The following characteristics should help to ensure accuracy as well as ease of interpretation:

a) Keep it simple. A fundamental single-line diagram should be made up of short, straight lines and components, similar to the manner in which a block diagram is drawn. It should be relatively easy to get the overall picture of the whole electrical system.

All, or as much as possible, of the system should be kept to one sheet. If the system is very large, and more than one sheet is necessary, then the break should be made at voltage levels or at distribution centers.

b) Maintain relative geographic relations. In many cases, it is possible to superimpose a form of the one-line diagram onto the facility plot plan. This is very helpful toward a quick understanding of the location of the system's major components for operating purposes.

It may, however, be more difficult to comprehend the overall system operation from this drawing. Such a drawing could be used for relatively simple systems. For more complex systems, however, it should be used in addition to the fundamental single-line diagram.

c) Maintain the approximate relative positions of components when producing the single-line diagram. The drawing should be as simple as possible and should be laid out in the same relationship as an operator would view the equipment. The diagram does not need to show geographical relationships at the expense of simplicity.

NOTE: A site plan with equipment locations may be required to accompany the single-line diagram.

d) Avoid duplication. Each symbol, figure, and letter has a definite meaning. The reader should be able to interpret each without any confusion. In this regard, equipment names should be selected before publishing the document; then, these names should be used consistently.

e) Show all known factors. All details shown on the diagram are important. Some of those important details are as follows:

-Manufacturers' type designations and ratings of apparatus;
-Ratios of current and potential transformers and taps to be used on multi-ratio transformers;
-Connections of power transformer windings;
-Circuit breaker ratings in volts, amperes, and short-circuit interrupting rating;
-Switch and fuse ratings in volts, amperes, and short-circuit interrupting rating;
-Function of relays. Device functions used should be from IEEE Std C37.2-1991;
-Ratings of motors, generators, and power transformers;
-Number, size, and type of conductors;
-Voltage, phases, frequency, and phase rotation of all incoming circuits. The type of supply system (wye or delta, grounded or ungrounded) and the available short-circuit currents should be indicated.

f) Future plans . When future plans are known, they should be shown on the diagram or explained by notes.

g) Other considerations. Refer to IEEE Std 141-1993 for further discussion of single line diagrams.


The selection of any circuit breaker, for any given duty, is ultimately based on an assessment of its ability to perform the following basic functions:

a) To carry the required full-load current without overheating (i.e., it should have the correct current rating),

b) To switch and isolate or disconnect the load from the source at the given system voltage (i.e., it should have the correct voltage rating),

c) To interrupt any possible abnormally high operating current or short-circuit current likely to be encountered during operation (i.e., it should have the correct interrupting rating), and

d) To be able to perform these functions over an acceptably long period of time under the operating and environmental conditions that will actually prevail in the application (i.e., it should have the correct mounting provisions, enclosure, and accessories for operation in the environment in which it is to be applied).

The degree to which a circuit breaker can satisfy these requirements is a measure of its applicability for a function. A circuit breaker's rating indicates these capabilities to the user because rating is established by proof testing. Hence, an understanding of how a circuit breaker of any given type is tested will give insight to its applicability for any function.

The role of standards
The primary vehicle for ensuring commonality in performance among circuit breakers of the same rating produced by different manufacturers is a product standard. Standards represent the consensus of manufacturers about what a given product should be able to do as a minimum.

Standards establish the design tests that each manufacturer must perform and pass in order to claim a rating and to be in compliance with that standard. Some standards include requirements for periodic follow-up testing which, in effect, continues to sample the capabilities of newly manufactured circuit breakers.

This assures that they maintain the capabilities of their product ratings. Standards provisions also provide for monitoring the quality of the materials used in the construction of circuit breakers and the quality of the workmanship in the manufacturing process.

As stated previously, standards requirements for the different classes of circuit breakers establish a basis for minimum performance. Circuit breakers may prove by test to perform better than their product ratings indicate, but they can never be permitted to perform worse.

The user, however, may never assume that a circuit breaker can perform better than its rating indicates and should realize that there are manufacturing variations among mass-produced products. The levels of performance required by the standards for the minimum acceptable performance of different classes of circuit breakers will be the primary references in the discussion that follows.


Coordination is a systematic application of current actuated devices in a power system, which in response to a fault or overload will remove only a minimum amount of equipment from service. The objective is to minimize the equipment damage.

A coordination study provides data useful for selection of instrument transformers, protective relay characteristics and settings, fuse ratings, and other information pertinent to provision of optimum protection and selectivity in coordinating these devices.

Planning and Data Collection.
The following data and initial planning steps are required before a coordination study is started:
• Single-line diagram of the electrical system with details of equipment ratings.
• Load flow data and short-circuit data. The maximum and minimum available short-circuit currents, both for phase and ground faults at each relay location in the system.
• Time-current curves, setting ranges, type of characteristics of the protective devices, instrument transformer connections and ratios.
• Power and voltage ratings and winding connections of all power transformers.
• Normal and emergency switching conditions.
• Transformer impedance data, generator fault decrement curves, equivalent impedances of the interconnected networks, conductor sizes, type and configurations and method of installations.

Coordinating Time Intervals.
When plotting coordination curves, certain time intervals must be maintained between curves of various protective devices in order to ensure the desired selectivity. These intervals take into account the circuit breaker interrupting time, relay overtravel and an arbitrary safety factor to take into account current transformer errors and tolerances in the relay characteristics.

For relayed medium-voltage circuit breakers, interrupting time five cycles, and very inverse and extremely inverse electromagnetic relays, a CTI of 0.4 s is adequate. For solid static relays this can be reduced to 0.3 s because relay overtravel is eliminated.

Relayed circuit breakers with electromagnetic relays can be coordinated with downstream fuses with 0.2 s CTI, which can be reduced to 0.1s with static relays. Coordination between fuses for a time duration of less than 0.01s should not be evaluated on a time-current basis.

Two series connected instantaneous devices will coordinate if the maximum let-through I2t of the downstream device is less than minimum I2t let-through of the upstream devices. Coordination between instantaneous relays without an intervening impedance is generally not possible.

An Example of Protection and Coordination in an Industrial Distribution System. Figure 11 shows the phase overcurrent coordination of the protective devices for 1862.5 kW (2500 hp) motor, 2.4 kV main and feeder breakers, and motor contactor interrupting ratings. The 1862.5 kW (2500 hp) motor is controlled by NEMA 2 motor starter, consisting of a 700 A vacuum contactor and a 650 A type R fuse.

The selected fuse should be the smallest whose minimum melting time characteristics does not cross the motor overload relay for currents less than the adjusted locked rotor current of the motor. The adjusted locked rotor current is taken 10% higher than the actual locked rotor current to account for system voltage variations and manufacturing tolerances.

In order to coordinate with the selected motor fuse, the pick-up settings on overload relays of feeder breaker L serving 2.4 kV control center are set at 1920 A. This exposes the circuit breaker to 160% of its continuous current rating; however, this compromise is acceptable because practically low level of short-circuit currents will not be sustained and each load at the control center has its own overcurrent protection.

