Preventing arc flash incidents by design
Understanding electrical distribution equipment and its associated risks can help reduce incident energy levels and save lives.
Anyone involved with electrical distribution systems—either as a design engineer, commissioning agent, or contractor—for more than five years probably has been directly involved in an arc flash incident or has heard of one.
According to NFPA 70E: Standard for Electrical Safety in the Workplace, arc flash is a “dangerous condition associated with the release of energy caused by an electrical arc.” It is measured in terms of arc flash incident energy, which is used to determine the appropriate level of personal protective equipment (PPE), and in terms of an arc flash protection boundary.
An arc flash is the result of an electric current passing through air as the result of conductor failure, equipment failure, or the accidental connection between voltage sources such as dropping a tool across buses in distribution equipment. The flash is immediate but the resultant release of energy can cause severe injury, and possibly death. There is a potential for a tremendous amount of heat to be released, which can result in overpressures, as well as flying debris. The energy released can cause temperatures exceeding 35,000 F, which can vaporize steel, copper, and aluminum. Inhaling these vaporized metals could be fatal. Injuries or fatalities could occur if personnel are in the area in front of an arc flash, which could send projectiles such as parts of metal buses away from the blast. Also, molten metal can cause significant burns, and the sudden air pressure increase can knock personnel off their feet.
Each year, more than 2,000 people are treated in burn centers for injuries from arc flash incidents. Many injuries caused by arc flash incidents can be prevented. Not working on or around energized equipment may the simplest way to avoid injury. Scheduling maintenance outages may seem like a bother, but will easily offset the loss of production, unscheduled outages, and equipment damage that may occur with an arc flash incident.
Arc flash hazard analysis
Sometimes working on live electrical equipment may be necessary. Appropriate PPE is required when working on or around energized electrical equipment. An arc flash hazard analysis is required by NFPA 70E. This determines the arc flash boundary, the incident energy at the working distance, and the level of PPE that must be used within the arc flash boundary.
Procedures for performing an arc flash hazard analysis can be found in IEEE 1584: Guide for Performing Arc Flash Calculations. Some of the factors that determine the amount of incident energy include:
· The available fault current at the circuit: This is the amount of current that could flow into the circuit in the event of a fault. This is calculated in the short-circuit analysis. Factors that determine fault current are the available fault current of the power source (typically available from the local power utility), the impedance of the transformers that supply the circuit, length and type of conductors in the circuit, and motor contribution. At first, it may seem counterintuitive, but higher fault currents may actually reduce the flash hazard because they will decrease the overcurrent device clearing time, which reduces the flash hazard.
· The operating characteristics of the overcurrent protective devices in the circuit: These vary with the type of device used. These characteristics are determined by simple fixed settings on thermal magnetic breakers, fuse melting curves, and multiple pickup settings on relays and solid-state breakers. Settings for adjustable devices are determined in a coordination study.
· Equipment labeling requirements: For distribution equipment that has a main overcurrent protective device, two labels may be required: one label for the energy level at the line side of the device and another on the load side of the device. Energy levels at the line side and load side may be significantly different. This differential should be identified to provide maintenance personnel with information regarding potential arc flash hazards. Even with the main breaker opened, the line side of the main is still energized.
After the arc flash hazard analysis is completed, warning labels are printed and affixed to the electrical equipment. The labels should include the level of PPE required, the flash hazard boundary, the flash hazard, the shock hazard, and approach distances.
After the study is completed and the labels are installed, work on this equipment or routine maintenance will likely be required at some point. Work that may be required could be thermal scans to check for equipment hot spots, racking out a breaker for routine maintenance, or installing a new circuit breaker to serve new loads.
Regardless of the work to be done, personnel must follow the appropriate safety procedures. Observe the label to determine the proper PPE level, gear up, and carefully proceed to perform the necessary work. PPE may be a simple as safety glasses, gloves, and untreated cotton—or it could include a full face shield and protective suit. For minor or simple maintenance tasks, the temptation may be to proceed without proper PPE to save a few minutes. But even with the most careful work, accidents can happen, and the potential for serious, life-threatening injuries still exists. Therefore, it is critical that personnel working on electrical distribution equipment be trained in proper procedures, and that they wear the appropriate PPE.
Reducing incident energy levels at a location where electrical work is to be performed reduces the level of PPE required when working on energized circuits at that location. However, energy incident level reduction does not eliminate PPE requirements.
Figure 1 shows a time-current curve (TCC) for a 1,600 A, 480 V, solid-state trip-unit circuit breaker with adjustable long time pick-up, long time delay (LTD), short time pick-up (STPU), short time delay (STD), and instantaneous settings. These settings allow the breaker’s operating characteristics to be adjusted. The settings in Figure 1 were selected to achieve selective coordination with upstream and downstream overcurrent protective devices to isolate the potential fault as close to the fault as possible. With these settings, the 85% arcing fault level of 3,009 A will last approximately 19 sec, and the 100% arcing fault level of 3,540 A will last approximately 13 sec. Using the maximum arcing exposure time of 2 sec as recommended in IEEE 1584 results in an arc flash hazard of 22 Cal/cm2 and requires Category 3 PPE as indicated on the warning label (see inset). Energy levels above 1.2 Cal/cm2 can cause a temperature rise that will result in second-degree burns on exposed human skin.
