Chapter 30

Distribution Short-Circuit Protection

30.1 Basics of Distribution Protection 30-2

Reach • Inrush and Cold-Load Pickup

30.2 Protection Equipment 30-5

Circuit Interrupters • Circuit Breakers • Circuit Breaker Relays • Reclosers • Expulsion Fuses • Current-Limiting Fuses

30.3 Transformer Fusing 30-19

30.4 Lateral Tap Fusing and Fuse Coordination 30-23

30.5 Station Relay and Recloser Settings 30-24

30.6 Arc Flash 30-25

30.7 Coordinating Devices 30-29

Expulsion Fuse–Expulsion Fuse Coordination • Current-Limiting Fuse Coordination • Recloser–Expulsion Fuse Coordination • Recloser–Recloser Coordination • Coordinating Instantaneous Elements

30.8 Fuse Saving versus Fuse Blowing 30-34

Industry Usage • Effects on Momentary and Sustained Interruptions • Coordination Limits of Fuse Saving • Long-Duration Faults and Damage with Fuse Blowing • Long-Duration Voltage Sags with Fuse Blowing • Optimal Implementation of Fuse Saving • Optimal Implementation of Fuse Blowing

30.9 Other Protection Schemes 30-40

Time Delay on the Instantaneous Element (Fuse Blowing) • High–Low Combination Scheme • SCADA Control of the Protection Scheme • Adaptive Control by Phases

30.10 Reclosing Practices 30-42

Reclose Attempts and Dead Times • Immediate Reclose

30.11 Single-Phase Protective Devices 30-47

Single-Phase Reclosers with Three-Phase Lockout

References 30-49

Tom A. Short

Electric Power Research Institute

Overcurrent protection or short-circuit protection is very important on any electrical power system, and the distribution system is no exception. Circuit breakers and reclosers, expulsion fuses, and current-limiting fuses—these protective devices interrupt fault current, a vital function. Short-circuit protection is the selection of equipment, placement of equipment, selection of settings, and coordination of devices to efficiently isolate and clear faults with as little impact on customers as possible.

Of top priority, good fault protection clears faults quickly to prevent

• Fires and explosions
• Further damage to utility equipment such as transformers and cables

Secondary goals of protection include practices that help reduce the impact of faults on the following:

• Reliability (long-duration interruptions): In order to reduce the impact on customers, reclosing of circuit breakers and reclosers automatically restores service to customers. Having more protective devices that are properly coordinated assures that the fewest customers possible are interrupted and makes fault-finding easier.
• Power quality (voltage sags and momentary interruptions): Faster tripping reduces the duration of voltage sags. Coordination practices and reclosing practices impact the number and severity of momentary interruptions.

30.1 Basics of Distribution Protection

Circuit interrupters should only operate for faults, not for inrush, not for cold-load pickup, and not for transients. Additionally, protective devices should coordinate to interrupt as few customers as possible.

The philosophies of distribution protection differ from transmission-system protection and industrial protection. In distribution systems, protection is not normally designed to have backup. If a protective device fails to operate, the fault may burn and burn (until an upstream device is manually opened). Of course, protection coverage should overlap, so that if a protective device fails due to an internal short circuit (which is different from failing to open), an upstream device operates for the internal fault in the downstream protector. Backup is not a mandatory design constraint (and is impractical to achieve in all cases).

Most often, we base distribution protection on standardized settings, standardized equipment, and standardized procedures. Standardization makes operating a distribution company easier if designs are consistent. The engineering effort to do a coordination study on every circuit reduces considerably.

Another characteristic of distribution protection is that it is not always possible to fully coordinate all devices. Take fuses. With high fault currents, it is impossible to coordinate two fuses in series because the high current can melt and open both fuses in approximately the same time. Therefore, close to the substation, fuse coordination is nonexistent. There are several other situations where coordination is not possible. Some low-level faults are very difficult—some would say impossible—to detect. A conductor in contact with the ground may draw very little current. The “high-impedance” fault of most concern (because of danger to the public) is an energized, downed wire.

30.1.1 Reach

A protective device must clear all faults in its protective zone. This “zone” is defined by the following:

• Reach: The reach of a protective device is the maximum distance from a protective device to a fault for which the protective device will operate.

Lowering a relay pickup setting or using a smaller fuse increases the reach of the protective device (increasing the device’s sensitivity). Sensitivity has limits; if the setting or size is too small, the device trips unnecessarily from overloads, from inrush, from cold-load pickup.

We have several generic or specific methods to determine the reach of a protective device. Commonly, we estimate the minimum fault current for faults along the line and choose the reach of the device as the point where the minimum fault current equals the magnitude where the device will operate. Some common methods of calculating the reach are as follows:

• Percentage of a bolted line-to-ground fault: The minimum ground fault current is some percentage (usually 25%–75%) of a bolted fault.
• Fault resistance: Assume a maximum value of fault resistance when calculating the current for a single line-to-ground fault. Common values of fault resistance used are 1–2, 20, 30, and 40 Ω. Rural Electrification Administration (REA) standards use 40 Ω.

Other options for determining the reach are as follows:

• Point based on a maximum operating time of a device: Define the reach as the point giving the current necessary to operate a protective device in a given time (with or without assuming any fault impedance). Example: The REA has recommended taking the reach of a fuse as the point where the fuse will just melt for a single line-to-ground fault in 20 s with a fault resistance of 30 or 40 Ω.
• Point based on a multiplier of the device setting: Choose the point where the fault current is some multiple of the device rating or setting. Example: The reach of a fuse is the point where the bolted fault current is six times the fuse rating.

None of these methods are exact. Some faults will always remain undetectable (high-impedance faults). The trick is to try to clear all high-current faults without being overly conservative.

Assuming a high value of fault resistance (20–40 Ω) is overly conservative, so avoid it. For a 12.47 kV system (7.2 kV line to ground), the fault current with a 40 Ω fault impedance is less than 180 A (this ignores the system impedance—additional system impedance reduces the calculated current even more). Using typical relay/recloser setting philosophies, which say that the rating of the recloser must be less than half of the minimum fault current, a recloser must be less than 90 A, which effectively limits the load current to an unreasonably low value. In many (even most, for some utilities) situations this is unworkable. Faults with arc impedances greater than 2 Ω are not common (see Short, 2004), so take the approach that the minimum fault is close to the bolted fault current. On the other hand, high-impedance faults (common during downed conductors) generally draw less than 50 A and have impedances of over 100 Ω. The 40 Ω rule does not guarantee that a protective device will clear high-impedance faults, and in most cases would not improve high-impedance fault detection.

30.1.2 Inrush and Cold-Load Pickup

When an electrical distribution system energizes, components draw a high, short-lived inrush; the largest component magnetizes the magnetic material in distribution transformers (in most cases, it is more accurate to say remagnetizes since the core is likely magnetized in a different polarity if the circuit is energized following a short-duration interruption). The transformer inrush characteristics important for protection are as follows:

• At a distribution transformer, inrush can reach peak magnitudes of 30 times the transformer’s full-load rating.
• Relative to the transformer rating, inrush has higher peak magnitudes for smaller transformers, but the time constant is longer for larger transformers. Of course, on an absolute basis (amperes), a larger transformer draws more inrush.
• Sometimes inrush occurs, and sometimes it does not, depending on the point on the voltage waveform at which the reclosing occurs.
• System impedance limits the peak inrush.

The system impedance relative to the transformer size is an important concept since it limits the peak inrush for larger transformers and larger numbers of transformers. If one distribution transformer is energized by itself, the transformer is small relative to the source impedance, so the peak inrush maximizes. If a tap with several transformers is energized, the equivalent connected transformer is larger relative to the system impedance, so the peak inrush is lower (but the duration is extended). Several transformers energizing at once pull the system voltage down. This reduction in voltage causes less inrush current to be drawn from each transformer. For a whole feeder, the equivalent transformer is even larger, so less inrush is observed. Some guidelines for estimating inrush are as follows:

• One distribution transformer: 30 times the crest of the full-load current.
• One lateral tap: 12 times the crest of the full-load current of the total connected kVA.
• Feeder: 5 times the connected kVA up to about half of the crest of the system available fault current.

At the feeder level, inrush was only reported to cause tripping by 15% of responders to an IEEE survey (IEEE Working Group on Distribution Protection, 1995). When a three-phase circuit is reclosed, the ground relay is most likely to operate since the inrush seen by a ground relay can be as high as the peak inrush on the phases (and it is usually set lower than the phase settings). An instantaneous relay element is most sensitive to inrush, but the instantaneous element is almost always disabled for the first reclose attempt. The ground instantaneous element could operate if a significant single-phase lateral is reconnected.

Transformers are not the only elements that draw inrush; others include resistive lighting and heating elements and motors. Incandescent filaments can draw eight times normal load current. The time constant for the incandescent filaments is usually very short; the inrush is usually finished after a half cycle. Motor starting peak currents are on the order of six times the motor rating. The duration is longer than transformer inrush with durations typically from 3 to 10 s.

Cold-load pickup is the extra load following an extended interruption due to loss of the normal diversity between customers. Following an interruption, the water in water heaters cools down and refrigerators warm up. When the power is restored, all appliances that need to catch up energize at once. In cold weather, following an extended interruption, heaters all come on at once (so it is especially bad with high concentrations of resistance heating). In hot weather, houses warm up, so all air conditioners start following an interruption.

Cold-load pickup can be over three times the load prior to the interruption. As diversity is regained, the load slowly drops back to normal. This time constant varies depending on the types of loads and the duration of the interruption. Cold-load pickup is often divided into transformer inrush which last a few cycles, motor starting and accelerating currents which last a few seconds, and finally just the load due to loss of diversity which can last many minutes. Figure 30.1 shows the middle-range time frame with motor starting and accelerating currents.

It is important to select relay settings and fuse sizes high enough to avoid operations due to cold-load pickup. Even so, cold-load pickup problems are hard to avoid in some situations. A survey of utilities reported 75% having experienced cold-load pickup problems (IEEE Working Group on Distribution Protection, 1995). When a cold-load pickup problem occurs at the substation level, the most common way to reconnect is to sectionalize and pick the load up in smaller pieces. For this reason, cold-load pickup problems are not widespread—after a long interruption, utilities usually sectionalize anyway to get customers back on more quickly. Two other ways that are sometimes used to energize a circuit are to raise relay settings or even to block tripping (not recommended unless as a very last resort).

In order to pick relays, recloser settings, and fuses, we often plot a cold-load pickup curve on a time–current coordination graph along with the protection equipment characteristics. Points can be taken from the curves in Figure 30.1. It is also common to choose one or two points to represent cold-load pickup. Three-hundred percent of full-load current at 5 s is a common point.

Distribution protective devices tend to have steep time–overcurrent characteristics, meaning that they operate much faster for higher currents. K-link fuses and extremely inverse relays are most commonly used, and these happen to have the steepest characteristics. This is no coincidence; a distribution protective device must operate fast for high currents (most faults) and slow for lower currents. This characteristic gives a protective device a better chance to ride through inrush and cold-load pickup.

30.2 Protection Equipment

30.2.1 Circuit Interrupters

All circuit interrupters—including circuit breakers, fuses, and reclosers—operate on some basic principles. All devices interrupt fault current during a zero crossing. To do this, the interrupter creates an arc. In a fuse, an arc is created when the fuse element melts, and in a circuit breaker or recloser, an arc is created when the contacts mechanically separate. An arc conducts by ionizing gasses, which leads to a relatively low-impedance path.

After the arc is created, the trick is to increase the dielectric strength across the arc so that the arc clears at a current zero. Each half cycle, the ac current momentarily stops as the current is reversing directions. During this period when the current is reversing, the arc is not conducting and is starting to de-ionize, and in a sense, the circuit is interrupted (at least temporarily). Just after the arc is interrupted, the voltage across the now-interrupted arc path builds up. This is the recovery voltage. If the dielectric strength builds up faster than the recovery voltage, then the circuit stays interrupted. If the recovery voltage builds up faster than the dielectric strength, the arc breaks down again. Several methods used to increase the dielectric strength of the arc are discussed in the following paragraphs. The general methods are as follows:

• Cooling the arc: The ionization rate decreases with lower temperature.
• Pressurizing the arc: Dielectric strength increases with pressure.
• Stretching the arc: The ionized-particle density is reduced by stretching the arc stream.
• Introducing fresh air: Introducing de-ionized gas into the arc stream helps the dielectric strength to recover.

An air blast breaker blasts the arc stream into chutes that quickly lengthen and cool the arc. Blowout coils can move the arc by magnetically inducing motion. Compressed air blasts can blow the arc away from the contacts.

The arc in the interrupter has enough resistance to make it very hot. This can wear contact terminals, which have to be replaced after a given number of operations. If the interrupter fails to clear with the contacts open, the heat from the arc builds high pressure that can breach the enclosure, possibly in an explosive manner.

In an oil device, the heat of an arc decomposes the oil and creates gasses that are then ionized. This process takes heat and energy out of the arc. To enhance the chances of arc extinction in oil, fresh oil can be forced across the path of the arc. Lengthening the arc also helps improve the dielectric recovery. In an oil circuit breaker, the contact parting time is long enough that there may be several restrikes before the dielectric strength builds up enough to interrupt the circuit.

Vacuum devices work because the dielectric strength increases rapidly at very low pressures (because there are very few gas molecules to ionize). Normally, when approaching atmospheric pressures, the dielectric breakdown of air decreases as pressure decreases, but for very low pressures, the dielectric breakdown goes back up. The pressure in vacuum bottles is 10−6–10−8 torr. A vacuum device only needs a very short separation distance (about 8–10 mm for a 15 kV circuit breaker). Interruption is quick since the mechanical travel time is small. The separating contacts draw an arc (it still takes a current zero to clear). Sometimes, vacuum circuit breakers chop the current, causing voltage spikes. The arc is a metal vapor consisting of particles melted from each side. Contact erosion is low, so vacuum devices are low maintenance and have a long life. Restrikes are uncommon.

SF6 is a gas that is a very good electrical insulator, so it has rapid dielectric recovery. At atmospheric pressures, the dielectric strength is 2.5 times that of air, and at higher pressures, the performance is even better. SF6 is very stable, does not react with other elements, and has good temperature characteristics. One type of device blows compressed SF6 across the arc stream to increase the dielectric strength. Another type of SF6 interrupter used in circuit breakers and reclosers has an arc spinner which is a setup that uses the magnetic field from a coil to cause the arc to spin rapidly (bringing it in contact with un-ionized gas). SF6 can be used as the insulating medium as well as the interrupting medium. SF6 devices are low maintenance, have short opening times, and most do not have restrikes.

