A historical perspective and comparative study of the standards IEEE 1584 and NFPA 70E

BY RAVEL F. AMMERMAN, P.K. SEN, & JOHN P. NELSON

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

HE EXPOSURE TO hazards associated with electrical arcing phenomena when working

on energized equipment is a topic of significant interest to industrial plant personnel. This article provides an overview of the current arc-flash standards, focusing on the methods used to calculate incident energy levels in a system. A thorough sensitivity analysis of the arc-flash hazard incident energy calculations currently adopted by the IEEE 1584 standard leads to some possible conservative simplification of the equations. These simple equations could be used for a quick first-cut assessment of the incident energy levels present in a system. A case study using data from a typical petrochemical application provides a comparison of the National Fire
Digital Object Identifier 10.1109/MIAS.2009.932345

42

CREATAS

1077-2618/09/$25.002009 IEEE

OSHA initially used the National Protection Association (NFPA) 70E Electrical Code (NEC) as a basis for and IEEE 1584 arc-flash incident electrical regulations. Because the NEC energy equations and the results obOSHA INITIALLY largely does not address employee tained using the proposed simplified safety, it became apparent that a new calculations. USED THE standard was needed. As a result, on 7 Awareness of the various hazards NATIONAL January 1976, a new NFPA electrical caused by an arc flash has increased standards development committee was significantly over the past two decaELECTRICAL formed. This group was given the task des. The regulations, standards, reof assisting OSHA in preparing stansearch, and application guidelines CODE (NEC) AS dards specifically addressing electrical focus on reducing the exposure of safety. The Committee on Electrical personnel to burn injuries associated A BASIS FOR Safety Requirements for Employee with arc-flash events in low-voltage Workplaces published the first edition (LV) and medium-voltage (MV) appliELECTRICAL of NFPA 70E in 1979. The initial cations. Currently, the NFPA 70E REGULATIONS. edition covered installation safety Standard for Electrical Safety in the requirements. Three subsequent ediWorkplace [1] and the IEEE 1584 tions over the next decade added secGuide for Performing Arc-Flash Hazard tions on safety-related work practices Calculations [2], [3] have the same goal, protecting individuals who must work on or near and safety-related maintenance requirements. OSHA used energized electrical equipment. However, the philosophi- this work to create many of its regulations applying to cal approaches used by these two groups to estimate the electrical safety. Title 29 of the Code of Federal Regulations (CFR), Subarc-flash hazards are different. For many who are required to apply and follow the standards, the arc-hazard calcula- part S, addresses electrical safety requirements that are nections and the interpretations, at times, can be confusing, essary for the practical safeguarding of employees in their particularly when there are discrepancies among the meth- workplaces [4]. It was not until 1991 that OSHA added ods being used. It has been widely accepted in the power words acknowledging arc flash as an electrical hazard. The industry that there is a need to perform additional research fifth edition of the NFPA 70E, published in 1995, became and refine the arc-flash calculation methods to more effec- the first standard specifically addressing the arc-flash hazard. This printing included requirements for protective tively manage the hazard. This article focuses primarily on the arc-flash incident clothing and defined a flash-protection boundary. The next energy calculations currently being used. Following a two revisions focused on detailed arc-flash hazard analysis, summary of the history of arc-flash hazard research and a providing more specifications regarding the arc-flash probrief review of the calculations, a sensitivity analysis is tection boundaries and incident energy calculations. NFPA performed, which leads to a potential simplification of the 70E-2004 includes sample calculations of flash protection incident energy calculations presented in the standards. It boundaries in Annex D. It is important to note, as quoted should be emphasized that the sensitivity analysis and the on page 70E-98 of the standard: This annex is not a part of recommendations suggested in this article are based exclu- the requirements of this NFPA document but is included sively on existing test data collected and made available in for informational purposes only [1]. In addition to the NFPA 70E standard and the OSHA IEEE 1584 and the equations presented as a part of the NFPA 70E-2004 and IEEE 1584-2002 standards. Finally, Title 29 (CFR), in 2002, the NEC started requiring the calculations are provided to compare the incident energy use of labels warning workers about potential arc-flash equations and to validate the proposed simplified hazards. This same year an IEEE working group comapproach for estimating energy levels. A number of graphs pleted the publication of the standard IEEE 1584-2002: and charts are added to enhance the understanding and Guide for Performing Arc-Flash Hazard Calculations. The new standard presented models for estimating incident possibly simplify the future application guidelines. energy levels based on a large amount of test data. As seen Evolution of Arc-Flash Standards from this brief summary, until recently, the arc-flash hazard has not been widely acknowledged. Extensive research and testing performed of late has led to a better underHistorical Perspective of the Development standing of the arc-flash hazard. The next section highof Arc-Flash Regulations and Standards On 29 December 1970, the Occupational Safety and Health lights, chronologically, some of the most significant Act was signed into law. The general duty clause mandates contributions to the body of knowledge pertaining to arcthat each employer shall furnish to each of his employees, ing phenomena and the associated hazards. employment and a place of employment which are free from recognized hazards that are causing or are likely to cause Significant Milestones in Arc-Flash Research death or serious physical harm to his employees [4]. The 1) Lee: In 1982, The Other Electrical Hazard: ElectriOccupational Safety and Health Administration (OSHA), cal Arc Blast Burns [5] was published. This article given the responsibility of providing for worker safety, iniis considered by many to be one of the most importiated the development of Federal regulations, including tant research contributions on arcing phenomena in those that targeted identifying the electrical hazards and open air. This article was significant in that it implementing safe work practices. quantified the potential burn hazards and educated

