1
Power Systems: A General Overview
In this chapter we present an overview of the structure of a modern power system, from the low voltage distribution networks with which we are partly familiar, to the high voltage transmission system bringing energy from remote electrical generators.
Before we begin, we need to define the different potentials which we will encounter throughout the power system. We use abbreviations to denote different voltage levels, as outlined in Table 1.1. However, as is often the case, there appears to be no universally accepted definition, so you may see slightly different definitions used elsewhere.
Table 1.1 Voltage definitions.
| LV (low voltage) | <1000 V AC |
| MV (medium voltage) | 1–35 kV AC |
| HV (high voltage) | 35–230 kV |
| EHV (extra high voltage) | >230 kV |
1.1 Three‐phase System of AC Voltages
The alternating voltage distributed to our homes has (ideally) a sinusoidal form. Sinusoidal waveforms are chosen because, when pure, they contain only one frequency; that is, they should contain no supply frequency harmonics (i.e. multiples of the fundamental frequency). Unfortunately, due to the increasing number of non‐linear loads connected to the power system and the non‐sinusoidal currents they consume, harmonic voltage distortion is becoming an increasing problem. This is particularly so in the LV network and, as a result, the AC voltage we receive often contains some harmonic distortion. However, despite this it is still approximately sinusoidal.
Power distribution systems universally use a system of three‐phase sinusoidal voltages, with each phase displaced from the next by 120°, as shown in Figure 1.1. These potentials are generated by rotating a magnetic field having a sinusoidal spatial flux distribution inside a machine having three sets of fixed stator windings. The sinusoidal magnetic flux cutting each stator conductor induces a time‐varying sinusoidal potential within it and the addition of the induced potential in each conductor in the winding produces the associated phase voltage.
Figure 1.1 Three‐phase alternating voltages.
In mainland Europe, the UK, China, India and the Middle East, the supply frequency is 50 Hz, whereas in the USA, Canada and parts of South America and Asia, 60 Hz is used.
Phase voltages are measured with respect to the neutral terminal, located at the star point of the winding, and are often represented by colours, in order to distinguish one phase from another. Red, white and blue (R, W, B) are frequently used, but in some locations red, yellow and blue (R, Y, B) are preferred. Alternatively, the letters A, B and C, or U, V and W are often used instead. For convenience throughout this text, we will use A, B, C, or a, b, c.
Low voltage customers are generally supplied from a three‐phase distribution transformer, although single‐phase transformers are often used for smaller loads. Figure 1.2a shows the transformer winding arrangement most commonly used. The primary windings or the medium voltage (MV) windings are delta connected (Δ) and are supplied from the three‐conductor MV bus. The secondary or the low voltage (LV) windings are connected in a wye (Y) or star configuration, within which the LV potentials are induced.
Figure 1.2 (a) Typical distribution transformer winding arrangement (b) Low voltage representation.
Each low voltage phase is referenced to a common neutral terminal (n), and while the three‐phase voltages each have the same magnitude, they differ from one another in phase by 120°. The magnitude of the phase voltages varies throughout the world, but in many countries a phase potential around 230 V is used.
The sequence in which the phase voltages reach their maximum value is important. Just as there are only two possible rotational directions for the magnetic field within a machine, (clockwise and anticlockwise), there are also only two possible phase sequences; ABC and ACB. The phase sequence of the supply determines the direction of rotation of polyphase motors. Reversing the phase sequence of a three‐phase supply, by swapping any two phases, will thus reverse the direction of rotation of any machine connected to it.
Some loads, such as three‐phase motors, do not always require a neutral connection; instead they operate from the potential difference between the phase voltages. These potentials are known as line voltages (or line‐to‐line voltages) and are illustrated in Figure 1.3. The line voltage amplitudes are greater by a factor of √3 than the phase voltages from which they are derived. Thus if the phase voltage is 230 V then the associated line voltage is about 400 V (often expressed as 230/400 V), and this is used to designate the system potential.
Figure 1.3 Phase and line voltages.
It is also worth noting that there is a 30° phase shift between phase voltages and the associated line voltage. For example, the a phase voltage leads the ac line voltage by 30°, i.e. it passes through its maximum 30° before the AC line voltage peaks. Similarly, the b phase voltage leads the ba line voltage by 30°, and finally the c phase voltage leads the cb line voltage by a further 30°, as shown in Figure 1.3.
1.2 Low Voltage Distribution
Most European countries use 230/400 V for LV distribution, or in some cases 240/415 V instead. The majority of residential customers receive a single‐phase supply (i.e. one phase conductor and the neutral) which terminate at the customer’s premises via a service drop (or service connection). These conductors frequently run from the overhead LV network, to a roof or wall mount on the customer’s dwelling, and from there to a service fuse, which can be used for isolation purposes. The supply cables or customer mains then run via the electricity meter to the customer’s switchboard. Customers who have an underground LV supply are likely to have mains running from a distribution kiosk on the street to their premises.
Commercial and industrial low voltage customers in Europe, Australia and New Zealand are usually provided with a 230/400 V three‐phase supply (three‐phase conductors and the neutral) from an LV transformer arrangement like that in Figure 1.4. These customers may therefore run single‐phase loads between any phase conductor and neutral, as well as line connected three‐phase loads. The four‐wire wye (or star) 230/400 V connection is probably the most common LV distribution arrangement, and it is used in many countries worldwide.
Figure 1.4 Typical 230/400 V three‐phase supply.
In some countries, large commercial or industrial LV loads can also be supplied with a three‐wire 400/690 V supply. In the USA and Canada, however, residential customers are often supplied with two voltages: 120 V for lighting and low‐current loads and 240 V for larger single‐phase loads, such as water heaters or power tools. These voltages are often generated within the same single‐phase transformer using a centre‐tapped winding as shown in Figure 1.5. The centre tap (or neutral terminal) is earthed (connected to ground potential) and therefore the potentials of the two active conductors are 180° apart, that is, they are symmetrical with respect to the earthed neutral conductor. Such single‐phase transformers are usually relatively small, (<100 kVA) and generally supply fewer than ten residences.
Figure 1.5 Single‐phase transformer using a centre‐tapped winding.
In older parts of North America other customers are supplied with a three‐phase four‐wire supply from a delta connected secondary winding (Figure 1.6), where one winding has been fitted with a grounded centre tap. This arrangement provides two 120 V supplies for lighting and low‐current loads, while also providing a three‐phase 240 V supply for high‐current loads such as water and space heating. Because one phase lies 208 V away from ground while the others are only 120 V from it, this arrangement is known as a high leg delta connection.
Figure 1.6 Three‐phase four‐wire supply from a delta connected secondary winding.
One advantage of the high leg delta arrangement is that, if necessary, it can be provided in an open delta configuration from two single‐phase transformers rather than three, as shown in Figure 1.7; however, the maximum three‐phase load that can be applied in this situation is only 1/√3 or 57.7% of the rating with three transformers present. In other parts of the USA, residential customers are provided with a three‐phase dual‐voltage supply derived from a delta/wye transformer (Δ/Y) with a phase voltage of 120 V and a line voltage of 208 V.
Figure 1.7 Open delta configuration from two single‐phase transformers.
Commercial and industrial customers in the USA frequently require a three‐phase supply with line voltages greater than either 208 or 240 V; for these a conventional delta/wye (Δ/Y) connected transformer is used, having phase/line voltages of either 265/460 V or 277/480 V.
Finally, there is one LV winding arrangement that has historically been used in the USA for mainly three‐phase loads, which deserves special comment. It is known as a corner grounded delta, or a grounded b phase system (Figure 1.8). This configuration is no longer used in new installations but it was used for many years in irrigation and oil pumping applications and is occasionally still in use today. One advantage of the corner grounded delta is its ability to continue to supply load when operating as an open delta, as described above. This feature provided the possibility of doing maintenance on each single‐phase transformer in turn, from a three‐phase bank, without having to disconnect the entire load. Today, three‐phase transformers are more reliable and are not assembled from single‐phase devices, and therefore the corner grounded delta has lost its popularity.
Figure 1.8 Corner grounded delta, or a grounded b phase system.
There are some disadvantages associated with this configuration that should also be mentioned. Firstly, for reasons of safety it is necessary to identify the grounded phase at all points along the line. Secondly, a phase‐to‐ground fault in a corner grounded delta network is essentially a phase‐to‐phase fault, and in 480 V systems quite large fault currents can flow through one pole of the circuit breaker. Interrupting such single‐phase fault currents can be challenging, especially if the circuit breaker chosen does not have a sufficiently high single pole interruption rating.
Many moulded case circuit breakers used in LV applications have a considerably higher three‐phase fault interruption capacity than a single‐phase one. A bolted phase‐to‐phase fault (i.e. one with zero impedance) can generate as much as 87% of the current available in a three‐phase fault, and the breaker chosen must be capable of clearing this level of fault current. Not all moulded case circuit breakers have the ability to interrupt single‐phase currents of this magnitude. Circuit breakers used in corner grounded delta networks must therefore be carefully chosen with the necessary single pole interruption capacity in mind.