Since interrupters work on the principle of the dielectric strength increasing faster than the recovery voltage, the X/R ratio can make a significant difference in the clearing capability of a device. In an inductive circuit, the recovery voltage rises very quickly since the system voltage is near its peak when the current crosses through zero. Asymmetry increases the peak magnitude of the fault current. For this reason, the capability of most interrupters decreases with higher X/R ratios. Some interrupting equipment is rated based on a symmetrical current basis while other equipment is based on asymmetrical current. Whether based on a symmetrical or asymmetrical basis, the interrupter has asymmetric interrupting capability.

30.2.2 Circuit Breakers

Circuit breakers are often used in the substation on the bus and on each feeder. Circuit breakers are available with very high interrupting and continuous current ratings. The interrupting medium in circuit breakers can be any of vacuum, oil, air, or SF6. Oil and vacuum breakers are most common on distribution stations with newer units being mainly vacuum with some SF6.

Circuit breakers are tripped with external relays. The relays provide the brains to control the opening of the circuit breaker, so the breaker coordinates with other devices. The relays also perform reclosing functions.

Circuit breakers are historically rated as constant MVA devices. A symmetrical short-circuit rating is specified at the maximum rated voltage (for more ratings information, see ANSI C37.06-1997; IEEE Std. C37.04-1999; IEEE Std. C37.010-1999). Below the maximum rated voltage (down to a specified minimum value), the circuit breaker has more interrupting capability. The minimum value where the circuit breaker is a constant-MVA device is specified by the constant K:

$\text{Symmetrical}\text{interrupting}\text{capability}=\{\begin{array}{cc}{V}_R\cdot I_R& \text{for}\frac{V_R}{K}<V\le V_R\\ K\cdot I_R& \text{for}V\le \frac{V_R}{K}\end{array}$

where

IR is the rated symmetrical rms short-circuit current operating at VR

VR is the maximum rms line-to-line rated voltage

V is the operating voltage (also rms line-to-line)

K = voltage range factor = ratio of the maximum rated voltage to the lower limit in which the circuit breaker is a constant MVA device

Newer circuit breakers are rated as constant current devices (K = 1).

Consider a 15 kV class breaker application on a 12.47 kV system where the maximum voltage will be assumed to be 13.1 kV (105%). For an ANSI-rated 500-MVA class breaker with VR = 15 kV, K = 1.3, and IR = 18 kA, the symmetrical interrupting capability would be 20.6 kA (15/13.1 × 18). Circuit breakers are often referred to by their MVA class designation (1000 MVA class for example). Typical circuit breaker ratings are shown in Table 30.1.

Circuit breakers must also be derated if the reclose cycle could cause more than two operations and if the operations occur within less than 15 s. The percent reduction is given by

$D=d_1(n-2)+\frac{d_1(15-t_1)}{15}+\frac{d_1(15-t_2)}{15}+\cdots$

where

D is the total reduction factor, %

d1 = calculating factor, % $=\{\begin{array}{ll}3\%\hfill & \text{for}{I}_R<18\text{kV}\hfill \\ \frac{I_R}{6}\hfill & \text{for}{I}_R>18\text{kA}\hfill \end{array}$

n is the total number of openings

tn is the nth time interval less than 15 s

The interrupting rating is then (100 − D)IR. The permissible tripping delay is also a standard. For the given delay period, the circuit breaker must withstand K times the rated short-circuit current between closing and interrupting. A typical delay is 2 s.

Continuous current ratings are independent of interrupting ratings (although higher continuous ratings usually go along with higher interrupting ratings). Standard continuous ratings include 600, 1200, 2000, and 3000 A (the 600 and 1200 A circuit breakers are most common for distribution substations).

A circuit breaker also has a momentary or close and latch short-circuit rating (also called the first-cycle capability). During the first cycle of fault current, a circuit breaker must be able to withstand any current up to a multiple of the short-circuit rating. The rms current should not exceed 1.6K × IR and the peak (crest) current should not exceed 2.7K × IR.

The circuit breaker interrupting time is defined as the interval between energizing the trip circuit and the interruption of all phases. Most distribution circuit breakers are five-cycle breakers. Older breakers interrupt in eight cycles.

Distribution circuit breakers are three-phase devices. When the trip signal is received, all three phases are tripped. All three will not clear simultaneously because the phase current zero crossings are separated. The degree of separation between phases is usually one-half to one cycle.

30.2.3 Circuit Breaker Relays

Several types of relays are used to control distribution circuit breakers. Distribution circuits are almost always protected by overcurrent relays that use inverse time overcurrent characteristics. An inverse time–current characteristic means that the relay will operate faster with increased current.

The main types of relays are as follows:

• Electromechanical relays: The induction disk relay has long been the main relay used for distribution overcurrent protection. The relay is like an induction motor with contacts. Current through the CT leads induces flux in the relay magnetic circuit. These flux linkages cause the relay disk to turn. A larger current turns the disk faster. When the disk travels a certain distance, the contacts on the disk meet stationary contacts to complete the relay trip circuit. An instantaneous relay functionality can be provided: A plunger surrounded by a coil or a disk cup design operates quickly if the current is above the relay pickup. Most electromechanical relays are single phase.
• Static relays: Analog electronic circuitry (like op-amps) provide the means to perform a time–current characteristic that approximates that of the electromechanical relay.
• Digital relays: The most modern relay technology is fully digital based on microprocessor components.

Electromechanical relays have reliably served their function and will continue to be used for many years. The main characteristics that should be noted as it affects coordination are overtravel and reset time. Overtravel occurs because of the inertia in the disk. The disk will keep turning for a short distance even after the short circuit is interrupted. A typical overtravel of 0.1 s is assumed when applying induction relays. An induction disk cannot instantly turn back to the neutral position. This reset time should be considered when applying reclosing sequences. It is not desirable to reclose before the relay resets or ratcheting to a trip can occur.

Digital relays are slowly replacing electromechanical relays. The main advantages of digital relays are as follows:

• More relay functions: One relay performs the functions of several electromechanical relays. One relay can provide both instantaneous and time–overcurrent relay protection for three phases, plus the ground, and perform reclosing functions. This can result in considerable space and cost savings. Some backup is lost with this scheme if a relay fails. One option to provide relay backup is to use two digital relays, each with the same settings.
• New protection schemes: Advanced protection schemes are possible that provide more sensitive protection and better coordination with other devices. Two good examples for distribution protection are negative-sequence relaying and sequence coordination. Advanced algorithms for high-impedance fault detection are also possible.
• Other auxiliary functions: Fault location algorithms, fault recording, and power quality recording functions.

Digital relays have another advantage: internal diagnostics with ability to self-test. With digital technology, the relay is less prone to drift over time from mechanical movements or vibrations. Digital relays also avoid relay overtravel and ratcheting that are constraints with electromechanical relays (though some digital relays do reset like an electromagnetic relay).

Digital relays do have disadvantages. They are a relatively new technology. Computer technology has a poor reputation as far as reliability. If digital relays were as unreliable as a typical personal computer, we would have many more interruptions and many fires caused by uncleared faults. Given that, most digital relays have proven to be reliable and are gaining more and more acceptance by utilities.

Just as computer technology continues to advance rapidly, digital relays are also advancing. While it is nice to have new features, technical evolution can also mean that relay support becomes more difficult. Each relay within a certain family has to have its own supporting infrastructure for adjusting the relay settings, uploading and downloading data, and testing the relay. Each relay requires a certain amount of crew learning and training. As relays evolve, it becomes more difficult to maintain a variety of digital relays. The physical form and connections of digital relays are not standardized. As a contrast, electromechanical relays change very little and require a relatively stable support infrastructure. Equipment standardization helps minimize the support infrastructure required.

The time–current characteristics are based on the historically dominant manufacturers of relays. Westinghouse relays have a CO family of relays, and the General Electric relays are IAC (see Table 30.2). Most relays (digital and electromechanical) follow the characteristics of the GE or Westinghouse relays. For distribution overcurrent protection, the extremely inverse relays are most often used (CO-11 or IAC-77).

The time–current curves for induction relays can be approximated by the following equation (Benmouyal and Zocholl, 1994):

$t=TD\left(\frac{A}{M^p-1}+B\right)$

where

t is the trip time, s

M is the multiple of pickup current (M > 1)

TD is the time dial setting

A, B, p are curve shaping constants

With actual induction disk relays, the constants A and B change with the time dial setting, but with digital relays, they stay constant.

Standardized characteristics of relays have been defined by IEEE (IEEE Std. C37.112-1996). This is an attempt to make relay characteristics consistent (since the relay curve can be adjusted to almost anything in a digital relay). The equations for the standardized inverse relay characteristics are shown in Table 30.3. Figure 30.2 compares the shapes of these curves. The standard allows relays to have tripping times within 15% of the curves. The standard also specifies the relay reset time for 0 < M < 1 as

$t=TD\left(\frac{t_r}{M^2-1}\right)$

where tr is the reset time, s for M = 0

30.2.4 Reclosers

A recloser is a specialty distribution protective device capable of interrupting fault current and automatically reclosing. The official definition of a recloser is as follows:

Automatic circuit recloser: A self-controlled device for automatically interrupting and reclosing an alternating-current circuit, with a predetermined sequence of opening and reclosing followed by resetting, hold closed, or lockout.

Like a circuit breaker, interruption occurs at a natural current zero. The interrupting medium of a recloser is most commonly vacuum or oil. The insulating medium is generally oil, air, a solid dielectric, or SF6. The recloser control can be electronic, electromechanical (the relay for tripping is electromechanical, and the reclosing control is electronic), or hydraulic. A hydraulic recloser uses springs and hydraulic systems for timing and actuation.

The interrupting rating of a recloser is based on a symmetrical current rating. The interrupting current rating does not change with voltage. There is an exception that some reclosers have a higher interrupting current if operated at a significantly lower voltage than the rating. Smaller reclosers with a 50–200 A continuous rating typically have interrupting ratings of 2–5 kA (these would normally be feeder reclosers). Larger reclosers that could be used in substations have continuous current ratings as high as 1120 A and interrupting ratings of 6–16 kA. Historically, reclosers with series coil types had coil ratings of 25, 35, 50, 70, 100, 140, 200, 280, 400, and 560 A (each rating is approximately 1.4 times higher than the next lower rating).

Reclosers are tested at a specified X/R ratio as specified in ANSI/IEEE C37.60-1981. A typical test value is X/R = 16. While a lower X/R ratio at the point of application does not mean you can increase the rating of a recloser, the recloser must be derated if the X/R ratio is larger than that specified.

There are some other differences with recloser ratings vs. circuit breaker ratings (Cooper Power Systems, 1994). Reclosers do not have to be derated for multiple operations. Reclosers do not have a separate closing and latching (or first-cycle) rating. The symmetrical current rating is sufficient to handle the asymmetry during the first cycle as long as the circuit X/R ratio is lower than the tested value.

Reclosers have many distribution applications. We find reclosers in the substation as feeder interrupters instead of circuit breakers. An IEEE survey found that 51% of station feeder interrupting devices were reclosers (IEEE Working Group on Distribution Protection, 1995). Reclosers are used more in smaller stations and circuit breakers more in larger stations. Three-phase reclosers can be used on the main feeder to provide necessary protection coverage on longer circuits, along with improved reliability. Overhead units and padmounted units are available. Reclosers are available as single-phase units, so they can be used on single-phase taps instead of fuses. Another common application is in autoloop automation schemes to automatically sectionalize customers after a fault.

Since reclosers are devices built for distribution circuits; some have features that are targeted to distribution circuit needs. Three-phase units are available that can operate each phase independently (so a single-phase fault will only open one phase). Some reclosers have a feature called sequence coordination to enhance coordination between multiple devices.

The time–current characteristics of hydraulic reclosers have letter designations: A, B, and C. The A is a fast curve that is used similarly to an instantaneous relay element, and the B and C curves have extra delay (“delayed” and “extra delayed”). For a hydraulically controlled recloser, the minimum trip threshold is twice the full-load rating of the trip coil of the recloser and is normally not adjustable. On electronically controlled reclosers, the minimum trip threshold is adjustable independently of the rating (analogous to setting the pickup of a time–overcurrent relay).

30.2.5 Expulsion Fuses

Expulsion fuses are the most common protective device on distribution circuits. Fuses are low-cost interrupters that are easily replaced (when in cutouts). Interruption is relatively fast and can occur in a half of a cycle for large currents. An expulsion fuse is a simple concept: A fusible element made of tin or silver melts under high current. Expulsion fuses are most often applied in a fuse cutout. In a fuse tube, after the fuse element melts, an arc remains. The arc, which has considerable energy, causes a rapid pressure buildup. This forces much of the ionized gas out of the bottom of the cutout (see Figure 30.3), which helps to prevent the arc from reigniting at a current zero. The extreme pressure, the stretching of the arc, and the turbulence help increase the dielectric strength of the air and clear the arc at a current zero. A fuse tube also has an organic fiber liner that melts under the heat of the arc and emits fresh, nonionized gases. At high currents, the expulsion action predominates, while at lower currents, the deionizing gases increase the dielectric strength the most.

The “expulsion” characteristics should be considered by crews when placing a cutout on a structure. Avoid placement where a blast of hot, ionized gas blown out the bottom of the cutout could cause a flashover on another phase or other energized equipment. Implement and enforce safety procedures whenever a cutout is switched in (because it could be switching into a fault), including eye protection, arc resistant clothing, and, of course, avoiding the bottom of the cutout.

The speed ratio of a fuse quantifies how steep the fuse curve is. The speed ratio is defined differently depending on the size of the fuse (IEEE Std. C37.40-1993):

$\text{Speed}\text{ratio}\text{for fuse}\text{ratings}\text{of}\text{1}00\text{A and under}=\frac{\text{melting}\text{current}\text{at}\text{0}\text{.1}\text{s}}{\text{melting}\text{current}\text{at}\text{300}\text{s}}$

$\text{Speed}\text{ratio}\text{for}\text{ratings}\text{above}\text{1}00\text{A}=\frac{\text{melting}\text{current}\text{at}\text{0}\text{.1}\text{s}}{\text{melting}\text{current}\text{at}\text{600}\text{s}}$

Industry standards specify two types of expulsion fuses, the most commonly used fuses. The “K” link is a relatively fast fuse, and the “T” is somewhat slower. K links have a speed ratio of 6–8. T links have a speed ratio of 10–13. The K link is the most commonly used fuse for transformers and for line taps. The K and T fuse links are standardized well enough that they are interchangeable among manufacturers for most applications.