The coordination between fuse and 50 MVA interrupting rating of the vacuum contactor for a drop out time of 0.02 cycles is not achieved, and there is a possibility of the contactor clearing a fault current exceeding its interrupting rating.

The remedial measures for this situation can be (1) delaying the opening of the motor contactor, (2) connecting the 1862.5 kW (2500 hp) motor to 4.16 kV system, or (3) devising a special design of the 1862.5 kW (2500 hp) motor to reduce the locked rotor current permitting a lower fuse size.

The example illustrates the judgment that a protection engineer should make in accepting compromises in a given situation for arriving at an acceptable engineering solution.

Recent Trends
Recent trends in protective relaying are being dictated by advancements in electronics, microprocessor technology, programming, and packaging. It would have been impossible to detect an impeding bearing failure in a motor using electromagnetic devices; however, neural network techniques to characterize the current spectra associated with a normal state of a motor and load makes the detection possible by monitoring the changes in the bearing frequencies as reflected in the current spectra.

Developments in high-impedance fault detection (HIFDs) fault localization systems, charge comparison type of current deferential relaying and adaptive relaying are further examples. More knowledge-based systems and new algorithms will be applied to protective relaying, coordination, service restoration, and remedial control actions.

Multifunction microprocessor-based relays make it possible to integrate a number of protective functions, metering data, fault location, remote communication, and data logging in a single modular package unit.

As an example, most of the generator protective functions are available in a single unit with added facilities of self-diagnostics, communications, and fault data capture.

Figure 1.
Phase overcurrent device coordination: 1862.5 kW (2500 hp) motor, 2.4 kV main and feeder breakers, and fused motor contactor. 1:Motor full load current = 545 A; 2: Motor relay pickup = 646 A; 3: Motor thermal damage curve; 4: Motor locked rotor current; 5: Motor adjusted locked rotor current; 6: 650 A motor fuse characteristics; 7: dropout time variations, vacuum contactors; 8: dropout time air-break contactors; 9: Inrush current of the motor control center (largest motor); 10: breaker K overcurrent relays pickup = 2800 A; 11: breaker L overcurrent relays pickup = 1920 A; 12: three-phase sym. Short circuit current after 6 cycles = 27.27 kA, with in-plant generator only in service = 11.74 kA; 13: Interrupting kiloampere vacuum contactor, 14: Interrupting kiloampere air-break contactor; 15: transformer let-through current = 57.20 kA asym.


How Differential Relay Protect Your System?

Differential relays provide high-speed (1 to 2 cycles), sensitive, and inherently selective protection. These will not provide protection for turn-to-turn winding faults in generators, motors, and transformers because of the small increment in the current produced by such faults, which remain below the pickup sensitivity of the relays.

An overcurrent relay can be used to provide differential protection when it is so connected that external fault currents through the current transformers balance out and do not give rise to a current in the relay operating coil. A phase or ground fault within the protected zone results in current unbalance and operates the relay.

This scheme is limited by current transformer saturation at high magnitudes of external fault currents. Partial differential protection of a motor uses core balance transformers, which circle phase and neutral leads so that under an external fault situation the magnetic fluxes in the core of the transformer balance out and current transformer saturation is avoided.

Percentage differential relays are used for protection of transformers, bus, motor or generator. Figure 2 shows the basic connections of a percentage differential relay and its characteristics. Load and external fault current circulates through the restraint coils, and no current flows through the operating coil, except as a result of current transformer errors.

For a fault in the protected zone, the difference current flows through the operating coil to actuate the relay. For a fixed restraint relay, the operating current required to overcome restraint is a fixed percentage of the restraint current, whereas in a variable restraint relay the current to operate the relay increases with the magnitude of fault current.

The number of relay input restraint elements will vary with the design and application. For transformer differential protection, harmonic restraint may also be applied to make the relay insensitive to transformer inrush currents, which are rich in harmonics. An instantaneous trip unit is included for high-magnitude internal faults.

High-impedance differential relays are primarily used for bus protection. A high-impedance relay is connected across the current transformer secondaries, which are paralleled together with proper polarity.

An external fault results in currents circulating between the current transformers and creates a low voltage across the relay, which is set to operate above this value. For an internal fault, the resulting secondary voltage exceeds this set value.

The current transformers must be of the same ratio. The system can easily accommodate expansion, when more circuits are added.

Pilot differential relays are applied to short transmission line protection of approximately 40 km (25 mi) length or less, where a metallic, microwave, or fiber-optic communication circuit is available to compare the system conditions at two ends of the transmission line.

The protection is analogous to differential protection of transformers and machines. Composite filters are used to convert three-phase currents at each end into a single-phase voltage.

These single-phase voltages are compared at each line terminal over the pilot channels to determine whether the fault is inside or outside the protected zone. The pilot channels are continuously monitored for open and short circuits.

Transfer trip facilities are usually added with additional relays. The series resistance and shunt capacitance of the pilot wires and the voltages developed under fault conditions are of concern.

Drainage reactors, neutralizing and insulating transformers, and surge arresters are needed. These complications can be avoided and security enhanced by using fiber-optic or microwave channels and interfaces, which are not affected by fault-induced voltages.


The FMEA for power distribution systems amounts to the determination and listing of those component outage events or combinations of component outages that result in an interruption of service at the load point being studied according to the interruption definition that has been adopted.

This analysis must be made in consideration of the different types and modes of outages that components may exhibit and the reaction of the system’s protection scheme to these events.

The primary result of the FMEA as far as quantitative reliability evaluation is concerned is the list of minimal cut-sets it produces. A minimal cut-set is defined to be a set of components that, if removed from the system, results in loss of continuity to the load point being investigated and that does not contain as a subset any set of components that is itself a cut-set of the system.

In the present context, the components in a cut-set are just those components whose overlapping outage results in an interruption according to the interruption definition adopted.

An important nonquantitative benefit of the FMEA is the thorough and systematic thought process and investigation that it requires. Often weak points in system design will be identified before any quantitative reliability indexes are computed. Thus, the FMEA is a useful reliability design tool even in the absence of the data needed for quantitative evaluation.

The FMEA and the determination of minimal cut-sets are most efficiently conducted by considering first the effects of outages of single components and then the effects of overlapping outages of increasing numbers of components. Those cut-sets containing a single component are termed first-order cut-sets.

Similarly, cut-sets containing two components are termed second- order cut-sets, etc. In theory the FMEA should continue until all the minimal cut-sets of the system have been found. In practice, however, the FMEA can be terminated earlier, since high-order cut-sets have low probability compared to lower-order cut-sets.

A good rule of thumb is to determine minimal cut-sets up to order n + 1 where n is the lowest-order minimal cut-set of the system. Since most power distribution systems have at least some first-order minimal cut-sets, the analysis can usually be terminated after the second-order minimal cutsets have been found.