1.2.1 Voltage Tolerance
The LV distribution voltage levels mentioned above are all subject to a prescribed tolerance. This is partly because the MV system voltage fluctuates as a function of changes in the reflected LV load. Although the MV system potential is regulated by the HV/MV transformer (through the use of an on‐load tap‐changer), voltage drops that occur throughout the MV network, as well as within the MV/LV transformer itself, are directly reflected into the LV network since the MV/LV transformer usually does not have on‐load tap‐change facilities, that is, it operates at a fixed transformation ratio.
The International Electrotechnical Commission (IEC) standard IEC 60038, entitled ‘Standard Voltages’ defines a ±10% tolerance1 on the nominal supply voltage for 50 Hz LV systems, and +10% −5% tolerance1 for 60 Hz systems. It also allows a 3% voltage drop for lighting circuits and 5% for all other circuits within a customer’s premises. Collectively this means that for 50 Hz systems the Utilisation Voltage seen at outlets within a customer’s premises may vary from the nominal voltage by +10% −13% for lighting circuits, and +10% −15% for all other circuits. For 60 Hz systems, the overall tolerance is +10% −8% for lighting circuits and ±10% for all other circuits. Table 1.2 shows the typical ranges of common LV distribution voltages.
Table 1.2 Typical distribution voltage ranges.
| System frequency (Hz) |
Nominal voltage | Highest supply or utilisation Voltage | Lowest supply voltage | Lowest utilisation voltage for lighting | Lowest utilisation voltage for other loads |
| 50 | 230/400 | 253/440 | 207/360 | 200/348 | 196/340 |
| 60 | 120/208 | 127/220 | 109/190 | 106/184 | 104/180 |
| 60 | 230/400 | 253/440 | 219/380 | 212/368 | 207/360 |
(IEC 60038, Copyright © 2009 IEC Geneva, Switzerland. www.iec.ch)
1.2.2 Load Balance
It is important that the total connected load is shared almost equally across all three phases, that is, it is balanced. This is generally achieved in a residential distribution system by ensuring that the number of customers connected to each phase is roughly equal, given that residential loads are generally similar and they also have comparable daily load profiles. Whether a small commercial or industrial customer’s load is balanced usually depends largely on how well the single‐phase load is distributed, since three‐phase loads, (motors for example) tend to be inherently balanced. On the other hand, very large industrial customers (base load customers) are required to present a balanced load to the network as part of their connection agreements.
1.3 Examples of Distribution Transformers
An engineer can learn much by observing equipment in the field. In particular, reading equipment nameplates can reveal a lot about an installation. We will consider two examples of distribution transformers: Figure 1.9 shows a public pole‐mounted LV distribution transformer in China, and Figure 1.10 shows an Australian ground‐mounted distribution transformer enclosure (also known as a pad‐mounted transformer). The pole‐mounted device is fed from three‐wire 10 kV droppers on the left‐hand pole, while the four‐wire 400 V LV supply leaves the transformer as an aerial bus via the right‐hand pole. MV drop‐out fuses in series with the incoming droppers protect the high voltage side, and exposed LV fuse elements mounted on the LV isolator on the right protect the low‐voltage side. The LV conductors also pass through a set of metering current transformers (in the metal enclosure), so that the energy supplied by this transformer can be measured.
Figure 1.9 A 500 kVA pole‐mounted distribution transformer (China).
Figure 1.10 A 500 kVA ground‐mounted transformer (Australia).
The pad‐mounted transformer enclosure shows little from the outside; internally, however, it is actually a miniature 11 kV:415 V substation. It contains both MV and LV switchgear, located in opposite ends of the enclosure, with the distribution transformer mounted in between. This particular unit is dedicated to one customer and is fed from an underground 11 kV ring main; it supplies a large building on an industrial site. Both the MV and LV cabling is underground, and therefore this installation is aesthetically more pleasing than the pole‐mounted one.
1.4 Practical Magnitude Limits for LV Loads
There are significant advantages in moving to higher potentials for larger loads, particularly the associated reduction in both supply current and electrical losses. This change also permits smaller section conductors to be used throughout the distribution network, which can provide considerable savings. It is for these reasons that commercial and industrial LV customers operate on 400 or 460 V rather than 230 or 120 V.
However, as industrial loads increase in size, eventually a limit is reached where it is no longer economic to provide the service from an LV supply. This limit generally occurs in industries where the site load is considerably larger than that usually found in the public network, for one of the following reasons. Firstly, very large phase currents demand very large supply conductors, which can become expensive, especially if they have to run a considerable distance from the distributor to the customer. Secondly, depending on the available transformer capacity, it may be necessary for the distribution company to install a dedicated transformer to cater for the requirements of particular customers. Thirdly, for particularly large loads and low impedance LV supplies, the prospective fault currents may become too large for conventional LV switchgear (circuit breakers) to handle.
The third point requires further clarification. Moulded case circuit breakers, frequently used in LV applications have several current ratings. For the purpose of this discussion we will consider only two: the continuous current rating and the service fault interruption capacity. The continuous current rating is the current that the breaker can continuously carry and interrupt safely; it is the nominal rating of the device (although it may be higher than the current at which the breaker is actually set to trip). The service fault interruption capacity is the value of the (balanced) three‐phase fault current that the breaker is capable of interrupting. This is the practical limit of the ability of the device to clear a downstream three‐phase fault.
The continuous current ratings of LV circuit breakers can be as large as 6000 A, but breakers of this size are generally for special applications. The current rating of LV circuit breakers used in general customer applications seldom exceed around 3200 A, and for these devices the service fault interruption capacity is frequently of the order of 50–70 kA.
A practical limit to the size of an LV load arises when the prospective three‐phase fault current cannot be interrupted by the selected LV circuit breaker. For example, consider a 2.2 MVA Δ/Y distribution transformer delivering 400 V to an LV bus. The rated secondary current for such a transformer is 3175 A (=2.2 MVA/(400 V√3)), so a 3200 A breaker would just suffice in this application (although it is not good engineering practice to operate either the transformer or the circuit breaker continuously at the limit of their ratings).
Assuming that the MV bus impedance is small with respect to that of the transformer, then the latter will limit any three‐phase fault current. As we shall see in a later chapter, the impedance of a transformer (sometimes called the voltage impedance) is usually expressed in per cent, with typical values being 4–20%. The percentage impedance is defined as that fraction of the nominal supply voltage which, when applied to a transformer with short circuits on all phases, results in the rated current flowing in each phase. Therefore the three‐phase fault current that would flow with the transformer fully excited, can be calculated by dividing the rated current by the transformer’s impedance. For example, if our transformer has an impedance of 5% (0.05) then the three‐phase fault current will be approximately 3175/0.05 = 63.5 kA, and this is the current that the LV breaker must be able to successfully clear in the event of such a fault.
Choosing a 3200 A breaker with a 70 kA service fault interruption capacity would therefore be acceptable in this instance, but a circuit breaker with only a 50 kA interruption capacity clearly would not. If it were deemed necessary to use two such transformers on the same LV bus, then neither of these circuit breakers would be acceptable, unless the transformer impedances were increased substantially.
Because of the limitations imposed by switchgear capacities and the size of the LV busbars that would be required, industrial customers generally segregate their load into smaller portions, fed individually from smaller transformers. This is often dictated by the location of equipment across a customer’s site and by a desire to minimise the load lost in the event of a transformer outage.
As an alternative to using one large transformer, a customer may take several independent LV supplies from the distribution network, perhaps at different locations across a site. Or instead, energy may be taken directly from the MV bus, and distributed to one or more private MV‐LV substations. While this demands a substantial investment in electrical infrastructure, and the ability for maintenance staff to operate both LV and MV equipment, the MV tariffs available may make this approach attractive.
Often the distribution company will provide a direct MV connection to a customer’s own MV‐LV transformer, either pole or ground mounted (Figures 1.9 and 1.10). This arrangement can be used to provide the customer with a dedicated LV supply, while obviating the need for MV distribution across their site. In both these cases the MV supply will pass through a metering unit, containing the necessary voltage and current transformers so that the energy used can be appropriately metered.
The VA threshold at which a customer will be directed to the MV network varies from country to country; in France this can be as little as 250 kVA, in some parts of the USA it can be as high as 2 MVA, while in Australia it is of the order of 1 MVA.
1.5 Medium Voltage Network
The purpose of the medium voltage distribution network is to transport electricity from the transmission or sub‐transmission substations to public customers via the LV network, or to industrial customers supplied directly at MV potentials. It represents a major infrastructure investment on the part of the distributor, and its design has a significant influence on the level of service and cost experienced by customers.
The IEC standard IEC 60038 ‘Standard Voltages’ specifies the MV potentials shown in Table 1.3 for both 50 Hz and 60 Hz systems. It defines two voltage levels, the nominal system voltage which is used to identify an MV system, and the highest voltage for equipment which is the highest voltage which may occur under normal operating conditions. Some or all of the voltages shown in Table 1.3 are used by MV distributors throughout their networks. It should be noted that although the nominal system voltage is usually used to identify a system voltage, equipment must be designed to operate continuously at the highest voltage for equipment. For example, a device designed for 11 kV must be able to tolerate a supply potential as high as 12 kV.