Two time–current curves are published for expulsion fuses: the minimum melt curve and the maximum total clear curve. The minimum-melt time is 90% of the average melt time to account for manufacturing tolerances. The total clearing time is the average melting time plus the arcing time plus manufacturing tolerances. Figure 30.4 shows the two published curves for 50 A K and T fuse links. The manufacturer’s minimum melt curves for fuses less than or equal to 100 A normally start at 300 s, and those over 100 A start at 600 s.

The time–current characteristics for K and T links are standardized at three points (ANSI C37.42-1989). The minimum and maximum allowed melting current is specified for durations of 0.1, 10, and either 300 s (for fuses rated 100 A or less) or 600 s (for larger fuses).

Published fuse curves are for no loading and an ambient temperature of 25°C. Both loading and ambient temperature change the fuse melting characteristic. Load current causes the most dramatic difference, especially when a fuse is overloaded. Figure 30.5 shows the effect of loading on fuse melting time. Figure 30.6, Figure 30.7, Figure 30.8, and Figure 30.9 show time–current curves for K and T links.

For operation outside of this ambient range, the fuse melting time changes. The melting characteristic of tin fuse links changes 3.16% for each 10°C above or below 25°C, so a fuse operating in a 50°C ambient will operate in 92% of the published time $\text{(1}00\%-(25\text{/}10\text{)3}.\text{16}\%)$ . Silver links are less sensitive to temperature (0.9% melting change for each 10°C above or below 25°C).

The I2t of a fuse is often needed to coordinate between fuses. Table 30.4 shows the minimum melt I2t of K and T links estimated from the time–current curves at 0.01 s. This number is also useful to estimate melting characteristics for high currents below the published time–current characteristics, which generally have a minimum time of 0.01 s.

The 6, 10, 15, 25, 40, 65, 100, 140, and 200 A fuses are standard ratings that are referred to as preferred fuses. The 8, 12, 20, 30, 50, and 80 A fuses are intermediate fuses. The designations are provided because two adjacent fuses (for those below 100 A) will not normally coordinate. A 40 and a 30 A fuse will not coordinate, but a 40 and a 25 A fuse will coordinate up to some maximum current. Most utilities pick a standard set of fuses to limit the number of fuses stocked.

K or T links with tin fuse elements can carry 150% of the nominal current rating indefinitely. It is slightly confusing that a 100 A fuse can operate continuously up to 150 A. Overloaded fuses, although they can be safely overloaded, operate significantly faster when overloaded, which could cause miscoordination. In contrast to tin links, silver links have no continuous overload capability.

Other nonstandard fuses are available from manufacturers for special purposes. One type of specialty fuse is a fuse even slower than a T link that is used to provide better coordination with upstream circuit breakers or reclosers in a fuse-saving scheme. Another notable type of specialty fuse is a surge-resistant fuse that responds slowly to fast currents (such as surges) but faster to lower currents. These achieve better protection on transformers for secondary faults and faster operation for internal transformer failures while at the same time reducing nuisance fuse operations due to lightning.

Expulsion fuses under oil are another fuse variation. These “weak-links” are used on CSP (completely self-protected) transformers and some padmounted and vault transformers. Since they are not easily replaced, they have very high ratings—at least 2.5 times the transformer full load current and much higher if a secondary circuit breaker is used.

Transformers on underground circuits use a variety of fuses. For padmounted transformers, a common fuse is the replaceable Bay-O-Net style expulsion fuse. The time–current characteristics of this fuse do not follow one of the industry standard designations.

30.2.5.1 Fuse Cutouts

The cutout is an important part of the fuse interrupter. The cutout determines the maximum interrupting capability, the continuous current capability, the load-break capability, the basic lightning impulse insulation level (BIL), and the maximum voltage. Cutouts are typically available in 100, 200, and 300 A continuous ratings (ANSI standard sizes [ANSI C37.42-1989]).

Cutouts are rated on a symmetrical basis. Cutouts are tested at X/R ratios of 8 to 12, so if the X/R ratio at the application point is higher than the test value, the cutout should be derated. The fuse line holder determines the interrupting capability, not the fuse link.

Most cutouts are of the open variety with a removable fuse holder that is placed in a cutout with a porcelain-bushing-type support. We also find enclosed cutouts and open-link cutouts. Open links have a fuse link suspended between contacts. Open links have a much lower interrupting capability (1.2 kA symmetrical).

Many cutouts available are full-rated cutouts that can be used on any type of system where the maximum line-to-line voltage is less than the cutout rating. Cutouts are also available that have slant voltage ratings, which provide two ratings such as 7.8/15 kV that are meant for application on grounded circuits (IEEE Std. C37.48-1997). One cutout will interrupt any current up to its interrupting rating and up to the lower voltage rating. On a grounded distribution system, in most situations, any cutout can be applied that has the lower slant rating voltage greater than the maximum line-to-ground voltage. On a 12.47Y/7.2 kV grounded distribution system, a 7.8/15 kV cutout could be used. If the system were ungrounded, a full-rated 15 kV cutout must be used. For three-phase grounded circuits, the recovery voltage is the line-to-line voltage for a line-to-line fault rather than the line-to-ground voltage (requiring a higher voltage rating). In this case, the slant-rated cutouts are designed and tested so that two cutouts in series will interrupt a current up to its interrupting rating and up to the higher voltage rating. The two cutouts share the recovery voltage (even considering the differences in the melting times of the two fuses). On grounded systems, there are cases where slant-rated cutouts are “under-rated”—any time that a phase-to-phase fault could happen that would only be cleared by one cutout. This includes constructions where multiple circuits share a pole or cases where cutouts are applied on different poles.

Most cutouts used on distribution systems do not have load break capability. If the cutout is opened under load, it can draw an arc that will not clear. It is not an uncommon practice for crews to open cutouts under load (if it draws an arc, they slam it back in). Cutouts with load-break capability are available, usually capable of interrupting 100–300 A. Cutouts with load-break capability usually use an arc chute. A spring pulls the arc quickly through the arc chute where the arc is stretched, cooled, and interrupted. A load-break tool is available that can open standard cutouts (with no load break capability of its own) under load up to 600–900 A. Utilities also sometimes use solid blades in cutouts instead of a fuse holders; then crews can use the cutout as a switch.

30.2.6 Current-Limiting Fuses

Current-limiting fuses (CLFs) are another interrupter having the unique ability to reduce the magnitude of the fault current. CLFs consist of fusible elements in silicon sand (see Figure 30.10). When fault current melts the fusible elements, the sand melts and creates a narrow tube of glass called a fulgerite. The voltage across the arc in the fulgerite greatly increases. The fulgerite constricts the arc. The sand helps cool the arc (which means it takes energy from the arc). The sand does not give off ionizable gas when it melts, and it absorbs electrons, so the arc has very little ionizable air to use as a conductor. Without ionizable air, the arc is choked off, and the arc resistance becomes very high. This causes a back voltage that quickly reduces the current. The increase in resistance also lowers the X/R ratio of the circuit, causing a premature current zero. At the current zero, the arc extinguishes. Since the X/R ratio is low, the voltage zero and current zero occur very close together, so there will be very little transient recovery voltage (the high arc voltage comes just after the element melts). Because the current-limiting fuse forces an early current zero, the fuse can clear the short circuit in much less than one half of a cycle.

Current-limiting fuses are noted for their very high fault-clearing capability. CLFs have symmetrical maximum interrupt ratings up to 50 kA; contrast that to expulsion fuses which may have typical maximum interrupt ratings of 3.5 kA in oil and 13 kA in a cutout. Current-limiting fuses also completely contain the arc during operation and are noiseless with no pressure buildup.

Current-limiting fuses are widely used for protection of equipment in high fault current areas. Table 30.5 shows the percentages of utilities that use CLFs. The major reason given for the use of CLFs is safety, and the second most common reason is high fault currents in excess of expulsion fuse ratings.

There are three types of current-limiting fuses (IEEE Std. C37.40-1993):

• Backup: A fuse capable of interrupting all currents from the maximum rated interrupting current down to the minimum rated interrupting current.
• General purpose: A fuse capable of interrupting all currents from the maximum rated interrupting current down to the current that causes melting of the fusible element in 1 h.
• Full range: A fuse capable of interrupting all currents from the rated interrupting current down to the minimum continuous current that causes melting of the fusible element(s), with the fuse applied at the maximum ambient temperature specified by the manufacturer.

Current-limiting fuses are very good at clearing high-current faults. They have a much harder time with low-current faults or overloads. For a low-level fault, the fusible element will not melt, but it will get very hot and can melt the fuse hardware resulting in failure. This is why the most common CLF application is as a backup in series with an expulsion fuse. The expulsion fuse clears low-level faults, and the CLF clears high-current faults. Current-limiting fuses have very steep melting and clearing curves, much steeper than expulsion links. Many current-limiting fuses have steeper characteristics than I2t. At low currents, heat from the notches transfers to the un-notched portion; at high currents, the element melts faster because heat cannot escape from the notched areas fast enough to delay melting.

General-purpose fuses usually use two elements in series: one for the high-current faults and one for the low-current faults. General-purpose fuses could fail for overloads, so restrict their application to situations where overloads are not present or are protected by some other device (such as a secondary circuit breaker on a transformer).

Full-range fuses provide even better low-current capability and can handle overloads and low-level faults without failing (as long as the temperature is within rating).

Current-limiting fuses can be applied in several ways, including

• Backup current-limiting fuse in series with an expulsion fuse in a cutout
• Full-range current-limiting fuse in a cutout
• Backup CLF under oil
• Full-range (or general-purpose) fuse under oil
• CLF in a dry-well canister or insulator

The best locations for use on distribution systems are close to the substation. This is where they are most appropriate for limiting damage due to high fault currents and where they are most useful for reducing the magnitude and duration of a voltage sag.

Some of the drawbacks of current-limiting fuses are summarized as follows:

• Voltage kick: When a CLF operates, the rapidly changing current causes a voltage spike (V = Ldi/dt). Usually, this is not severe enough to cause problems for the fuse or for customer equipment.
• Limited overload capability: A backup or general-purpose fuse does not do well for overloads or low-current faults. A full-range fuse performs better but could still have problems with a transient overcurrent that partially melts the fuse.
• Coordination issues: A current-limiting fuse may be difficult to coordinate with expulsion fuses or reclosers or other distribution protective devices. CLFs are fast enough that they almost have to be used in a fuse-blowing scheme (fuse saving will not work because the fuse will be faster than the circuit breaker).
• Cost: High cost relative to an expulsion fuse.

Current-limiting fuses limit the energy at the location of the fault. This provides safety to workers and the public. Arc damage to life and property occurs in several ways:

• Pressure wave: The fault arc pressure wave damages equipment and personnel.
• Heat: The fault arc heat burns personnel and can start fires.
• Pressure buildup in equipment: An arc in oil causes pressure buildup that can rupture equipment.

All of these effects are related to the arc energy and all are greatly reduced with current-limiting fuses. Distribution transformers are a common application of current-limiting fuses to prevent them from failing violently due to internal failures.

30.3 Transformer Fusing

The primary purpose of a transformer fuse is to disconnect the transformer from the circuit if it fails. Some argue that the fuse should also protect for secondary faults. The fuse cannot effectively protect the transformer against overloads.

Engineers most commonly pick fuse sizes for distribution transformers from a fusing table developed by the utility, transformer manufacturer, or fuse manufacturer. These tables are developed based on criteria for applying a fuse such that the fuse should not have false operations from inrush and cold-load pickup.

One way to pick a fuse is to plot cold-load pickup and inrush points on a time–current coordination graph and pick a fuse with a minimum melt or damage curve that is above the cold-load and inrush points. Most fusing tables are developed this way. A fuse should withstand the cold-load and inrush points given in Table 30.6. The inrush points are almost universal, but the cold-load pickup points are more variable (and they should be since cold-load pickup characteristics change with predominant load types). An example application of the points given in Table 30.6 for a 50 kVA, 7.2 kV single-phase transformer which has a full-load current of 6.94 A is shown in Figure 30.11. The cold-load pickup and inrush points are plotted along with K links. The minimum melt time and the damage time (75% of the minimum melt time) are shown. Use the damage curve to coordinate. For this example, a 12 A K link would be selected; the 1 s cold-load pickup point determines the fuse size. Since this point lies between the damage and minimum melt time of the 10 K link, some engineers would pick the 10 K link (not recommended).

Some utilities have major problems from nuisance fuse operations (especially utilities in high-lightning areas). A nuisance operation means that the fuse must be replaced, but the transformer was not permanently damaged. Nuisance fuse operations can be over 1% annually. Some utilities have thousands of nuisance fuse operations per year. A utility in Florida had a region with 57% of total service interruptions due to transformer interruptions, and 63% of the storm-related interruptions required only re-fusing (Plummer et al., 1995). During a storm, multiple transformer fuses can operate on the same circuit. There are differences of opinion as to what is causing the nuisance operations. Some of the possibilities are as follows:

• Inrush: Transformer inrush may cause fuse operations even though the inrush points are used in the fuse selection criteria. Reclosing sequences during storms can cause multiple inrush events that can heat up the fuse. In addition, voltage sags can cause inrush (any sudden change in the voltage magnitude or phase angle can cause the transformer to draw inrush).
• Cold-load pickup: This is the obvious culprit after an extended interruption (many of the nuisance fuse operations have occurred when there is not an extended interruption).
• Secondary-side transformer faults: Secondary-side faults that self-clear can cause some nuisance fuse events.
• Lightning current: Lightning current itself can melt small fuses. Arrester placement is key here since the lightning current flows to the low impedance provided by a conducting arrester. If the fuse is upstream of the arrester (which would be the case on a tank-mounted arrester), the lightning surge current flows through the fuse link. If the fuse is downstream, then little current should flow through the fuse.
• Power-follow current through gapped arresters: Following operation of a gapped arrester, a few hundred amps of power follow current flows in a gapped silicon carbide arrester until the gap clears (usually just for a half cycle if the gap is in good shape).
• Transformer saturation from lightning currents: Lightning can contain multiple strokes and long-duration components that last from 0.1 to 2 s. These currents can saturate distribution transformers. Following saturation, the transformer becomes a low impedance and draws high current from the system through the fuse (Hamel et al., 1990).
• Animal faults: Across transformer bushings or arresters.

