Most people have experienced some type of electrical power loss. Some power losses are localized, involving only one—or perhaps several—circuits. Other power losses are spread over a wide area and involve an entire facility or several facilities. Regardless of how robust an electrical system may be, electrical faults happen. It's in the interest of everyone that the fault affects only a limited area of the electrical system, especially where life safety is concerned. That is why NFPA 70: National Electrical Code (NEC) requires selective coordination of protective devices, such as circuit breakers and fuses, in emergency systems.
What is selective coordination?
Although the definition has changed slightly over the years, NEC 2014—the most recent version of the code—defines selective coordination as: "Localization of an overcurrent condition to restrict outages to the circuit or equipment affected, accomplished by the selection and installation of overcurrent protective devices (OCPDs) and their ratings or settings for the full range of available overcurrents, from overload to the maximum available fault current, and for the full range of OCPD opening times associated with those overcurrents."
In other words, the electrical engineer must carefully consider OCPDs so that the protective device closest to the fault opens first and quickly enough to prevent the upstream devices from tripping (see Figure 1). In Figure 1, the fault occurs downstream near the load. Selective coordination requires that only "circuit breaker one" clear the fault, allowing the other breakers in the switchboard identified as "NSG" to continue normal operation. The other circuit breakers in the switchgear would not see the fault; as a result, the rest of the loads would continue to be fed without interruption.
NFPA 110-2016: Standard for Emergency and Standby Systems also requires coordination in Article 6.5: "The OCPDs in the emergency power supply system shall be coordinated to optimize selective tripping of the circuit OCPDs when a short circuit occurs."
It seems very reasonable to have the protective devices selectively coordinated. Nobody likes power outages. The problem is that everything comes at a cost, and selective coordination makes for a more expensive electrical system. Because NEC is primarily interested in the safety of humans, it requires selective coordination only in systems that directly affect life safety. To what extent one needs to coordinate the protective devices depends on the authority having jurisdiction. In some locations, the requirement is to coordinate to 0.01 second and other locations to 0.1 second.
As with everything else, NEC turned attention to selective coordination as a matter of life safety. The coordination requirement was first introduced in 1993 in relation to elevators (Article 620: Elevators, Dumbwaiters, Escalators, Moving Walks, Platform Lifts, and Stairway Chairlifts). In 2005, the requirement was expanded to include emergency systems (Article 700), legally required standby systems (Article 701), and health care facilities (Article 517). The definition of selective coordination was also added in 2005 to better define the requirement. In 2008, Article 708: Critical Operations Power Systems was added to the requirements. Now, the requirement for selective coordination includes Articles 695: Fire Pumps and Article 645: Information Technology Equipment, as well.
When is selective coordination required?
Selective coordination is not required everywhere. It is only required where power continuity is critical to human life or when an interruption of power can cause hazardous conditions.
NEC Article 700.28 states: "Emergency system(s) overcurrent devices shall be selectively coordinated with all supply-side OCPDs. Selective coordination shall be selected by a licensed professional engineer or other qualified persons engaged primarily in the design, installation, or maintenance of electrical systems. The selection shall be documented and made available to those authorized to design, install, inspect, maintain, and operate the system."
It makes for a lively discussion when someone says selective coordination is only required on the emergency-power-source side of the electrical system. The reasoning behind this line of thought is that if the normal power is interrupted, the emergency system will take over, thereby providing the necessary power. The problem is that if the normal power protective device is not coordinated with the downstream protective device, the upstream device could trip first. The automatic-transfer device will ask the emergency system to energize and transfer over when the emergency power is established. In this case, the emergency system will be feeding a fault, which could force the whole system out of service.
NEC Article 701.27 states: "Legally required standby system(s) overcurrent devices shall be selectively coordinated with all supply-side OCPDs." The protective device upstream of the automatic-transfer device is also part of these "supply-side OCPDs." As such, it should be coordinated with the emergency system protective devices downstream of the automatic-transfer device (see Figure 2).
How selective coordination is accomplished
Selective coordination can be done by employing several methods. One such method would be by comparing time-current curves (TCCs) of the given protective devices (see "TCCs explained"). In the past, engineers would construct these curves by hand. Needless to say, this method was time-consuming and not as accurate as using modern software.
Using specific software that provides TCCs enables the engineer to be more precise in his or her approach. One can choose to use software that can be purchased, such as SKM or ETAP, or software provided for free by the manufacturers of protective devices. Choosing one over the other has to do with several factors. Sometimes, clients request that certain software is used because they have confidence in that software and/or they may have already invested in it. Other times, the design engineer will use the software with which he or she is more familiar. Regardless, choosing the free manufacturer's software has one downfall: The only protective devices in the library of that software are the ones made by that manufacturer.
When using generic software, the first thing to do is build the electrical system single-line diagram. This diagram can then be used for many different analyses, such as arc flash or load flow. For the short-circuit analysis to be accurate, the single-line elements must be selected appropriately. The utility-supplied fault duty is an important input because it is directly connected to the short-circuit current availability in the system. If not known, a good number to use as the utility contribution is 500 MVA.
