew from the point of the blast.

People near arc flash hazards may be subject to severe burns, ruptured eardrums, collapsed lungs, and forces that can violently knock them back.

There are several approaches to remediating arc flash hazards. One approach is to use engineering methods to reduce the amount of short-circuit current associated with the electric arc, as well as to reduce the duration of the arc.

Reducing the amount of energy associated with the electric arc and the duration of the arc results in fewer injuries. Another approach to remediating arc flash hazards is to require people working on or near exposed energized conductors to wear personal protective equipment (PPE) (e.g., heavy jackets, protective hoods, face shields, and gloves) designed specifically for arc flash hazards. A third approach to remediating arc flash hazards involves using procedural or administrative methods to warn people of the potential for the occurrence of arc flash hazards if one is going to work on or be near exposed energized conductors that could experience an electric arc. Procedural or administrative methods may include applying warning labels on or near exposed energized conductors. For example, a typical label could state that if one is going to work on equipment having exposed energized conductors, then workers have to wear authorized PPE and maintain a predetermined distance from the conductor materials that could be subject to an arc flash hazard.

The procedures and policies set forth in ANSI/ AIHA Z10-2012 Occupational Health and Safety Management Systems (AIHA, 2012) can be used as part of an electrical safety program to mitigate the hazards of arc flash and electrical shock. The actions described in the previous paragraph are part of the "Hierarchy of Control" set forth in Z10-2012:

"A. Elimination;

Substitution of less hazardous materials, processes, operations or equipment;

Engineering;

Warnings;

Administrative controls, and Personal protective equipment."

The above-described approaches to remediating arc flash hazards are not always effective for arc flashes caused by electric arcs which produce an incident energy that exceeds…. Some design techniques which reduce the possibility of such high levels of incident energy are discussed in this Section. Design for safety is critical because engineering methods that reduce the amount of energy associated with the electric arc and the duration of the arc may not be available for use with levels greater than…. Furthermore, currently available PPE may not provide protection against blasts that can arise from energy levels that are greater than . Similarly, the hazards associated with blasts that can arise from energy levels that are greater than essentially make the use of warning labels ineffectual for their intended purposes. Because the above-described approaches to remediating arc flash hazards are not suitable for arc flashes caused by electric arcs having energy that exceeds, workers need to shut down the equipment and work on it while de-energized.

However, shutting down equipment and working on it while de-energized is not always an optimal solution, especially in facilitates where it is desirable to keep equipment operating 24 h a day, 7 days a week.

For these reasons, arc flash hazard analysis has developed as a field within electrical engineering. There are three major aspects to this:

1. Safety by prevention. Systems may be designed to minimize arc flash hazards.

Protective device selection and adjustments are made to minimize exposure time and incident energy.

2. Operational safety. Design of work practices and procedures in order to reduce or eliminate work on energized electrical conductors. This includes methods such as arc flash labels, energized work permits, lock-out-tag-out, and application of safety grounds.

3. Safety by protection. The use of PPE such as clothing, face protection, eye protection, and hearing protection. These steps are used to protect the worker after the arc flash hazard has been minimized as much as possible by other methods.

Once arc flash hazard analysis became a requirement for many industrial power system studies, the reduction of arc flash hazards became an important concern. Inevitably, the factors that lead to reductions in arc flash hazards do not always lead to an improvement in other areas, in particular, protective device coordination.

These examples cover several particular cases where an arc flash hazard analysis of an older industrial plant yielded several cases of buses where the incident energy was above , making an "Extreme Danger" label necessary.

By changing protective device settings and, if necessary, the devices themselves, the incident energy can be reduced. Methods to maintain coordination, if possible, are discussed.

An arc flash hazard analysis usually begins with a study of a facility based on the procedures and methodology of the National Fire Protection Association (NFPA) Standard 70E (NFPA, 2012), and IEEE 1584-2002 (IEEE, 2002b).

However, an arc flash hazard analysis is normally performed in conjunction with a suite of other power systems studies, such as

1. Load flow

2. Short circuit

3. Protective device coordination

The results of an arc flash hazard analysis are labels to be placed on equipment, which give, among other things,

1. The flash protection boundary, which is the distance within which a person must wear PPE. This is the distance at which exposed skin will be at risk of a just-curable second-degree burn. It is not the distance at which there is no risk of injury.

This boundary must be reasonable, such that people who are not working on the equipment can still perform their functions.

For example, it should not extend beyond the fence of an outdoor substation or prohibit opening the door of an indoor substation.

2. The working distance for which the incident energy is calculated and the PPE is specified.

Persons or body parts closer than this distance will require additional protection.

3. The incident energy, in , to a person at the working distance specified. If the incident energy is above , no approach is possible without de-energization of the equipment.

When no arc flash study is performed, hazard/ risk categories are determined with the help of Table 130.7[C] (16) in NFPA 70E-2012 (NFPA, 2012). This table should be used with caution, because it was prepared using many assumptions, which are detailed in the footnotes. The effect of upstream current limiting fuses (CLF) is not taken into account.