Several of these causes may add to the total. Nuisance fuse operations have occurred when circuits were out of service. This means that lightning is the cause since any type of inrush would require the system to be energized. Detroit Edison found that 70%–80% of fuse operations were due to lightning (Gabrois et al., 1973). Lightning and inrush events are the most likely cause of nuisance fuse operations. Heavily loaded transformers are more susceptible to nuisance fuse operations because of the preheating of the fuse (a heavily loaded transformer is more susceptible to cold-load pickup as well).

Another method of choosing the transformer fuse size that gives “looser” fusing is the ×2 method (Burke, 1996). Choose a fuse size larger than twice the transformer full load current. A 50 kVA, 7.2 kV single-phase transformer, which has a full-load current of 6.94 A, should have a fuse bigger than 14 A (the next biggest standard size is a 15 A fuse). This applies for any type of fuse (K, T, or other). The factor of two provides a safety margin so that transformer fuses do not operate for inrush or cold-load pickup, and it helps with lightning.

The fusing ratio is the ratio of the fuse minimum melt current to the transformer full-load current (some sources also define a fusing ratio as the ratio of fuse-rated current to transformer-rated current which is different from this definition by a factor of two). Tight fusing means the fuse ratio is low. Relatively low fusing ratios have been historically used which has led to the nuisance fuse problems. The tighter fusing given using the Table 30.6 approach results in fusing ratios of 2–4. The looser ×2 method gives a fusing ratio of at least 4 (since the fuse rating is multiplied by two, and the minimum melting current at 300 s is twice the fuse rating). The fusing ratio for the 50 kVA, 7.2 kV transformer with the 12-K fuse is 3.46, and it is 4.32 with the 15-K fuse.

Another strategy that is especially useful in high-lightning areas: Use a standard fuse size for all transformers up to a certain size. This also helps ensure that the wrong fuse is not applied on a given transformer. A standard fuse size of at least 15 T or 20 K results in few nuisance fuse operations (IEEE Std. C62.22-1997). At 12.5 kV, a 20 K fuse should protect a 5 kVA transformer almost as well as it protects a 50 kVA transformer. It may lose some secondary protection relative to a smaller fuse, and a small portion of evolving faults will not be detected as soon; but other than that, there should not be much difference. If fuses get too big, they may start to bump up against tap fuse sizes and limit fuse options for lateral taps.

If looser fusing is used, some argue that overload protection of transformers is lost. Countering that argument, overload protection with fuses is not really possible if the transformer is used for its most economic performance (which means overloading a transformer at peak periods). To avoid nuisance fuse operations from load, we must use a fuse big enough so that thermal overload protection is impossible. It is also argued that most transformer failures start as failures between turns or layers and that a smaller, faster fuse detects this more quickly. Tests have indicated that a smaller fuse might not be much better than a larger fuse at detecting interwinding failures (Lunsford and Tobin, 1997) (pressure-relief valves limit tank pressures very well during this type of failure). All together, the arguments for a smaller fuse are not enough to overcome the concerns with nuisance fuse operations. If overload protection must be used, use a surge resistant fuse (it has a slower characteristic for high-magnitude, short-duration currents).

A few utilities practice group fusing where a lateral fuse provides protection to all of the transformers on the tap. If the transformer failure rate (including bushing faults) is low enough, then this practice will not degrade the overall frequency of interruptions significantly. One of the major disadvantages of this approach is that an internal transformer failure on a tap may be very hard to find. This drives up repair time (so the duration reliability numbers suffer but not necessarily the frequency indices). Also, the beneficial feature of being able to switch the transformer with the fused cutout is lost if group fusing is used.

Widely used, completely self-protected transformers (CSPs) have an internal weak-link fuse; an external fuse is not needed (although they may need an external current-limiting fuse to supplement the weak link).

Transformer bushing faults often caused by animals can have different impacts depending on fusing practices. A fault across a primary bushing operates an external transformer fuse. If the transformer is a CSP or group fusing is used, the upstream tap fuse operates (so more customers are affected).

Current-limiting fuses are regularly used on transformers in high fault current areas to provide protection against violent transformer failure. NEMA established tests which were later adopted by ANSI (ANSI C57.12.20-1988) for distribution transformers to be able to withstand internal arcs. Transformers with external fuses are subjected to a test where an internal arcing fault with an arc length of 1 in. (2.54 cm) is maintained for 1/2 to one cycle. It was thought that 1 in. (2.54 cm) was representative of the length that arcs could typically achieve. The current is 8000 A. Under this fault condition, the transformer must not rupture or expel excessive oil. Note that this test does not include all of the possible failure modes and is no guarantee that a transformer will not fail with lower current. For example, a failure with an arc longer than 1 in. has more energy and ruptures the transformer at a lower level of current.

Table 30.7 shows rupture limits for several types of transformers based on tests for the Canadian Electrical Association. If fault current values exceed those given in this table, consider using current-limiting fuses to reduce the chance of violent failures (the CEA report considers the limits provisional and suggests that more tests are needed). At arc energies within this range, the failure probability is on the order of 15%–35%. Note that the 2.5 kA limit for pole-mounted transformers is much less than the ANSI test limit of 8 kA. Based on a series of tests with internal 2 in. (5 cm) arcs, Hamel et al. (2003) recommend considering current-limiting fuses for pole-type transformers when the short-circuit current exceeds 1.7 kA.

For transformers with an internal fuse, completely self-protected (CSP) transformers or padmounted transformers, the arcing test is done to the rating of the fuse which is generally much lower than 8000 A. Table 30.8 shows the maximum fault current ratings based on the ANSI tests. If the available line-to-ground fault current exceeds these values, then consider current-limiting fuses to reduce the possibility of violent failures. Not all utilities use current-limiting fuses in these situations, and in such instances, internal faults have failed transformers violently, blowing the cover.

If a transformer is applied in a location where the available line-to-ground fault current is higher than shown in Table 30.8, use current-limiting fuses.

30.4 Lateral Tap Fusing and Fuse Coordination

Utilities use two main philosophies to apply tap fuses: fusing based on load and standardized fusing schedules. With fusing based on load, we pick a fuse based on some multiplier of peak load current. The fuse should not operate for cold-load pickup or inrush to prevent nuisance operations. As an example, one utility sizes fuses based on 1.5 times the current from the phase with the highest connected kVA. With standardized fuse sizes, a typical strategy is to apply 100 K links at all taps off of the mains (even if a tap only has one 15-kVA transformer). If using second-level fusing, use 65 K links for these and 40 K fuses for the third level. There is no clear winner; each has advantages and disadvantages:

• Fusing based on load: This tends to fuse more tightly. High-impedance faults are somewhat more likely to be detected. Nuisance fuse operations are more likely, especially with utilities that tightly fuse laterals. We are more likely to have load growth cause branch loadings to increase to the point of causing nuisance fuse operations. Fusing based on load helps on circuits that have covered wire because a smaller fuse helps protect against conductor burndown (taps that are more heavily loaded usually have a larger wire, which resists burndown).
• Standardized fuse sizes: It is simple: we spend less time coordinating fuses, we do not constantly check loadings, and utilities have less inventory. There is also less chance that the wrong fuse is installed at a location. A disadvantage of this approach is that larger fuses than needed are used at many locations, resulting in higher fault damage at the arc location, longer voltage sags, and more stress on in-line equipment.

Coordinating lateral tap fuses is generally straightforward. The fuse must coordinate with the station recloser or circuit breaker relays. Station ground relays are usually set to coordinate with the largest tap fuse. On the downstream side, a tap fuse should coordinate with the largest transformer fuse. This usually is not a problem.

In addition to sizing a fuse to avoid nuisance operations and coordinating with upstream and downstream protectors, we size fuses to ensure that the fuses provide protection to the line section that they are protecting. The reach of the fuse must exceed the length of the line section. Several methods are used to quantify the reach of a fuse:

• Where the fuse will clear a bolted single line-to-ground fault in 3 s
• Where the bolted single line-to-ground fault current is six times the fuse rating
• Where the fuse will clear a single line-to-ground fault with a 30 Ω resistance in 5 s

In most situations, typical fuse sizes provide sufficient reach by any of these methods. The first two methods are the best; the 30 Ω resistance is overly conservative and difficult to apply.

Reliability needs dictate the number of fuses used. The most common application for line fuses is at tap points. Occasionally, utilities fuse three-phase mains, but a recloser is more commonly used for this purpose. In the southwest United States, in areas with few trees and little lightning, fuses may be rarely used. This is the exception, not the rule. Most utilities fuse most taps off the main line. Some go further and provide several levels of fusing, especially utilities with heavy tree coverage. Returns diminish: Too many fuses leads to situations where fuses do not coordinate, and the extra fusing does not increase reliability significantly. Cutouts themselves contribute to causing faults by providing an easy location where animals, trees, and lightning cause faults, especially if they are poorly installed.

30.5 Station Relay and Recloser Settings

The main feeder circuit breaker relays (or recloser) must be set so that the circuit breaker coordinates with downstream devices, coordinates with upstream devices, and does not have trips from inrush or cold-load pickup. Station relays almost always use phase and ground time–overcurrent relays.

Table 30.9 shows typical settings used by several utilities. Many utilities try to use standardized relay settings at all distribution stations. This has the advantage that relays are less likely to be set wrong, and there is less engineering effort in a coordination study. Some other utilities set each relay based on a coordination study.

Differences exist about the meaning of “peak load.” Some utilities base it on the maximum design emergency load (which is typically something like 600 A). Others use the designed normal load (typically 400 A). Others may base it on some percentage of the total connected distribution transformer kVA.

Instantaneous relay settings vary more than phase relay settings. Several utilities also either disable or do not use an instantaneous relay setting. The instantaneous relay pickup ranges from one to almost 10 times the phase relay pickup.

One reasonable set of base pickup settings is

• Phase TOC (time–overcurrent) relay: Use two times the normal designed peak load on the circuit.
• Ground TOC relay: Use 0.75 times the normal designed peak load on the circuit.
• Instantaneous phase and ground relays: Use two times the TOC relay pickups.

Settings any less than this are prone to false trips from cold-load pickup and inrush.

In addition to avoiding nuisance trips, the relays (or recloser) must provide protection to its line section (to the end of the line or to the next protective device in series). Ensure that the relay has sufficient reach at the minimum operating current of the relay.

For a phase relay, sufficient reach is achieved by ensuring that 75% of the bolted line-to-line fault current at the end of the circuit is greater than the relay’s pickup (its minimum operating current). So, if the line-to-line fault current at the end of the circuit is 1000 A, the pickup of the relay should be no more than 750 A. We use the line-to-line fault current because the two types of faults not seen by the ground relay are the three phase and the line to line. Of these, the line-to-line fault has the lower magnitude. The 75% factor provides a safety margin and allows some fault impedance. Another approach is to ensure that the line-to-ground fault current at the end of the circuit is less than the minimum operating current. The line-to-ground fault current is less than the line-to-line fault current, which provides the safety margin.

For a ground relay, ensure that the relay pickup is less than 75% of the line-to-ground fault current at the end of the line or to the next protective device. The ground relay must also coordinate with the largest lateral fuse.

Feeders dedicated to supplying secondary networks, either grid or spot networks, have similar settings as feeders supplying radial loads. Two main differences are related to loading and the ground relay setting. The pickup settings of station circuit breakers may have to account for higher peak loads. Some utilities have phase relay pickups that are similar to radial circuits, from 600 to 800 A, but some have settings above 900 A with one utility having a 1680 A setting (Smith, 1999). Also, if feeders are supplying only network load, and all network transformers are connected delta—grounded wye—the unbalanced current seen by the ground relay is small. Utilities can set a low ground relay setting; some have settings ranging from 40 to 80 A (Smith, 1999). The main limitation on lowering the setting further is that during line-to-ground faults, the unfaulted circuits will backfeed the fault through the zero-sequence capacitance of that circuit. Lower ground-relay settings also help detect turn-to-turn or layer-to-layer faults within the primary windings of network transformers.

30.6 Arc Flash

Arc flash is another situation often requiring overcurrent coordination, much like protecting against conductor burndown or transformer damage. For arc flash, the goal is to greatly reduce the chance of burns to workers if an arc occurs near a worker. We want a protective device to clear the fault before a fault arc could cause incident energy in excess of the rating of the clothing.

The severity of an arc flash event is normally quantified as the incident energy that would reach a worker, normally given in terms of cal/cm2. Flame-resistant (FR) clothing systems have an arc thermal performance value (ATPV) rating based on ASTM test standards (ASTM F1959, 2006). This rating is the incident energy in cal/cm2 on the clothing surface that has a 50% probability of causing a second-degree skin burn. The goal of an arc flash analysis is to ensure that workers have an ATPV protection sufficient to handle the incident energy that might be expected in a given work scenario. Out of 14 responses to an Electric Power Research Institute (EPRI) survey, utility minimum ATPV ratings ranged from 4 to 8.7 cal/cm2 with a median value of 5.4 cal/cm2.

A number of approaches are available for estimating arc flash. The two most commonly cited methods are the ARCPRO program (Kinectrics, 2002) and the IEEE 1584 method (IEEE Std 1584-2002). The IEEE 1584 method is based on curve-fit regressions to mainly three-phase arc-in-a-box tests and is most suitable for arcs in equipment and other arc-in-a-box scenarios. ARCPRO is based on a single arc in open air, so it is most suitable for overhead, open-air scenarios.

IEEE 1584-2002 was developed by the IEEE Industry Applications Society, a society focusing on industrial and commercial power. IEEE 1584 is the most widely adopted approach to arc flash analysis. The method for estimating arc flash incident energies is based on tests performed at several short-circuit labs. From this test set, regression was used to find equations to best fit the test data. IEEE 1584 assumes a three-phase fault and is mainly geared toward arc-in-a-box evaluations. Above 15 kV, the IEEE 1584 guide and companion spreadsheet default to the Lee method. The Lee method is the oldest arc flash calculation method. It is a simple model of a single, open-air arc. For medium-voltage applications, the Lee method produces unrealistically high predictions of incident energies. For 25 and 35 kV scenarios, consider using the 15 kV results from the IEEE spreadsheet as suggested in Short (2011). Figure 30.12 shows time-current curves based on IEEE 1584 that are suitable for medium-voltage switchgear. A protective device should clear before the time indicated on the appropriate curve. For different clothing, the curves can be shifted. For 4 cal/cm2 clothing, the protective device must operate twice as fast as with 8 cal/cm2 clothing.