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

43

44

very relevant article in 1987, Pressures Developed from Arcs [6]. The pressure effects of an arc incident are quantified in this publication. 2) Doughty et al. [7]: Testing Update on Protective Clothing and Equipment for Electric Arc 1800 Exposure [7], published in Electric Arc 1997, details the incident Research energy levels associated with LV arc-flash events and was the Developments first to describe how an event Leading to Regulations and Standards is intensified when the arc iniAddressing the Arc-Flash Hazard 1970 tiates within electrical equipNational Electrical Code ment enclosures. OSH Act Used for the Basis of 3) Doughty et al. [8]: Predicting 29 December 1970 OHSA Regulations Incident Energy to Better Manage the Electric Arc Hazard on NFPA 70E NEC Does Not Address 600 V Power Distribution SysCommittee Worker Safety Formed to Assist OSHA tems [8] was published in New Standard Needed 1976 2000. This article semiempirically quantified the incident NFPA 70E energy calculations for LV sysEdition 1 tems and is the source of the inPart I cident energy calculations used in the NFPA 70E standard. 1980 4) Jones et al.: In 2000, Staged NFPA 70E Arc-Flash Edition 2 Tests to Increase Awareness of Research Part II Added Arc-Flash Hazards in Electrical Equipment [9] was also pubNFPA 70E lished. Experimental investigaRalph Lee Edition 3 tions, using mannequins, were Arc-Blast Burns Part III Added conducted to improve the understanding of how humans can NFPA 70E Ralph Lee be adversely affected by arcEdition 4 Pressure from Arc flash incidents. Minor Revisions Blasts 5) IEEE Standard 1584: The first edition of IEEE Guide for Per1990 forming Arc-Flash Hazard CalcuOSHA lations [2] was issued in 2002. Subpart S Arc-Flash Added This standard used extensive test data to develop empirical Testing Update on NFPA 70E equations derived from statistiProtective Clothing and Edition 5 cal analysis. Tests data were Equipment for Electric Arc Arc-Flash Added Exposure made available from various sources and are included as an 2000 Predicting Incident Energy to appendix to the standard. An Better Manage the Electric article written by Gammon and Arc Hazard on 600-V Power NFPA 70E Matthews, IEEE 1584-2002, Distribution Systems Edition 6 Incident Energy Factors and SimArc-Flash ple 480-V Incident Energy EquaRequirements NEC Requires ArcStokes and Sweeting tions [10], includes a thorough Expanded Flash Hazard Warning Electric Arcing Burns statistical analysis and summary of the IEEE 1584 test data. IEEE 1584-2002 NFPA 70E 6) Stokes and Sweeting: Electric ArcIEEE 1584-2004a Effects of Insulating Guide for Performing Edition 7-2004 Barriers in Arc-Flash ing Burn Hazards [11], pubArc-Flash Calculations Testing lished in 2006, provides a critical evaluation of the testing method2006 ology, in particular, the electrode Collaborative Effort orientation used to assess the Development of Definitive Industry Standard arc-flash hazard for the IEEE 1 1584 standard development. In Historical development of arc-flash standards. addition, this article included an

personnel about the safety implications. Lee established the curable burn threshold for the human body as 1.2 cal=cm2 . Lee also published a second