Table 1.3 Medium voltage levels.
| Australia, NZ & Europe (50 Hz) | North America (60 Hz) | ||
| Nominal system line voltage | Highest voltage for equipment | Nominal system line voltage | Highest voltage for equipment |
| 3.3 kV | 3.6 kV | 4.16 kV | 4.4 kV |
| 6.6 kV | 7.2 kV | 12.47 kV | 13.2 kV |
| 11 kV | 12 kV | 13.2 kV | 13.97 kV |
| 22 kV | 24 kV | 13.8 kV | 14.52 kV |
| 33 kV | 36 kV | 24.94 kV | 26.4 kV |
| — | — | 34.5 kV | 36.5 kV |
(IEC 60038, Copyright © 2009 IEC Geneva, Switzerland. www.iec.ch)
While networks may adopt potentials within the range outlined in Table 1.3, some jurisdictions provide a tighter tolerance in the case of large MV customers, typically the nominal system voltage ±5%.
The voltage chosen depends upon the length of the feeder concerned and the total load (MW) that it must supply. Since losses are proportional to the square of the currents flowing, they can be minimised by choosing as high a system voltage as possible; however, equipment costs rise steeply with system voltages above about 33 kV, and therefore an economic compromise must be made as to the choice of voltage and the acceptable network efficiency.
In addition, the voltage drop along an MV feeder must be maintained within prescribed limits. IEC60038 specifies that the highest MV voltage and the lowest MV Voltage should not differ from the nominal voltage by more than ±10% for 50 Hz systems and +5% −10% for 60 Hz systems. Given these tolerances the practical limit to voltage drop along an MV feeder will generally be less than 10%, although this may often be relaxed under contingency conditions.
The voltage drop along a feeder is a function of both the line resistance and its reactance, and therefore for a given power transfer, the line voltage drop will be influenced by the average power factor along the line. As the line power factor gets smaller (worse), the line current will become larger and with it the voltage drop and also the line losses.
The line’s reactance to resistance, (X/R) ratio also has a significant influence on the voltage drop. Large X/R ratios mean that a transmission line’s impedance is largely inductive, and therefore lagging currents will contribute significantly to the cumulative voltage drop along the line. For this reason the transmission of energy at high power factors improves the apparent capacity of a line.
For a given power flow, it can easily be shown that the line loss is inversely proportional to the square of the system voltage. Figure 1.11a shows the maximum power that can be transmitted along a 33 kV MV feeder, subject to a 7.5% voltage drop, as a function of line length. Four curves are shown: the upper two apply for 120 mm2 conductors with 0.9 and 0.8 power factors, while the lower two apply for 50 mm2 conductors with the same power factors. Clearly, as the line length increases, the power transmitted must be reduced if the voltage drop constraint is to be met. Further, as the average power factor becomes worse, the apparent line capacity diminishes and the addition of shunt capacitance at regular intervals can be used to improve MV voltage regulation.
Figure 1.11b shows the quadratic nature of the relationship between transmitted power on a 10 km line and system voltage, for the same voltage drop requirement. The choice of the transmission voltage clearly depends on both the power to be transmitted and the distance over which it must be sent. In reality, the problem is a little more complex since feeders rarely deliver all their power to an end‐connected load. The sending end power tends instead to be absorbed progressively by loads distributed along the line’s length, as shown in Figure 1.12. In order to extend the transmission distances available from MV feeders, autotransformer based regulators or step regulators are frequently inserted at intervals along a feeder to restore the voltage drop.
Figure 1.11 (a) Transmitted power vs line length (b) Transmitted power vs system voltage.
Figure 1.12 Voltage profile of MV feeder trunk and spurs.
Figure 1.12 shows an example of voltage profiles of an MV feeder trunk and spurs, as functions of distance from the source substation. Close to the substation the trunk feeder potential falls quite steeply, since the load carried is heavy. A regulator located about 13 km from the substation produces a sharp rise in the voltage downstream, but the potential of the remaining spurs continues to fall, with the most heavily loaded decaying fastest. The lower voltage planning limit in this case is 94% (0.94 per unit), and the potential on one heavily loaded spur falls below this limit at about 27 km from the substation. Assuming that this voltage profile represents the worst case, and depending on how frequently it arises, a new regulator may be required at a suitable upstream location. Voltage profiles like this are useful tools for managing feeder performance and planning network upgrades.
1.5.1 Three‐Wire and Four‐Wire MV Circuits
The circuit arrangements used for MV distribution vary considerably from one country to another, the main differences being the number of conductors and the earthing arrangements used on the MV side of the HV/MV transformer.
In North America and associated countries, a four‐wire system is used. The HV/MV transformer is generally a delta/wye configuration (Δ/Y), generally with the MV star point directly (solidly) grounded at the transformer. The MV network consists of four wires: the three‐phase conductors and the neutral. The neutral is grounded at regular intervals of about 250 m, thereby ensuring that the neutral conductor is close to ground potential along the route of the feeder. This permits the use of single‐phase MV/LV distribution transformers energised from a phase potential for residential loads, in addition to three‐phase transformers for commercial and industrial loads.
In the USA, single‐phase MV transformers are available in 10, 15, 25, 37.5, 50 and 100 kVA capacities, and often have dual primary voltage ratings so that they can also operate from the local line voltage if required. The use of single‐phase transformers like these provides a significant advantage where light loads are supplied from branch or spur lines running off a main trunk feeder. In this situation, the branch line can consist of only one phase conductor and the neutral, resulting in a considerable saving in material costs. (Similar techniques are also used in other countries where two‐phase spur lines supply single‐phase MV/LV transformers.)
The solidly earthed neutral means that phase‐to‐earth faults can generate large currents, limited only by the network and local ground impedances. These currents therefore tend to diminish with the distance from the parent substation. Unfortunately, the single‐phase loads on the MV network also result in substantial currents in the neutral, to the point where it can be difficult to discriminate between the two. As a result, it is necessary to break the line up into protection zones and to use a distributed protection system for each one, capable of discriminating between earth fault currents and residual load current, within each zone.
This arrangement means that the protection system is highly dependent on the configuration of the network, making it difficult to temporarily rearrange during fault recovery or indeed during major upgrades. In North America, the MV network is mainly an aerial one, emanating radially from a parent MV substation.
By contrast the MV network in most of Europe, Australia and New Zealand is based on three‐wire distribution. As before, HV/MV transformers are frequently delta/Wye (Δ/Y) connected, although Y/Y connected transformers with the HV star point open, are used as well. The MV star point is usually grounded, sometimes directly and sometimes through an impedance designed to limit the magnitude of earth fault currents. In some countries, the star point is not grounded at all, and the MV system is constrained around ground potential by the effects of stray phase‐to‐earth capacitances.
An advantage of three‐wire MV distribution lies in the fact that there are no phase‐to‐neutral loads and therefore no neutral current flows, except for a small residual capacitive current, flowing back to the transformer’s star point. (This is the result of unbalanced stray capacitances that exist between the phases and ground.) MV/LV transformers are either three‐phase, for larger LV loads or line connected single‐phase units, for small residential loads.
The lack of neutral current makes the detection of phase‐to‐ground fault current relatively easy, either by directly measuring the neutral current, or by a residual overcurrent measurement on the phase conductors themselves. Both these measurements can be made at the parent substation, which considerably simplifies the protection system. In particular, this protection approach to three‐wire MV networks is tolerant of changes to the network, thus facilitating network reconfiguration during fault recovery. In this respect, three‐wire MV distribution networks can provide a higher level of service.
While impedance earthing is effective in reducing the magnitude of earth fault currents, it can also lead to significant over‐voltages throughout the network. This is because the neutral potential is not tightly constrained and is therefore able to rise in potential during fault conditions.
1.5.2 MV Network Topologies
There are several topologies used in MV networks, and the use of each depends upon the density of the connected load and its relative importance. In rural areas where there are relatively few customers connected to each distribution transformer, the failure of part of the local MV network or of a distribution transformer itself, may not inconvenience many customers, and if the distribution company is prompt in carrying out repairs, the inconvenience will be brief. (Many distribution companies take advantage of this and allow LV distribution transformers to run to failure, thus avoiding the expense of a transformer monitoring and replacement programme.)
By contrast, in more densely populated areas, the number of customers supplied from an entire MV feeder can be substantial, and system faults may cause serious interruptions to supply, including critical loads such as hospitals and health centres. Because of this, parts of the MV network are often designed to provide a high level of redundancy that can be exploited during fault recovery so as to minimise the number of customers ultimately affected. The cost of this approach must be weighed against the benefits to the community, and in general the more redundancy provided, the more expensive the network becomes.
There are many different network arrangements in use worldwide and while we can’t discuss them all, we will consider three fundamental topologies; radial feed, open loop feed, (or ring main feed) and mesh networks.
