Preventing arc flash in mission critical facilities
To address arc flash problems, we turn to codes and standards. NEC 240.87 is an important leap in arc flash safety for the electrical industry, along with NFPA 70E and IEEE 1584: IEEE Guide for Performing Arc Flash Hazard Calculations.
- Evaluate the key codes and standards that define arc flash safety, including NFPA 70E: Standard for Electrical Safety in the Workplace, IEEE 1584, and NFPA: National Electric Code (NEC) 240.87.
- Explain arc flash calculations and studies, with best practices of mitigating arc flash incidents.
- Demonstrate the importance of arc flash calculations and studies in mission critical facilities.
"Failure is not an option." This quote from Eugene Kranz, a flight director for NASA during the Apollo 13 space mission, defines the concept of mission critical in five words. Yes, the Apollo 13 mission may be an extreme example of mission critical, where every decision and every action made was essential to the survival of the astronauts aboard that unstable spacecraft. However, for those owning, managing, or running a mission critical facility, the operation of their electrical distribution system is essential to the survival of their business or organization. Whether that mission critical system is necessary for the protection of human life, such as an emergency-operation center, or essential for business continuity, such as a data center, they often share the same "failure is not an option" philosophy.
When describing the electrical distribution system for a mission critical facility, two of the key components are availability and reliability. The electrical system must be available when called upon to function (24/7) and it must not fail while in operation. Based on this "must not fail" philosophy, most of the protection systems for mission critical facilities traditionally have been designed to keep the system operating. As such, protection devices are set as high as possible to prevent them from tripping and de-energizing the critical load. This philosophy guards against dropping the critical load, but it does not protect the equipment or more importantly, the personnel working on the equipment from potential hazards such as arc flash. This article looks at the factors that determine the intensity of an arc flash hazard and how arc flash is affected by different protection schemes used in mission critical facilities. The article also investigates methods to mitigate arc flash hazards for mission critical facilities.
NFPA 70E-2015: Standard for Electrical Safety in the Workplace defines an arc flash hazard as "a dangerous condition associated with the possible release of energy caused by an electrical arc." An electrical arc occurs when the electrical current diverges from its intended path and travels through the air from one conductor to another or to ground. Arcing faults generate large amounts of heat that can severely burn human skin and set clothing on fire, making them extremely dangerous and potentially lethal. In addition to the intense heat, an arc fault can also create an explosive blast. During an arc fault, the high temperatures vaporize the electrical conductor, changing it from a solid state to gas vapor and causing it to expand outward with an explosive force. This explosive force can cause destruction to equipment, start fires, and injure employees working on the equipment as well as any surrounding bystanders.
NFPA 70E and IEEE 1584: IEEE Guide for Performing Arc Flash Hazard Calculations are the two standards used by the industry for guidelines and analysis regarding arc flash and arc flash safety. Arc flash hazards are usually expressed as a unit of incident energy (cal/cm2). Incident energy is a measure of thermal energy at a working distance from an arc fault. The three parameters used in an arc flash study to determine the incident energy and the severity of an arc flash injury are arcing current, working distance, and arcing time.
Arcing current: As previously indicated, an electrical arc occurs when the electrical current diverges from its intended path and travels through the air from one conductor to another or to ground. Arcing current is the current released during this fault condition. The magnitude of that arcing fault current is used in the calculation of incident energy. Data centers tend to be consumers of immense electric power; therefore, they have very sizable distribution systems. In the past, and even today, many data centers are being designed with large, multimodule uninterruptible power supply systems and large transformers to reduce the amount of equipment. These larger systems have a substantially higher let-through current that can allow increased amounts of fault current downstream in the distribution system.
Mission critical facilities also tend to have generators for backup power with closed-transition capabilities that allow the load to transfer from generator to utility without interruption (or vice versa). During this closed-transition period, the generator and utility operate in parallel and both contribute to the fault current, which results in increased amounts of fault current. In general, the more fault current available, the higher the arc flash thermal energy and the more dangerous the potential hazard becomes. Because of the larger systems and closed-transition transfers, data centers often have a higher magnitude of available fault current throughout the distribution system, leading to higher arc flash thermal energy. The higher the arc fault thermal energy becomes, the more dangerous the arc flash hazard becomes to those working on the equipment.
