Arc flash calculations are required to determine the incident energy level a worker could be exposed to in calories/cm2. This is outlined by NFPA 70E - 2004, Electrical Safety Requirements for Employee Workplaces, states, “A hazard analysis shall be done in order to protect personnel from the possibility of being injured by an arc flash. The analysis shall determine the Flash Protection Boundary and the personal protective equipment that people within the Flash Protection Boundary shall use.” In addition, IEEE Standard 1584, Guide for Performing An Arc-Flash Hazard Calculations.
Annex D in NFPA 70E offers the same incident energy level equations that appear in IEEE Std. 1584, but the latter goes into more detail. NFPA 70E doesn't specify that the IEEE Std. 1584 method has to be used. The incident energy level could be determined using one of the several arc-flash software programs that are currently available on the market.
As understanding of the arc-flash hazard has grown, several methods for calculating the hazard have been developed. Three of these methods will be examined in this section—the theoretical model, the equations and tables used in NFPA 70E-2004, and the calculation methods presented in IEEE Std 1584™.
In addition to these two methods, there are other methods or tools available for calculation of hazard levels including various Windows and DOS-based shortcut calculator programs, IEEE 1584-based calculators on equipment manufacturer’s web sites, or equipment-specific equations. Even IEEE 1584 presents two alternate calculation methods for many situations—the general equations and the simplified equations for circuit breakers and fuses. For a given system location, one can calculate several different values for incident energy levels or for the hazard boundary distance. While the calculations may be close to one another in many situations, this may not always be the case. How can one be sure which method produces the best results for a given situation? No single calculation method is applicable to all situations, but several principles may be followed to ensure that the best results are obtained in a given situation:
Out with the old, in with the new. A hazard analysis is a relatively new science, and as a result the available methods have changed significantly as understanding of the phenomenon has grown over the past 20 years. Newer test results, industry standards, and calculation methods are more likely to accurately represent the actual hazard levels than older methods. They should be used in preference to older methods that may be based on smaller sets of test data or may be applicable over a smaller range of system conditions.
Use device-specific equations rather than general equations. While the general equations in IEEE 1584 are based on lab testing over a wide range of system conditions, the testing cannot possibly accurately characterize the performance of every available protective device in every possible situation.
First, the engineer must determine which circuit breaker acts to clear the fault. Depending on exactly where in the panel the fault initiates, any of the three devices might initially act to clear the fault. Typically, the worst case scenario will be for the fault to occur on the line-side of the panel’s main circuit breaker, in which case it must be cleared by the upstream feeder device (“A”). This breaker, which would normally be set to selectively coordinate with device “B”, should have the longest tripping time of the three devices shown for a given value of fault current. Even if the arcing fault initiates on the load-side of branch circuit breaker “C”, the fault could easily propagate to the line-side of the other devices in the same enclosure. Therefore, to ensure that the calculations reflect the maximum energy level to which a worker might be exposed, the trip characteristics of device “A” should be considered.
What value of fault current should be considered—the available bolted fault current at the switchboard containing device “A”, or the available fault current at the lighting panel itself? Suppose that 100 kA bolted fault current is available at the switchboard, but the panel is located 100 feet away. The impedance of 100 feet of #3/0 AWG conductor drops the available bolted fault current at the panel to approximately 28 kA. Since the concern in this case is over arcing faults at the lighting panel, this is the value of bolted fault current that should be used as an input to the IEEE 1584 equations. IEEE 1584 is then used to calculate the arcing fault current level, approximately 15 kA. The device’s trip characteristics must be consulted in order to determine its clearing time at 15 kA, and then IEEE 1584 is used to calculate the incident energy level and protection boundary at the panel. In some situations, the best practice may be to calculate two incident energy levels and protection boundaries for a single piece of equipment. For example, consider a lineup of 480 V drawout switchgear with a main circuit breaker and several feeder circuit breakers. The circuit breaker cubicles are more physically separated from one another than circuit breakers are in a typical electrical panel, so propagation of a fault from a feeder to the line-side of the main would be expected to be more difficult. If a fault were to occur when a feeder circuit breaker was racked in or out, then the main circuit breaker would be expected to clear the fault. However, when the main circuit breaker is racked in or out, then the upstream protective device—possibly a fuse or relay on the primary side of an upstream transformer—would be called upon to clear the fault. In this case, the upstream protective device may act relatively slowly, which could mean that workers are exposed to a much higher level of hazard when racking the main than when racking a feeder. In cases such as this, or in other situations when workers may potentially be exposed to hazards in a section of gear on the lineside of the main (i.e., in a fire pump section), more than one calculation per piece of equipment may be warranted.