Transients generate substantial radiated energy, electric and magnetic fields, and transient currents within the substation grounds. Any of these phenomena can couple into poorly executed control wiring systems, but none of them can couple to an appreciable degree into well-executed control wiring systems.

The obvious and correct approach to GIS control wiring is to enclose the entire control system in a Faraday cage, i.e., within a metal enclosure. This is much simpler than it sounds, as will be described below.

A Faraday cage is a metal enclosure that fully surrounds the system, offering protection from EMI. In the case of control wiring, the system is typically a sensor (e.g., a gas density relay), the attached control wiring, the local control wiring cabinet in which the control wiring is terminated, the control wiring from the local cabinet to the substation control room, and the relay or computer racks in the control room.

Each of these elements is usually well shielded. The sensor is usually housed in a metal case that sits on the GIS. The control wiring is usually shielded by a solid copper shield or several layers of braid.

The local control cabinet is metal and well shielded, as are the computers or relay racks in the substation control room. The problem, therefore, is not to shield the individual elements, which are all usually well shielded, but to ensure the continuity of the shield from one element to the next. To this end, it is necessary to understand something about the flow of high-frequency currents in metals.

From 1 MHz to 100 MHz, the skin depth of current in copper varies from about 70 mm to 7 mm, respectively, so that almost no current flows in the conductor more than 0.25 mm below its surface. Since the copper cable shield, sensor enclosure, and local cabinet are all thick compared to the skin depth in this frequency range, independent currents can flow on the inner and outer surfaces of the cable shield.

A large switching-induced transient current could be flowing on the outside surface of the shield with negligible current flowing on the inside surface. No coupling will occur to the sensor or control wiring withinthe shield so long as the current is not allowed to cross over from the outside of the shield to the inside.

The key to proper control wiring practice for GIS is effecting connections between shielding elements that provide shield continuity and avoid such crossover. When a cable enters a control cabinet, the cable shield should be terminated immediately on the control cabinet enclosure as the cable conductors enter the cabinet.

A long pigtail termination of the cable shield after the cable has entered the cabinet is poor practice, as this brings the transient on the control cable shield within the control cabinet where it can couple to all of the conductors therein.

Coaxial termination of the cable shield on the cabinet forces the shield currents to flow on the outside of the metal cabinet, which shields the conductors within from the shield currents. A range of connectors and cables is suitable for coaxial termination of cable shields.

Cable with a solid copper shield offers best performance, and such cables do not necessarily cost more than cable with a less effective braided shield. Some components in GIS require careful design to avoid the coupling of transients into the control wiring system.

Voltage transformers (VT) are of special concern, as they effect a connection between the high-voltage conductor and the control wiring system. The interwinding capacitance in a magnetic VT can result in unacceptable coupling of transients from the GIS conductor to the low side of the VT unless an electrostatic shield is employed between the windings.


The IEC 60870-5 standards address the basic goals of telecontrol systems and their particular environmental conditions. IEC 60870-5 does not define one particular protocol profile; but rather like EPRI/UCA, it specifies a number of frame formats and services that may be provided at different layers.

IEC 60870-5 is based on a three-layer Enhanced Performance Architecture (EPA) reference model for efficient implementation within RTUs, meters, relays, and other IEDs. Additionally, IEC 60870-5 defines basic application functionality for a user layer, which is situated between the OSI Application Layer and the application program.

This user layer adds interoperability for such functions as clock synchronization and file transfers. The following descriptions provide the basic scope of each of the five documents in the base IEC 60870-5 telecontrol transmission protocol specification set.

Standard profiles are necessary for uniform application of the IEC 60870-5 standards. Such profiles have been and are being created. The T101 profile is described in detail following the description of the applicable standards.

IEC 60870-5-1 (1990) specifies the basic requirements for services to be provided by the data link and physical layers for telecontrol applications.

In particular, it specifies standards on coding, formatting, and synchronizing data frames of variable and fixed lengths that meet specified data integrity requirements. ¾ IEC-60870-5-2 (1992) offers a selection of link transmission procedures using a control field and optional address field; the address field is optional because some point-to-point topologies do not require either source or destination addressing.

IEC 60870-5-3 (1992) specifies rules for structuring application data units in transmission frames of telecontrol systems. These rules are presented as generic standards that may be used to support a great variety of present and future telecontrol applications. This section of IEC 60870-5 describes the general structure of application data, and describes basic rules to specify application data units without specifying details about information fields and their contents.

IEC 60870-5-4 (1993) provides rules for defining information data elements and a common set of information elements, particularly digital and analog process variables that are frequently used in telecontrol applications.

IEC 60870-5-5 (1995) defines basic application functions that perform standard procedures for telecontrol systems, which are procedures that reside beyond Layer 7 (application layer) of the ISO reference model.

These utilize standard services of the application layer. The specifications in IEC 60870-5-5 (1995) serve as basic standards for application profiles that are then created in detail for specific telecontrol tasks.

Each application profile will use a specific selection of the defined functions. Any basic application functions not found in a standards document but necessary for defining certain telecontrol applications should be specified within the profile. Examples of such telecontrol functions include station initialization, cycle data transmission, data acquisition by polling, clock synchronization, and station configuration.


The foundations for poles are just as important as the structure above ground. The pole back fill should be capable of withstanding structure reactions. Pole-setting equipment should be moved clear of the structure site prior to back filling.

Differences in ground elevation at each pole location, and pole length tolerances permittedby ANSI O5.1-1987 [9] should be considered to ensure a level structure. The tops of poles should not be cut. If cutting is necessary, the pole top should be covered with a mastic-type cap.

Under no circumstances should the butt of any pole be cut. The design engineer should specify a minimum hole depth. The actual hole depths required to obtain a level structure are the responsibility of the installing contractor.

Digging operations should not be too far in advance of the setting operation. Holes open too long may deteriorate due to ground water seepage and/or heavy rains and increase the chance for accidents. Unattended pole holes should be temporarily covered. All Local, State, and Federal safety regulations must be met.

Structure Alignment
When the structure is set and the load line completely released, the structure should remain plumb and level. If the structure is not plumb or the crossarm is not level, additional material will have to be placed under one pole. The additional material should be approved by the design engineer.

Pole Holes
All holes should be in the correct locations and large enough to provide a minimum of 6 in of space for tamping around the pole to the full depth of the hole. Pneumatic tamping equipment is recommended to expedite the setting operation.

The poles should be placed to prevent damage to the structure grounding materials. Poles not required to be raked should be set plumb and in alignment. Unless otherwise specified, structures at angles should be set to bisect the line angle.

The holes may be back filled with earth excavated from the hole, provided this material can be properly compacted. Frozen material for back fill should not be permitted.

The back fill should be compacted to a dry density not less than the natural in-place dry density of the surrounding earth. Since the measurement of the density may not be practical, no more than one shoveler should be utilized for three tampers.

Front-end loaders are not recommended during back filling. Back fill should be banked and tamped around the poles to a height of 12 in above the natural ground surface.