When energized maintenance is required for this equipment or downstream equipment, the energy level and required PPE may be reduced by setting the LTD to 0.5, the STPU to 1.5, and the STD to instantaneous, with the resulting TCC shown in Figure 2 and its corresponding warning label shown in the inset These settings reduce the arcing fault durations (the 85% and 100% fault levels) to 0.02 sec. This means the breaker will clear the fault more quickly. The flash boundary is reduced from 105 in. to 15 in., and the flash hazard has been reduced from 22 Cal/cm2 to 0.86 Cal/cm2. Note that some breaker designs may not allow adjusting these points while energized.
Reducing incident energy
After the maintenance has been performed, the original settings can be restored. Please note that adjusting the trip characteristics of an energized breaker may also cause a nuisance trip to occur. Before performing this work, the facility staff should be made aware of this possibility and informed that a temporary power outage could ensue. If this happens, the breaker could be immediately reset so the outage would be no longer than a few sec, but for critical facilities, such as hospitals and data centers, even this temporary outage could be detrimental to the facility operation.
Many manufacturers offer maintenance switches that can decrease the time an arcing fault is allowed to exist, thereby reducing the incident energy exposure. This function is usually enabled with a keyed switch, often located out of the hazard boundary. Some manufacturers offer an arc fault detection circuit, which typically uses a photoelectric sensor to differentiate between an overload, a fault, and an arc flash. Others offer zone-selective interlocking between the levels of the overcurrent protective devices. Using this feature can reduce clearing time if a fault occurs, thereby reducing the incident energy and the level of required PPE.
Another method to reduce exposure during device operation is with a remote operating station located outside the flash hazard boundary area. The station allows personnel to operate the device without having to wear PPE. Remote operating stations typically provide positive feedback—usually an indicating lamp—that verifies the open or closed status of the device.
For many critical facilities such as data centers, dual power paths to the equipment are provided. This allows the electrical distribution equipment to be de-energized while maintaining operation. Dual paths require additional distribution equipment—almost a mirror image of all the distribution components. This increases initial installation costs, but when properly designed and installed, complete isolation of any component in the electrical distribution system is possible. Dual paths give personnel the ability to perform maintenance and testing on de-energized equipment.
Understanding arc flash and its potential hazards, calculating risk mitigation, knowing the importance of labeling, and the proper use of PPE can maintain the effective use of electrical distribution equipment through equipment maintenance and upgrades—and ultimately save lives.
Learning from experience
Being an electrical engineer for more than 25 years, I’m aware of at least 15 arc flash incidents and have been directly involved in three. The following is an account of my experience with one of those incidents.
Our design-build team was adding a new distribution section onto an existing service entrance switchboard that supports a critical infrastructure distribution system. Installing the new section required the switchboard to be de-energized. Shutdown arrangements were made, the method-of-procedure (MOP) was in place, and the new section was on-site and ready to be connected. Work was to begin at 1 a.m. on a Sunday and to be completed by 6 a.m.—more than enough time (we thought) to accomplish our task.
As 1 a.m. approached, we had portable generators running, flashlights in hand, all the equipment was powered down, and we were ready to begin. The first step in the MOP was to trip the main disconnect using the ground fault test relay. We pressed the trip button; nothing happened. Because the equipment was more than 20 years old and probably had not been tested in some time, the consensus was that the ground fault relay probably had a blown fuse and could be repaired after the main was opened.
Because there was a limited amount of time to perform the work, the decision was made to manually open the main. The main was a bolted pressure-fused disconnect, so the electrician tripped the “open” lever. Nothing happened. He recharged the trip spring, tried to open it again, but the switch did not open, but we did detect some movement. He tried again, but it still did not open. Looking back, we should have stopped here, delayed the project, and called in a service technician for the equipment. But, there was a lot of pressure to finish the work and delaying the project was not an option.
A few access panels were removed to allow switch inspection. The trip spring was recharged and we tried to open the switch again, but still with no success. The operating mechanism did move a small amount, so the decision was made to keep trying. As we kept trying, the operating mechanism moved a little each time, so we thought the switch would open with just a few more tries. We left the panels off for easy inspection. We tried a few more times and there was a large flash, with lots of smoke and noise. Luckily, no one was injured, but we were a bit shaken.
Upon further examination, we determined that because the switch had not been operated in several years, the lubricant on the operating linkage had deteriorated. Cycling the switch, along with the deterioration of the lubricant, had resulted in a mechanical failure of the linkage, and a piece of the linkage had fallen across the load side bus causing the arc flash. Although no one was injured, the flash damaged the main switchboard, which necessitated using a portable generator while the switchboard was being repaired.
Young is senior associate and the electrical department manager at Bala Consulting Engineers Inc. in King of Prussia, Pa. His expertise is in critical power distribution systems and on-site power generation systems. He is a member of NFPA and the 7x24 Exchange, and a U.S. Air Force veteran.