ARCPRO is a commercial program for analyzing arc flash incident energies, developed by Kinectrics. The ARCPRO algorithm is based on the work of Bingwu and Chengkang (1991), but it is not completely described in any peer-reviewed paper. The ARCPRO model assumes the following (Cress, 2008): a vertical free burning arc in air; an arc length much greater than arc diameter; a one-arc column, either phase–phase or phase–ground; no electrode region heat transfer; and an optically thin plasma and gas. Cress (2008) reported that ARCPRO was verified with over 300 test points for arc energy and incident energy for currents from 3 to 25 kA, arc durations from 4 to 35 cycles, distances from 8 to 24 in. (20–60 cm), and with gaps from 1 to 12 in. (2.5–30 cm).

The key input parameters for an arc flash study are as follows:

• Working distance: Distance from the worker to the fault arc is an important input. Incident energies drop significantly with distance, normally as a power of 1/d1.5 to 1/d2.
• Arc length: Some arc flash models like ARCPRO include an arc length. The arc energy increases almost linearly with arc length. Arc lengths and arc voltages are primarily a function of gap spacings, not the system driving voltage. The arc length is different than the shortest gap between energized conductors or between an energized conductor and ground. Because a fault arc can move and elongate, the arc length is normally longer than the gap length. IEEE 1584 includes an arc gap internally in calculations.
• Fault current: For medium-voltage arc flash, engineers commonly assume a bolted fault. For low-voltage arc flash (under 1000 V), the arc impedance will reduce the fault current appreciable, so the arcing current is needed. IEEE 1584 suggests the following equation to estimate arcing current:

${\log }_{10}{I}_a=K+0.662{\log }_{10}{I}_{bf}+0.0966V+0.000526G+0.5588V{\log }_{10}{I}_{bf}-0.00304G{\log }_{10}{I}_{bf}$

where

Ia is the arc current, kA

Ibf is the bolted fault current, kA

K = −0.153 for open configurations, −0.097 for box configurations (enclosed equipment)

V is the system voltage, kV

G is the distance between buses, mm (IEEE assumes 32 mm for low-voltage switchgear.)

• Duration: The duration is based on the clearing time of the upstream protective device(s) based on its time–current characteristics. For very long or indeterminate clearing times, a worker self-extraction time is sometimes assumed as a maximum duration to consider. IEEE 1584 mentions 2 s for this duration. Even 2 or 3 ft (0.7–1 m) of movement away from the fault or to the side will dramatically reduce incident energies.
• Fault type: Whether the fault is a single phase–ground fault, a phase–phase fault, or a three-phase fault will affect the fault current and the duration. For faults in equipment, three-phase faults are commonly assumed as it is likely for the fault to expand to involve all three phases.

The 2007 National Electrical Safety Code (NESC) (IEEE C2-2007) requires an arc flash assessment be performed on systems above 1000 V. They do not provide specifics in general but do offer a table with default assumptions based on an ARCPRO analysis for open-air, single-phase-to-ground faults. The NESC table 410-1 footnotes specify a 15 in. (38 cm) separation distance from the arc to the employee for glove work and arc lengths as follows: 1–15 kV = 2 in. (5 cm), 15.1–25 kV = 4 in. (10 cm), and 25.1–36 kV = 6 in. (15 cm). Figure 30.13 shows time–current curves for 8 cal/cm2 clothing based on ARCPRO with the 15 in. (38 cm) working distance along with various arc lengths. Because Short and Eblen (2011) found that the default arc lengths in the NESC are low, longer arc lengths are also given.

As an overcurrent protection problem, two related assumptions are made for arc flash:

1. The incident energy increases linearly with time. If you double the duration, the incident energy doubles. All of the modeling approaches make this same assumption.
2. At lower currents and longer durations, the main impact is due to lower current. Most models have almost a linear relationship between current and incident energy: if the current doubles, the incident energy doubles.

At longer durations, these assumptions are uncertain. As in Figure 30.13, incident energies often become more of a concern where fault currents are lower and durations are longer. Arc flash models have been mostly tested with durations less than 0.5 s. At longer durations, several factors can come into play: the arc can move and/or elongate (increasing energy), the arc may involve additional phases (increasing energy), and the arc may self-extinguish (decreasing energy).

Another consideration is arc impedance. At low voltages, this is important to consider. Above 1000 V, a bolted fault is most appropriate. The 30–40 ohm impedance given by the Rural Electrification Association (REA Bull. 61-2, 1978) is too large; see also Dagenhart (2000) and Burke (2006). The arc in an arc flash scenario involves relatively low arc resistances. A 3 ft (1 m) arc has a voltage of about 1400 V. If the fault current at that point in the line was 1000 A, then the arc resistance is about 1.4 Ω. A 1 ft (0.3 m) arc with the same fault current has a resistance of 0.47 Ω. Most fault arcs have resistances of 0–2 Ω, so that can be used as guidance to find the minimum fault current.

For low-voltage arcs, arc sustainability is an important variable. Below 250 V, it is rare to find conditions where high-current arcs can sustain in an arc flash scenario. At 480 V, arc sustainability is highly dependent on equipment. In the 2012 NESC, a table is provided for clothing to protect workers for different types of equipment. This is partially based on tests documented by Eblen and Short (2010; EPRI 1018693, 2009; EPRI 1020210, 2009). For cases where a coordination study is needed at 480 V (like network protectors), IEEE 1584 is the most appropriate tool.

Arc flash analysis is still in its infancy, and further advances in modeling and protection are expected. Arc flash is equipment specific. In some cases, existing methods do not adequately predict performance. Short and Eblen (2011) show test results for a medium-voltage pad-mounted switch with incident energies three times that predicted by IEEE 1584. They provide an equation to estimate the incident energies. The higher incident energies were caused by the bus configuration which projected arcs out the front of the enclosure (Figure 30.14).

A number of relaying options are available to reduce incident energies, depending on the application. Common approaches include reducing clearing times by enabling a fast trip and disabling reclosing. Other options are available to coordinate relaying times with clothing capabilities and with work practices.

30.7 Coordinating Devices

Several details often arise when coordinating specific devices. Normally, we want to ensure that the downstream device clears before the upstream device operates over the range of fault currents available at the downstream device. Time–current characteristics of both device normally show us how well two devices coordinate. Because of device differences, some combinations require slightly different approaches. We discuss some of the common combinations in the sections that follow.

30.7.1 Expulsion Fuse–Expulsion Fuse Coordination

When coordinating two fuses, the downstream fuse (referred to as the protecting fuse) should operate before the upstream fuse (the protected fuse). To achieve this goal, ensure that the total clear time of the protecting fuse is less than the damage time of the protected fuse. The damage time is 75% of the minimum melt time. An example for coordinating a 100 K link with a 65 K link is shown in Figure 30.15. Above a certain current, the two fuses do not coordinate; the protected fuse could suffer damage or melt before the protecting fuse can clear. For high fault currents, coordination is impossible because both fuses can open. The example shows that above 2310 A, the total clear curve of the 65 K is above the damage curve of the 100 K link. Utilities live with this common miscoordination. Table 30.10 lists the maximum coordination currents between K links. In cases where fuses do not coordinate, why have the second fuse? The second fuse still has some value; it adds another sectionalizing point (for a fuse in a cutout), and for a downstream fault, it identifies the fault location to a smaller area. Also, the downstream fuse may operate without damaging the upstream fuse. The amount of damage to the upstream fuse depends on the point of the waveform where the fault occurs (the extra 1/2 + cycle waiting for a current zero causes the extra heating to the protected fuse).

30.7.2 Current-Limiting Fuse Coordination

Coordinating two current-limiting fuses is similar to coordinating two expulsion links. Plot the time–current characteristics and ensure that the maximum clearing time of the load-side fuse is less than 75% of the minimum-melting time of the source-side fuse over the range of fault currents available at the load-side fuse. The 75% factor accounts for damage to the source-side fuse. Unlike expulsion links, current-limiting fuses can coordinate to very high currents. For coordination at higher currents than are shown on published time–current characteristics (operations faster than 0.01 s), ensure that the maximum clearing I2t of the load-side fuse is less than 75% of the minimum-melt I2t of the source-side fuse. Manufacturers provide both of these I2t values for current-limiting fuses.

Coordinating an expulsion link with a current-limiting fuse follows similar principles. Because the melting and clearing characteristics of current-limiting fuses are so much steeper than those of expulsion links, coordination is sometimes difficult; the operating characteristic curves are more likely to cross over. A load-side current-limiting fuse coordinates over a wide range of fault current. For a source-side current-limiting fuse, the clearing-time limitations of expulsion links (to about 0.8 cycles) prevent coordination at high currents. For currents above this value, either both will operate, or just the current-limiting fuse will operate.

Backup current-limiting fuse coordination requires special attention. To ensure that the CLF does not try to operate for currents below its minimum interrupting rating, the intersection of the expulsion fuse’s total clearing curve and the backup fuse’s minimum-melting curve must be greater than the maximum interrupting rating of the backup fuse. Normally, we select backup current-limiting fuses for use with expulsion links based on matched-melt coordination. Select a backup current-limiting fuse that has a maximum melting I2t below the maximum clear I2t of the expulsion element. Also, check the time–current curves of the devices. The expulsion link should always clear for fault currents in the low-current operating region, especially below the minimum interrupting current of the current-limiting fuse.

With matched-melt coordination, the expulsion fuse always operates, including when the backup current-limiting fuse operates. In overhead applications with an expulsion fuse in a cutout, the dropout of the expulsion fuse provides a visible indication when the fuse(s) operate. Also, the backup fuse is unlikely to have full voltage across it.

The maximum melting I2t of expulsion links is not provided from curves or data. To estimate this, take the minimum melting I2t calculated from the minimum-melt curve at 0.0125 s, and multiply by 1.2 for tin links or 1.1 for silver links. The multiplier allows for conservatism in minimum-melt curves and for manufacturing tolerances.

Somewhat less conservatively, experience has shown that fuses coordinate well if the maximum melt I2t of the expulsion link does not exceed twice the minimum melt I2t of the backup fuse (IEEE Std. C37.48-1997). We can tighten up the backup fuse because, under most practical situations, the backup fuse lets through significantly more I2t than its minimum-melt value.

Manufacturers of backup current-limiting fuses normally provide coordination recommendations for their fuses, but review of the coordination approach is sometimes appropriate. Backup fuses often use a “K” nomenclature signifying the K link that it coordinates with. For example, a “25 K” backup link coordinates with a K link rated at 25 A or less. Figure 30.16 shows the time–current curves of a 40 K expulsion link and the curves of one manufacturer’s 40 K backup current-limiting fuse. This graph extends below the normal cutoff time of 0.01 s to show how the fuses coordinate at high currents. This example shows that the backup fuse does not coordinate using the strict matched-melt criteria (the maximum melting time of the expulsion link is more than the minimum melting time of the backup fuse). The minimum-melt I2t of the backup fuse is 1.6 times the minimum melt I2t of the backup fuse, so it meets the relaxed matched-melt criteria since this ratio is less than two.

The time–current curve crossover coordination allows a smaller backup current-limiting fuse. As before, the intersection of the expulsion fuse’s total clearing curve and the backup fuse’s minimum-melting curve must be greater than the maximum interrupting rating of the backup fuse. We do not try to ensure that the backup fuse always melts. We can use a smaller fuse, which reduces the I2t let through and reduces energy to faults. The backup CLF operates for a wider range of short-circuit currents. With a smaller fuse, the backup fuse can operate before the expulsion link melts for high fault currents. Utilities often use time–current curve crossover coordination for under-oil backup current-limiting fuses. In addition to lowering energy to faults, crossover coordination extends the range of current-limiting fuse protection to larger transformers.

For transformer protection, overload and secondary faults are also considerations for backup current-limiting fuses. Secondary faults at the terminals of the transformer should not damage or melt the backup CLF. One way to do this is to ensure that at the total clearing time of the expulsion link with the bolted secondary fault, the backup fuse’s minimum melting current is at least 125% of the secondary fault current (Hi-Tech Fuses, 2002). Also, overload should not damage or melt the backup CLF.

30.7.3 Recloser–Expulsion Fuse Coordination

Normally, we want the recloser’s fast curve (A curve) to clear before downstream fuses operate. This saves the fuse for temporary faults (we discuss fuse saving in more detail later in this chapter). Select the delayed curve (B, C, …) to be above the clearing time of downstream fuses. A permanent fault downstream of a fuse should blow the fuse, not lockout the recloser.

To open the recloser before the fuse blows, Cooper (1990) recommends adjusting the A curve by multiplying the time by a factor of 1.25 for one fast operation, a factor of 1.35 for two fast operations with a reclosing time greater than or equal to 1 s, and a factor of 1.8 for two fast operations with a reclosing time from 25 to 30 cycles. For applications with two or more delayed operations, the fast curve coordinates for fault currents up to the point where the adjusted A curve crosses the expulsion fuse’s minimum-melting curve.

On hydraulically controlled reclosers, the trip-coil rating determines the recloser’s “pickup.” Beyond that, hydraulically controlled reclosers have limited curve selections and no adjustments. Figure 30.17 shows average clearing curves for a single-phase Cooper 4E hydraulically controlled recloser overlayed on top of two K fuse links. For this example, only a limited range of fuses coordinate for low and high fault currents. Fuses larger than a 65 K have significant overlap in the low-current area, leaving more chance that the recloser could lock out for a fault on a lateral tap. The slower delayed curves, such as the C curve shown, reduce the chance of miscoordination for lower fault currents. For the fast-trip A curve, the 40 K link only coordinates with the fast curve for fault currents up to 360 A; smaller links are worse. Since K links are significantly steeper than these recloser curves, we must expect limited coordination for certain combinations. In this instance, T links coordinate over a wider current range because their time–current characteristics match the slope of the recloser characteristics more accurately. Miscoordination is more problematic in the low-current region. If the recloser locks out for faults downstream of a fuse, more customers are interrupted, and crews have a harder time finding the fault (more area to patrol).

Reclosers with electronic controls and relayed circuit breakers offer more flexibility. We can tailor tripping characteristics to coordinate over a wider range of currents. Three-phase reclosers have a ground-trip element that can increase the sensitivity of the recloser and also coordinate better with downstream fuses.

30.7.4 Recloser–Recloser Coordination

For coordinating two reclosers, the curve separation we need depends on the type of recloser. For hydraulically controlled reclosers that are series coil operated, both operate if there is less than a two-cycle separation; both may operate for a separation of 2–12 cycles, and both coordinate properly if there are more than 12 cycles of separation. For hydraulically controlled reclosers that use high-voltage solenoid closing (larger reclosers), we need 8 cycles of separation for coordination (if it is less than 2 cycles, both devices operate). This data is for Cooper reclosers (Cooper Power Systems, 1990).