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

extensive list of literature on n System conditions electric arcs. The authors suggest n available short-circuit current that this body of knowledge has n X/R ratio IN 2006, THE IEEE largely been overlooked in the n prefault voltages AND NFPA development of the current IEEE n loading. 1584 standard. Several discussion n Protective devices (timecurrent AGREED TO articles were published which characteristic) provided additional analysis of n the first upstream device COLLABORATE the issues being debated. Clon the second upstream device. sure to Discussions of Electric n System grounding ON A JOINT Arcing Burn Hazards [12] pubn Electrical electrodes and potential lished by Stokes and Sweeting arc lengths RESEARCH further documented their concerns. n spacing between phases INITIATIVE TO 7) Wilkins et al.: Effect of Insulatn spacing between phases and ing Barriers in Arc-Flash Testground INCREASE THE ing [13] was published in n orientation 2008. The authors of this article n insulated versus noninsulated. UNDERSTANDING used vertical conductors termin Size and shape of enclosures nated in insulating barriers for n Atmospheric conditions OF ARC-FLASH their testing methodology. The n ambient temperature nature of the arc is very similar n barometric pressure PHENOMENA. to what is observed when the n humidity. electrodes are oriented horizonn Arc conditions tally, thus reinforcing the work n randomness of the arc of Stokes and Sweeting. Lang presented additional n interruption of the arc information regarding the evaluation of alternate n arc plasma characteristics test configurations in February 2007 at the 14th n other unidentified factors. Annual IEEE/IAS Electrical Safety Workshop held n Dissipation of energy in Calgary, Alberta, Canada [14]. n heat n latent heat of vaporization Future Development of Arc-Flash Standards n light In 2006, the IEEE and NFPA agreed to collaborate on a n sound joint research initiative to increase the understanding of n pressure wave. arc-flash phenomena. This effort also plans to include n Other miscellaneous factors. working with the international community for global Reviewing this long list of variables, it is obvious that adoption of such a standard. It is the hope that the part- determining the precise arc-flash incident energy to which nership between these two organizations will lead to a a worker may be exposed is extremely difficult, if not definitive industry standard regarding arc-hazard analy- impossible. Only an estimate of a workers potential incisis and mitigation. Figure 1 summarizes some of the dent energy exposure can be established. Consequently, it more significant events in the development of arc-flash is wise to have the calculations be on the conservative or hazard standards and regulations, provided for future safer side when protecting personnel. reference only.
NFPA 70E Arc-Flash Incident Energy Calculations

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

One of the most important and essential elements of an arc-flash hazard analysis is the estimation of the incident energy. These calculations help predict the amount of energy available during an arc-flash event. Incident energy is typically expressed in J=cm2 or cal=cm2. The calculations detailed by NFPA 70E-2004 and IEEE 1584-2002 are used to establish the flash protection boundary, i.e., the distance from an arc source that would cause the onset of a second-degree burn. The energy required to produce a curable, second-degree burn on unprotected skin has been established as 5.0 J=cm2 (or 1.2 cal=cm2 .
Factors Influencing Incident Energy Levels

The sixth edition of NFPA 70E, Standard for Electrical Safety Requirements for Employee Workplaces [15], includes a set of equations used to calculate the available incident energy for LV systems (600 V and below). The seventh edition, NFPA 70E-2004, Standard for Electrical Safety in the Workplace, moved the incident energy calculations to Annex D. The IEEE 1584-2002 methods for computing incident energy are also included in the annex. The calculations are used to establish the personal protective equipment (PPE) required for a worker. The incident energy calculations, based on fault current, working distance, and protective equipment clearing times are as follows: EMA 5, 271 D1:9593 A 3 tA 0:0016 F2 0:0076 F 0:8938, EMB 1038:7
3 tB 0:0093 F 0:3453 F 5:9675,

To help the reader fully comprehend the complexity of these types of calculations, a comprehensive list of the factors influencing the incident energy is provided below. This has been known and well recognized by researchers over the years.

(1) (2)

D1:4738 B 2

45

where EMA , maximum open air incident energy (cal=cm2 ); EMB , maximum 20 in cubic box incident energy (cal=cm2 ); DA and DB , distance from arc electrodes (in) (for distances 18 in and greater); tA and tB , arc duration (s); F, short-circuit current (kA) (for the range of 1650 kA).
IEEE 1584

Incident Energy Calculations

lg (En ) K1 K2 1:081 lg (Ia ) 0:0011 G En 10

lg (En )

(6) (7)

The IEEE 1584-2002 standard was developed using test data compiled from several laboratories. The calculations, which were derived statistically, are used to predict the incident energy an employee could experience when working on energized equipment. These equations also help establish the boundary distances for workers not wearing the proper PPE. This article does not address the charts and simplified equations that were developed for Class L and RK1 LV fuses or for the materials presented that deal with certain types of LV circuit breakers. This article focuses specifically on the incident energy equations that are described later. Compared with the NFPA 70E calculations, the IEEE 1584 equations are more complicated, involving an increased number of variables. It is also apparent that the IEEE 1584 calculations accommodate a wider range of voltage and bolted fault current levels. The equations for the incident energy calculations are summarized as follows:
Arcing Current Calculations
n

where En , normalized incident energy ( J=cm2 ); K1 , 0.792 for open configurations and 0.555 for box configurations; K2 , 0 for ungrounded and high-resistance grounded systems and 0.113 for grounded systems; G, gap between conductors (mm) (Table 1). E Cf En (t=0:2) (610x =Dx ), (8)

where E, incident energy (cal=cm2 ); Cf , calculation factor 1.0 for voltages above 1 kV and 1.5 for voltages below 1 kV; En , normalized incident energy ( J=cm2 ); t, arcing time (s); D, distance from the possible arc point to the person (mm); x, distance exponent (Table 1).
Lee Method

For cases outside the ranges established for use in both the NFPA 70E and IEEE 1584 standards, the Lee method is to be used. The Lee model is presented below:  (9) E 5:12 3 105 V Ibf (t D2 ), where E, incident energy (cal=cm2 ); V, system voltage (kV); t, arcing time (s); Ibf , bolted 3/ fault current (kA); D, distance from the possible arc point to the person (mm). Figure 2 provides a comparison between the NFPA 70E and IEEE 1584 standards. This side-by-side appraisal of the standards includes a summary of variables needed to calculate the incident energy and the conditions for which the calculations are applicable.