The capacity of the main transformer(s) is the next key piece of information in determining the short-circuit current availability along with the impedance. The impedance of the transformers is usually expressed as a percentage. Emergency generators also play a big role in the fault current availability. In addition to generator capacity, subtransient reactance is important information. The subtransient reactance of the generator plays the same role as the impedance of the transformer and provides resistance to the current flow, thereby lowering the current availability. Cable impedances also play a role, but engineers typically choose to be conservative and use shorter wire lengths than those in the studies. Shorter wires have smaller impedance to the current flow because the impedance is a function of the wire length. After running the short-circuit analysis and establishing the maximum available current, the coordination study can be performed.
By looking at the TCC, engineers can clearly see how long it takes for a protective device to clear the fault for a given amount of fault current. By looking at the curves for different devices, the engineer can easily tell if the protective devices are coordinated and, if so, determine the time difference between pickups.
However, a short-circuit analysis must be completed before starting the coordination study. From the short-circuit analysis, one can determine the momentary (first-cycle) short-circuit current, interrupting duty, 30-cycle short-circuit current, and ground-fault current.
The momentary short-circuit value is the greatest current value available in the system. While performing the study, sometimes it can be noted that two certain protective devices coordinate fully to a certain current value and then their curves converge. It is important to determine if the convergence happens for current values greater than the available current. If that's the case, the devices are considered to be coordinated, because only the current values equal to and less than the available short-circuit current values are of importance. In other words, if two protective devices are selectively coordinated up to 15,000 amps, but not beyond that value, then the fault current availability in the system is considered. If the available fault current is less than 15,000 amps, these two devices are considered to be fully coordinated.
A good method to selectively coordinate the protective devices is using the coordination tables published by the manufacturer. These tables are reliable because the published numbers are derived through tests. The fault current availability is still needed, but if no software is available, it can be easily calculated by hand. The easiest way is to assume infinite availability on the utility side and then use the main transformer's capacity and impedance to calculate the fault current for the system. However, do not forget to sum the contribution of the generator(s) to that of the utility if there is a closed transition from generator power to the normal utility power. When other methods fail, using these protective-device coordination tables is the only reliable way to get the job done.
What selective coordination looks like
To perform a coordination study, it is necessary to have the TCCs of several devices on one plot for comparison. Most of the time, more than one plot is needed to study all the devices in one lineup (in series). In this case, it is important to ensure that the next plot includes one of the devices of the previous plot. In other words, if the study starts at the utility going downstream, the leftmost device on the first plot will be the device on the rightmost position in the second plot.
Referring to Figure 3, a certain current value is chosen. In this worst-case scenario, this is where the curves are closest. Starting at time (t) = 0.01 second, a vertical line is drawn so that it intersects the curves. The dashed line is shown in Figure 3. Because this is a vertical line, there is only a time difference between the curves. The first intersected curve shows the first device that should operate in case of a fault because that device is closest to the fault. The time difference between this device and the one immediately upstream of it can be measured. That time-difference measurement is the coordination time between the two devices. There should be enough time difference between the devices so that the downstream device clears the fault before the upstream device commits to tripping. In the Figure 3 example, there appears to be a separation of 0.12 second, approximately seven cycles. Because the static breaker has an operating time of 0.08 second, that curve's time interval is long enough to protect the fuse, even from a partial melting.
As a general rule, if the devices are coordinated at 0.01 second then there is total coordination, which is hard to achieve. If the devices are coordinated at 0.1 second then there is partial coordination, because the area between 0.01 and 0.1 second is not coordinated.
What to look for when choosing protective devices?
It's easier to coordinate devices that are the same type (i.e., coordinating fuses with each other is easier than coordinating fuses with circuit breakers). But when dealing with devices with curves that appear as bands on the TCC, make sure that these device bands do not intersect. It is best to allow a safety margin so that when using fuses, the upstream fuse doesn't partially melt.
Circuit breakers equipped with an electronic trip unit are easier to coordinate. However, they are more expensive. With cost in mind, breakers that are 400 amps and smaller are usually thermal-magnetic molded-case circuit breakers (MCCBs), where the thermal element provides overload protection and the magnetic trip provides the fault protection. The problem with small MCCBs is that the long delay and instantaneous settings are not adjustable.
When it comes to protective relays, the coordination time interval very much depends on the type of relay. If electromechanical relays are used, the coordination time interval recommended by IEEE 242-2001: IEEE Recommended Practice for Protection and Coordination of Industrial and Commercial Power Systems (IEEE Buff Book) is between 0.3 and 0.4 second. When using static relays (electronic relays, which have no moving parts), the interval can be reduced to between 0.2 and 0.3 second. The reason for this difference in settings is that electromechanical relays are slower than static relays because they operate with an induction disk (see "Coordination resource").
NEC requires selective coordination, but it does not specify the degree of separation between protective devices.