Working distance: Arc flash energy is established at the working distance from the arcing fault. The working distance is the distance between the arcing source and the worker’s head and chest. The IEEE 1584a standard defines a working distance for each equipment type and voltage class. NFPA 70E Table D.4.3 provides typical working distances for different equipment and voltages.
The farther away a person can be from a fault, the lower the incident energy exposure. The 24/7 availability and reliability components of a mission critical facility require a comprehensive preventive maintenance program. Often, this preventive maintenance program demands that the equipment be operated or worked on live. For example, using infrared for scanning cable lugs to ensure all connections are tight. Working on equipment live puts the personnel working on the equipment close to the fault; therefore, they are at risk of severe injury.
When working on or near a potential source of an arc flash, personal protective equipment (PPE) must be worn. PPE typically refers to the protective clothing and equipment designed to protect the wearer’s body from injury by reducing the workers’ exposure to incident energy. The rating and type of PPE worn is based on the calculated exposure to incident energy. NFPA 70E Table H.3(b) provides guidance on the selection of arc-rated clothing and other PPE based on the incident-energy exposure.
Another issue with data centers is that electrical equipment, such as power distribution units and remote power panels, is often located in a data center’s white space, which is occupied by unqualified electrical personnel. In addition to those working on the equipment, the explosive force caused by an arc flash can injure bystanders in close proximity to the arc flash. For this reason, an arc flash boundary is calculated in addition to the working distance. This boundary defines the distance from exposed live parts within which an unprotected person could receive second-degree burns. During equipment servicing, those without PPE should not be permitted within this boundary. Locating electrical equipment inside the data center, rather than in dedicated galleries adjacent to the data center, puts the information technology equipment at risk of damage and those employees working in the data center at risk of injury.
Arcing time: Arcing time is the time duration of the arc flash. While arcing time is not as obvious a characteristic of arc flash as the first two parameters, it is arguably the most important—especially in data centers. The duration of the arc flash is determined by the time required for the upstream protection device to trip and clear the fault condition. The equations in IEEE 1584-2002 indicate a linear relationship between arc flash incident energy and time, meaning the longer the duration of the arc flash, the higher the incident energy produced. Therefore, the longer the arc is allowed to propagate, the more intense the heat of the arc flash becomes and the longer the person working on the equipment is exposed to that intense heat. At the same time, the faster the fault is cleared and the arc extinguished, the less time the high temperatures have to vaporize the conductor, greatly reducing the chances of explosion.
Something worth noting is the difference between arc fault current and bolted fault current. Bolted fault currents (two conductors being bolted together with little or no impedance) and short circuit values normally are calculated at maximum. Bolted fault currents are used for sizing the interrupting rating of equipment and setting the protection devices. Arc fault currents are calculated based on the assumption that there is a small gap between the conductors that is bridged by something causing the arc. Because of the impedance caused by the gap and the bridging component, the arc fault current is usually lower than the bolted fault current. Lower current means a longer time period before the protection device clears the fault.
The failure-is-not-an-option mentality of a typical mission critical facility and/or data center has pushed the configuration and coordination of the protection system toward keeping the system operating. In these situations, protection devices are set as high as possible to prevent them from tripping and de-energizing the critical load. The instantaneous settings on the switchgear main breakers are turned off to eliminate the possibility of a downstream fault taking out the entire switchgear. High instantaneous devices or devices with built-in time delays are used in an attempt to avoid nuisance tripping. Although this philosophy keeps the electrical system operating as long as possible, it also extends the duration of the arc flash by delaying the time to clear the fault. This delay in clearing the fault exposes the person working on the equipment to more hazards and increases the potential severity of the injury.
A protection device’s tripping characteristics are defined by its time-current curves, where the vertical axis represents clearing time in seconds and the horizontal axis represents the arcing short circuit current. The correlation between the instantaneous trip region and the arcing fault current is important in determining the arc flash duration. The instantaneous region is the band at the bottom of the curve where a fault of that magnitude will trip immediately with no delay (see Figure 1). An arcing fault happening in this region will cause the protection device to clear the arc flash in the shortest possible duration, thus reducing the severity of the hazard. If the arc fault happens with a magnitude to the left of or less than the instantaneous region, the protection device will operate in its time-delay region, extending the duration of the arc flash. Depending on the protection device used, this could result in a clearing time of 2 seconds. Either setting the instantaneous trip value of the protection device too large for the available arc fault current, or turning off the instantaneous trip setting, as is often done with data centers, forces that device to operate in the time-delay region, increasing the potential severity of the hazard.