30.7.5 Coordinating Instantaneous Elements

Coordinating instantaneous relay elements or recloser fast curves is difficult. By the nature of an instantaneous element, two in series will both operate if the short-circuit current is above the pickup of both relays.

The most common way to coordinate two instantaneous elements is to raise the pickup of the upstream element. Find a setting where the instantaneous relay will not operate for faults downstream of the second protective device. The upstream relay cannot operate if its pickup is above the available fault current at the location of the downstream element. For this strategy, the instantaneous pickup on the element must be higher than its time–overcurrent pickup. This rules out hydraulic reclosers, which have the same pickup for the fast (A) curve and the delayed curves (B and C), but is not a concern with electronic reclosers because they have the same flexibility as relayed circuit breakers.

Rather than using an instantaneous relay element, we can perform the “fast trip” function with a time–overcurrent relay with a fast characteristic. Now, we might be able to coordinate the fast curve of a line recloser with the substation circuit breaker or recloser.

As another way to coordinate two instantaneous elements, use a time delay on the upstream instantaneous element. Choose enough time delay, 6–10 cycles, to allow the downstream device to clear before the station device operates.

Even with coordinated fast curves (either using a delay or using a fast TOC curve), nuisance momentary interruptions occur for faults cleared by a downstream line recloser. Consider a station recloser R1 and a downstream line recloser R2 each with one fast curve (A) and two delayed curves (B). If a permanent fault occurs downstream of R2, R2 will first operate on its A curve. If the fast curves of R1 and R2 are coordinated, R1 will not operate. After a delay, R2 recloses. The fault is still there, so R2 operates on its delayed curve (its B curve). Now, R1 does operate because it is on its A curve which operates before R2’s B curve. After R1 recloses, R2 should then clear the fault on its B curve, which should operate before R1’s B curve. The fault is still cleared properly, but customers upstream of R2 have extra momentary interruptions.

A more advanced form of coordination called sequence coordination removes this problem. Sequence coordination is available on electronic reclosers and also on digital relays controlling circuit breakers. With sequence coordination, the station device detects and counts faults—but does not open—for a fault cleared by a downstream protector on the fast trip. If the fault current occurs again (usually because the fault is permanent), the station device switches to the time–overcurrent element because it counted the first as an operation. Using this form of coordination eliminates the momentary interruption for the entire feeder for permanent faults downstream of a feeder recloser. On a relay or recloser that has sequence coordination, if the device senses current above some minimum trip setting and the current does not last long enough to trip based on the device’s fast curve, the device advances its control-sequence counter as if the unit had operated on its fast curve. So when the downstream device moves to its delayed curve, the upstream device with sequence coordination also is operating on its delayed curve. With sequence coordination, for the fast curves, the response curve of the upstream device must still be slower than the clearing curve of the downstream device.

30.8 Fuse Saving versus Fuse Blowing

Fuse saving is a protection scheme where a circuit breaker or recloser is used to operate before a lateral tap fuse. A fuse does not have reclosing capability; a circuit breaker (or recloser) does. Fuse saving is usually implemented with an instantaneous relay on a breaker (or the fast curve on a recloser). The instantaneous trip is disabled after the first fault, so after the breaker recloses, if the fault is still there, the system is time coordinated, so the fuse blows. Because most faults are temporary, fuse saving prevents a number of lateral fuse operations.

The main disadvantage of fuse saving is that all customers on the circuit see a momentary interruption for lateral faults. Because of this, many utilities are switching to a fuse blowing scheme. The instantaneous relay trip is disabled, and the fuse is always allowed to blow. The fuse blowing scheme is also called trip saving or breaker saving. Figure 30.18 shows a comparison of the sequence of events of each mode of operation. Fuse saving is primarily directed at reducing sustained interruptions, and fuse blowing is primarily aimed at reducing the number of momentary interruptions.

30.8.1 Industry Usage

Until the late 1980s, fuse saving was almost universally used. As power quality concerns grew, some utilities switched to a fuse blowing mode. An IEEE survey on distribution protection practices that is done periodically has shown a decrease in the use of fuse saving as shown in Table 30.11.

Another survey done by Power Technologies, Inc., in 1996 showed a mixture of practices at utilities as shown in Table 30.12. A few used fuse blowing because they indicated that fuse saving was not successful. Many of the “mixed practices” utilities decided on a case-by-case basis. Many of these normally used fuse saving but switched to fuse blowing if too many power quality complaints were received.

30.8.2 Effects on Momentary and Sustained Interruptions

The change in the number of momentary interruptions can be estimated simply by using the ratio of the length of the mains to the total length of the circuit including all laterals. For example, if a circuit has 5 mi of mains and 10 mi of laterals, the number of momentaries after switching to fuse blowing would be 1/3 of the number of momentaries with fuse saving (5/(5 + 10) = 1/3). This assumes that the mains and laterals have the same fault rate; if the fault rate on laterals is higher (which it often is because of less tree trimming, etc.), the number of momentaries is even less. Note how dramatically we can reduce momentaries by using fuse blowing. No other methods can so easily eliminate 30%–70% of momentaries. The effect on reliability of going to a fuse blowing scheme is more difficult to estimate. Fuse blowing increases the number of fuse operations by 40%–500% (Dugan et al., 1996; Short, 1999; Short and Ammon, 1997). This will increase the average frequency of sustained interruptions by 10%–60%. Note that there are many variables that can change the ratios. One example is given in Figure 30.19.

Note that the effect on sustained interruptions is not equally distributed. Customers on the mains see no difference in the number of permanent interruptions. Customers on long laterals may have many more sustained interruptions with a fuse-blowing scheme.

30.8.3 Coordination Limits of Fuse Saving

One of the main reasons that utilities have decided not to use fuse saving is that it is difficult to make it work. Fuses clear quickly relative to circuit breakers, so where fault currents are high, the fuse blows before the breaker trips. This results in a fuse operation and a momentary interruption for all customers on the circuit. K links, the most common lateral fuses, are fast fuses. Most distribution circuit breakers take five-cycles to clear. For fuse saving to work, the breaker must open before the fuse blows, so the fuse needs to survive for the time it takes the instantaneous relay to operate (about one cycle) plus the five-cycles for the breaker. As an illustration, Figure 30.20 shows the limit of coordination of a five-cycle breaker and a 100 K fuse. Fuse saving only coordinates for faults below 1354 A. Smaller fuses have lower current limits. Note that the breaker time is coordinated with the damage time of the fuse.

Table 30.13 and Table 30.14 show the limits of coordination of several common lateral fuses for a standard circuit breaker (five-cycle) and for a fast breaker/relay combination (three-cycle circuit breaker and one-cycle relay). Also shown are translations of these fault currents into distances from the substation at 12.47 kV (assuming an 8 kA fault level at the substation). Note that only the larger fuses shown (greater than 100 A) will coordinate for significant portions of the feeder. Smaller fuses used as second- and third-level fuses do not coordinate over the length of most feeders. The situation is even worse at higher voltages. At 24.94 kV, the distances in Table 30.13 and Table 30.14 are doubled, so fuse saving is more difficult to achieve at higher voltages. Reclosers are faster than standard five-cycle breakers—the four-cycle total operating time in Table 30.14 is representative of many reclosers.

If smaller K links are used such as 100 K and 65 K fuses (the most common lateral fuses), then fuse saving is not going to work very well. In that case, why use it? There is no sense in having a momentary every time a fuse blows (which is what will happen since the circuit breaker is not fast enough to save the fuse).

30.8.4 Long-Duration Faults and Damage with Fuse Blowing

Fuse blowing has drawbacks: faults on the mains can last a long time. With fuse saving, main-line faults normally clear in 5 to 7 cycles (0.1 s) on the first shot with the instantaneous element. With fuse blowing, this same fault may last for 0.5–1 s. Much more damage at the fault location occurs during this extra time. Some of the problems that have been identified are as follows:

• Conductor burndowns: At the fault, the heat from the fault current arc burns the conductor enough to break it, dropping it to the ground.
• Damage of inline equipment: The most common problem has been with inline hot-line clamps. If the connection is not good, the high-current fault arc across the contact can burn the connection apart.
• Station transformers: Extra duty on substation transformers.
• Evolving faults: Ground faults are more likely to become two- or three-phase faults.
• Underbuilt: Faults on underbuilt distribution are more likely to cause faults on the transmission circuit above due to rising arc gases.

A fault current arc will expand after it is initiated. It has been found that the growth of the arc is generally in the vertical direction, and the growth is primarily a function of time and not of current or voltage (Drouet and Nadeau, 1979). The growth of the arc means that a 0.1 s fault on the instantaneous trip (with fuse saving) is less likely to involve other phases or other circuits than a 0.2–1 s fault on the time-delay trip (with fuse blowing).

30.8.5 Long-Duration Voltage Sags with Fuse Blowing

With fuse blowing, voltage sags last longer, especially for faults on the three-phase mains, which have to be cleared by phase or ground time–overcurrent elements. An example is shown in Figure 30.21 where voltage sag magnitudes and durations are shown for faults at various distances from a substation using fuse blowing. For the same circuit with fuse saving, all of the faults would have cleared in 0.1 s. For a fault at the substation, the duration triples. For a fault one mile (1.6 km) from the substation, the duration quadruples. The situation is worse for phase-to-phase faults and three-phase faults because they must be cleared by the phase relays which are generally slower.

30.8.6 Optimal Implementation of Fuse Saving

In order to get a fuse-saving scheme to work, it is necessary to get the substation protective device to open before fuses operate. We can achieve this in several ways:

1. Slow down the fuse: Use big, slow fuses (such as a 140 or 200 T) near the substation to ensure proper coordination.
2. Faster breakers or reclosers: If three-cycle circuit breakers are used instead of the normal five-cycle breakers, fuse saving coordination is more likely. Some reclosers are even faster than three-cycle breakers.
3. Limit fault currents:
   1. Open station bus ties: An open bus tie will reduce the fault current on each feeder and make fuse saving easier. This is the normal operating mode for most utilities.
   2. Use a transformer neutral reactor: A neutral reactor reduces the fault current for single-phase faults (all faults on single-phase taps).
   3. Use line reactors: This reduces the fault current for all types of faults. This has been an uncommon practice. An added advantage, reactors reduce the impact of voltage sags for faults on adjacent feeders.
   4. Specify higher impedance transformers.

We can employ other strategies to limit the impact of momentary interruptions:

• More downstream reclosers: Extra downstream devices will reduce the number of momentaries for customers near the substation. It is important to coordinate reclosers with the upstream device (including sequence coordination).
• Single-phase reclosers.
• Immediate reclose.
• Switch to a fuse blowing scheme on poor feeders: For a feeder with many momentaries, disable the instantaneous relay for a time period. Identify poorly performing parts of the circuit during this time. The blown branch fuses provide a convenient fault location method. Once the poor performing sections are identified and improved, switch the circuit back to fuse saving.

30.8.7 Optimal Implementation of Fuse Blowing

Several strategies can optimize a fuse blowing scheme:

• Fast fuses (or current-limiting fuses): If smaller or faster fuses are used, faults clear faster, so voltage sag durations are shorter. Current-limiting fuses also limit the magnitude and duration of the sag. Be careful not to fuse too small, or fuses will operate unnecessarily due to loading, inrush, and cold-load pickup. Note that if smaller fuses are used, it is difficult to switch back to a fuse-saving scheme.
• Covered wire or small wire: Watch burndowns on circuits with covered wire or small wire that is protected by the station circuit breaker or recloser. If either of these cases exists, use a modified fuse-blowing scheme with a time-delayed instantaneous element (see the next section).
• Use single-phase reclosers on longer laterals: A good way of maintaining some of the reliability of a fuse-saving scheme is to use single-phase reclosers instead of fuses on longer taps. Then, temporary faults on these laterals do not cause permanent interruptions to those customers.
• More fuses: Add more second- and third-level fuses to segment the circuit more.
• Track lateral operations: Temporary faults on fused laterals cause sustained interruptions. In order to minimize the impacts on lateral customers, track interruptions by lateral. Identify poorly performing laterals, patrol poor sections, then add tree trimming, animal guards, etc.

30.9 Other Protection Schemes

30.9.1 Time Delay on the Instantaneous Element (Fuse Blowing)

An alternative implementation of a fuse-blowing scheme is to use a time delay on the instantaneous trip (rather than removing the instantaneous trip; a definite-time overcurrent relay also could do the same function) (Engelman, 1990). Faults do not last as long as they would if the relay went to a time–overcurrent element; there is less chance of wire burndowns, and voltage sags are of shorter duration for faults on the mains. A common delay is 0.1 s.

An example implementation is shown in Figure 30.22 where a 0.1 s delay is added to the instantaneous. A 100 K fuse link is also shown. For the 100 K link, the scheme is actually a mixture of fuse saving and fuse blowing. For fault currents above roughly 1700 A, it is a fuse-blowing scheme (the fuse clears before the instantaneous relay operates). For currents below 1000 A, the scheme is a fuse-saving scheme (the circuit breaker trips before the fuse is damaged). Between 1000 and 1700 A, one or both devices operate.

Another option sometimes used with this scheme is a high-set instantaneous. The high-set instantaneous has no time delay and is set to clear faults close to the station. This removes the most damaging faults quickly (because they are the most likely to cause damage and cause the most severe voltage sags).

Using a time delay is a better fuse-saving scheme than just removing the instantaneous relay. The disadvantage, and the reason that it is not implemented as much, is that it is usually more difficult and costly to implement. For electromechanical relays, another timer relay must be added, and the relay scheme must be engineered. Many digital relays ease the implementation since they have this time delay option available.

30.9.2 High–Low Combination Scheme

Another option is to use fuse blowing at the substation and fuse saving at downstream reclosers (Burke, 1996).

• Substation fuse blowing: Fault currents are high near the substation, so it is difficult to get fuse saving to work here.
• Recloser fuse saving: Fault currents are lower downstream, and reclosers are faster, so fuse saving should work well here.

The high–low scheme is easy to implement. The station instantaneous trip is eliminated. Reclosers are operated with a fast trip (the A curve). Most are already in this mode, so no changes are necessary here.