System voltage less than 1,000 V: lg (Ia ) K 0:662 lg(Ibf ) 0:0966 V 0:000526 G 0:5588 V lg (Ibf )

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

0:00304 G lg (Ibf ):
n

(3)

System voltage more than 1,000 V: lg (Ia ) 0:00402 0:983 lg (Ibf ), Ia 10

lg (Ia )

(4) (5)

46

Sensitivity Analysis of IEEE 1584 Incident Energy Calculations where Ia , arcing current (kA); K, 0.153 for open configura- Many companies follow the IEEE 1584 methodology when tions and 0.097 for box configurations; Ibf , bolted 3/ fault calculating incident energy levels, because it was developed current [symmetrical rms (kA)]; V, system voltage (kV); G, using a large number of test data and encompasses a wider gap between conductors (mm) (Table 1); lg, log with a base 10. range of voltage and current. This method is believed by many to provide more accurate results. On the other hand, the TABLE 1. FACTORS FOR EQUIPMENT AND VOLTAGE CLASSES. increased complexity of the reTypical Gap Between Distance System Voltage quired calculations suggests that a Equipment Type Conductors (mm) Exponent (x) (kV) computer program should be used to manage the equations effec0.2081 Open air 1040 2.000 tively. The IEEE 1584-2002 stanSwitchgear 32 1.473 dard comes equipped with a set of spreadsheet calculators to assist MCC and panels 25 1.641 with an arc-flash study. These calCable 13 2.000 culators are not always easy to follow and at times can be confusing. >15 Open air 102 2.000 Some companies rely heavily on Switchgear 13102 0.973 commercial software packages to Cable 13 2.000 help estimate incident energy levels within their facility. >515 Open air 13153 2.000 In an effort to provide a clear Switchgear 153 0.973 understanding of the IEEE 1584 incident energy calculations, the Cable 13 2.000 sensitivity of the equations to the ,

NFPA 70E - 2004 Standard for Electrical Safety in the Workplace

IEEE 1584 - 2002 Guide for Performing Arc-Flash Hazard Calculations

System Application Limits Parameter Working Distance Type of Installation Voltage Level Range of Current Time of Arc Exposure NFPA 70E-2004 457 mm (18 in) or more Open Air or Cubic Box 208600 V 1650 kA Unlimited IEEE 1584-2002 457 mm (18 in) or more Open Air, Cubic Box and Cable Bus 208 V15 kV 0.7106 kA Unlimited

Incident Energy Equations (Incident Energy in cal/cm2) System Voltage Under 1,000 V Maximum open-air incident energy (cal/cm2) E = 5271 D 1.9593 t [0.0016
2 Ibf

0.0076 Ibf + 0.8938]

lg Ia = K + 0.662 lg (I bf) + 0.0966 V + 0.000526 G + 0.5588 V (lg (Ibf)) 0.00304 G (lg (Ibf)) 1,000 V < System Voltage < 15 kV lg (Ia) = 0.00402 + 0.983 lg (Ibf) Ia =10 lg(Ia) lg (En) = K1 + K2 + 1.081 lg (Ia) + 0.0011 G En = 10 lg(En) Normalized incident energy (J/cm2) E = Cf En (t / 0.2) (610x/ Dx )

Maximum 20 cubic box incident energy (cal/cm2) E = 1038.7 [0.0093 where D is in inches D 1.4738 t
2 I bf

0.3453 Ibf + 5.9675]

Required Variables and Constants Constants K: 0.153 for open configurations 0.097 for box configurations K1: 0.792 for open configurations (no enclosure) 0.555 for box configurations (enclosed equipment) K2: 0 for ungrounded and high-resistance grounded systems 0.113 for grounded systems C f: calculation factor 1.0 for voltages above 1 kV 1.5 for voltages below 1 kV Distance Factors and Typical Conductor Gaps
System Voltage (kV) Equipment Type Open Air Switchgear MCC and Panels Cable Open Air Switchgear Cable Open Air Switchgear Cable Typical Gap Between Conductors (mm) 1040 32 25 13 102 13102 13 13153 153 13 Distance Factor 2.000 1.473 1.641 2.000 2.000 0.973 2.000 2.000 0.973 2.000

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

Required Variable V: System Voltage (kV) Ibf: Fault Current (kA) t: Arcing Time (sec) D: Working Distance (mm) Open/Enclosed Equipment G: Conductor Gap (mm) K: Grounded, Ungrounded, and High-Resistance Grounded Systems Cf: Calculation Factor x: Distance Factor