Safety awareness and the overall safety of employees have become top priorities in the engineering industry. Ideally, the safest and most widely recommended practice is to work on de-energized equipment only. However, in a mission critical facility, such as a data center, this is not always an option. As a result, more owners are commissioning arc flash studies to understand the hazards associated with arc faults and the PPE required to safely work on live electrical equipment.
For many owners of older data centers, arc flash study results are indicating that they have high levels of incident energy and thermal energy within the distribution system. This high level of energy is requiring personnel to wear 40 cal/cm2 PPE protection (safety glasses, hearing protection, leather footwear, hard hat, arc-rated gloves, arc-rated flash suit, arc-rated flash suit hood). Depending on the functions being performed, this level of PPE protection could become cumbersome, causing an additional level of risk to the critical load. In other cases, the energy level is so high that there is no PPE available to provide protection; in which case, the equipment can’t be worked on live.
Due to the critical nature of the facility, the majority of electrical distribution equipment will require maintenance and adjustments while energized. Therefore, something must be done to reduce the incident energy level to within the range of a hazardous risk category so that it can be worked on with the proper protective equipment and clothing. In the 2014 version of NFPA 70: National Electrical Code (NEC), Article 240.87 now requires arc flash reduction on all overcurrent devices 1,200 amps or higher. The process of taking steps to minimize the level and severity of the arc flash hazard, as well as the risk associated with the probability that an arc flash event will occur, is known as arc flash mitigation. There are several arc flash mitigation solutions that can be implemented in mission critical facilities to reduce the arc flash incident energy and the potential hazard.
Protection device settings
Whether it is a new design or existing facility, the preferred method and least costly solution for mitigation is a re-evaluation of the coordination study. Incident energy is the result of short circuit current and clearing time under arcing fault conditions. Therefore, small changes in arcing fault current and trip settings can significantly affect the amount of incident energy. The goal of the coordination study is to examine the electrical distribution system and set the protection devices so that only the device closest to the fault opens, which isolates the fault from the rest of the system. This allows the rest of the system to remain in operation. Arc flash levels must be kept in mind when performing a coordination study so that there is a balance between arc flash reduction and breaker selectivity. The primary goal of this re-evaluation is to ascertain if the time-dial settings can be lowered enough to reduce the calculated incident energy without affecting the integrity of the critical distribution system.
Another method of mitigation is to select or replace older equipment, such as breakers and relays, with units that have quicker clearing times. For example, this could mean using solid-state trip units and replacing thermal-magnetic breakers with devices that have electronic trip units, which provide more adjustability and allow quicker operation and better coordination between devices. Breakers that have adjustable instantaneous settings and no built-in time delays should be used where possible so that they don’t extend the duration of the arc flash. Digital relays that have an instantaneous element—unlike most electromechanical overcurrent relays—can also be used on the medium-voltage systems to allow for a quicker clearing time.
Zone-selective interlocking eliminates intentional delay without sacrificing coordination. In a well-coordinated system, longer delays and higher pickups are selected on upstream devices to allow downstream devices to pick up first. This delay extends the arc flash and exposes the system to higher levels of incident energy. Zone-selective interlocking allows the electronic trip devices to communicate with each other so that a fault will be isolated and cleared by the nearest upstream device with no intentional time delay. For example, if a fault occurred at the switchgear, the main breaker would trip instantly instead of going through its normal time-delay progression. It should be noted that equipment manufacturers recommend only using zone-selective interlocking as a means for arc flash mitigation on switchgear constructed according to UL 1558: Standard for Metal-Enclosed Low-Voltage Power Circuit Breaker Switchgear, with full barriers between cubicle sections.
Another form of zone protection is current differential protection, most commonly used with medium-voltage systems. This uses current transformers to measure and compare the incoming and outgoing currents. Because this protection system is independent and associated only with a particular zone, it is not required to be time-coordinated with other systems, allowing for tripping without additional delay.