30.9.3 SCADA Control of the Protection Scheme

Another option is to use SCADA to change back and forth between fuse saving and fuse blowing, getting some of the benefits of both schemes. Fuse blowing is the normal operating mode, but operators could switch to a fuse-saving scheme during storms. This avoids clear-sky momentaries while at the same time improving storm restoration. Several factors make fuse saving better during storms:

• Faults are more likely to be temporary during storms (lightning, wind).
• Customers are more forgiving about momentaries during storms.
• Interruptions due to fuse operations last longer during storms (because crews have many repairs to perform). If fuses are blown due to temporary faults, this increases the number of repair locations. Saving fuses reduces the number of interruptions crews will have to address.

In order for SCADA control of fuse saving to work best, we must design the system for fuse saving to work: larger, slower fuses for laterals close to the substation, faster circuit breakers (or use substation reclosers), or possibly even using grounding reactors in the substation to limit fault currents. Likewise, we must design for fuse blowing, so avoid using tree wire (or go to delayed instantaneous relaying rather than removing the instantaneous trip).

Control is more readily available in the substation because the SCADA infrastructure may already be in place. If so, the cost of the SCADA system has already been justified, and this added functionality could be piggybacked on the existing system if there are free channels available. It is feasible to use automation technology to implement remote control of feeder reclosers, but the cost of the communication equipment may not justify having this functionality.

For SCADA control, microprocessor-controlled relays are not needed. A SCADA channel can be used to control a blocking relay on the instantaneous elements of the feeder relays. Alternatively, the SCADA channel could control the delay on the instantaneous relay element (no delay: fuse saving, with delay: fuse blowing). One SCADA channel could control the fuse saving/blowing status of all of the distribution feeders in a station. Alternatively, we could control each feeder independently.

30.9.4 Adaptive Control by Phases

Various protection schemes are classified as adaptive. An adaptive approach to a fuse blowing mode is to adjust the scheme depending on how many phases are faulted:

• Two- or three-phase fault: Use the instantaneous; the fault is assumed to be on the three-phase mains. Tripping quickly reduces the duration of voltage sags for faults on the mains.
• Single-phase fault: Use fuse blowing (time delay curves or delayed instantaneous relay).

Adaptive control requires microprocessor-based relays. This is not a common scheme, and the expense and complexity are difficult to justify unless the chosen relay comes with this functionality.

30.10 Reclosing Practices

Automatic reclosing is a universally accepted practice on most overhead distribution feeders. On overhead circuits, 50%–80% of faults are temporary, so if a circuit breaker or recloser clears a fault and it recloses, most of the time the fault is gone, and customers do not lose power for an extended period of time.

On underground circuits, since virtually all faults are permanent, we do not reclose. A circuit might be considered underground if something like 60%–80% of the circuit is underground. Utility practices vary considerably relative to the exact percentage (IEEE Working Group on Distribution Protection, 1995). A significant number of utilities treat a circuit as underground if as little as 20% is underground while some others put the threshold over 80%.

The first reclose usually happens with a very short delay, either an immediate reclose which means a 1/3–1/2 s dead time (discussed later) or with a 1–5 s delay. Subsequent reclose attempts follow longer delays. The nomenclature is usually stated as 0–15–30 meaning there are three reclose attempts: the first reclose indicated by the “0” is made after no intentional delay (this is an immediate reclose), the second attempt is made following a 15 s dead time, and the final attempt is made after a 30 s dead time. If the fault is still present, the circuit opens and locks open. We also find this specified using circuit breaker terminology as O-0 s-CO-15 s-CO-30 s-CO where “C” means close and “O” means open. Other common cycles that utilities use are 0–30–60–90 and 5–45.

With reclosers and reclosing relays on circuit breakers, the reclosing sequence is reset after an interval that is normally adjustable. This interval is generally set somewhere in the range of 10–2 min. Only a few utilities have reported excessive operations without lockout (IEEE Working Group on Distribution Protection, 1995).

30.10.1 Reclose Attempts and Dead Times

Three reclose attempts is most common as shown in Figure 30.23. More reclose attempts give the fault more chance to clear or burn free. Returns diminish; the chance that the third or fourth reclose attempt is successful is usually small. Additional reclose attempts have the following negative impacts on the system:

• Additional damage at the fault location: With each reclose into a fault, arcing does additional damage at the fault location. Faults in equipment do more damage. Cable faults are harder to splice, wire burndowns are more likely, and oil-filled equipment is more likely to rupture. Arcs can start fires. Faults (and the damage the arcs cause) can propagate from one phase to other phases.
• Voltage sags: With each reclose into a fault, customers on adjacent circuits are hit with another voltage sag. It can be argued that the magnitude and duration of the sag should be about the same, so depending on the type of device, if the customer equipment survived the first sag, it will probably ride through subsequent sags of the same severity. If additional phases become involved in the fault, the voltage sag is more severe.
• Through-fault damage to transformers: Each fault subjects transformers to mechanical and thermal stresses. Virginia Power changed their reclosing practices because of excessive transformer failures on their 34.5 kV station transformers due to through faults (Johnston et al., 1978).
• Through-fault damage to other equipment: Cables, wires, and especially connectors suffer the thermal and mechanical stresses of the fault.
• Interrupt ratings of breakers: Circuit breakers must be derated if the reclose cycle involves more than one reclose attempt within 15 s. This may be a consideration if fault currents are high and breakers are near their ratings. Reclosers do not have to be derated for a complete four-sequence operation. Extra reclose attempts increase the number of operations, which means more frequent breakers and reclosers maintenance.
• Ratcheting of overcurrent relays: An induction relay disc turns in response to fault current. After the fault is over, it takes time for the disk to spin back to the neutral position. If this reset is not completed, and another fault occurs, the disk starts spinning from its existing condition, making the relay operate faster than it should. The most common problem area is miscoordination of a substation feeder relay with a downstream feeder recloser. If a fault occurs downstream of the recloser, the induction relay will spin due to the current (but not operate if it is properly coordinated). Multiple recloses by the recloser could ratchet the station relay enough to falsely trip the relay. The normal solution is to take the ratcheting into account when coordinating the relay and recloser, but in some cases modification of the reclosing cycle of the recloser is an option. Another option is to use digital relays, which do not ratchet in this manner.

Given these concerns, the trend has been to decrease the number of reclose attempts. We try to balance the loss in reliability against the problems caused by extra reclose attempts. A major question is how often are the extra reclose attempts successful. Table 30.15 shows the success rate for one utility in a high-lightning area. Table 30.16 shows a second utility with similar reclosing practices but quite different reclose success (more lockouts and lower success rates for the first two reclose attempts). Reclose success rates change based on the types of faults most commonly seen in a region. Another data point with a broader distribution of utilities is obtained in the EPRI distribution power quality study. Table 30.17 shows the number of momentary interruptions (reclose attempts) that do not lead to sustained interruptions (lockouts). The key point is that it is relatively uncommon (but not rare) for the third or fourth reclose attempt to be successful.

We may block reclosing in some cases. It is common to block all reclose attempts when workers are doing maintenance on a circuit to provide an extra level of protection (an instantaneous relay element is also commonly enabled in this situation). Another situation is for very high-current faults. A high-set instantaneous relay covering just the first few hundred feet of circuit detects faults on the substation exit cables. If it operates, reclosing is disabled. This practice is done to reduce the damage for a failure of one of the station exit cables.

The duration of the open interval—the dead time between reclose attempts—is also a consideration. For a smaller number of reclose attempts, use longer delays to give tree branches and other material more time to clear.

Operator practices must also be considered as part of the reclosing scheme. Not uncommonly, an operator manually recloses the circuit breaker after a feeder lockout (especially during a storm). This sends the breaker or recloser through its whole reclosing cycle along with all of the bad effects (like more equipment damage and more voltage sags) with very little chance of success.

Some engineers and field personnel believe that the purpose of the extra reclose attempts is to burn the fault clear. This is dangerous. Faults regularly burn clear on low-voltage systems (<480 V), rarely at distribution primary voltages. Faults can burn clear on primary systems. The most common example is that tree branches or animals can be burned loose. The problem with this concept is that, just as easily, the fault burns the primary conductor, which falls to the ground causing a high-impedance fault. Fires and equipment damage are also more likely with the “burn clear” philosophy.

To reduce the impacts of subsequent reclose attempts, we could switch back to an instantaneous operation after the first time–overcurrent relay operation. If the fault does not clear after the first time–overcurrent relay operation, it means the fault is not downstream of a fuse (or a recloser). The reason to use a time overcurrent relay is to coordinate with the fuse. Since the fuse is out of the picture, why not use a faster trip for subsequent reclose attempts? While not commonly done, we could implement this with digital relays. The setting of the “subsequent reclose” instantaneous relay element should be different than the first-shot instantaneous. Set the pickup at the pickup of the time–overcurrent relay. Because of inrush on subsequent attempts, we may use a fast time–overcurrent curve or an instantaneous element with a short delay (something like five-cycles).

As an example, if a utility uses a 0–15–30–90 s reclosing cycle, the system is subjected to five faults if the system goes through its complete cycle. With the instantaneous operation enabled on the first attempt and disabled on subsequent attempts, we have a very high total duration of the fault current. For a CO-11 ground relay with a time dial of 3, a 2 kA fault clears in roughly 1 s. For the reclosing cycle to lock out, the system has a total fault time of 4.1 s (one 0.1 s fault followed by four 1 s faults). If the instantaneous operation is enabled for reclose attempts 2 through 4, the total fault duration is 1.4 s (one 0.1 s fault followed by a 1 s fault and three 0.1 s faults). This greatly reduces the damage done by certain faults.

30.10.2 Immediate Reclose

An immediate reclose (also called an instantaneous or fast reclose) means having no intentional time delay (or a very short time delay) on the first reclose attempt on circuit breakers and reclosers. Many residential devices such as digital clocks, VCRs, and microwaves can ride through a 1/2 s interruption but not a 5 s interruption, so a fast reclose helps reduce residential complaints.

From a power quality point of view, a faster reclose is better. Some customers may not notice anything more than a quick blink of the lights. Many residential devices such as the digital clocks on alarm clocks, microwaves, and VCRs can ride through a 1/2 s interruption where they usually cannot ride through a 5 s interruption (a first reclose delay used by several utilities).

30.10.2.1 Effect on Sensitive Residential Devices

The most common power quality recorder in the world is the digital clock. Many complaints are due to the “blinking clocks.” Using an immediate reclose reduces complaints. Florida Power has reported that a reclosing time of 18–20 cycles nearly eliminates complaints (Dugan et al., 1996). Another utility that has successfully used the immediate reclose is Long Island Lighting Company (now Keyspan) (Short and Ammon, 1997). According to an IEEE survey, a time to first reclose of less than 1 s is the most common practice although the fast reclose practice tends to decline with increasing voltage (see Figure 30.24).

Clocks have a wide range of voltage sensitivity, but most digital clocks will not lose memory for a complete interruption that is less than 0.5 s. So, an immediate reclose helps residential customers ride through momentary interruptions without resetting many devices. Given the wide variation, some customers are sensitive to a 0.5 s interruption. Note that the immediate reclose helps with digital clock-type devices whether it be on radio alarm clocks, VCRs, or microwaves. Fast reclosing does not help with most computers or other computer-based equipment, limiting the power quality improvement of using the immediate reclose to residential customers (no help for commercial or industrial customers).

30.10.2.2 Delay Necessary to Avoid Retriggering Faults

Sometimes a delayed reclose is necessary if there is not enough time to clear the fault. A fault arc needs time to cool, or the reclose could retrigger the arc. Whether the arc strikes again is a function of voltage and structure spacings. A 34.5 kV utility (Vepco) added a delay to the first reclose because the probability of success of the first reclose was much less than normal for distribution circuits (Johnston et al., 1978). The success rate for the first attempt after an instantaneous reclose was 25% which is much less than the 70%–80% experienced by most utilities. Another item that added to the low success rate of Vepco’s 34.5 kV system is that they used a lot of armless design, and the combination of higher voltage and tighter spacings requires a longer time delay for the arc to clear.

With the following equation, we can find the minimum deionization time of an arc based on the line-to-line voltage (Westinghouse Electric Corporation, 1982):

$t=10.5+\frac{V}{34.5}$

where

t is the minimum deionization time, 60 Hz cycles

V is the rated line-to-line voltage, kV

The deionization time increases only moderately with voltage. Even for a 34.5 kV system, the deionization time is 11.5 cycles. This equation is a simplification (separation distances are not included) but does show that arcs rapidly deionize. Many high-voltage transmission lines successfully use a fast reclose. The reclose time for distribution circuit breakers and reclosers varies by design. A typical time is 0.4–0.6 s for an immediate reclose (meaning no intentional delay). The fastest devices (newer vacuum or SF6 devices) may reclose in as little as 11 cycles. This may prove to be too fast for some applications, so consider adding a small delay of 0.1–0.4 s (especially at 25 or 35 kV).

On distribution circuits, other things affect the time to clear a fault besides the deionization of the arc stream. If a temporary fault is caused by a tree limb or animal, time may be needed for the “debris” to fall off the conductors or insulators. Because of this, with an immediate reclose use at least two reclose attempts before lockout. For example, use a 0–15–30 s cycle (three reclose attempts), or if you wish to use two reclose attempts, use a 0–30 or 0–45 s cycle (use a long delay before the last reclose attempt).

30.10.2.3 Reclose Impacts on Motors

Industrial customers with large motors have concerns about a fast reclose and damage to motors and their driven equipment. The major problem with reclosing is that the voltage on a motor will not drop instantly to zero when the utility circuit breaker (or recloser) is opened. The motor has residual voltage, where the magnitude and frequency decay with time. When the utility recloses, the utility voltage can be out of phase with the motor residual voltage, severely stressing the motor windings and shaft and its driven load. The decay time is a function of the size of the motor and the inertia of the motor and its load.

Motors in the 200–2000 hp range typically have open-circuit time constants of 0.5–2 s (Bottrell, 1993). The time constant is the time it takes for the residual voltage to decay to 36.8% of its initial value. Reclose impacts are worse with

• Larger motors
• Capacitor banks—excitation from the capacitor banks can greatly increase the motor decay time
• Synchronous motors and generators—much larger time constants makes synchronous machines more vulnerable to damage than induction machines

On the vast majority of distribution circuits, reclosing impacts will not be a concern because of the following:

• Motors on contactors will drop out. Also, larger motors and synchronous motors normally have an undervoltage relay to trip when voltage is lost.
• Most utility feeders do not have individual motor loads larger than 500 hp.
• Even with feeders with large industrial customers, the non-motor load will be large enough to pull the voltage down to a safe level within the time it takes to do a normal immediate reclose (0.4–0.6 s).