The use of CLFs with a clearing time of ½ cycle (8.3 ms in a 60 Hz system, 10 ms in a 50 Hz system) or less can significantly reduce the incident energy. Specific fault currents, trip times, working distances, and arc gaps were used for the different voltage ranges covered.

Arc flash hazard levels depend on fault current magnitude and duration, . The power going into the arc resistance is a function of , where is the arc resistance. This is multiplied by time, which is the arc duration. Since arcs, currents, and resistances are in constant variation, this is really a time-varying integral whose units are those of energy (joules or calories). When this energy is radiated to a person's skin, they will be exposed to incident energy, which is what cases the burn.

Mitigation techniques for reducing incident energy include (Hodder et al., 2006; Brown and Shapiro, 2009). The factors that can be controlled include the following:

1. Selection of suitable time-current characteristics for protective devices.

2. Modification of circuit breaker and relay settings.

3. Pickup. This is the minimum current at which a device actuates. A lower pickup provides arc fault protection for a greater range of fault currents.

4. Time delay. A shorter time delay reduces the time to trip and lowers .

5. Instantaneous pickup. The operating time is typically the minimum possible for the circuit breaker being used. Lower instantaneous pickup settings reduce arc flash hazard.

6. Zone-selective interlocking.

7. Current-limiting low-voltage circuit breakers.

8. Fused circuit breakers.

9. Arc blast containment systems.

10. Optical sensors for arcing fault detection.

Protective devices in a power system are coordinated in time and in current pickup in order to provide for an orderly shutdown in case of a fault and to prevent blackouts.

Changes in protective device settings solely to reduce arc flash hazards will inevitably result in a loss of coordination, resulting in failure to operate or delayed operation during faults, and unnecessary blackouts. Examples of changing a slow fuse for a fast fuse can be found in (Doan and Sweigart, 2003) and (Sutherland, 2009a).

Resolution of combined arc flash and coordination problems requires evaluation of multiple options. The solution of one problem may cause the other problem to reappear in another place. The presence of old equipment with unreliable characteristics complicates the assessment. Published time-current curves and breaker-opening times should not always be relied on. Proposed solutions may not always be implemented, and the engineer should be prepared for a long drawn out process before all problems are resolved.

A more extensive list of "design changes" and "overcurrent upgrades" is given in (Hodder et al., 2006).

This is slightly different from the problem of the design of limited arc energy distribution systems (Das, 2005) although many of the same principles can be used.

2. Factors Affecting the Severity of Arc Flash Hazards

There are many formulae and procedures for calculating arc flash hazard results. These have been evaluated, and it has been found that the IEEE Standard 1584-2002 (IEEE, 2002b) method for up to 15 kV and the Lee method (Lee, 1971) for voltages above 15 kV give the most accurate results (Ammerman, Sen, and Nelson, 2009). The empirical method has been used to find the best curve fits for the variables on the basis of laboratory tests. Theory-based approaches have not been shown to be as accurate. The following variables are utilized:

fault current, kA. This is the current for a three-phase zero-impedance short circuit with no arcing present, and thus a zero-voltage fault also. Bolted fault current for typical electrical systems are limited by the impedance of the generators or utility system (sources), and by the impedance of the lines, cables, and transformers that bring the current to the load (distribution system). Additional fault current may be contributed by synchronous and induction motors, which produce rapidly decaying fault currents due to magnetic field transients and load and motor inertia. For the IEEE Standard 1584-2002 (IEEE, 2002b) model, the available bolted fault currents ranged from 0.70 to 106 kA. Typical secondary low-voltage systems have bolted fault currents ranging from 1 to 50 kA. Bolted fault current is the primary factor in determining both arcing current and incident energy.

The system voltage is the nominal voltage based on system transformation ratios. The actual voltage during a short circuit may be considerably less. Tests for IEEE Standard 1584-2002 were conducted for the voltage range of 208-15 kV. All tests were performed for three-phase, alternating current (AC) systems. The arcs at 208 V were not always self-sustaining, so the caveat was added that the model does not apply for voltages of 240 V and below, fed by a single transformer rated less than 125 kVA. System voltage is a small factor in determining arcing current.

The spacing between electrodes is dependent on the type of equipment and the system voltage. Typical values are 25 mm for low-voltage panelboards and MCCs, 32 mm for low-voltage switchgear, 104 mm for 5 kV switchgear, and 152 mm for 15 kV switchgear. For the tests in IEEE-1584-2002, vertical bus sections were used, both in open air and enclosed in a box.

This is representative of switchgear, motor control center (MCC), and panelboard construction. The electrode gap was a small factor in determining both arcing current and incident energy.

Box gap. The spacing between the electrodes and the back of the box used in testing were representative of actual distances used in equipment. The box gap was not a factor in determining arcing current or incident energy.

The presence or absence (in open conductor configurations) of a box gap was a small factor in determining the arcing current. See factor described below.