Optical light sensing
Similar to zone interlocking, the concept behind optical light sensing is to isolate and clear the fault by tripping the nearest upstream device with no intentional time delay. This system uses a combination of light sensors, which detect the flash of light associated with the arc fault, and relays that detect the high fault current. When both conditions are present, the relay will quickly clear the fault by opening the upstream device without any delay. For protection, both conditions must be present. A burst of light or high fault current, such as inrush alone, will not activate the system.
Arc resistant equipment
Arc resistant equipment is designed to provide protection to those workers surrounding the equipment from internal arcing faults under normal operating conditions. Arc resistant equipment normally provides the required protection by venting the hot gasses and explosive material away from the work area surrounding the equipment. The overall enclosure and doors have an increased level of strength to force those gasses to pass through the vents that are the paths of least resistance, to prevent doors and covers from blowing off during the event. Normal operating conditions are often defined as opening and closing breakers and switches and inserting and removing withdrawable components (see Figure 2). This type of mitigation does not protect the equipment (it still gets damaged) and only works with the doors closed.
The maintenance-mode switch is normally an external switch wired to the circuit breaker that allows the operator maintaining that piece of equipment to modify the trip settings of the device to a lower setting. The lower and faster setting is intended to reduce the incident energy levels downstream of the device. Because this switch sacrifices coordination, this switch is often alarmed so that it is not left in maintenance mode after the work is complete.
One very effective mitigation technique is to increase the distance between the worker and the fault. Using remote operators, such as hard-wired control switches, programmable logic controller-based human-machine interface screens, and supervisory control and data acquisition systems allows the worker to be completely outside the room and outside the arc flash boundary when breakers and switches are operated. Portable devices are also available to perform the same breaker operation on existing breakers and remote racking for withdrawable components.
Infrared viewing windows
Infrared viewing windows are installed in equipment to allow the operator to scan cable connections and other key components of live equipment with infrared thermography devices without having to remove the equipment covers that expose them to hazardous energy (see Figures 3 and 4). Although this option may not be directly considered an arc flash mitigation technique, it does make the maintenance process safer.
As previously mentioned, availability and reliability are two key components used in the description of mission critical data centers. In the past, the failure-is-not-an-option mentality drove the design of the data center and the configuration of its protection system. As the commitment to safety and employee well-being becomes more prevalent, arc flash mitigation is likewise attracting an increasing focus. As such, it is important that arc flash safety is taken into consideration during the design of the distribution system.
It is recommended that a preliminary arc flash study be performed during the design process to identify potential problem areas. The primary goal is to provide the safest environment possible without affecting the integrity or the reliability of the critical distribution system. Ultimately, the safest possible practice is to only work on de-energized equipment. Therefore, whenever possible, there should be a means within the electrical distribution system to allow for the critical load to be transferred to a bypass source, so that portions of the system can be taken offline for maintenance.
Kenneth Kutsmeda, PE, LEED AP, is an engineering design principal at Jacobs in Philadelphia. For 20 years, he has been responsible for engineering, designing, and commissioning power distribution systems for mission critical facilities. His project experience includes data centers, specialized research and development buildings, and large-scale technology facilities containing medium-voltage distribution. He is a member of the Consulting-Specifying Engineer editorial advisory board and a 2010 40 Under 40 winner.
Case study: Reducing energy levels
An arc flash study was performed on a 20-year-old, 14-MW data center with power distribution units and remote power panels (RPPs) located within the white space. The arc flash study determined that the RPP in the white space had 17 cal/cm2 of incident energy, which required Level 3 personal protective equipment (PPE) and a 91-in. arc flash boundary (see Figure 5). It was determined that the thermal-magnetic molded-case circuit breakers used for the main breaker in the RPP have had a fixed, long time delay and a fixed, high instantaneous trip setting. A high instantaneous trip setting was commonly used in data centers to avoid nuisance tripping caused by inrush current.
Due to the safety of the nonelectrical personnel working in the white space and the cost of the equipment in the data center, the owner requested an investigation into potential mitigation techniques. One technique presented was to replace the main protection device in the RPP with a device that had a lower instantaneous setting. Replacing the protection device with another thermal-magnetic molded-case breaker with a slightly lower instantaneous setting reduced the incident energy to 0.17 cal/cm2 of incident energy, which requires only a Level 0 PPE (see Figure 6). The reduced setting on the RPP main circuit breaker also coordinated well with the downstream devices, so it had no negative impact on the integrity of the system.