Excessive water should be pumped out, leaving not more than 6 in of water in the bottom of the hole, and 6 in of granular material should be placed to firm up the bearing surface. Care should be exercised where pumping will cause excessive sluffing of the bottom of the hole.

Casing should be used where moving water and/or gravel is encountered, working the casing down as the material and/or water is removed.


The protective functions noted in the various generating station configurations provide both primary and backup protection for the generating station as well as additional protection schemes which could also be applied.

These protective functions are listed below with a reference to the subclause in the text that discusses their application in detail. Also included is a discussion of the various tripping modes used in generating stations.

Protective Devices

Device Function                      Subclause
21 Distance relay. Backup for system and generator zone phase faults. Device 21
requires a time delay for coordination.
24 Volts/hertz overexcitation protection for the generator and its associated step-up 
and auxiliary transformers.
27 Undervoltage relay. 
32 Reverse-power relay. Motoring protection. 
40 Loss of field protection
46 Stator unbalanced current protection. Negative sequence relay. 
49 Stator thermal protection. 
50N Instantaneous overcurrent relay used as current detector in a breaker failure scheme.
50/51 Time overcurrent relays with instantaneous element. High-side bank overcurrent relays providing phase-fault protection for unit auxiliary transformer and backup protection for failure of UAT low-side bank breaker.
50/51GN Time overcurrent relay with instantaneous element. Primary and/or backup protection for generator ground faults.
51 Time overcurrent relay. Detection of turn-to-turn faults in generator windings. 
51TN1 Time overcurrent relay. Provides backup protection for transmission ground faults when applied to GSU neutral. Protects for ground faults on the unit  auxiliary bus when applied to UAT neutral.
51TN2 Time overcurrent relay. Provides backup protection for GSU ground faults when applied to GSU neutral. Protects for faults in the low-side of the UAT down to the low-side bank breaker when applied to UAT neutral. Provides backup for failure of low-side breaker to trip.
Device Function Subclause
51 UAT Time overcurrent relays connected to current transformers in UAT low-side bank breaker. Protects for phase faults on unit auxiliary bus.
51V Voltage controlled or voltage-restrained time overcurrent relay. Backup for system and generator zone phase faults.
53 Exciter or dc generator relay. 
59 Overvoltage protection.
59BG Zero-sequence voltage relay. Ground fault protection for an ungrounded bus.
59GN Voltage relay. Primary ground fault protection for a generator. 
60 Voltage balance relay. Detection of blown potential transformer fuses. 
62B Breaker failure timer. 
63 Fault pressure relay. Detects transformer faults. 
64F Voltage relay. Primary protection for rotor ground faults. 
71 Transformer oil or gas level. 
78 Loss of synchronism protection. This protection is optional. Applied when, during a loss of synchronism, the electrical center is in the step-up transformer or in the generator zone. Alternate locations are shown for this protection. A study should be made to determine which location is best for the detection of an out-of-step condition.
81 Frequency relay. Both under frequency and overfrequency protection may be required.
86 Hand-reset lockout auxiliary relay.
87B Differential relay used for bus protection.
87G Differential relay. Primary phase-fault protection for the generator. 
87GN Differential relay. Sensitive ground-fault protection for the generator. 
87T Differential relay. Primary protection for the GSU or UAT transformer. May be used to provide phase fault backup for the generator in some station arrangements. The zone may be extended to cover the generator bus using cts from the generator and Unit Auxiliary Transformer when lowside cts are not available.
87U Differential relay for overall unit and transformer.


It is common practice to ground all types of generators through some form of external impedance. The purpose of this grounding is to limit the mechanical stresses and fault damage in the machine, to limit transient voltages during faults, and to provide a means for detecting ground faults within the machine.

A complete discussion of all grounding and ground protection methods may be found in IEEE Std C62.92.2-1989 and IEEE Std C37.101-1993.

The methods most commonly used for generator grounding will be discussed in this guide. They are listed in the following four broad categories:

a) High-impedance grounding
b) Low-resistance grounding
c) Reactance grounding
d) Grounding-transformer grounding

Solid grounding of a generator neutral is not generally used since this practice can result in high mechanical stresses and excessive fault damage in the machine. According to ANSI C50.13-1989, the maximum stresses that a generator is normally designed to withstand is that associated with the currents of a three-phase fault at the machine terminals.

Because of the relatively low zero-sequence impedance inherent in most synchronous generators, a solid phase-to-ground fault at the machine terminals will produce winding currents that are higher than those for a three-phase fault.

Therefore, to comply with this guide, generators shall be grounded in such a manner to limit the maximum phase-to-ground fault current to a magnitude equal to, or less than, the three-phase fault current.

Generators are not often operated ungrounded. While this approach greatly limits the phase-to-ground fault currents, and consequently limits damage to the machine, it can produce high transient over voltages during faults and also makes the fault location difficult to determine.


Delayed autoreclosing may need to be considered when the upstream protection is provided by electromechanical relays or fuses and the circuit protection is provided by microprocessor-based relays, unless the microprocessor-based relays can be set to mimick the reset characteristic of the electromechanical relays.

Without this time-delay reset feature on the microprocessor-based relay, it is possible to have the upstream device operate incorrectly, resulting in an overtrip. As an example of this, the low-set instantaneous trip on a distribution feeder is eliminated to improve power quality by eliminating momentary service interruptions.

If an instantaneous autoreclose is used after a time delayed trip, an additional time margin needs to be used between the operating times of protective devices in order to maintain coordination of the feeder overcurrent relays and an upstream electromechanical relay or fuse.

By delaying an upstream protective device to coordinate with the back-to back operation of the feeder relay, coordination is maintained with the instantaneous reclose. Delaying the autoreclosing eliminates this problem by allowing all devices time to rest before the next fault.

Delayed autoreclosing is used on circuits that have automatic sectionalizers to allow proper coordination with the distribution circuit breaker. The time-delay autoreclosing of the distribution circuit breaker needs to be set to match the programmed time intervals of the sectionalizer switches to allow successful isolation of the faulted line section.

Distribution circuits that have customer-owned generators connected to them present a special problem. In most cases, it will be necessary to delay autoreclosing to allow the customer generator to be disconnected before the circuit is re-energized from the utility source.

Removal of the customer generation is normally accomplished by the operation of an underfrequency, undervoltage, or reverse power relay, which tend to have longer tripping times. As the operating time for these devices may be slower to remove the connected generator than the relays that detected and cleared the fault, autoreclosing times could need to be extended to allow these devices to operate or the function be disabled.

In cases where the connected generator is comparable to the load, it may be necessary to provide additional security against energizing the generator out of synchronism. This additional security can be provided by dead-line autoreclosing logic, synchronizing check, or transfer trip protection.


There is never a reason to autoreclose an electrical circuit breaker following a trip unless there is reason to believe that the fault is no longer present on the circuit. Historically, when distribution circuit breakers would trip and result in a circuit outage, the circuit was patrolled before the circuit breaker was closed.

This practice delayed restoration. Records were kept of these events. It was discovered that for 85–90% of the occurrences, no permanent faults were found.

