Conducting an arc flash study
Consulting engineers have an opportunity to provide real-time arc flash diagnostics remotely to their customers.
According to the Institute of Electrical and Electronics Engineers (IEEE), electrical injuries in the workplace─in particular, “arc flash” accidents─result in the death of a facility worker every 48 hours.
In fact, during a seven-year study conducted by the U.S. Dept. of Labor’s Bureau of Statistics, 2,576 U.S. workers died and another 32,807 sustained lost-time injuries─missing an average of 13 days away from work─due to electrical shock or burn injuries. These statistics were validated in a second study involving more than 120,000 employees that determined arc flash injuries accounted for 77% of all recorded electrical injuries.
Arc flash testing
The problem is that an arc flash incident can occur, without warning, anywhere the voltage is above 120 V. Although NFPA 70E Article 130.3 requires that an arc flash study be conducted at least every five years or whenever a major modification occurs, a study is a painstaking, labor-intensive undertaking, and one that may be outdated almost as soon as it is completed.
Such a study may also be expensive, increasing in cost by these factors: size of the facility for data collection, revision of one-line drawings, short-circuit and protective device coordination studies, warning label installation, and arc flash training. For all of their good intentions, companies may be slower to spend money on inspections until the need for them is obvious.
The rise and fall of arcing current
Most of the arc flash studies done today use commercially available software that calculates arcing energy using the algorithm presented in IEEE-1584. The software performs a short-circuit calculation to determine available bolted fault current. Once the arcing current is derived from the bolted fault current, the arcing current is then used to calculate fault-clearing time. Finally, the arcing energy is calculated using IEEE-1584 formulae.
Arcing energy essentially depends on arcing current and fault clearing time. An increase in arcing current will cause higher energy, providing that fault clearing time is constant. An example of that is shown Figure 1.
However, arcing current often falls in the I2t region of solid-state protective devices. Such a case is shown in Figure 2. Here, if arcing current is reduced, fault clearing time will increase, which will cause an increase in arcing energy.
The examples in Figures 1 and 2 illustrate the necessity to evaluate several network configurations. Different configurations (scenarios) will produce different short-circuit levels, which translate to different arcing currents, fault clearing times, and, finally, different arcing energies and arc flash boundaries.
High versus low short-circuit currents
The short-circuit level in the network depends primarily on:
- Network configuration (tie-breaker position, generators are running, etc.)
- Available fault from the utility
- State of the large motors in the network (running or not).
Sometimes (but not very often) an electric power utility will provide information about minimum and maximum short-circuit current. If this information is available, it is wise to analyze at least two scenarios:
- Maximum utility contribution + all motors are running
- Minimum utility contribution + none of the motors are running.
The above scenarios are bound to find the minimum and maximum short-circuit results throughout the network.
If the facility has standby generators, it is expected that the minimum short circuit will be obtained when systems are fed from generators. Therefore, an engineer may want to evaluate the following scenarios:
- Maximum utility contribution + all motors are running + generators are on or off (depending on the way the system is run)
- Utility off + generators are running + none of the motors are running.
However, things get more complex if there are several points of coupling within the utility and the coupling transformers are of different sizes. One such system is depicted in Figure 3. The system is a simplification of many networks the author has encountered in the past.
The system in Figure 3 has 500 and 2000 kVA transformers feeding the separate switchgears. There is also an abundance of local generation. Due to the high impedance of the 500 kVA transformer, the fault level at the BUS-1 is significantly lower than the fault level at the BUS-2 (i.e., 10 versus 35 kA). If the system is fed from generators, the fault level at both buses is between these two extremes (e.g., 20 kA). Therefore, the following conclusions can be drawn:
- BUS-1 has a maximum fault level when fed from the generators.
- BUS-1 has a minimum fault level when fed from the utility.
- BUS-2 has a maximum fault level when fed from the utility.
- BUS-2 has a minimum fault level when fed from the generators.
