doi: 10.1007/978-3-030-35386-5 29abstract before the advent of facts controllers, breaker-switched...

Controllers Using the Saturation of Ironfor AC Network Control 10David J. Young

Contents1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3562 The Saturation Characteristic of Iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357

2.1 The Basic Static Var Compensator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3592.2 The Magnetic Constant Voltage Transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361

3 Harmonics in Saturated Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3623.1 Harmonics in a Single-Phase Self-Saturated Reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3623.2 Harmonics in Three-Phase Self-Saturated Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3633.3 Reduction of Harmonics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3663.4 Magnetic Frequency Multipliers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368

4 The Magnetic Amplifier or Transductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3694.1 100 MVATransductor for Alternator Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3714.2 Tertiary-Connected Transductor for Dynamic Var Balancing in a 132/275/400 kV

Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3744.3 Magnetically Controlled Shunt Reactors (MCSR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

5 Development of Effective Compensation for Arc Furnaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3765.1 Characteristic Features of Arc Furnaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3765.2 Experimental Arc Furnace Compensation by Transductor . . . . . . . . . . . . . . . . . . . . . . . . . . . 3765.3 Experimental Arc Furnace Compensation by Self-Saturated Reactor . . . . . . . . . . . . . . . 3785.4 Commercial Applications of Saturated Reactors for Arc Furnace Compensation . . . 3815.5 Compensation by Decoupling Transformer-Reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382

6 Three-Phase Self-Saturated Reactors with Harmonic Compensation . . . . . . . . . . . . . . . . . . . . . . 3836.1 The Twin Tripler Saturated Reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3836.2 The Treble Tripler Saturated Reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3856.3 Slope Correction for Saturated Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387

7 Applications of Self-Saturated Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3887.1 Disturbances Caused by Industrial Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3887.2 Compensation for Long Transmission Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3887.3 Commercial Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3907.4 Static Var Compensation for the 2000 MW HVDC Cross-Channel Link . . . . . . . . . . . 392

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398

D. J. Young (*)Stafford, UKe-mail: [email protected]

© Springer Nature Switzerland AG 2020B. R. Andersen, S. L. Nilsson (eds.), Flexible AC Transmission Systems, CIGRE GreenBooks, https://doi.org/10.1007/978-3-030-35386-5_29

355

http://crossmark.crossref.org/dialog/?doi=10.1007/978-3-030-35386-5_29&domain=pdf

mailto:[email protected]

https://doi.org/10.1007/978-3-030-35386-5_29#DOI

Abstract

Before the advent of FACTS controllers, breaker-switched capacitors and induc-tors were used to provide a stepwise balance of vars (reactive power) fortransmission and distribution systems; synchronous compensators were installedat some substations to provide a continuously variable reactive power output, butthese machines were expensive and needed regular maintenance.

Prior to the availability of power electronic devices, Dr. E S Friedlanderdeveloped the first static voltage stabilizers, based on the properties and charac-teristics of saturated iron. Similarly to synchronous compensators, these “staticvar compensators” (SVCs) had a continuously variable output, but they werecapable of a much faster response than synchronous compensators and had otheradvantages. For over two decades, they found widespread use in transmission anddistribution systems. This chapter describes saturated reactor-based var control-lers and provides application examples.

1 Introduction

The control of reactive current is vital to the satisfactory and efficient operation ofelectricity supply systems. For many years it was necessary to switch capacitors andinductors into and out of operation to provide a balance of reactive power (vars) forrelatively steady system conditions. Whereas frequent or cyclic switching of shuntcapacitors is not uncommon in distribution and industrial applications, it is generallyundesirable to use frequent switching in transmission networks. In situations wherevariability of reactive compensation was important, synchronous compensators hadoffered a solution and were being used more widely as transmission networksdeveloped. Such machines were expensive, required substantial civil works andauxiliaries, and needed regular maintenance and refurbishment.

In the 1960s a breakthrough in respect of the dynamic reactive support of ACnetworks by means of static devices instead of rotating machines was spearheadedby Dr. Erich Friedlander (Friedlander 1966). These early controllers were based onthe properties and characteristics of saturated iron and were capable of a much fasterresponse than synchronous compensators. They made it possible to overcome manyof the limitations experienced when using conventional reactive power controlmethods. Their application was facilitated by the availability of much improved,grain-oriented, transformer steels and by developments in capacitor designs, whichhad led to both lower losses and lower price/kvar.