Selective-coordination best practices
Selective coordination is required by code in emergency systems where life safety is a concern. Properly coordinating protective devices requires plenty of thought and tedious work. Sometimes, even if it's not required by code, the client will deem certain processes very critical, so selective coordination becomes a very important part of the electrical system design.
The easiest way to selectively coordinate protective devices is to start all the way downstream, and then, moving upstream, allow for proper separation between the devices in series. However, utility-imposed restrictions force device coordination to be initiated upstream. The problem with starting downstream is that utility companies dictate where the devices at the main substation (whether utility- or client-owned) will be set, and they will have the final word. This means that there is an upper limit imposed; thus, it is prudent to start upstream, at the substation, and work your way down.
Prior to beginning the coordination study, preliminary steps include:
If software is used to perform the selective coordination study, input all the data in the system first and run a fault analysis. Creating TCCs with software is a very simple procedure, but achieving the desired level of coordination can be difficult.
An important aspect to keep in mind is that protective devices in series, such as "breakers two and three" in Figure 1, do not need to be coordinated because tripping of either breaker yields the same result, which is the loss of NSG. However, "breaker five" must be selectively coordinated with "breaker six" and "breaker seven." Breakers in parallel with "breaker one," for example, do not need to be coordinated with breaker 1 because they wouldn't see the same fault as breaker 1. But each of them would need to coordinate with breaker 2. These breakers are not on the emergency branch, but depending on the type of loads supported by the electrical system (mission critical being among them), these breakers would need to be selectively coordinated as well.
It is advisable to work closely with the client and vendors to maximize selective coordination efforts. If no other information is available, use the manufacturer's coordination tables. These tables are reliable and depict data achieved through testing.
Selective coordination of protective devices is desired everywhere, but it can get costly and that is the reason it is only required where the continued operation of electrical systems is a must. Selective coordination assures that in the case of a fault in the electrical system, the protective device closer to the fault trips. The rest of the electrical system will operate normally. Electrical design professionals typically use TCCs to coordinate protective devices, which can be generated by hand or using specific software. It is very helpful to use protective-device coordination tables from the manufacturer when TCCs do not yield a satisfying result. The IEEE Buff Book is a great reference and should be consulted when analyzing electrical systems and trying to coordinate its protective devices.
ABOUT THE AUTHOR
Eduard Pacuku is an electrical project engineer at Jacobs, where he spends the majority of his time designing electrical distribution systems for universities, health care facilities, and data centers.
Protective devices are primarily circuit breakers, fuses, and relays. For medium-voltage systems, protection is typically provided by using protective relays in conjunction with breakers. For low-voltage systems, protective devices could be breakers or fuses. A TCC is an operating curve of a protective device plotted on a logarithmic scale (whether paper or electronically displayed graph) with the "Y" axis representing time and the "X" axis representing current. Usually, the time range is from 0.01 to 1,000 seconds and the current range is chosen to cover the highest current available in the system. Any desired current range can be obtained by multiplying it by a factor of 10 because the TCC is a logarithmic plot. The electrical system designer must ensure that the multiplying factor is shown on the curve so that whoever is reading the TCC has the proper reference. For example, if the current axis shows a value of 100 but the multiplying factor is 102, the current really is 10,000 amps. In addition, devices with different voltage ratings often appear on the same TCC. Therefore, the reference voltage is also very important and must be noted on the TCC as well.
In Figure 3, the TCCs are shown differently for different devices. The curve for the protective relay is a single line while the curves for the fuse and low-voltage breaker are bands. The reason for this difference is that in the case of the fuse and low-voltage breaker, there is a delay caused by tolerances, operating times, ambient temperature, and conditions immediately prior to the overload or fault. The region to the left (and under) the curve shows the area that is accepted as normal operation, and the device does not react. The area to the right and above is the tripping area, where the device recognizes the situation as abnormal and reacts to break the overload or fault. For a given current value, start at time (t) = 0.01 second on the X axis, and move vertically until the first device curve is intersected. The intersection point reveals the amount of time the device will wait to operate for the given current. The farther to the right of the plot, the shorter the tolerance. For sufficiently large currents, that delay becomes very small, approaching zero. Continue vertically until the end line of the band is intersected, then calculate the difference between the two intersections to determine the device tolerance and operating time.
The IEEE 242-2001: IEEE Recommended Practice for Protection and Coordination of Industrial and Commercial Power Systems (IEEE Buff Book) is an excellent resource for anyone needing help in setting protective devices and coordinating them. Table 15-3 in the Buff Book shows the recommended curve time intervals between different protective devices. For example, consider the fuse in the first row under the "Downstream" column and proceed to the right. If a fuse is upstream of the given fuse, there should be clearance between the bands of each fuse curve. The same is true if the upstream device is a low-voltage circuit breaker (row 1, column 3). However, if the upstream device is an electromechanical relay, there must be a 0.22-second separation between the curves (or 13 cycles). That is a large separation that could create trouble as more devices are introduced into the study. Today, most relays are static relays and, as seen in the last column of this table, only 0.12-second separation is required between the fuse curve and the relay curve.