Radial Feed Lines
In low population density areas, MV distribution is generally achieved using aerial radial networks. These are relatively inexpensive but tend to suffer from a lack of redundancy and are more prone to faults through contact with lightning, foliage or animals. Each customer fed from a radially connected feeder has only one source of supply, and the failure of part of the upstream network means an outage until the fault can be repaired. Radial feed arrangements tend to be tree‐like in structure, with many branch feeders emanating from a major trunk feeder. Although there is no redundancy in terms of source of supply, these networks generally have the ability to automatically detect and isolate faulty network segments. Thus a minimum number of customers will experience power loss until repairs can be made.
Faults to aerial networks arise for several reasons: lightning strikes or wildlife and foliage impinging on the line, are but a few. Many line faults are cleared by the fault current during the fault or during a subsequent re‐energisation of the line. A modern MV feeder is equipped with numerous devices that collectively are able to detect, clear or isolate faults and then restore supply, while causing a minimum of disruption to customers.
Recloser Circuit Breakers
Reclosers (or automatic circuit reclosers, ACRs) are intelligent electronic devices (IEDs), containing current and voltage transformers; they are capable of measuring and storing both operational and fault record data and can communicate with the MV network control centre, from where they can be operated remotely (Figure 1.13). The prime function of a recloser is to segment the line in the event of a ‘permanent’ fault that cannot be cleared by several reclosing attempts. In the event of a fault, the nearest upstream recloser will initially trip on the fault current and after a brief period will reclose for a short time, in an effort to clear whatever may have caused the fault. This process occurs a maximum of four times before the recloser permanently locks itself open. (The number of reclose attempts may be remotely altered as required, if necessary. It is usual to reduce this number to zero in times of severe fire danger, or when live‐line work is being carried out.)
Figure 1.13 Recloser, bypass switches and isolator switches.
For example, a downstream transformer fault can usually be cleared by the operation of its local protection fuse(s). An animal carcass or foliage on the line can often be blown clear by subsequent applications of the fault current itself, but if the recloser attempts fail and the fault current remains, the recloser will lock itself out in the open position, thereby isolating the faulty portion of the network.
It is common to use several reclosers in series at intervals along a line and in this event it is necessary for downstream devices to be graded with those upstream. They must be programmed so that the downstream devices trip earlier and possibly on a lower current threshold than those upstream, so that the number of customers disrupted is minimised.
Reclosers can also be used to reconfigure a network under remote control from the network control centre. This is frequently done during fault recovery, where normally open points in the network are temporarily closed, so that power can be restored to sections which would otherwise suffer an outage.
As an example, Figure 1.14 shows a typical radial network, consisting of two trunk feeders originating from an HV/MV substation. This drawing is an example of a one‐line diagram, where one line is used to represent all three phases, and the neutral conductor if required. The outgoing feeders supply many MV/LV distribution transformers, each of which is protected by a set of fuses. At the fork in the left‐hand feeder are two reclosers, R1 and R5. Should a ‘permanent’ fault occur downstream of R5, this device will lock itself open, isolating the fault, while R1 remains closed, having seen no fault current. This situation can be further improved through the uses of sectionalisers or load break switches.
Figure 1.14 Radial MV network.
Sectionalisers or Load Break Switches
Sectionalisers are also IED devices and are thus capable of being remotely controlled and providing operational data to a network control centre (Figure 1.15). They are used in conjunction with reclosers to isolate a faulty section of line. While they can interrupt load currents, they have limited fault current making and breaking abilities. In particular, they do not have the fault current interruption capabilities of reclosers.
Returning to our example in Figure 1.14, a permanent fault on the line downstream of S3, will ultimately see R5 lock out, resulting in the loss of all the associated load. In order to prevent this, sectionaliser S3 measures fault current pulses created by the reclose attempts of R5, counting the number of reclose cycles. During the open interval of the penultimate reclose cycle of R5, S3 opens, thus isolating the faulty circuit so that when R5 recloses for its final time no fault current is seen. R5 consequently remains closed and thus supplies the S2 load. Therefore by working together, recloser R5 and sectionaliser S3 collectively act to isolate only the faulty portion of the network, while maintaining supply to the majority of customers without the need for inter‐device communication.
Figure 1.15 Sectionaliser or load break switch.
Consider now the feeder sections downstream of R3 and R4 in Figure 1.14. Because there is significant load downstream of these reclosers, provision has been made for a backup supply from the adjacent feeder in the event of a fault. For example, in the case of a permanent fault on the line between R2 and R4, recloser R2 will ultimately lock out, isolating the faulty line segment, but this will see the loss of a considerable amount of load, normally supplied through R4. If recloser R4 is remotely opened and sectionaliser S1 is closed, the network can be reconfigured so that these customers are maintained via R3 and R1 until the fault can be repaired. Therefore by temporarily reconfiguring the network, much of the load can be restored immediately. In many instances it is possible to program a normally open sectionaliser with a set of rules which will be followed in the event of a disruption to the supply, thereby permitting automatic reconfiguration of the network, without the need for remote intervention.
Voltage Regulators
The voltage drop constraints mentioned earlier can be overcome to a large extent by inserting step voltage regulators in the line. Two such regulators are included in Figure 1.14 on the long line sections running from reclosers R1 and R2. These are generally autotransformers equipped with on‐load tap‐changers and can automatically regulate the downstream voltage over a range of about ±10%, thus virtually removing the effects of line voltage drop.
On long lines with end‐connected loads and a mid‐line regulator, it is also possible to configure these devices to regulate the voltage at the receiving end of the line, a technique known as line drop compensation. This is achieved by measuring both the current flowing and the voltage at the output terminals of the regulator, and from knowledge of the line’s impedance it is possible to calculate and therefore regulate the voltage at the receiving end.
Figure 1.16 shows an example of a three‐phase wye/wye (Y/Y) connected 22 kV autotransformer, fitted with a tertiary delta winding and a 16‐position on‐load tap‐changer. It is capable of regulating line voltages from 18,700 V to 22,550 V, and has a line current rating of 131 A and a VA rating of 5 MVA. Although the nameplate rating suggests that the transformer will deliver 5 MVA, this is really just the nominal rating of the line at 131 A. The maximum VA contribution delivered by this transformer is:
Figure 1.16 Three‐phase 22 kV autotransformer regulator.
Thus a relatively small autotransformer can regulate a much higher capacity feeder.
Figure 1.17 shows a pair of single‐phase, 330 kVA, 150 A step voltage regulators, arranged in an open delta configuration on a 22 kV line. In this configuration these are able to regulate a 3.8 MVA feeder; they are also based on an autotransformer design, complete with an on‐load tap‐changer. Single‐phase regulators of this type can be easily replaced or maintained, without the need to remove the line from service through the inclusion of bypass and isolation switches as part of the installation. They can also be installed singly on two‐wire feeders, in pairs in open delta configurations and in threes in either a three‐wire closed delta or a four‐wire grounded wye arrangement. It is common to series connect line regulators on long MV feeders.
Figure 1.17 Single‐phase 22 kV step regulators (open delta connection).
Open Loop Feed
The open loop feed arrangement, also known as a ring main arrangement, is used in more highly populated areas where many customers are supplied from each MV transformer, and consequently where supply reliability is important. This arrangement is shown in Figure 1.18 and it includes two possible MV supply routes to each downstream substation. In each loop there is one open point which defines the path of supply to each distribution transformer. Ideally, each loop will originate from a different busbar within the parent MV substation, or possibly from a different MV substation, ensuring a further level of supply redundancy.
Figure 1.18 Open loop feed. (Extracted from Cahier Technique No. 203, ‘Basic selection of MV public distribution networks’ with kind permission from Schneider Electric).
Should a fault develop in a section of one loop, the open point can be relocated and the fault isolated by opening the loop in a second location as well. Therefore most if not all of the customers can have their supply rapidly restored. Open loop feeder arrangements are frequently implemented as part of public underground cable distribution networks. Private MV networks in large buildings, educational campuses, hospitals and other public facilities also frequently use this distribution system, enabling multiple MV/LV transformers to be supplied from within a customer’s premises. Each node in the loop requires a small suite of MV switchgear, sufficient to provide for at least one incoming switch from the loop, one outgoing switch to the loop, and an MV circuit breaker through which the local transformer is supplied.
A switchboard of this kind, used in a hospital’s private 11 kV network, is shown in Figure 1.19. The incoming feeder enters via the extreme left‐hand cubicle and the outgoing one departs from the cubicle adjacent to it. These cubicles are provided for isolation purposes only; they contain a load break switch rated at 630 A and insulated with sulphur hexafluoride gas (SF6), as well as an interlocked earth switch, capable of earthing the incoming cable. The right‐hand cubicles provide feeds to two local distribution transformers via SF6 insulated circuit breakers. The operating mechanism for these can be seen protruding from the lower left of each cubicle. Incorporated into each circuit breaker is a protection current transformer, and the associated protection relay can be seen above the mimic panel.
Figure 1.19 11 kV ring main switchboard.