For Friedlander the nonlinear saturation characteristic of iron presented an oppor-tunity instead of a limitation. The study of nonlinearities had been at the heart of hisdoctoral thesis on “relaxation oscillations” (Friedlander 1926). Friedlander regularlymade use of models to check the correlation between theory and practice and toidentify what might have been overlooked either in the theory or the application. Heworked for several years under Dr. Reinhold Rudenberg at Siemens AG in Berlin,

356 D. J. Young

but later joined The General Electric Company of England (GEC), initially at thecompany’s Research Laboratories at Wembley and subsequently as the Consultant atits group of electrical engineering factories in Witton, Birmingham. At GEC, he wasable to develop controllable saturated reactors (transductors) and self-saturatedreactors for application in power systems. These could be called the first FACTScontrollers (even though they preceded EPRI’s acronym by about 40 years). Newapplications of saturated reactor SVCs for transmission networks declined whenSVCs based on the use of power thyristors were introduced and were demonstratedto be economic and reliable.

This chapter describes how various reactive power controllers were developedunder Friedlander’s supervision and provides examples of the application of theseSVCs in transmission and distribution networks.

2 The Saturation Characteristic of Iron

The economic application of electrical power has only become possible because ironhas a remarkable magnetic property; its relative permeability, μ, is several thousandtimes that of a vacuum, air, water, or any other commonly occurring material. A highlevel of flux density, B, can be obtained in iron and steel with only a very smallmagnetizing force, H, as illustrated in the well-known B-H characteristic Fig. 1a. Asthe flux density is increased beyond a certain point, the characteristic becomesnonlinear because the relative permeability starts to reduce very quickly. Iron isdescribed as being saturated when the value of μ decreases to 1, the same value as air.When iron is used in alternating current applications, in which the flux reversesdirection in each half cycle, there is a hysteresis effect, and the change of flux densityfollows a different path when it reverses direction Fig. 1b. The area enclosed by theB-H curve represents the small amount of energy dissipated in the iron due to theflux reversal, the hysteresis loss. When iron is used in electrical equipment such as intransformers, the flux density in normal operation is kept comfortably below thetypical core saturation level of about 2 tesla (T).

The B-H curve needs to be drawn on a different scale, Fig. 1c, to show thecharacteristic when iron is driven deeply into saturation; the slope of the B-H curve inthe saturated region becomes the same as air with μ equal to 1. On this scale, the “knee”of the characteristic shows a sharp change at the transition between the unsaturated andthe saturated states.

When a sinusoidal voltage is applied to a winding around a closed iron core, theflux induced in the iron will also vary sinusoidally. When the applied voltage is low,the flux density is also low, and the magnetizing current is very small. When thevoltage becomes sufficiently high, the crest of the flux wave will exceed thesaturation flux, and this causes a large pulse of magnetizing current to flow in thewinding, Fig. 2. This current has the shape of a truncated sine wave and containsfundamental and odd harmonic components.

10 Controllers Using the Saturation of Iron for AC Network Control 357

Fig. 3 illustrates the most important fundamental, third, fifth, and seventh har-monic frequencies in a typical current wave. As the value of the applied voltageincreases, the duration of the current wave is more prolonged, and the proportions ofthe harmonic components decrease relative to the fundamental. An iron core whichis driven into saturation only by an alternating voltage applied to the winding isdescribed as a self-saturated reactor (usually abbreviated to “saturated reactor”).

Fig. 4 shows that the relationship between the fundamental frequency current andthe voltage applied to the winding is similar to the saturation curve, though it has asofter knee and a continuous slight curvature above the knee point. Nevertheless, thefundamental characteristic approximates closely to a straight line between about10% and 100% of the maximum magnetization current; the point at which this lineintersects the axis gives the “saturation voltage”Vs and the slope of the characteristicis the air reactance or “saturated inductive reactance” Xs.

Fig. 1 Magnetic characteristics of iron

current

H (I)

B

crest fluxsaturationflux

Fig. 2 Sinusoidal flux wave and current pulse

358 D. J. Young

2.1 The Basic Static Var Compensator

The equivalent circuit of a self-saturated reactor is represented by the saturationvoltage in series with the saturated reactance as indicated in Fig. 5a. This hassimilarities to the characteristic and equivalent circuit of a synchronous compensatorwith fixed excitation voltage Fig. 5b. For this reason, a saturated reactor provides aninertia-less static equivalent to a synchronous compensator and was the first kind of“static var compensator” (CIGRE TB 25 1986).