Because of this, we can safely implement an immediate reclose on almost all distribution circuits. One exception is a feeder with an industrial customer that is a majority of the feeder load, and the industrial customer has several large induction or (especially) synchronous motors. Another exception is a feeder with a large rotating distributed generator. In both of these cases, delay the first reclose or, alternatively, use line-side voltage supervision (if voltage is detected downstream of the breaker, reclosing is blocked to prevent an out-of-phase reclosing situation).

30.11 Single-Phase Protective Devices

Many distribution protective devices are single phase or are available in single-phase versions including reclosers, fuses, and sectionalizers. Single-phase protective devices are used widely on distribution systems; taps are almost universally fused. On long single-phase taps, single-phase reclosers are sometimes used. Most utilities also use fuses for three-phase taps. The utilities that do not fuse three-phase taps most often cite the problem of single-phasing motors of three-phase customers. Some utilities use single-phase reclosers that protect three-phase circuits (even in the substation).

Single-phase protective devices on single-phase laterals are widely used, and the benefits are universally accepted. The fuse provides an inexpensive way of isolating faulted circuit sections. The fuse also aids in finding the fault.

Using single-phase interrupters helps on three-phase circuits—only one phase is interrupted for line-to-ground faults. We can easily estimate the effect on individual customers using the number of phases that are faulted on average as shown in Table 30.18. Overall, using single-phase protective devices cuts the average number of interruptions in half. This assumes that all customers are single phase and that the customers are evenly split between phases.

Service to three-phase customers downstream of single-phase interrupters generally improves, too. Three-phase customers have many single-phase loads, and the loads on the unfaulted phases are unaffected by the fault. Three-phase devices may also ride through an event caused by a single-phase fault (although motors may heat up because of the voltage unbalance as discussed in the next section). Single-phase protective devices do have some drawbacks. The main concerns are

• Ferroresonance
• Single-phasing of motors
• Backfeeds

Ferroresonance usually occurs during manual switching of single-pole switching devices (where the load is usually an unloaded transformer). It is less common for ferroresonance to occur downstream of a single-phase protective device that is operating due to a fault. The reason for this is that if there is a fault on the opened phase, the fault prevents an overvoltage on the opened phase. Also, any load on the opened section helps prevent ferroresonant overvoltages. Because ferroresonance will be uncommon with single-phase protective devices, it is usually not a major factor in protective device selection. Still, caution is warranted on small three-phase transformers that may be switched unloaded (especially at 24.94 or 34.5 kV).

With single-phase protective devices, backfeeds can create hazards. During a line-to-ground fault where a single-phase device opens, backfeed through a three-phase load can cause voltage on the load side of an opened protective device. Backfeeds can happen with most types of three-phase distribution transformer connections (even with a grounded-wye–grounded-wye connection). The important points to note are as follows:

• The backfeed voltage is enough to be a safety hazard to workers or the public (e.g., in a wire down situation).
• The available backfeed is a stiff enough source to maintain an arc of significant length. The arc can continue causing damage at the fault location during a backfeed condition. It may also be a low-level sparking and sputtering fault.

Based on these points, single-phasing can cause problems from backfeeding. Whether this constrains use of single-phase protective devices is debatable. Most utilities do use single-phase protective devices, usually with fuses, on three-phase circuits.

Under single-phasing, motors can overheat and fail. Motors have relatively low impedance to negative-sequence voltage; therefore, a small negative-sequence component of the voltage produces a relatively large negative-sequence current. Consequently the effect magnifies; a small negative-sequence voltage appears as a significantly larger percentage of unbalanced current than the percentage of unbalanced voltage.

Loss of one or two phases is a large unbalance. For one phase open, the phase-to-phase voltages become 0.57, 1.0, 0.57 for a wye–wye transformer and 0.88, 0.88, 0.33 for a delta–wye transformer. In either case, the negative-sequence voltage is 0.66 per unit. With such high unbalance, a motor overheats quickly. The negative-sequence impedance of a motor is roughly 15%, so for a 66% negative-sequence voltage, the motor draws a negative-sequence current of 440%.

Most utility service agreements with customers state that it is the customer’s responsibility to protect their equipment against single phasing. The best way to protect motors is with a phase-loss relay. Nevertheless, some utilities take measures to reduce the possibility of single-phasing customers’ motors, and one way to do that is to limit the use of single-phase protective devices. Other utilities are more aggressive in their use of single-phase protective equipment and leave it up to customers to protect their equipment.

30.11.1 Single-Phase Reclosers with Three-Phase Lockout

Many single-phase reclosers and recloser controls come with a controller option for a single-phase trip and three-phase lockout. Three-phase reclosers that can operate each phase independently are also available. For single-phase faults, only the faulted phase opens. For temporary faults, the recloser successfully clears the fault and closes back in, so there will only be a momentary interruption on the faulted phase. If the fault is still present after the final reclose attempt (a permanent fault), the recloser trips all three phases and will not attempt additional reclosing operations.

Problems of single-phasing motors, backfeeds, and ferroresonance disappear. Single-phasing motors and ferroresonance cause heating, and heating usually takes many minutes for damage to occur. Short-duration single-phasing occurring during a typical reclose cycle does not cause enough heat to do damage. If the fault is permanent, all three phases trip and lock out, so there is no long-term single phasing. A three-phase lockout also reduces the chance of backfeed to a downed wire for a prolonged period.

Single-phase reclosers are available that have high enough continuous and interrupting ratings that utilities can use them in almost all feeder applications and many substation applications.

Another consideration with single-phase reclosers vs. three-phase devices is that a ground relay is often not available on single-phase reclosers. A ground relay provides extra sensitivity for line-to-ground faults. Not having the ground relay is a tradeoff to using single-phase devices. Even if a ground relay is available on a unit with single-phase tripping, if the ground relay operates, it trips all three phases (which defeats the purpose of single-phase tripping).

References

Amundson, R. H., High voltage fuse protection theory & considerations, IEEE Tutorial Course on Application and Coordination of Reclosers, Sectionalizers, and Fuses, 1980. Publication 80 EHO157–8-PWR.

ANSI C57.12.20-1988, American national standard requirements for overhead-type distribution transformers, 500 kVA and smaller: High-voltage, 67,000 volts and below; low-voltage, 15,000 volts and below.

ANSI C37.42-1989, American national standard specifications for distribution cutouts and fuse links.

ANSI C37.06-1997, AC high-voltage circuit breakers rated on a symmetrical current basis—Preferred ratings and related required capabilities.

ANSI/IEEE C37.60-1981, IEEE standard requirements for overhead, pad mounted, dry vault, and submersible automatic circuit reclosers and fault interrupters for AC systems.

ASTM F1959, Standard Test Method for Determining the Arc Rating of Materials for Clothing, ASTM International, West Conshohocken, PA, 2006.

Benmouyal, G. and Zocholl, S. E., Time–current coordination concepts, in Western Protective Relay Conference, Spokane, WA, 1994.

Bingwu, G. and Chengkang, W., The gasdynamic and electromagnetic factors affecting the position of arc roots in a tubular arc heater, Acta Mechanica Sinica, 7(3), 199–208, 1991.

Bottrell, G. W., Hazards and benefits of utility reclosing, in Proceedings of Power Quality, 1993.

Burke, J. J., Philosophies of distribution system overcurrent protection, in IEEE/PES Transmission and Distribution Conference, Los Angeles, CA, 1996.

Burke, J. J., High impedance faults (40 ohms is a fallacy), IEEE Rural Electric Power Conference, Albuquerque, NM, April 9–11, 2006.

CEA 288 D 747, Application Guide for Distribution Fusing, Canadian Electrical Association, Ottawa, Ontario, Canada, 1998.

Cook, C. J. and Niemira, J. K., Overcurrent protection of transformers—Traditional and new philosophies for small and large transformers, in IEEE/PES Transmission & Distribution Conference and Exposition, 1996. Presented at the training session on “Distribution overcurrent protection philosophies.”

Cooper Power Systems, Comparison of Recloser and Breaker Standards, Reference data R280-90-5, February 1994.

Cooper Power Systems, Electrical Distribution—System Protection, 3rd edn., 1990.

Cress, S. L., Arc modeling and hazard assessment, EPRI Distribution Arc Flash Workshop, Chicago, IL, April 2008.

Dagenhart, J., The 40-ohm ground-fault phenomenon, IEEE Transactions on Industry Applications, 36(1), 30–32, 2000.

Distribution Line Protection Practices Industry Survey Results IEEE Power System Relaying Committee Report, December 2002. http://www.pes-psrc.org/Reports/DISTRIBUTION%20LINE%20PROTECTION%20PRACTICES%20report%20final%20dec20.pdf

Drouet, M. G. and Nadeau, F., Pressure waves due to arcing faults in a substation, IEEE Transactions on Power Apparatus and Systems, PAS-98(5), 1632–1635, 1979.

Dugan, R. C., Ray, L. A., Sabin, D. D., Baker, G., Gilker, C., and Sundaram, A., Impact of fast tripping of utility breakers on industrial load interruptions, IEEE Industry Applications Magazine, 2(3), 55–64, May/June 1996.

Eblen, M. L. and Short, T. A., Arc flash testing of typical 480-V utility equipment, IEEE IAS Electrical Safety Workshop, Memphis, TN, February 1–5, 2010. Paper ESW2010-05.

Engelman, N. G., Relaying changes improve distribution power quality, Transmission and Distribution, pp. 72–76, May 1990.

EPRI 1001665, Power Quality Improvement Methodology for Wire Companies, Electric Power Research Institute, Palo Alto, CA, 2003.

EPRI 1018693, Distribution Arc Flash: Analysis Methods and Arc Characteristics, Electric Power Research Institute, Palo Alto, CA, 2009. http://my.epri.com/portal/server.pt?Abstract_id=000000000001018693

EPRI 1020210, Distribution Arc Flash: 480-V Padmounted Transformers and Network Protectors, Electric Power Research Institute, Palo Alto, CA, 2009. http://my.epri.com/portal/server.pt?Abstract_id=000000000001020210

EPRI TR-106294-V2, An Assessment of Distribution System Power Quality: Volume 2: Statistical Summary Report, Electric Power Research Institute, Palo Alto, CA, 1996.

Gabrois, G. L., Huber, W. J., and Stoelting, H. O., Blowing of distribution transformer fuses by lightning, in IEEE/PES Summer Meeting, Vancouver, British Columbia, Canada, 1973. Paper C73-421-5.

Hamel, A., Dastous, J. B., and Foata, M., Estimating overpressures in pole-type distribution transformers. Part I: Tank withstand evaluation, IEEE Transactions on Power Delivery, 18(1), 113–119, 2003.

Hamel, A., St. Jean, G., and Paquette, M., Nuisance fuse operation on MV transformers during storms, IEEE Transactions on Power Delivery, 5(4), 1866–1874, October 1990.

Hi-Tech Fuses, Bulletin FS-10: Application of trans-guard EXT and OS fuses, in Hi-Tech Fuses, Hickory, NC, 2002.

IEEE C2-2007, National Electrical Safety Code.

IEEE Std 1584-2002, IEEE Guide for Performing Arc-Flash Hazard Calculations.

IEEE Std. C37.40-1993, IEEE standard service conditions and definitions for high-voltage fuses, distribution enclosed single-pole air switches, fuse disconnecting switches, and accessories.

IEEE Std. C37.112-1996, IEEE standard inverse–time characteristic equations for overcurrent relays.

IEEE Std. C37.48-1997, IEEE guide for the application, operation, and maintenance of high-voltage fuses, distribution enclosed single-pole air switches, fuse disconnecting switches, and accessories.

IEEE Std. C62.22-1997, IEEE guide for the application of metal-oxide surge arresters for alternating-current systems.

IEEE Std. C37.04-1999, IEEE standard rating structure for AC high-voltage circuit breakers.

IEEE Std. C37.010-1999, IEEE application guide for AC high-voltage circuit breakers rated on a symmetrical current basis.

IEEE Working Group on Distribution Protection, Distribution line protection practices—Industry survey results, IEEE Transactions on Power Delivery, 10(1), 176–186, January 1995.

IEEE Working Group on Distribution Protection, Distribution line protection practices—Industry survey results, IEEE Transactions on Power Delivery, 3(2), 514–524, April 1988.

Johnston, L., Tweed, N. B., Ward, D. J., and Burke, J. J., An analysis of Vepco’s 34.5 kV distribution feeder faults as related to through fault failures of substation transformers, IEEE Transactions on Power Apparatus and Systems, PAS-97(5), 1876–1884, 1978.

Kinectrics, User’s Guide for ARCPRO, version 2.0, Toronto, Ontario, Canada, 2002.

Lunsford, J. M. and Tobin, T. J., Detection of and protection for internal low-current winding faults in overhead distribution transformers, IEEE Transactions on Power Delivery, 12(3), 1241–1249, July 1997.

Plummer, C. W., Goedde, G. L., Pettit, E. L. J., Godbee, J. S., and Hennessey, M. G., Reduction in distribution transformer failure rates and nuisance outages using improved lightning protection concepts, IEEE Transactions on Power Delivery, 10(2), 768–777, April 1995.

REA Bull. 61-2, Guide for Making a Sectionalizing Study on Rural Electric Systems, Rural Electrification Administration, U.S. Department of Agriculture, Washington, DC, 1978.

Short, T. A., Fuse saving and its effect on power quality, in EEI Distribution Committee Meeting, pp. 87–100, March 29–31, Phoenix, AZ, 1999.

Short, T. A., Electric Power Distribution Handbook, CRC Press, Boca Raton, FL, 2004.

Short, T. A., Arc-flash analysis approaches for medium-voltage distribution, IEEE Transactions on Industry Applications, 47(4), 1902–1909, 2011.

Short, T. A. and Ammon, R. A., Instantaneous trip relay: Examining its role, Transmission and Distribution World, 49(2), 60–63, February 1997.

Short, T. A. and Eblen, M. L., Medium-voltage arc flash in open air and padmounted equipment, IEEE Rural Electric Power Conference (REPC), Chattanooga, TN, April 10–13, 2011.

Smith, D. R., Network primary feeder grounding and protection, in Draft 3 of Chapter 10 of the IEEE Network Tutorial, 1999.

Smithley, R. S., Normal relay settings handle cold load, Electrical World, pp. 52–54, June 15, 1959.

Westinghouse Electric Corporation, Applied Protective Relaying, 1982.