It generally became accepted to autoreclose these distribution circuit breakers. With the advent of additional protective devices available to the distribution engineer such as fuses, sectionalizers, and reclosers with which coordination was necessary, multiple autoreclose attempts were chosen.

In many areas, three autoreclose attempts were chosen. This results in four trips to lockout. This practice continued for several years.

As time went on, load increased and it became necessary that distribution source transformer size increased as well as the number of supplied feeders. It is known that when transformers are subjected to any fault on the secondary that the transformer windings are stressed.

If the transformer was not designed for the exposure that is encountered in distribution operation, it is possible that autoreclosing into a fault that would allow the transformer to contribute its maximum available short circuit current could result in deformation of the windings and subsequent arc damage to the transformer core and mounting structure.

Often, repeated occurrences of these stress levels resulted in transformer failure. The practice of some utilities is to block autoreclosing for close-in faults or for faults with a fault current magnitude in excess of the transformer design capability, in an effort to mitigate the cumulative effect of these severe faults.

Observation of fault events resulted in the conclusion by a number of utilities that the third autoreclose attempt was seldom successful. As source transformer size and distribution voltage increased, many engineers decided to remove the last autoreclose attempt as a means of reducing the exposure to through fault events.


There has been a tendency to attribute disturbances and failures to power surges, a term often used by the media but rather ill-defined. The ambiguity results in part from an unfortunate dual definition of the word surge.

a) To some people, a surge is indeed the phenomenon being discussed here, that is, a transient voltage or current lasting from microseconds to at most a few milliseconds, involving voltages much higher than the normal (two to ten times).

b) To other people, a surge is a momentary overvoltage, at the frequency of the power system, and lasting for a few cycles, with voltage levels slightly exceeding the five to ten percent excursions that are considered normal occurrences.

The term swell has been adopted by this recommended practice to describe this second type of overvoltage; perhaps one day it will supplant the usage of surge for that meaning. It would be a mistake to attempt protection against these long-duration power frequency swells with a surge protective device that is designed to absorb large but short impulses of energy.

There is a growing recognition that the horror tales of surge protective device failures are more likely
to be caused by swells rather than by large surges.

Nonlinear loads draw nonsinusoidal currents from the power system, even if the power system voltage is a perfect sine wave. These currents produce nonsinusoidal voltage drops through the system source impedance, which distort the sine wave produced by the power plant generator.

A typical nonlinear load is a dc power supply consisting of rectifiers and a capacitor-input filter, such as used in most computers, drawing current only at the peaks of the voltage sine wave. This current has a high third harmonic content that has also created a new concern, that of insufficient ampacity of the neutral conductor in a three-phase system feeding power supplies.


As a general principle, it would be desirable to determine the efficiency of an HVDC converter station by the direct measurement of its energy losses. However, there are practical difficulties that prevent such a measurement, including the following:

1) Attempts to determine the station losses by subtracting the measured output power from the measured input power must recognize that such measurements have inherent inaccuracy, especially if performed at high voltage dc.

Moreover, the losses of an HVDC converter station at full load are generally less than 1% of the transmitted power. Therefore, the difference between the measured input and output power is a small difference between two large quantities and, as such, is not likely to be a sufficiently accurate indication of the actual station losses.

2) In some special circumstances, it may be possible to arrange a temporary test connection in which the two converters are operated from the same ac source and are also connected together via their dc terminals. In this connection, the ac source only provides enough power to make up the losses in the circuit.

However, it must also provide var support and commutating voltage for the two converters. Here again, there are practical measurement difficulties.

Because of the problems described above, it is recommended that the total station losses be calculated from the losses of the individual equipment. The equipment losses should be determined according to Section 4.

Summation of Equipment Losses
The total losses of the HVDC converter station should be arrived at as the sum of the losses of each piece of equipment. It is important to note that the actual losses in each piece of equipment will depend on the ambient conditions under which it operates, as well as on the operating conditions or duty cycles to which it is applied.

Therefore, in order for the summation of the individual losses to be a sufficiently accurate representation of the actual total HVDC converter station losses, the ambient and operating conditions for each piece of equipment must be defined, based on the ambient and operating conditions of the entire HVDC converter station.

For some equipment or components, losses or electrical characteristics are measured at the factory under standardized ambient and operating conditions. In these cases, the results should be related to the actual conditions in the HVDC converter station through well-recognized calculation procedures.

The fundamental principle of this recommended practice is that the determination of the equipment losses is, in fact,
based on physical tests on the actual equipment or components. Therefore, the sum of the equipment losses does give
a dependable measure of the losses for the entire HVDC converter station.


For certain applications, the most cost-effective solution for poor power factor, excessive voltage distortion, and IEEE Std 519-1992 violations is to install one or more larger harmonic filters at a distribution bus or busses.

Generally, an automatic harmonic filter bank will be installed on the secondary of each main transformer in the plant requiring power factor and harmonic compensation. Placement of multiple banks on a common low-voltage system can create problems by changing network harmonic flows and thereby increase the potential for overloading some of the filter banks.

Therefore, this practice is not generally recommended. Where power factor correction is most important, systems tuned to the 2nd harmonic or below can generally be safely applied in this manner. Parallel resonance at the 3rd harmonic must be carefully evaluated.

Caution must be exercised when a harmonic filter is electrically close to the main and is tuned to the 7th harmonic or higher. In this case, the potential exists to absorb large amounts of harmonic current from the utility distribution system.

Harmonic filters should be designed assuming the distribution system will have up to 3% voltage distortion at the harmonic nearest the tuning frequency per IEEE Std 519-1992. If the harmonic filter reactors are not equipped with taps, as is common with low-voltage filters, a good practice is to over specify the thermal rating so that additional capacitance may be added to detune the filter in the event a harmonic overload occurs.

If altering the tuning results in unacceptable filtering of in-plant harmonics, the utility can generally help to identify methods for reducing the available harmonic current. Possible utility-side solutions include the following:
Changing the size or status of capacitor banks to alter the impedance characteristics of the system
— Enforcing IEEE Std 519-1992 limits on customers with excessive harmonic injection
— Circuit reconfiguration to isolate harmonic injectors
— Medium-voltage harmonic filters

During lightly loaded conditions, harmonic filters that are fixed on the bus can produce an overvoltage condition. The maximum per-unit voltage rise caused by the harmonic filter can be estimated as approximately equal to the harmonic filter power system frequency current (i.e., fundamental current) divided by the system three-phase short-circuit current at the harmonic filter location (see IEEE Std 1036-1992).

If overvoltage is a concern, an automatically switched harmonic filter should be considered. Switched harmonic filter(s) comprise a number of steps, each of which is an individually tuned harmonic filter.

Reactive current controllers (sometimes referred to as var controllers) that can switch steps in and out automatically as system reactive current (i.e., power factor) changes are readily available. Other switching alternatives include the use of current relays, time-of-day controllers, voltage controllers, or other sensing devices. Switching times become more important as the harmonic filter is tuned closer to its rated frequency.