Grounding type. Power distribution systems may be operated with the system neutral connected to earth ground by a low-resistance cable, in which case the system is said to be solidly grounded. Solid grounding is the norm for electric utility transmission and distribution and for industrial and commercial low-voltage distribution in the United States. Grounding the neutral through a small resistor, called low-resistance grounding, is used in many industrial power systems at medium voltage in order to limit equipment damage due to ground faults. Low-resistance grounding typically limits ground fault current to the range of 100-400 A, and requires special ground fault protection relays to operate. For systems where service continuity is paramount, high-resistance grounding may be used. In this system, ground fault current is limited to 5-10 A, and an alarm is produced when a ground fault is detected. This may be investigated and removed at a time which is convenient, with the system still operating. However, if a second ground fault should occur, the resulting phase-to-phase fault may shut the system down.

The effectiveness of a grounding system, whether solid grounding, depending on ground rods and ground grids, or impedance grounded system, may be measured using the zero-sequence impedance ratios (Central Station Engineers, 1964, p. 464). A system may be said to be "effectively grounded" when eq.1 where…

= zero-sequence reactance for a given point of the system

= positive-sequence reactance for a given point of the system

= zero-sequence resistance for a given point of the system.

For hand calculation, the user may simply examine the single-line diagram and determine whether a system is solidly grounded or resistance grounded. For computer calculation, the zero-sequence impedance method or any other which gives equivalent results should be utilized. The grounding type was a small factor in determining the arcing current. See factor described below.

Fault type. In three-phase systems, most faults will quickly escalate into three-phase faults; therefore, this type was used in the tests for IEEE 1584-2002. However, because single-phase systems do exist and may experience arcing faults, further testing to extend the model is underway.

This is the time from arc inception to arc interruption. Arc interruption is usually performed by a fuse or circuit breaker. The mode of operation of these devices is to create a controlled arc within a safe enclosure, which is drawn out and extinguished, thus halting the flow of current.

Arcing time may range from 0.0 s up to several seconds. Typically, arc flash calculations will limit the maximum arcing time calculated to 2s. After this time, all persons are assumed to have been moved away from the vicinity of the arc. Arcing time is a linear factor in the calculation of incident energy.

Frequency. Tests for both 50 and 60 Hz systems were made for IEEE 1584-2002. No significant difference was found between the two. Tests for direct current (DC) and for other frequencies of AC, such as 400 Hz used in aircraft electrical systems, will be performed at a later date.

Electrode Materials tested for IEEE 1584-2002 included copper and aluminum, the most common conductor materials. No significant effect of electrode materials was found.

…from arc location to worker, in mm.

The incident energy is inversely proportional to the distance squared in open air, and raised to the distance exponent, , in an enclosure.

The working distance is either determined in the field, or standardized for different types of equipment at different voltages. Typically, an arm's length, 455 mm (18 in.), is used for low-voltage MCCs and panelboards, while longer distances are used for low-voltage switchgear, 610 mm (24 in.), and medium-voltage switchgear, 910 mm (36 in.).

The following factors were used in the calculations:

The incident energy is inversely proportional to the distance squared in open air, and raised to the distance exponent, , in an enclosure.

This is a statistical factor used in IEEE 1584-2002 to bring the curve-fitting results into a 95% confidence level.

For system voltages below 1 kV, while for system voltages above 1 kV.

For systems where the nominal voltage is less than 1 kV, the calculations for incident energy may not give the worst-case results using the arcing times determined with the calculated arcing current. For these systems, a second calculation is taken at 85% of calculated arcing current, with correspondingly longer arcing times (dependent on the protective device characteristics). The worst-case incident energy from these two calculations is used in the final result.

for enclosure type in arcing current calculation:

for open conductors (no enclosure).

for box enclosure.

for enclosure type in incident energy calculation.

for open conductors (no enclosure).

for box enclosure.

for grounding type in incident energy calculation.

for ungrounded and high-resistance grounded systems.

for grounded systems.

to produce a just-curable second-degree burn.

This can be converted to calories per square centimeter, which is a commonly used set of units in the field, by dividing by 4.184. Thus…

The following are calculated values:

current, kA, is the current in the electrical arcing fault:

eq.2

E = Incident energy, J/cm². This is the radiant energy from the arc to which a person is exposed to at distance D for time t. The IEEE 1584-2002 procedure has two steps. The first is to compute normalized incident energy at , and .

eq.3

This normalized value is converted in the second step to the actual time and distance.

eq.4

Lee Method for incident energy (Lee, 1982).

The IEEE 1584-2002 model extends only to 15 kV systems and for conductor gaps 13 mm < G < 152 mm. The more conservative Lee model may be used outside this range.

eq.5

The IEEE 1584-2002 model is eq.6 The Lee model is eq.7

3. Example Arc Flash Calculations

Given 480V switchgear, in a solidly grounded system, with available bolted fault current of 40 kA, bus gap is 32mm, and working distance is 610mm. The interrupting time is 0.3 s.

Next, compute normalized incident energy at , and ...

This normalized value is converted in the second step to the actual time and distance.

The flash-protection boundary.

cont. to part 2 >>