NFPA 70E

IEEE 1584

0.2081

>15

> 515

Lee Method: E = 5.12 105V Ibf (t / D 2) Used to predict the open-air incident energy levels in cases where working voltages or conductor gaps fall outside of the range of the NFPA 70E and IEEE 1584 Standards

2
Summary of incident energy calculations.

47

assorted variables was investigated. This is to be emphasized; no effort has been made in this article to validate the equations presented in the IEEE Standard 1584. The approach is demonstrated on the equations derived for LV and MV systems. A discussion of three cases follows: 1) system voltages 480 V and below, 2) system voltages below 600 V, and 3) system voltages over 1,000 V. These voltage values were selected so that some comparison could be made with the experimental results provided in the IEEE Standard 1584.
System Voltages 480 V and Below

Equations (3) and (5) were evaluated for different conductor gap distances on a 480-V system with a box configuration as shown in Figure 3. Figure 3 reveals that a worst case relationship between the arcing current and the bolted

ONE OF THE MOST IMPORTANT AND ESSENTIAL ELEMENTS OF AN ARC-FLASH HAZARD ANALYSIS IS THE ESTIMATION OF THE INCIDENT ENERGY.

fault current, for voltages less than 480 V, in closed configurations, could be approximated by the simplified equation of a straight line given by (as shown in Figure 3) Ia 0:6 Ibf : (10)

Arcing Versus Bolted Fault Current (480 V, Box Configuration) 70 Arcing Fault Current (kA) 60 50 40 30 20 10 0 0 20 40 60 Bolted Fault Current (kA) 80 100 la = 0.6 lbf G = 10 mm G = 25 mm G = 40 mm

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

Sensitivity to gap between conductors: 480 V and below.

Using information extracted from the IEEE 1584 test results database, arcing current and three-phase bolted fault current data corresponding to voltages less than 480 V are plotted (scatter plot) in Figure 4. A line representing (10) is superimposed over the IEEE 1584 test data. Test points on the scatter plot are grouped into three shaded areas as identified. Group 3 is of particular interest because of the wide variation in arcing current levels observed for fault currents more than 100 kA. This is attributed to the test procedure, which investigated a broad range of conductor gap distances varying from 7 to 32 mm. Figure 3 reveals that the incident energy equation models the increasing influence of conductor gap distance at higher fault current levels. The simplified approach does not account for this because only the worst case was considered, i.e., a conductor gap of 10 mm. Nevertheless, all of the data points fall beneath this line, confirming that this simple linear equation could be used as a worst case approximation of the relationship between the arcing current and bolted fault current, for voltages less than 480 V, in closed configurations. It will be interesting to see whether this simplification could be proven by additional testing. Next, the sensitivity of the normalized incident energy to the conductor gap distance was evaluated using (6) and (7). Results for ungrounded or high-resistance grounded systems are shown in Figure 5. Ungrounded systems are featured in this study because our investigation confirmed other literature, which states that typically an

Arcing Versus Bolted Fault Current IEEE 1584 Test Data Under 480 V Sample Size: 48 60 Normalized Energy (J/cm2) Arcing Fault Current (kA) 50 la = 0.6 lbf 40 30 20 1 10 0 0 20 60 40 80 Bolted Fault Current (kA) 100 2 3 30 25 20

Normalized Incident Versus Arcing Current (480 V, Ungrounded) G = 10 mm G = 25 mm G = 40 mm

En = 0.43 la 15 10 5 0 0 10 20 30 40 Arcing Current (kA) 50 60

48

4
Sensitivity to gap between conductors: 480 V and below.

Scatter plot of IEEE 1584 LV test data: 480 V and below.

ungrounded system results in the incident energy level thats about 30% greater than that of a solidly grounded system [16]. Taking the worst case, as depicted on the graph later, in a similar approach as before, a simplified equation for the incident energy is derived as follows: En 0:43 Ia : (11)

Taking the worst case shown on the graph, as before, a simplified expression for the incident energy is derived: En 0:43 Ia : (15)

Finally, using the procedure previously described, the incident energy calculation derived for this case is summarized as E 4:14 Ibf (t):
System Voltage More Than 1,000 V

Last, the incident energy is calculated using (8). Continuing to simplify the equations using the worst case approach, the working distance (D) is set to 457 mm (18 in) and the value of 1.641 is selected for the distance exponent (x) for G 25 mm. It should be noted that selecting the worst case condition for the distance exponent could be problematic. It has been shown that the procedure adopted to calculate the incident energy using a distance exponent can give anomalous results [17], and it should be further investigated. Because this is a LV system being evaluated, 1.5 is used for the calculation factor (Cf ). The incident energy calculation shown with the values selected for the variables follows: E 1:5 En (t=0:2) (610=457)1:641 : (12)

(16)