Temporary duties should be evaluated with automatic harmonic filters to ensure steps do not overload before enough compensation is added. Steps should typically be switched in at 10 s to 15 s intervals. If the controller has the capability, the step removal interval should be greater than 1 min to avoid unnecessary “seeking” or “hunting” by the controller. Note that some controllers include algorithms to avoid “seeking” or “hunting.”


Each phase of each filter step should be protected by fuses. Fuses should be current limiting, rated for the available fault current at the fuse location. Fuses should be UL-listed Class J or T, CSA-rated HRC 1, or equivalent.

The current rating of the fuses should be a minimum of two times the capacitor current calculated from its rated reactive power and its rated voltage. The voltage rating of the fuses should be greater than or equal to the system voltage.

Fuses internal to the capacitor should not be accepted as the primary means of filter protection. In a harmonic filter assembly containing more than one capacitor per reactor, a single set of fuses, one per phase, should be provided.

The current rating of the fuses should be at least equal the total filter current including all harmonics, with margin (typically 35%) to cover contingency conditions. Fuses should be located after the harmonic filter main lugs or main disconnect and before the contactor/ reactor/capacitor assemblies.

In addition to fusing, some harmonic filters may be protected by devices that detect phase loss or thermal overload and trip the step/unit off line. Such devices are not meant to replace fuses.

The fuses should be rated for the following conditions:
a) Maximum system voltage
b) Maximum continuous filter current, including fundamental and harmonics
c) Interrupting rating, equal to or greater than the available short-circuit current at the fuse location
d) Sized to limit the fault current to a level consistent with the capabilities of the harmonic filter components
e) Sized to withstand inrush currents when the harmonic filter is energized

Circuit breakers
Circuit breakers may be used in place of fuses to provide the primary means of overcurrent protection (i.e., thermal/magnetic trip). Alternately, circuit breakers or molded case switches may serve in addition to fuses to provide a primary means of disconnect (i.e., manual switch) or to provide overload (i.e., thermal trip) and/ or short-circuit (i.e., magnetic trip) protection for the complete harmonic filter.

Circuit breakers should be rated for the following conditions:
a) Maximum system voltage
b) The harmonic filter current spectrum, including the fundamental and harmonics (Note that this current spectrum should be determined based on the maximum system operating voltage and maximum positive capacitor tolerance.

The heating of the circuit breaker may be greater at higher frequencies than at the fundamental frequency because of eddy currents and the skin effect. It is not sufficient to specify only the rms value of the circuit breaker current.)

c) Interrupting rating, equal to or greater than the available short-circuit current at the harmonic filter
d) Sized to limit the fault current to a level consistent with the capabilities of the harmonic filter
e) Sized to have sufficient short-time current rating to withstand inrush currents when the harmonic filter is energized
f) The number and the frequency of switching operations


Harmonic filter reactors for low-voltage applications are typically dry-type iron-core. No existing standard addresses harmonic filter reactors, but most manufacturers use IEEE Std C57.12.01-1998 as a guideline.

Cores are constructed from silicon sheet steel (such as M-6). The number 6 corresponds to the approximate power loss per pound of steel at a magnetic flux density of 1.5 T, i.e., M-6 has a loss of “0.6 W/lb” or 1.5 W/kg. M-6 is the typical grade of silicon steel used, but both lower and higher grades of steel are available.

A manufacturer may choose to use a lower grade steel and either let the harmonic filter reactor operate with a higher temperature rise or use more steel. Conversely, a higher grade of steel can be used and either the harmonic filter reactor may operate with a lower temperature rise or the harmonic filter reactor could be made smaller.

The construction may be from individual pieces of cut strip stock or E-I laminations. To create a harmonic filter reactor, it is necessary to have gaps in the core. These gaps are known as air gaps, but for physical integrity and rigidity the gaps are filled with hard insulation.

This insulation will have permeability similar to air. These gaps can be distributed (many small gaps) or a single larger gap. A single gap will use E-I laminations whereas a distributed gap will be made up of individual cut strips. The E-I construction requires less labor and can be clamped and wedged better than a distributed gap core.

However, the distributed gap core will significantly reduce fringing. This reduction in fringing helps control the effective cross-sectional area as well as stray fields that may result in localized heating of the coil. A “C” core may be used, but it will not offer an advantage for harmonic filtering.

Reactor coils may be constructed from sheet conductor or magnet wire. Sheet conductor may be more economical and easier for construction, but reactors with significant harmonic currents can incur heating problems.

If not properly designed, sheet copper windings can become annealed due to large localized current densities. Sheet conductor windings should be used only on harmonic filter reactors that have a distributed gap well within the boundaries of the coil. Magnet wire, which is less susceptible to localized heating, is commonly used for harmonic filter reactors.

In some cases, it may be desirable to have parallel strands of smaller gauge magnet wire to reduce heating. Although using parallel strands of such wire significantly complicates winding construction, the increase in winding complexity can be justified because coil losses may be significantly reduced.

Clamping is also very important. If a reactor is not properly clamped, the harmonic current can cause laminations to vibrate. Lack of proper clamping could result in a loud audible noise and the breakdown of the insulation coating on the laminations.

Laminations are clamped with insulated through-bolts, or bolts that go through clamps, and are external to the laminations. Clamps that bridge air gaps must be of nonferrous construction. Such nonmagnetic clamps are used to avoid shunting the air gap with a magnetic path.

Furthermore, ferrous components need to be as far from air gaps as possible to prevent inductive heating of the ferrous material. Coils are held in place with spacers and wedges. The harmonic filter reactor should be vacuum impregnated, preferably vacuum-pressure impregnated with a suitable varnish. Impregnation, however, should not be relied upon as the only mechanical means of support.


Capacitors are generally rated for the system line-to-line voltage (e.g., 240 V, 480 V). However, in a harmonic filter application, they should be selected to withstand overvoltages and overcurrents caused by fundamental and harmonic current flow through the series connected tuning.

IEEE Std 18-2002 requires shunt capacitors, under contingency conditions, to withstand continuous voltages up to 110% of rated rms voltage and continuous currents up to 135% of nominal rms current based on rated kvar and rated voltage.

When applied in harmonic filters, the normal voltage and current may exceed these levels even before system contingencies are considered. Consequently, capacitors selected for use in normal shunt capacitor applications may not be suitable for use in harmonic filters.

Harmonic filter capacitors should be selected based on their expected duty under normal and contingency conditions. The capacitor manufacturer should be consulted when specifying capacitors for harmonic filter applications.

Capacitors should be manufactured in accordance with UL 810-1995. Capacitors should be protected by a UL-listed/recognized protective device. For capacitors that do not contain a UL-listed/recognized protective device, the capacitors should be protected with external current limiting fuses or other external protective devices that are UL-listed.

Capacitor cells connected with a wiring harness may also be protected with UL listed current limiting fuses (even if they have internal pressure sensitive interrupters) to protect in the event of a failure of external resistors, bushings, or connecting wires.