On the front of each cubicle is a mimic diagram showing the function of the switchgear within. Figure 1.20 shows the mimic diagram on the transformer circuit breaker panel. The horizontal line represents the 11 kV bus, while the vertical one running from it represents the feed to the transformer, which passes through an isolator before entering the circuit breaker. This is operated by inserting a lever into the receptacle behind the clear panel on the lower right of Figure 1.20; an interlock mechanism prevents the isolator from being operated unless the circuit breaker is open.
Figure 1.20 Circuit breaker and mimic diagram.
The earth switch is used to ground the outgoing bus connection to the transformer during maintenance activities. It is operated using the same lever; this time it is inserted into the receptacle on the upper part of the mimic panel, shown in Figure 1.20. The interlock mechanism also prevents the earth switch from being closed unless the circuit breaker is open.
Also visible on the left‐hand side of Figure 1.20 is a viewing window, which permits the operator to view the state of the isolator. Viewing windows are also provided for the earth switch, and these can been seen on the breaker cubicle in Figure 1.19. Verification of the position of both these switches is a critical safety requirement when maintenance is to be carried out.
Mesh Networks
Mesh networks are used in cities where the population density is high and there are many MV/LV distribution transformers. Supply security is therefore important, and a mesh arrangement similar to that shown in Figure 1.21 is often used. This is really a high density application of the open loop arrangement discussed above. Cables are often run beneath footpaths or streets and service nearby buildings, either with an LV supply from a local distribution transformer, or with a private MV supply for the transformers within the building.
Figure 1.21 Medium voltage mesh network.
Figure 1.21 includes numerous ring main switchboards, supplying an MV/LV distribution transformer, each with a supply from both sides of the ring. Interconnectors may also run between strategic feeders providing a further level of redundancy. Protection zones are defined within sections of each feeder as shown, and should a fault occur in any one it will automatically be isolated from the network with the minimum loss of load.
1.6 Transmission and Sub‐Transmission Networks
While there is an increasing amount of embedded generation within today’s distribution networks, the bulk of the electrical energy is still generated by large coal, hydro, gas or nuclear‐powered machines. These are generally located near their respective fuel source, since it is generally more economical to transport electricity than either coal or gas in the quantities required. Likewise hydro generation must, by definition, occur where the water power is available, while the location of nuclear‐powered plant tends to be a compromise between safety concerns and the proximity to the electrical load.
It is therefore the job of the transmission system to transport the electrical energy from where it is generated to the load centres where it will be used. This can be further broken into transmission and sub‐transmission networks. The transmission network often extends over considerable distances and can carry very large quantities of energy. In most applications the transmission system consists of three‐phase open wire lines, supported by large transmission towers and operating at extra high voltages (EHV), typically 220 kV–1 MV. EHV transmission cables are used within cities where it is impractical to erect transmission towers, but their lengths are generally limited due to their expense and the large capacitive charging currents that must be supplied. With the advent of high power semiconductor devices in the 1970s and 1980s, high voltage DC links (HVDC) began to appear. Initially these tended to be between different AC networks, spanning borders, in the form of open wire lines, or as DC cables crossing waterways, where it was impractical to operate AC cables. More recently, however, HVDC links have begun to compete with AC transmission lines within AC networks. This is due to the absence of the capacitive charging effects, as well as the lack of an inductive voltage drop along the length of a DC line.
The sub‐transmission system generally operates at slightly lower voltages (66–132 kV) and supplies somewhat smaller loads. It is generally used to interlink major distribution substations within or near city boundaries, and provides redundant supply paths, so that the demand for electricity can be met in the event that part of the network is out of service, either through planned maintenance or through the unplanned operation of the protection system, in response to a fault.
The choice of supply voltage (either AC or DC) is an economic compromise between acceptable transmission losses and system cost. Generally, the higher the chosen transmission voltage the lower will be the total loss, but the more costly will be the terminal equipment which must operate at that potential. Open wire transmission lines being air insulated, do not substantially add to this cost, but the transmission towers they require certainly do. As the transmission potential rises, so do the electrical clearances required, both between the phase conductors themselves and over objects on the ground. These requirements lead to bigger and therefore more costly transmission towers. The requirement to carry two, three or four conductors per phase and the desire to increase the span between towers, further adds to this cost. Finally, although each transmission tower is quite large, it does not provide a span of more than a few hundred meters (see Figure 1.24), so long transmission lines requiring many towers are costly to build.
Transmission Redundancy and Network Planning
In the case of LV networks – and to a lesser extent with MV networks – the loss of a single feeder usually results in the loss of supply to relatively few customers. Therefore occasional outages can be tolerated, so long as their duration is kept short. This is definitely not the case in the EHV network, where transmission lines carry substantial amounts of energy and the unexpected loss of a line may result in severe disruption to the overall network. For this reason, transmission networks must be designed to provide a considerable degree of redundancy, so that the sudden loss of part of the network can be automatically accommodated without a loss of supply to customers.
Redundancy in this context effectively means the provision of excess network capacity together with sufficient network flexibility, in order that the entire load can be carried at times when parts of the network are out of service, either due to a fault or perhaps as a result of a planned maintenance outage. The decision as to how much redundancy should be provided is essentially an economic one. Since the provision of redundancy in the transmission network is expensive, the most likely contingencies should be provided for first, while those considered less likely may be deferred until money becomes available. Sometimes a particular contingency can remain unforeseen until it actually occurs, the recovery from which can involve a widespread loss of supply which can be very expensive, particularly if a black start is required. A black start is required following the total collapse of a power system. Generators must be restarted a few at a time and base load must be progressively restored in such a way that it matches the generation capacity available. The time required to fully restore power to all customers can be considerable, so it is far better to shed load early in order to avoid this possibility.
Often a degree of redundancy can be provided at a discounted cost, if it is carefully planned prior to construction. For example, transmission lines like those shown in Figures 1.22–1.24 can be built as either a single circuit line or a dual circuit line, without a proportional increase in cost. While the dual circuit option provides twice the transmission capacity and therefore appears an attractive option, both circuits share the same route and therefore failures due to environmental factors (such as ground fires) are equally likely to affect both circuits. Truly redundant circuits require route diversity in addition to spare capacity. Dual circuits sharing the same transmission towers are often used none the less, and provide the ability for one line to be taken out of service for maintenance, while the other carries the load.
Figure 1.22 275 kV single circuit transmission tower (France).
Figure 1.23 400 kV dual circuit transmission tower (France).
Figure 1.24 Each tower provides a relatively short span, so many towers are required for a long line.
Redundancy provision also includes the terminal equipment within each transmission substation. For example, the bus arrangements provided have a major impact on the substation’s flexibility; the provision of autotransformers linking one transmission potential with another also affects the ease with which the loss of one transmission line can be mitigated by the others in the substation.
As a result of their importance to network security, individual EHV transmission lines are also equipped with a far higher degree of electrical protection than their MV counterparts. Frequently, independent protection schemes are used simultaneously on the same line, achieved through the use of duplicate protection relays. Even the substation batteries powering these relays must be duplicated, so that there is no single point of failure in common to the two schemes.
Network planning is thus a complex task, which not only includes the provision of adequate redundancy, but must also anticipates future load growth so that the network can be adapted for changes in population density and generation profile. Transmission networks are therefore not only expensive to build, but also to maintain, since they require ongoing upgrading in order to meet community expectations.
Load Classifications and Demand Response
The connected load on an electrical system falls roughly into two classifications: base load and variable load. Base load is provided by large industrial customers, many of whom are connected to the network at either EHV or HV potentials. Aluminium, zinc and nickel refineries are typical base load customers, since they consume large amounts of electricity, usually at high potentials, and operate 24 hours per day. Paper and automotive manufacturers also fall into this class. Base load customers are usually required to maintain power factors, in excess of 0.95, and collectively they provide a stabilising effect on a power network. The power they consume generally has a flat profile, and usually a substantial proportion of the available generation must run continuously in order to meet this demand, thus providing a spinning reserve of deliverable power, which can quickly be made available should the load suddenly increase.
The variable load component consists of residential, commercial and small industrial loads, most of which are supplied at LV potentials. These loads are not ‘flat’, but instead exhibit a seasonal and cyclic daily profile, in addition to which they usually also consume significant reactive power. They are also largely responsible for creating the daily peaks in demand, which the transmission and distribution networks must have sufficient capacity to deliver. These peaks in daily demand generally correspond with meal times, particularly in the morning and evening. Meeting this demand is one of the main aims of the network planners, and is made worse during extreme weather events, either hot or cold, when the additional cooling or heating load substantially increases the maximum demand imposed on the system. The transmission, distribution and generation system must cope with such extreme loads, and it is during these events that the system may become overloaded, in which case the load must be artificially reduced.
Some distribution companies reduce the potential they supply to residential and commercial customers slightly during times of peak demand, in the expectation that the demand will fall sufficiently to avoid overloading the network. This works well in the case of resistive heating, where the power demanded is proportional to the square of the voltage, but it does not have the same effect with other load types. Others companies dynamically reduce the residential hot water heating load, using a control signal imposed on the network itself. A third load control mechanism is to reduce the speed at which inverter‐driven air conditioners operate, and hence the load they impose on the network. This is achieved by communicating with the air conditioner via a smart meter installed on the premises or through a direct internet connection. These mechanisms are all examples of demand response, whereby the demand on the electrical system is reduced during extreme load events.