A similar approach was used to evaluate the incident energy calculations for voltages more than 1,000 V using the IEEE 1584 equations listed as (4) and (5). Figures 911 detail the data used to develop the simplified equations. Interpreting the information on the graphs gives the equations that follow: Ia 0:95 Ibf , En 0:60 Ia : (17) (18)

Combining (10) to (12) results in a simplified form of the incident energy equation: E 3:11 (Ibf ) (t), (13)

To conclude this process, the incident energy for an ungrounded system is calculated using (8). The working
Arcing Versus Bolted Fault Current IEEE 1584 Test Data Under 1 kV Sample Size: 166
IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

Arcing Fault Current (kA)

where E, incident energy (cal=cm2 ); Ibf , bolted 3/ fault current (kA); t, arcing time (s).
System Voltages 600 V and Below

80 70 60 50 40 30 20 10 0 0

la = 0.8 lbf

A similar procedure was applied for different conductor gap distances on a 600-V system with a box configuration. Figures 68 summarize the information. As before, the data are grouped into three distinct regions. A scatter plot of the data confirms that as a worst case the relationship between the arcing current and bolted fault current, for voltages less than 1,000 V in a closed configurations, could be approximated by the simplified equation Ia 0:8 Ibf : (14)

2 1 20 60 80 40 Bolted Fault Current (kA) 100

Scatter plot of IEEE 1584 LV test data: 600 V and below.

Arcing Versus Bolted Fault Current (600 V, Box Configuration) 90 Normalized Energy (J/cm2) Arcing Fault Current (kA) 80 70 60 50 40 30 20 10 0 0 20 40 60 80 Bolted Fault Current (kA) 100 la = 0.8 lbf G = 10 mm G = 25 mm G = 40 mm 40 35 30 25 20 15 10 5 0

Normalized Incident Energy Versus Arcing Current (600 V, Ungrounded) G = 10 mm G = 25 mm G = 40 mm En = 0.43 la

10

20

30 40 50 60 Arcing Current (kA)

70

80

Sensitivity to gap between conductors: 600 V and below.

Sensitivity to gap between conductors: 600 V and below.

49

distance (D) is set to 457 mm (18 in), and the value of 2.000 is selected for the distance exponent (x). Because the voltage of the system being evaluated is greater than 1 kV, 1.0 is used for Cf . Equation (19) summarizes this approach: E 1:0 En (t=0:2) (610=457)2:000 : (19) Combining (17)(19) results in a simplified form of the incident energy equation: E 5:1 (Ibf ) (t): (20)

ONE OF THE DRAWBACKS OF USING THE NFPA 70E APPROACH IS THE POTENTIAL TO OVERPROTECT THE WORKERS.

(0.1 s), then the expressions become as follows: E 0:31 (Ibf ), E 0:41 (Ibf ): (21) (22)

Verification of Results Derived for the LV Case

To check the validity of the simplified approach presented in this article, average incident energies and bolted fault currents data from the IEEE 1584 test database are plotted for voltages under 1,000 V. The values selected correspond to arc durations of approximately six cycles (100 ms). If the time in (13) and (16) is set equal to 100 ms

Figure 12 shows that all the available data points fall below the lines representing the equations derived for the 480- and 600-V cases, indicating that the simplified approach results in conservative estimates for the incident energy levels in a LV system when compared with the available test data. As an additional means of verification, the simplified version of the equations presented in this article are compared with the results derived from the NFPA 70E and IEEE 1584 incident energy equations. NFPA 70E incident energy levels were derived using (2) and IEEE 1584 (3) and (5)(8) were used to calculate the incident energy levels. Figure 13 provides a summary of the calculations in a graphical form, confirming that

Arcing Versus Bolted Fault Current 45 Arcing Fault Current (kA) 35 30 25 20 15 10 5 0 0 5 10 15 20 25 30 35 Bolted Fault Current (kA) 40 45 la = 0.95 lbf Normalized Energy (J/cm2) 40 60 50 40 30 20 10 0

Normalized Incident Energy Versus Arcing current (Undergrounded) G = 23 mm G = 103 mm G = 53 mm G = 153 mm

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

En = 0.6 la

20

40 60 Arcing Current (kA)

80

100

11

System voltage more than 1 kV.

Sensitivity to gap between conductors: More than 1 kV.

Incident Energy (cal/cm2)

Arcing Fault Current (kA)

45 40 35 30 25 20 15 10 5 0 0 5

Arcing Versus Bolted Fault Current IEEE 1584 Test Data Under 1 kV Sample Size: 148

Incident Energy Versus Bolted Fault Current 40 35 30 25 20 15 10 5 0 0 20 40 60 80 Bolted Fault Current (kA) 100 E = 0.31 lbf (480 V) E = 0.41 lbf (600 V)

la = 0.95 lbf

10

15 20 25 30 35 Bolted Fault Current (kA)

40

45

50

10

12

Scatter plot of IEEE 1584 MV test data.