Consequently, an engineer needs to evaluate at least four scenarios to find the min/max fault levels.
Are maximum and minimum short circuits enough?
The next logical question is if maximum and minimum fault cases are bound to reveal the worst case arc flash. The answer is, unfortunately, no.
Let’s analyze BUS-2 from Figure 3. If one assumes that BUS-2’s protection (from Figure 3) is similar to the one shown in Figure 2, it should be expected that the worst case arc flash scenario will correspond to the minimum fault level at BUS-2. Minimum fault level at BUS-2 will be obtained when the system is fed from generators with no motors running. However, there is a large motor contribution to the fault from Chiller-1. This fault current will not flow through breaker CB-BUS2-1 responsible for clearing the fault; hence, it will not influence the fault clearing time. Therefore, the maximum arcing energy at BUS-2 will be obtained when:
- The system is fed from a generator—minimum fault current to the protective device causes maximum fault clearing time
- Chiller-1 is running—increased fault current at the bus (not branch CB-BUS2-1) will not influence fault clearing time, but will help increase the energy.
This example illustrates that there are special cases where it is not enough to find min/max fault regimes. Additional analysis is needed if one is to find the worst-case scenario.
Real-time arc flash analysis
Many modern mission critical facilities have state-of-the-art data acquisition systems. These systems gather a plethora of information about system configuration: position of switching equipment, state of generators, state of motors, etc. At any given point in time, all of the conditions discussed in previous sections are known. If all of these data are transferred to the facility model, arc flash calculation would give the exact results at the given time. Moreover, if the model is to follow the system in real time (every change in the system is immediately reflected in the model), multiple arc flash calculations would explore all the system configurations and therefore find the worst-case arc flash results.
In fact, within the past 18 months, a new software technology called “power analytics” has emerged that─though created primarily to ensure power reliability and energy efficiency─has also been proven to provide real-time arc flash assessments.
By enabling real-time model-based diagnostics of power systems infrastructure─incorporating the arc flash, power flow, short circuit, protective device coordination, etc., calculations from the original design model ─power analytics enables consultants and facility operators to glean unprecedented visibility into operational detail and previously unpredictable events.
Perhaps the most important aspect of real-time arc flash analysis is the opportunity for the user to understand how different system configurations impact arc flash hazard. For example, the user will realize that when the system is fed from standby generation the energy at a given panel is always above 20 cal/cm2, while if the system is fed from the utility the energy is well below 20 cal/cm2. Hence, he can decide to send the electrician to work on that panel when the system is not on generators.
Prior to entering an energized area and beginning daily tasks, a worker simply queries the system for a real-time arc flash status. It then responds with an up-to-date recommendation on the appropriate safety procedures and personal protective equipment (PPE) necessary to work in the vicinity. Recommendations are based upon IEEE 1584 and the NFPA 70E standards entitled, “IEEE Guide for Performing Arc-Flash Hazard Calculations,” and “Standard for Electrical Safety Requirements for Employee Workplaces,” respectively.
Consulting engineers now have a great opportunity to provide real-time arc flash diagnostics remotely to their customers. Every time a worker goes into an energized area, the need for safety is paramount. Thanks to the sophisticated abilities of today’s analytical software, consultants can not only provide better protection and service to their clients, but create an additional revenue stream as well. Real-time arc flash studies are much faster, automate the process that is painfully manual, and cost less.
While no systems are 100% foolproof in detecting arc flash hazards, the new generation of power analytics technology seems to be the most promising development yet. Why? Because this sophisticated software ensures that workers are forewarned to the greatest extent possible about how to protect themselves from arc flash, an all-too-common and often-surprising threat to workplace health and safety.
- Radibratovic is director of power engineering at EDSA Micro Inc. He has performed numerous power systems studies for mission critical facilities, data centers, manufacturing facilities, and oil refineries. He received both his doctorate and master’s degree from Georgia Institute of Technology in electrical and computer engineering.