However, perhaps the best way to control the load at times of extreme demand is via a price signal. It is during extreme weather events like these, that the market price for electricity is greatest. Base load customers (and many smaller customers too) are usually quite price sensitive, and where electricity contracts include an exposure to the market price, they will usually curtail their consumption at times when the price is high.
Building owners and industrial customers often take advantage of the services offered by demand response aggregators. These companies manage blocks of electrical load offered by their clients for load shedding purposes. In an extreme weather event or when network security is threatened the aggregator in conjunction with the network operator, notifies its clients of the need to curtail load, and manages the load reduction throughout the event. This may be achieved by temporarily altering heating or cooling set point temperatures in building management systems, or by the shedding of industrial load for intervals as short as half an hour. Through the load shed by its many clients, often on a rolling basis, the aggregator can offer the network operator substantial blocks of distributed load for shedding when needed. Payments are usually made by the network operator to the aggregator for the provision of this service, as well as for the actual load shed when an event occurs. Clients who can provide larger blocks of load or who can respond rapidly to a load shedding event are paid more than those who can supply less or require advanced notice. Since many extreme weather events can be predicted well in advance, the load can generally be progressively shed as an event unfolds, consistent with network requirements.
Demand response represents one of the major benefits of an electricity market; it flattens the overall load profile and reduces periods of peak demand, thereby deferring the need for network upgrades and capital expenditure.
1.6.1 Transmission System Operation
Unlike MV transmission lines, which must supply the active as well as any reactive power demanded by the loads they service, the transmission system is used primarily for the transmission of active energy from a source of generation to a load centre, or between transmission substations. There are two reasons for this.
Firstly the voltage drop along a transmission line increases substantially as the power factor becomes smaller and for very long lines this can be substantial. This effect is illustrated in Figure 1.25, which shows the effect of a worsening power factor on the fractional voltage drop, for a constant VA flow as a function of the X/R ratio of the line. As load power factor decreases, the line voltage drop increases, as it rapidly becomes dominated by the larger inductive impedance. This effect is particularly evident at higher X/R ratios.
Figure 1.25 Fractional voltage drop as a function of the load power factor, for a constant VA load.
While this drop can be corrected using series connected line regulators in MV networks, voltage regulation of this kind is much more expensive to achieve at transmission potentials.
Secondly, the current capacity of the transmission network represents an expensive asset, one that should not normally be used for the transmission of reactive power, which can easily be generated near the loads requiring it, through either the provision of network capacitors or the installation of power factor correction equipment within customers’ premises.
Reactive power flows along a transmission line as a result of a voltage difference between its ends, flowing towards the lower potential. In contrast, active power flows along a line when there is a phase difference between its ends, and flows towards the more lagging potential. Therefore depending on the conditions at each end of a line, it is quite possible to have these quantities simultaneously flowing in opposite directions.
EHV transmission lines, unlike their MV counterparts, run between substations at which the bus voltages can usually be independently controlled by adjusting the taps on network transformers or by adjusting the excitation of nearby generators. Therefore it is a relatively simple matter to reduce the reactive power flowing along a line to an acceptable level, by adjusting the voltage difference between its ends. This can generally be achieved almost independent of the active power flow required. For these reasons, most transmission systems are operated in such a way that they mainly transport active power.
At transmission potentials the power quality also tends to be high, meaning that, in addition to power factors close to unity, there will be relatively few harmonics, and the line currents will also be well balanced. For this reason the negative or zero‐sequence voltages present on a transmission bus should also be very small.
Frequency Control
It is almost taken for granted that the frequency in large power systems is particularly stable. However, the maintenance of a stable operating frequency requires a precise match between the power being generated and that consumed. Should this balance be upset, for example through a sudden increase in connected load, the system frequency will begin to fall. The rate at which it falls is proportional to the power mismatch and the energy stored in the inertia of the generators themselves. The imbalance will be absorbed from the rotational energy of the machine rotors at the expense of a fall in the speed of each machine until the governor reacts and increases the energy delivered by the prime mover. The energy collectively stored in the system’s rotating inertia is given by the equation:
where J is the collective inertia of the system, and ω(t) is the system frequency, which must now be considered to be a function of time. Differentiating this equation and putting t = 0 yields the following expression for the initial rate of change of frequency:
where ΔP is the initial mismatch between the power generated and that demanded. As shown, the initial rate of change of frequency is proportional to ΔP and inversely proportional to the collective machine inertia, J and the nominal system frequency ω0. Should an excess of generation capacity become available (as the result of a loss of load), then this energy will be absorbed by the machine inertia and the frequency will begin to rise.
The frequency is regulated by assigning a frequency control function to many of the generators operating within the system, so that when a fall in frequency is detected, the governor associated with each machine proportionally increases its machine’s output to restore the balance. Since the load on the network is rarely static, the frequency tends to oscillate slowly about its target level. Because each governor is particularly sensitive to small changes in frequency, the variations about the target are generally very small. Large networks with many generators collectively have a proportionally large rotating inertia and therefore are able to achieve a very tight operational range; typically within ±0.2% of the target frequency. On the other hand, smaller networks with fewer generators will see larger frequency swings as predicted by Equation (1.2). Therefore each governor must work harder to make up for any deficit (or excess) of generation, and as a result a wider frequency tolerance must be accepted.
Under‐Frequency Load Shedding
All networks can be vulnerable to a large and sudden loss of generation, particularly when one or more transmission lines are unexpectedly removed from service due to a fault or perhaps a protection mal‐operation. Depending on how much generation is lost, the remaining machines may not have sufficient capacity to make up the shortfall, or they may not be able to restore the power balance while maintaining the system frequency within prescribed limits. When this situation occurs, it is critical that the generation–load imbalance is rectified as quickly as possible, in order to avoid a complete network collapse and the need for a black start.
At such times under‐frequency load shedding is often used to help rapidly restore the balance. This is a process whereby large predefined blocks of MV or HV load are automatically tripped, in response either to the frequency falling below a predefined threshold, or when the rate of change of frequency becomes excessive. Equation (1.2) shows that the rate of change of frequency is proportional to the power deficit ΔP and therefore this derivative is an ideal control parameter.
Base load is ideal for this, since the inconvenience is felt by relatively few customers, and large amounts of load can be shed rapidly, through the action of local under‐frequency relays. For this reason, most large industries will be required to provide a portion of their load for an under‐frequency load shedding scheme, and acceptance of this will usually be a condition of the network connection agreement. If the frequency continues to fall, then residential and commercial load may also need to be shed, so as to avoid a system collapse.
Restoration of the load tripped usually occurs progressively and reasonably quickly. As soon as the generation shortfall has been rectified or the network is reconfigured to account for the equipment lost and the frequency can again be maintained within its prescribed limits, the load can be progressively restored. In this way, industrial production is disturbed minimally and importantly the security of the electrical system is preserved. Power networks generally employ several stages of under‐frequency load shedding, triggered as the frequency progressively falls, until such time as it finally begins to rise, and the power balance begins to recover. This enables critical loads, such as hospitals, to be shed last.
Voltage Control
The control of bus potentials within the transmission system is usually the responsibility of the associated transmission company, although where a regional electricity market is in operation, this task may also be overseen by the market operating company.
Voltage control is generally achieved by adjusting taps on EHV transmission transformers coupling MV generators to the transmission network as well as on autotransformers operating between different transmission and sub‐transmission potentials, or by adjusting the generator potentials directly. EHV to MV distribution transformers also are equipped with on load tap‐changers, to regulate the MV load voltage. Generally each transformer is equipped with a voltage regulator relay, controlling its tap‐changer so that the local bus voltage is automatically maintained within the prescribed limits, independent of the applied load.
The transmission, sub‐transmission and MV distribution networks are generally monitored remotely by network control operators, who can override the automatic voltage control function should the need arise. They generally also have the ability to remotely reconfigure the parts of the network in emergency situations, in order that supply may be maintained.
Both the HV and MV networks usually contain voltage support capacitors which are either automatically or manually switched into the network prior to periods of peak load. The effect of this is twofold: firstly, they cause the local bus voltage to rise slightly, typically by about 2–3%; secondly, they inject a source of reactive power into the local network sufficient to satisfy the reactive demand of the variable load, the increase in which is creating the peak load condition. This reactive injection also acts to preserve the reactive margin of the system, which may be regarded as the excess in reactive capacity required throughout a network to avoid the possibility of a voltage collapse.
1.7 Generation of Electrical Energy
The worldwide demand for electricity has increased substantially in the past two decades, most of which has come from developing nations. The People’s Republic of China, for example, has seen an almost exponential growth in electricity production this century, accompanied by a parallel increase in coal and wind generation. It is now the largest electricity producer in the world. In contrast, developed nations such as the USA, the UK, France and Germany have seen an almost constant electrical demand during the same period, as shown in Figure 1.26.