Validation of simplified approach.

the simplified approach gives conservative values in most cases. The NFPA hazard or risk categories are also shown on the figure for reference.
Incident Energy (cal/cm2)

Incident Energy Versus Bolted Fault Current IEEE 1584, NFPA 70E, and Simplified Method Calculation Comparison 35 E = 0.41 lbf 30 Hazard Risk Category 4 (600 V) E = 0.31 lbf (480 V)

Discussion of Results for the LV Case

Petrochemical System: A Case Study Data from a typical petrochemical power distribution network were used to perform an arc-flash comparative study using one of the commercial software packages available. Nominal voltages of 480 V, 4.16 kV, and 12 kV are present in the system, providing a good opportunity to compare both the LV and MV incident energy calculations. Figure 14 shows some of the results of the study. Recall that the NFPA 70E incident energy calculations are valid for voltage levels up to 600 V and for bolted fault currents between 16 and 50 kA. Therefore, the Lee method was used to estimate the energy values for the 12-kV bus. A working distance of 18 in was used for all the calculations. Using the calculations described in the arc-flash standards, the NFPA 70E approach typically produces more conservative estimates (higher values) for the incident energy in the cases explored for this study. One of the drawbacks of using the NFPA 70E approach is the potential to overprotect the workers. Using a sensitivity analysis, the alternative method proposed in this article produces more conservative estimates of the potential incident energy exposure than the IEEE 1584-2002 equations. This is to be expected as the method is based on some simplifications of the IEEE 1584-2002 equations. However, the estimates for incident energy gleaned from the approximations are less conservative than the NFPA 70E calculations.

Incident Energy (cal/cm2)

As observed in the preceding analysis, the sim20 plified approach provides Hazard Risk Category 3 conservative estimates of 15 arcing fault current and the incident energy levels. Caution applying this 10 simplified approach is advised at bolted faults 5 below 20 kA, because arc sustainability issues are 0 probable at these current 0 20 levels. Validation of the simplified approach in this section focused on the rela- Comparison of various results. tionship observed between incident energy and bolted fault current. It is well documented that the available fault current and timecurrent characteristics of the protective devices have the most significant effect on arc-flash hazard incident energy levels. Therefore, the next section, which features a case study, provides an analysis of the relationship between the incident energy and the arc duration.

25

IEEE 1584

NFPA 70E

Hazard Risk Category 2 Hazard Risk Category 1 g y Category

40 60 Bolted Fault Current (kA)

80

100

13

Incident Energy Versus Arc Duration (480 V Bus) 500 450 400 350 300 250 200 150 100 50 0 0 NFPA 70E IEEE 1584 Simplified (480 V) Simplified (600 V)

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS

20

40 60 80 Arc Duration (Cycles)

100

120

Incident Energy Versus Arc Duration (12 kV Bus) 1,000 900 800 700 600 500 400 300 200 100 0 0 20 40 60 80 Arc Duration (Cycles) 100 120 NFPA 70E IEEE 1584 Simplified (>1 kV)

Incident Energy (cal/cm2)