IEEE Std 18-200215 specifies that the terminal-to-case test for the internal insulation of indoor capacitors should be performed at 3 kV rms (capacitors rated 300 V or less) or 5 kV rms (capacitors rated 301 V to 1199 V) for 10 s.

The terminal-to-terminal test should be 10 s at 2 × rated rms voltage (ac test) or 4.3 × rated rms voltage (dc test). Each capacitor rated 600 V or less must be provided with a discharge resistor(s) to reduce the residual voltage from peak of rated to less than 50 V within 1 min of de-energization (5 min for capacitor units rated higher than 600 V) to meet the requirements of IEEE Std 18-2002 and Article 460-6 of NFPA 70-2002 (National Electrical Code).


Harmonic distortion on the power system is caused by nonlinear devices that produce distorted or nonsinusoidal waveforms. Examples include electronically controlled devices (such as rectifiers and power controllers), arcing loads (such as arc furnaces and arc welders), and magnetic devices to a lesser degree (such as rotating ac machinery and transformers).

Excessive harmonic voltage and/or current can cause damage to equipment and the electrical system. IEEE Std 519-199210 gives application guidelines. One of the common ways of controlling harmonic distortion is to place a passive shunt harmonic filter close to the harmonic producing load(s).

The harmonic-producing device can generally be viewed as a source of harmonic current. The objective of the harmonic filter is to shunt some of the harmonic current from the load into the filter  thereby reducing the amount of harmonic current that flows into the power system.

The simplest type of shunt harmonic filter is a series inductance/capacitance (LC) circuit. More complex harmonic filters may involve multiple LC circuits, some of which may also include a resistor.

Key filter design considerations include the following:
a) Reactive power (kilovar) requirements
b) Harmonic limitations
c) Normal system conditions, including ambient harmonics
d) Normal harmonic filter conditions
e) Contingency system conditions, including ambient harmonics
f) Contingency harmonic filter conditions

These considerations can be grouped into performance and rating criteria. The performance criteria relate to normal expected operating conditions and include capacitive reactive power requirements, harmonic limitations, normal system conditions, and normal harmonic filter conditions.

The rating criteria relate to unusual conditions that may place a more severe duty on the equipment. These unusual conditions include contingency system conditions and contingency harmonic filter conditions.

Under the contingency conditions, it may be acceptable to have a more relaxed harmonic limitation. 


Two classes of expenses determine the total cost of generating, transmitting, and distributing electric energy. They are:

1. Capital investment items: depreciation, interest on notes, property taxes, and other annual expenses arising from the electric utility’s capital investment in generating, transmitting, and distributing equipment, and in land and buildings,

2. Operation and maintenance items: fuel, payroll, renewal parts, workmen's compensation, rent for office space, and numerous other items contributing to the cost of operating, maintaining, and administering a power system.

In billing the individual consumer of electricity, the utility considers to what extent the total cost of supplying that consumer is determined by capital investment and to what extent it is determined by operation and maintenance expenses.

Furnishing power to some consumers calls for a large capital investment by the utility. With other consumers, the cost may be due largely to operation and maintenance. The following two examples illustrate these two
extremes of load.

1. In a certain plant, electricity is used largely to operate pumps, which run at rated load night and day. The power consumed by the pump motors is low and the plant shares a utility-owned transformer with several other consumers.

The amount of energy used each month is large because the pumps are running constantly. Therefore, the cost of supplying this consumer is largely determined by operating expenses, notably the cost of fuel. The capital investment items are relatively unimportant.

2. Another factory uses the same number of kilowatthours of energy per month but consumes all of it in a single eight-hour shift each day of the month. The average power is therefore three times greater than for the pump plant and the rating (and size) of equipment installed by the utility to furnish the factory with energy must also be about three times higher.

Costs rising from capital investment are a much greater factor in billing this consumer than in
billing the operator of the pump plant.

Demand is an indication of the capacity of equipment required to furnish electricity to the individual consumer. Kilowatthours or energy per month is no indication of the rating of equipment the utility must install to furnish a particular maximum power requirement during the month without overheating or otherwise straining its facilities.

What is needed in this case, is a measure of the maximum demand for power during the month. The demand meter answers this need. More importantly, the true demands that the equipment experiences are the maximum kilovoltamperes (kVA).

This takes into account the real power watts, and the reactive power VARs, as one quantity. With electronic meters, the meter calculates kVA demand from the coincident peak demand of the real and the reactive power. This represents the true maximum stress on the power equipment.



Kilowatt demand is generally defined as the kilowatt load averaged over a specified interval of time. In any one of the time intervals shown, the area under the dotted line labeled demand is exactly equal to the area under the power curve.

Since energy is the product of power and time, either of these two areas represents the energy consumed in the demand interval. The equivalence of the two areas shows that the demand for the interval is that value of power which, if held constant over the interval, will account for the same consumption of energy as the real power. It is then the average of the real power over the demand interval.

The demand interval during which demand is measured may be any selected period but is usually 5, 10, 15, 30, 60, and in similar increments up to 720 minutes. The demand period is determined by the billing tariff for a given rate schedule.

Demand has been explained in terms of power (kilowatts) and usually this information has the greater usefulness. However, demand may be expressed in kilovoltamperes reactive (kVAR), kilovoltamperes (kVA), or other suitable units.

Coincidental Demand—Many utility customers have two or more revenue meters that meter separate electrical loads. A common example is a large factory that has multiple meters at different locations. Assuming each revenue meter measures demand, then each meter would provide a maximum demand.

Coincidental demand is the maximum demand that is obtained when all metered loads are summed coincidentally. The summation of the individual demands must be performed on a demand interval basis.

In other words, when all measured demands from each individual meter are summed on each interval of the billing period, the maximum total demand obtained from the summation is the coincidental demand. The individually metered maximum demands typically do not occur at the same demand interval in which the coincidental demand occurs.

Therefore, the summation of the individually metered maximum demands will normally be higher than the demand that occurs at the demand interval in which the total coincidental demand occurs.

This is due to the variation in the time in which electrical equipment operates. The total coincidental peak demand is usually less than the sum of the individual maximum demands.

Aggregated Demand—Aggregated demand is similar to coincidental demand in that it is derived from the summation of multiple meters. Typically, aggregated demand is obtained from the aggregation of load profile data from multiple meters.

Totalized Demand—Totalization, as applied to revenue metering, is the addition of two or more metered electrical loads. Totalization is often requested by customers that have two or more metered loads. Benefits of totalization include the ability to obtain coincidental demands, simplified meter reading, and billing and subsequent accounting procedures.

Totalization is the algebraic sum of two identical energy values performed on a real time or near instaneous basis. Simple totalization could be the addition of the kilowatthour useage of two metered loads. Complex totalization could be the algebraic sum of multiple metered loads from different locations, some of which
could be negative values.

It is important to note that totalized demand is derived from totalized energy. Energy is summed on a near instantaneous basis. Because the totalizing device or software knows the time interval over which the demand is desired, totalized demand can then be obtained from the simple relationship, Demand  Energy/Time.