Figure 1.26 Annual electricity production by country. (Based on IEA data from © OECD/IEA 2016, IEA Energy Atlas, www.iea.org/statistics, Licence: www.iea.org/t&c; modified by the author).
Electricity has traditionally been generated thermally using steam produced through the burning of fossil fuels such as coal, oil or gas, or through the use of nuclear energy, to drive turbines. These fuels have a high calorific value which has permitted the construction of sizeable machines capable of producing electricity on a large scale and relatively cheaply. Thermal machines, however, require long start‐up and shut‐down intervals making them impossible to start up and shut down in response to daily cyclic load variations. They are most efficient when run continuously, supplying base load. In locations where these machines must also cater for the peaks in demand, sufficient spinning reserve2 must be maintained in order to meet the next peak load event. In between, these machines will be only partially loaded and will operate at slightly lower efficiencies.
Renewable Energy
For much of the twentieth century hydro‐powered generators represented the main source of renewable energy. The power available from these machines depends on the topography of the generation site, specifically the volume of water and the head available. Hydro generation schemes were among the first to be built and have complemented thermal generators for many years.
In recent years natural gas driven turbines operating on a similar principle to aircraft jet engines have also become popular. Both gas and hydro turbines offer the advantage of rapid response to changes in system load. They can be quickly shut down or partially unloaded during times of low demand, and therefore they often find application as peaking generators, operating largely to meet the peaks in the daily load cycle. They can also supply base load when required to do so, or undertake frequency regulation, by accepting or rejecting load in order to maintain the system frequency within its prescribed limits. Figure 1.27 shows an operational comparison between a machine supplying base load and a hydro machine used for frequency control.
Figure 1.27 Operational comparison between a base load machine and a frequency control machine.
In recent decades advances in technology have enabled various sources of renewable energy including solar, wind, wave and tidal energy to also become part of generation mix. While these technologies all offer the ability to offset the CO2 emissions characteristic of thermal generators, they do so at the expense of relatively low energy dense ‘fuels’. Wind turbines, for example, must present a large swept blade area to the wind in order to extract a moderate amount of energy. As a consequence, physically large turbines have relatively small ratings, typically between 2 and 4 MW, and many turbines are required to produce the same quantity of electrical energy available from a single thermal machine. Offshore wind farms are not limited by the same construction, environmental and noise constraints as land‐based installations and are therefore able to use considerably larger turbines to harvest energy from more reliable winds. The largest currently installed offshore machines have capacities of the order of 7 MW.
Similarly, the solar energy incident at the surface of the earth on a clear day is about 1 kW/m2, and given that the conversion efficiency presently available from commercial photovoltaic cells is only about 18%, the area required to collect 1 kW is a large 5.6 m2. While this limitation is acceptable in the case of small roof‐top installations in residential applications it represents one reason why large‐scale solar farms are relatively few. Despite this, there are several very large solar farms throughout Europe and many in the USA, with nameplate ratings as high as 600 MW on sites as large as 2500 hectares.
A significant improvement in the solar collector area required can be had by using a large number of mirrors (known as heliostats) to track and focus sunlight onto a receiver which uses solar energy to heat molten salts to produce steam for the generation of electricity. This technology has the advantage that sufficient thermal energy can be stored in molten salt to permit electricity to be generated both day and night. Concentrated solar power systems as they are known are becoming popular in locations where land is relatively inexpensive and the daily solar radiation is high.
Tidal generation systems use tidal flows to power turbines in a similar way to wind turbines. However, because the density of water is about 850 times greater than that of air, considerably more energy is recoverable from a tidal generator than can be had from a wind turbine of the same size, permitting useful quantities of energy to be harvested from relatively low flow velocities. The performance of tidal generators can be further improved by fitting a duct or shroud around the machine. This increases the water velocity through the turbine and since the power output is proportional to the cube of velocity, a shrouded design can provide a useful return from small tidal or river flows. While the harvesting of tidal energy is not a new idea, recent advances in technology have made the construction of large tidal machines possible, many more of which are currently planned.
Renewable sources such as wind, solar and tidal all share the characteristic of being variable. The wind speed is rarely constant for more than a few hours, and since the energy recovered from a wind turbine is also proportional to the cube of the wind speed, small speed variations can result in large changes in machine output, as shown in Figure 1.28.
Figure 1.28 Daily variations in a wind farm generation.
Solar intensity varies with the passage of clouds and with the diurnal movement of the sun. Figure 1.29 contrasts the power generated by a large PV array on a clear day with its performance on a cloudy one. Both these graphs highlight the uncontrollable nature of renewable energy sources and the fact that they cannot be relied upon during times of peak demand.
Figure 1.29 Typical photovoltaic solar installation performance.
1.7.1 Synchronous and Asynchronous Generation
The problem of global warming, and in particular the greenhouse effects of CO2, mean that the world’s reliance on traditional fossil fuel generation must be reduced. While in some developing countries coal‐fired generation is still being installed, in others older, less efficient and more polluting machines are progressively being withdrawn from service. This, together with the simultaneous increase in renewable energy systems, is creating a change in the generation mix, from one characterised by a few large thermal and hydro machines, towards one in which many smaller renewable generators are playing an increasingly significant part. The former may be generally classified as synchronous generators, while the latter are broadly asynchronous. The electrical differences between the two are significant, and the networks to which they are connected must adapt to the change in generation if they are to maintain their current levels of security and reliability, particularly under fault conditions.
Both fossil‐fuelled and hydro turbines are capable of delivering reliable and controllable power. These are connected to the local network through large synchronous generators which rotate at speeds synchronised to the network in both frequency and phase. These machines frequently produce hundreds of megawatts and collectively possess characteristics that are particularly suited to ensuring stable and reliable electricity networks. For example, their large rotors provide considerable inertia, and since their governors are sensitive to tiny variations in shaft speed, they can effectively respond to sudden load variations while maintaining a near constant system frequency.
Thermal and hydro turbines also have the capability of providing ample reserve torque (spinning reserve), and the synchronous generators they drive can provide the necessary active and reactive power for peak load events, network contingencies and voltage regulation purposes. The power quality provided by synchronous machines is also high. They generate sinusoidal waveforms of high purity and can provide significant voltage support during network faults.
Asynchronous Generation
Asynchronous generation is a common characteristic of renewable energy sources. The variable nature of these sources means that it may not be possible for renewable rotating machines to achieve synchronous speed while delivering useful power to the grid. For this reason, wind‐powered turbines generally use induction generators that operate over a range of speeds dependent on the wind energy available; they operate asynchronously. For example doubly fed induction generators (DFIG) employed on many wind turbines use a back‐to‐back voltage converter fed from the grid to supply the rotor windings with the necessary low frequency excitation. This allows the machine to appear to have synchronous characteristics and to deliver both active and reactive power to the network under normal conditions. Other wind‐powered machines also use induction generators rotating at asynchronous speeds and deliver energy to the grid via back‐to‐back AC‐DC and DC‐AC converters. Both these topologies are limited by the current capacity of their electronics, and each has the propensity to suffer considerable electronic damage in the event of a nearby fault. Their inbuilt protection systems have therefore been primarily designed to preserve the integrity of the machine rather than to provide fault support for the network.
Network Fault Level and Fault‐Ride‐Through Ability
When a fault occurs, the local network voltage becomes depressed and, depending on the proximity and impedance of the fault, it may even fall to zero. A strong network will inject a large current into a fault, in an effort to support the network voltage. Further from the fault the voltage depression is less pronounced, supported by the fault current flowing towards the fault from nearby sources.
The strength of a network determines the extent of the resulting voltage depression as well as the degree to which its effects propagate throughout the remainder of the system. It is generally measured by the magnitude of the three‐phase fault current at a particular location, known as the fault level and is usually expressed in amps (or kA), or sometimes in volt‐amps (or MVA).
Fault levels vary considerably throughout a network and depend on both the network voltage and the network impedance at a given location. They tend to be highest near large synchronous generators where the network impedance is very low, and generally become lower as the distance from the generation sources increases. Synchronous machines are capable of generating very large fault currents, many times their continuous current ratings, whereas asynchronous generators and inverter connected machines generally contribute far less fault current, often equal to or less than their nameplate rating.
The ability of a generator to assist in supporting the network during a fault depends upon its ability to ride through the fault. Generators without fault‐ride‐through (FRT) ability are likely to trip in response to significant voltage depressions and a strong demand for reactive current. A significant loss of generation capacity during fault conditions can lead to a frequency collapse, and if load shedding is either inadequate or too slow, a system blackout may result. Therefore FRT is an important requirement in maintaining a network’s transient stability.
Synchronous machines have the ability to ride through faults, and their high fault current contribution and the large quantities of reactive power they provide can help to support the system voltage during faults. Further, they can rapidly re‐establish normal operation once a fault is cleared.
Asynchronous machines powered by renewable sources, on the other hand, have traditionally had limited FRT capacity, disconnecting themselves from the network during moderate voltage depressions. This upsets the post‐fault load–generation balance, and removes a useful current contribution that would have helped support the network voltage during the fault.