14
Comparison of incident energy calculations.

51

52

Acknowledgments Conclusions The authors wish to acknowledge the This article has reviewed the incident THE Power Systems Energy Research Cenenergy equations used in the NFPA 70E ter (PSerc) for helping to support this standard as well as the empirically based CALCULATIONS research. PSerc is an Industry Univerequations of the IEEE 1584 document. A sity Cooperative Research Center (presimplified quick assessment approach has ARE USED TO viously a National Science Foundation been proposed for performing incident ESTABLISH THE Center). energy calculations. The question is then, which PERSONAL method should be used? This issue is References [1] Standard for Electrical Safety in the Workplace, complex, because research focused on PROTECTIVE NFPA 70E-2004. modeling arc-flash events and predict[2] IEEE Guide for Performing Arc-Flash Hazard ing incident energy levels is still in its Calculations, IEEE 1584-2002. EQUIPMENT infancy. As emphasized in this article, [3] IEEE Guide for Performing Arc-Flash Hazard real arc-flash exposures are very diffiCalculations-Amendment 1, IEEE 1584a-2004. REQUIRED FOR [4] Title 29 of the Code of Federal Regulations cult to predict because of their ran(29 CFR). Part 1910: General Industry; dom complex nature and the large A WORKER. Safety Standards for Electrical Systems and number of variables involved. Further Safety-Related Work Practices, Occupacomplicating matters are the varied tional Safety and Health Administration (OSHA) Standards, Washington, DC. working conditions and actual equipment configurations [5] R. Lee, The other electrical hazard: Electrical arc blast burns, IEEE encountered. Trans. Ind. Appl., vol. IA-18, no. 3, pp. 246251, May/June 1982. The incident energy information derived from an arc[6] R. Lee, Pressures developed from arcs, IEEE Trans. Ind. Appl., flash study is used to help develop strategies for minimizvol. IA-23, no. 4, pp. 760764, July/Aug. 1987. [7] R. L. Doughty, T. E. Neal, T. A. Dear, and A. H. Bingham, Testing ing burn injuries. Effectively modeling a large-scale power update on protective clothing and equipment for electric arc exposystem, analyzing and accurately calculating the energy sure, in IEEE PCIC Conf. Rec., 1997, pp. 323336. released during an arc-fault event is the cornerstone of an [8] R. L. Doughty, T. E. Neal, and H. L. Floyd, Predicting incident arc-flash hazard analysis. NFPA 70E incident energy calenergy to better manage the electric arc hazard on 600 V power culations are based on theoretical concepts and from moddistribution systems, IEEE Trans. Ind. Appl., vol. 36, no. 1, pp. 257269, Jan./Feb. 2000. els derived using very limited test data. Similarly, IEEE [9] R. A. Jones, D. P. Liggett, M. Capelli-Schellpfeffer, T. Macalady, L. 1584 includes a theoretically derived model developed for F. Saunders, R. E. Downey, L. B. McClung, A. Smith, S. Jamil, and three-phase, open-air systems, applicable for any voltage. V. J. Saporita, Staged tests to increase awareness of arc-flash hazards In other words, both standards use Lees article [5] as the in electrical equipment, IEEE Trans. Ind. Appl., vol. 36, no. 2, pp. 659667, Mar./Apr. 2000. theoretical basis for understanding the electrical arcing phenomena. Lees research includes many simplifying [10] T. L. Gammon and J. H. Matthews, IEEE 1584-2002, incident energy factors and simple 480-V incident energy equations, IEEE assumptions, most notably that the shape of the arc is not Ind. Appl. Mag., vol. 11, no. 1, pp. 2331, Jan./Feb. 2005. important [5]. Certainly, a methodology developed for open [11] A. D. Stokes and D. K. Sweeting, Electric Arcing Burn Hazair is not suitable for situations where the arc initiates ards, IEEE Trans. Ind. Appl., vol. 42, no. 1, pp. 134141, Jan/ Feb 2006. within an enclosure or in cases where the system buses are Sweeting, Closure tightly spaced. In an attempt to fill in the obvious gaps, [12] A. D. Stokes and D. K. IEEE Trans. Ind. to discussions of electric arcing burn hazards, Appl., vol. 42, no. 1, the IEEE 1584 standard also featured statistically derived pp. 146147, Jan./Feb. 2006. models for incident energy calculations based on a signifi- [13] R. Wilkins, M. Lang, and M. Allison, Effect of insulating barriers in arc flash testing, IEEE Trans. Ind. Appl., vol. 44, no. 5, cant amount of test data. Test results obtained for the pp. 13541359, Sept./Oct. 2008. IEEE 1584 Standard were compiled using a vertical orienalternate test configurations tation of the three-phase arcing electrodes. The effect of [14] M. Lang and K. Jones, An evaluation ofat the 14th Ann. IEEE IAS for future arc flash models, Presented different electrode orientations and the use of insulating Electrical Safety Workshop, ESW2007-14, Calgary, Alberta, Canada, 2007. barriers have been investigated, and the results indicate that a horizontal electrode configuration produces higher [15] Standard for Electrical Safety Requirements for Employee Workplaces, NFPA 70E-2000. incident energy levels [11][14]. A number of specific [16] C. St. Pierre, Putting arc-flash calculations into perspective, Elecitems were presented at the 2007 IEEE/IAS Electrical tric. Constr. Mainten., vol. 103, no. 6, pp. 4858, June 2004. Safety Workshop [14] recommending ways to incorporate [17] R. Wilkins, M. Allison, and M. Lang, Calculating hazards, IEEE Ind. Appl. Mag., vol. 11, no. 3, pp. 4048, May/June 2005. this research into subsequent versions of the arc-flash hazard standards. In the meantime, the quick first-cut approach developed in this article represents an effective way to estimate Ravel F. Ammerman (rammerma@mines.edu) and P.K. Sen the arc-flash hazard incident energy levels based on the are with the Colorado School of Mines, Golden, Colorado. current IEEE 1584 standard. Available fault current and John P. Nelson is with NEI Electric Power Engineering, Inc., the clearing time of the protective devices, as known, have in Arvada, Colorado. Ammerman is a Member of the IEEE. the greatest impact on the potential arc-flash hazard. The Sen is a Senior Member of the IEEE. Nelson is a Fellow of simplified versions of the IEEE 1584 incident energy the IEEE. This article first appeared as Arc Flash Hazard equations, proposed in this article, emphasize these rela- Incident Energy Calculations, A Historical Perspective and tionships. Furthermore, it has been demonstrated that Comparative Study: IEEE 1584 and NFPA 70E at the consistent results are obtained using this method. 2007 Petroleum and Chemical Industry Conference.

IEEE INDUSTRY APPLICATIONS MAGAZINE  MAY j JUNE 2009  WWW.IEEE.ORG/IAS