The selection of the rated component insulation level consists of the selection of standard insulation withstand voltages that provide sufficient margin above the system overvoltage stress.

The tests required to verify the component rated maximum voltages are defined by the relevant apparatus standards. The component low-frequency, short-duration withstand voltage is selected from the list of standard withstand voltages provided in 4.5. The standard BIL and BSL values are selected from the table in 4.6.

Low-frequency, short-duration withstand voltages
The following list of low-frequency, short-duration withstand voltages (rms values, expressed in kilovolts), are extracted from IEEE Std C57.12.00-1993 and IEEE Std C57.21-1990. The withstand value should be taken from this table.

10, 15, 19, 26, 34, 40, 50, 70, 95, 140, 185, 230, 275, 325, 360, 395, 460, 520, 575, 630, 690, 750, 800, 860, 920, 980, 1040, 1090

The relevant apparatus standards recognize low-frequency, short-duration withstand voltages other than
those listed above. Refer to these other standards for specific values.

Standard BIL and BSL
The BIL and BSL values should be taken from this table.

10, 20, 30, 45, 60, 75, 95, 110, 125, 150, 200, 250, 350, 450, 550, 650, 750, 825, 900, 975, 1050, 1175, 1300, 1425, 1550, 1675, 1800, 1925, 2050, 2175, 2300, 2425, 2550, 2625, 2675, 2800, 2925, 3050

Some apparatus standards recognize that a fixed relationship between BIL and BSL is appropriate for specific equipment and substation assemblies. Therefore the BSL may differ from the table values. In such cases, refer to the relevant apparatus standards for specific values.

Classes of maximum system voltage
The standard highest voltages are divided into the following two classes:
— Class I: Medium (1–72.5 kV) and high (72.5–242 kV) voltages: > 1 kV and </=242 kV
— Class II: Extra high and ultra high voltages: > 242 kV


The procedure for insulation coordination consists of

a) Determination of voltage stresses
b) Selection of the insulation strength to achieve the desired probability of failure

The voltage stresses can be reduced by the application of surge-protective devices, switching device insertion resistors and controlled closing, shield wires, improved grounding, etc.

Determination of the system voltage stress
System transient analyses that include the selection and location of the overvoltage limiting devices are performed to determine the amplitude, waveshape, and duration of system voltage stresses.

The overvoltage stress may be characterized either by
— The maximum crest values, or
— A statistical distribution of crest values, or
— A statistical overvoltage value [this is an overvoltage generated by a specific event on the system (lightning discharge, line energization, reclosing, etc.), with a crest value that has a 2% probability of being exceeded].

The results of the transient analysis should provide voltage stresses for the following classes of overvoltage:
— Temporary overvoltage (phase-to-ground and phase-to-phase)
— Switching overvoltage (phase-to-ground and phase-to-phase)
— Lightning overvoltage (phase-to-ground and phase-to-phase)
— Longitudinal overvoltage (an instantaneous combination of switching or lightning surge and a power-frequency voltage)

Comparison of overvoltages with insulation strength
To compare the overvoltages with the insulation strength, the insulation strength must be modified because of the (1) nonstandard waveshape of overvoltages and (2) nonstandard atmospheric conditions. The dielectric strength of insulation for surges having nonstandard waveshapes is assessed by comparison to the dielectric strength as provided by standard chopped wave tests.

The rules for the atmospheric correction of withstand voltages for external insulations are specified in IEEE Std 4-1995. For insulation coordination purposes, wet conditions are assumed and only the relative air density corresponding to the altitude needs to be taken into account.

In addition, a safety margin may be necessary based on consideration of
— Statistical nature of the test results
— Factory or field assembly of equipment
— Aging of insulation
— Accuracy of analysis
— Other unknown factors

The overall protective margin is derived from experience and further described in IEEE P1313.2


Switching overvoltages may have times-to-crest from 20–5000 ms and time to half value of less than
20 000 ms. They are generally a result of the following:

— Line energization,
— Faults and fault clearing,
— Load rejections, or
— Switching of capacitive or inductive currents.

In general, the time to crest (wavefront) is of more importance since the critical flashover voltage (CFO) is function of the wavefront. The minimum CFO occurs at the critical wavefront (CWF), which in ms is equal to about 50 times the strike distance in meters (m). For a wavefront smaller or greater than the CWF, the CFO increases.

The CFO increases by about 10% when the wavefront is in the order of 1000 ms to 2000 ms, which usually occurs when employing low-side transformer switching. The distribution of switching overvoltages is obtained using a transient program where the breakers are randomly closed or reclosed 200 to 400 times.

These overvoltages are then statistically analyzed to obtain a probability distribution, which approximates the data. Several distribution functions have been used, e.g., Gaussian, extreme-value Weibull. However, a Gaussian or normal distribution is used most frequently.

This distribution is defined by its 2% value, called the statistical switching overvoltage (E2), and by its standard deviations in pu of E2. The standard deviation pu is:

s/E2. To clarify the definition, E2 means that 2% of the switching overvoltages equal or exceed E2. The shape of the distribution may be affected by the surge controlling action of an arrester. Typically the gapless arrester modifies the reflected wave and reduces the voltage stress caused by the return wave.


Temporary overvoltages caused by load rejection are a function of the load rejected, the system topology after disconnection, and the characteristics of the sources, i.e., short-circuit power at the station, speed and voltage regulation of the generators, etc.

These overvoltages are especially important in the case of load rejection at the remote end of a long line due to the Ferranti effect. Primarily, it affects the apparatus at the station connected on the line side of the remote circuit breaker.

A distinction should be made between various system configurations when large loads are rejected. A system with relatively short lines and high short circuit power at terminal stations will have low overvoltages.

A system with long lines and low short circuit power at generating sites will have high overvoltages. In a symmetrical three-phase power system the same relative overvoltages occur phase-ground and phasephase.

The longitudinal temporary overvoltages depend on whether phase opposition is possible. Phase opposition can occur when the voltages on each side of the open switching device are not synchronized.

A description of the degree and duration of phase-ground and longitudinal overvoltages for two types of substations follows:

— System substation: In a moderately extended system, for a full load rejection, the temporary overvoltage is usually less than 1.2 pu. The duration depends on the voltage control operation and may be up to several minutes. In extended systems, the overvoltages may reach 1.5 pu or even more when Ferranti effects or resonance occur.

The duration may be in the order of seconds. The longitudinal overvoltage across a switching device is usually equal to the phase-ground overvoltage unless motors or generators, connected to the rejected side, produce phase opposition.

— Generator station: For a full load rejection, the overvoltage at the substation may reach up to 1.5 pu. The duration may be up to 3 s depending on the generator characteristics and control.

The longitudinal temporary overvoltage is the difference between the phase-ground operating voltage at one terminal and the phase-ground temporary overvoltage on the other terminal. In the case of phase opposition the longitudinal overvoltages could be as high as 2.5 pu. Shunt reactors, static VAR compensators, or special (grouped) arresters can control these overvoltages.

free counters