1.7.2 Effects of Renewable Energy Sources on Network Performance
When there were relatively few asynchronous sources in the power network, the loss of some under fault conditions was of little consequence. However, the increasing proportion of asynchronous sources at a time when synchronous generators are being retired has increased the requirements for FRT and fault current capacity from those machines remaining. This may even necessitate the provision of synchronous spinning reserve solely on this account.
Limited FRT ability is but one of the drawbacks associated with asynchronous machines. The variable nature of renewable energy and the low rotational inertia of asynchronous machines make them generally unsuitable as peaking generators. Further, as their penetration increases, the resulting reduction in system inertia makes the frequency control task more difficult. In addition, large converter installations used in wind or solar farms frequently inject considerable harmonic current into the local MV or HV bus, thereby generating a degree of harmonic voltage distortion, necessitating the installation of harmonic filters near the injection point.
In 2016, the European Commission published a code for the grid connection for both synchronous and asynchronous generators. This document prescribes the characteristics required of new generation equipment installed throughout the European Union. Chief among these is the requirement for all generation sources to possess FRT ability under a prescribed voltage profile, both during and post‐fault, and to provide reactive power during a fault, up to a specified fraction of the machine’s rated capacity.
In addition, the code requires that generators shall remain connected over a defined range of network frequencies and shall provide a degree of frequency support. This may require a reduction in power delivery during over‐frequency events, an increase during under‐frequency events, or the capacity for both, depending on the size of the machine. It also requires a voltage support characteristic for machines connected to the HV or EHV networks, in the form of a defined reactive power exchange with the network commensurate with the local network voltage. This means that the machine must consume reactive power from the network when the voltage exceeds the machine set point, and export it when the voltage falls below the set point.
Residential Photovoltaic Installations
Low‐voltage residential photovoltaic (PV) solar arrays with capacities up to 5 kW have become very popular in countries such as Australia, where government incentives and high feed‐in tariffs initially encouraged a rapid uptake of this technology. These arrays are interfaced to the LV network using small DC‐AC inverters. They require the network voltage as a synchronising signal, and shut down when this falls below a defined threshold, so they cannot support the network during fault conditions or periods of depressed voltage. The inverters also inject harmonic rich currents into the LV network which contribute to LV harmonic distortion. Distributed LV generation tends to be visible only in so far as it offsets the consumer demand; it is transparent to network operators and cannot be controlled.
These PV systems were often initially installed with little consideration of the problems that unplanned distributed generation might create. In suburbs where the uptake was dramatic, the high density of solar generation created a significant voltage regulation problem. This has been caused by the reverse flow of power from the LV to the MV network, via local distribution transformers. The LV network is designed for a power flow in the opposite direction, and the fixed tap setting on distribution transformers allows for a potential drop within the transformer as well as in the LV network and the customer’s premises. In the middle of the day when residential demand is low and solar radiation is high, the net power flow may be towards the MV network, causing these voltage drops to reverse. The LV network potential therefore rises and it may exceed the maximum level permitted. In many cases, this rise is sufficient to cause inverters to shut down, reducing the solar energy collected at a time when it is most abundant. Since most distribution transformers have off‐load tap‐changers, there is no easy way to rectify the situation. Further, the voltage experienced by customer equipment may vary considerably because the solar radiation can change dynamically throughout the day.
Regulators have begun to react to these issues, and new standards for low voltage inverters now demand a degree of both voltage and frequency support. For example, in Australia the 2015 edition of AS 4777, ‘Grid connection of energy systems via inverters Part 2: inverter requirements’, now requires both voltage and frequency support modes of inverter operation, in addition to a total harmonic current distortion of less than 5%. The voltage support mode is designed to permit an increase in the penetration of solar generation throughout the LV network, without creating the voltage regulation issue described above. This requires that the ratio of active to reactive power be modulated by the local voltage. Inverters must be capable of exporting up to 60% of nameplate rating as reactive power when the voltage is low and import up to this fraction when it is high.
1.7.3 Transmission Performance Improvement Through FACTS Devices
System strength has generally not been improved by the introduction of multiple asynchronous generators and inverter connected sources, but in recent decades various power electronic devices aimed at improving transmission network performance have become available. Collectively these fall into a category known as flexible AC transmission systems (FACTS). They can be used to increase or control the power flow in transmission lines and assist in the integration of renewable sources into a network by improving transient stability and providing voltage support during faults. FACTS devices can be broadly classified according to their method of connection as shunt compensators, series compensators or a combination of the two.
Shunt Compensators
SVCs – Static VAr compensators are devices that consist of switched capacitors in parallel with thyristor controlled reactors. Unlike a simple shunt capacitor, they can either consume or deliver reactive power to a bus in response to rapid changes in its voltage, and are therefore useful in improving network transient stability. Although the principle of operation is quite different, their characteristics are not dissimilar to traditional synchronous capacitors, but because they have no moving parts, they cannot provide inertial frequency support.
STATCOMS – Static compensators are based on a voltage source converter and are also able to exchange reactive power with a network in response to changes in network voltage. They are often used to maintain a desired power factor or regulate a bus voltage. When powered from a separate DC source, a STATCOM can also dynamically exchange active power with the bus, thereby damping power oscillations and improving transient stability.
Series Compensators
TCSC – The thyristor controlled series capacitor is connected in series with a transmission line and consists of a capacitor in parallel with a thyristor controlled reactor. By varying the thyristor firing angle the impedance of the combination can be progressively varied between that of an inductor and a capacitor. They are useful for controlling the power flow in transmission lines, damping power oscillations, limiting short circuit currents and providing voltage support.
Shunt‐Series Compensation
UPFC – The unified power flow controller is the most complex and versatile member of the FACTS family. It is applied to a transmission line, and consists of shunt and series connected DC‐AC converters, both supplied from a common capacitively supported DC bus. The series connected converter injects a voltage into the line that may be varied in both magnitude and phase, while the shunt connected converter absorbs real power from the line and delivers it to the DC bus supplying the series converter. Reactive power can flow bidirectionally from both converters, each independent of the other.
Depending on the phase of the injected potential, the UPFC can be used to regulate a line’s receiving end voltage, to insert series (capacitive) compensation into the line, to adjust the phase shift across the line, or a combination of all three. As a result, it can independently control the active and reactive power flows as well as damping power oscillations and improving the transient stability.
1.7.4 Future Transmission Network Planning and Augmentation
Traditionally transmission networks have been designed around the need to transport large quantities of energy from fossil‐fuelled power stations to load centres, or from one transmission substation to another. However, the location and capacity of new energy sources will be different from those presently in service; this will require transmission companies to augment their networks accordingly. In addition, it is likely that the proliferation of more residential PV systems, the introduction of storage batteries and smart metering, together with an improved understanding of time‐based electricity pricing, will see changes in future consumer demand profiles. Collectively these changes may mean that future network planning will no longer be largely driven by a need to provide sufficient infrastructure to meet peak demand; instead it may be targeted at accommodating new sources of generation into the network together with the voltage and frequency services necessary to support it.
Two things are certain: considerable expenditure will be required to move the world towards more sustainable generation, and the transmission and distribution networks of the future will look quite different to those of today.
1.8 Sources
- 1 IEC 60038 ‘Standard Voltages’ Edition 7, 2009–06, International Electrotechnical Commission, Geneva.
- 2 Schneider Electric, Cahier Technique No. 203. ‘Basic selection of MV public distribution networks’. 2001.
- 3 Schneider Electric/Square D ‘Corner‐grounded delta (Grounded B Phase) systems’. 2012.
- 4 Zhang X, Cao X, Wang W, Yun C, ‘Fault ride through study of wind turbines’, School of Electrical Engineering, Xinjiang University, Urumqi China, Journal of Power and Energy Engineering, 2013.
- 5 Australian Standard AS 4777.2, 2015, ‘Grid connection of energy systems via inverters Part 2: Inverter requirements’, Standards Australia, Sydney.
- 6 AEMO ‘National transmission network development plan 2016’, Australian Energy market Operator, Sydney.
Further Reading
- 1 Mohanty AK, Barik AK, ‘Power system stability improvement using FACTS devices’ International Journal of Modern Engineering Research, Vol. 1, Issue 2, pp. 666–672.
- 2 Gyugyi L, Schauder CD, Williams SL, Rietman TR, Torgerson DR, Edris A, ‘The unified power flow controller: A new approach to power transmission control’ IEEE Transactions on Power Delivery, Vol. 10, No. 2, April 1995.
- 3 Varma, Rajiv K ‘Elements of FACTS controllers’, IEEE Power Engineering Society.
- 4 Abdulrazzaq AA, ‘Improving the power system performance using FACTS devices’, Journal of Electrical and Electronics Engineering (IOSR‐JEEE), Volume 10, Issue 2, pp. 41–49, March‐April 2015.
- 5 Beck G, Breuer W, Povh D, Retzmann D, Telsch E, ‘Use of FACTS and HVDC for power system interconnection and grid enhancement’, Siemens Power Transmission and Distribution, Power–Gen Conference, Abu Dhabi, United Arab Emirates, 2006.