B

fundamental

linearizedfundamental

H (I)

B (V)

Fig. 4 B-H characteristic andlinearized fundamentalfrequency voltage-currentcharacteristic

Fundamental

3rd harmonic

5th harmonic

7th harmonic

Current pulse

Fig. 3 Harmonics in a current pulse


By itself, a saturated reactor will only absorb vars whenever the voltage at itsterminals is greater than its saturation voltage, whereas a synchronous compensatorcan also generate vars when its terminal voltage is below its excitation voltage. Theability to generate vars with a static compensator can be obtained by adding a shuntcapacitor, Fig. 5c.

In this arrangement, the equivalent saturation voltage isVs’ = Vs.Xc/(Xc – Xs) and the equivalent slope reactance isXs’ = Xc.Xs/(Xc – Xs)The switching duty on a circuit breaker is very easy when switching a static

compensator which comprises a shunt capacitor in parallel with a saturated reactor.This is because the trapped charge effect normally associated with capacitorswitching is eliminated. After separation of the breaker contacts and interruptionof the arc at current zero, the residual charge on the capacitor causes the reactor tosaturate again. The capacitor energy is then dissipated in a cyclic manner atsteadily decreasing frequency. Consequently the rate of voltage rise across thecontacts of the breaker as they separate is so low that re-striking of the arc cannotoccur.

2.1.1 Speed of ResponseFig. 6a shows a simple circuit with a self-saturated reactor connected in parallel withan inductive load. While the inductive load is in operation, the voltage on the loadbusbar, Vl, is lower than the open-circuit voltage, Vo, due to the voltage drop in thesupply reactance, Xo. If the saturation voltage, Vs, of the saturated reactor is veryslightly lower than Vl, the saturated reactor core will be unsaturated, and theinductive magnetizing current will be very small. At the instant when the load is

Fig. 5 Saturated reactor equivalent circuit and characteristic

360 D. J. Young

switched off, Vl instantaneously becomes equal to Vo. The initial increase in Vl

causes an increase of the rate of change of the magnetic flux in the saturated reactorcore. Because the flux now increases more rapidly, it reaches and substantiallyexceeds the saturation flux, Bs, and the saturated reactor draws an increased currentduring the same half cycle as the load current and voltage changes occur. There iseffectively no delay in the response of a self-saturated reactor to changes of loadcurrent and applied voltage. This is illustrated in Fig. 6b. As can also be seen, theload voltage becomes distorted.

2.2 The Magnetic Constant Voltage Transformer

The voltage on a typical supply system can vary over a considerable range, often asmuch as �5% or even �10% of its nominal value. A static compensator is able tocontribute to a reduction of the voltage variations at its point of connection, whetherthese variations are caused by changes of load or by changes of the supply voltage.Some sensitive loads do not operate correctly if supplied from a supply system inwhich the voltage can vary over a wide range; the magnetic constant voltagetransformer was developed to protect such sensitive loads from large variations ofsupply voltage (Friedlander 1935). A tapped reactor is connected in series with thesensitive load, with a saturated reactor connected to the tapping point Fig. 7a. In apractical application, the tapped reactor can consist of a linear reactor, (1 + n)2X0,with an autotransformer connected in parallel as shown in Fig. 7b. By matching the

Fig. 6 Undelayed response of a self-saturated reactor


slope reactance of the saturated reactor to a combination of the linear reactance andthe tapping ratio n: 1 of the reactor or transformer, the effects of the supply voltagevariations can be neutralized.

Fig. 7c shows the equivalent three-terminal star-impedance representation of thetapped reactor; the impedance of the branch to the tapping point is a negativereactance, �nX0. When the slope reactance, Xs, has an equal numerical value, thereactances will cancel, and the voltage at the star point of the equivalent circuit willbe the same as the saturation voltage, Vs, which is constant. This constant equivalentvoltage Vs will be applied to the load through an equivalent reactance, n(1 + n)X0 asin Fig. 7d. Variations of the system voltage are absorbed by variations of thesaturated reactor current and do not disturb the effectively constant voltage supply-ing the load.

3 Harmonics in Saturated Reactors

3.1 Harmonics in a Single-Phase Self-Saturated Reactor

When a sinusoidal voltage is applied to a self-saturated reactor, the waveform of themagnetizing current is not sinusoidal but takes the form of a current pulse (thewaveshape is the top part of a cosine wave) as illustrated above, in Figs. 2 and 3. Theharmonics in this waveshape can be studied in detail, for example, by FourierAnalysis. As described earlier, the wave is made up of a fundamental frequencycomponent together with smaller amplitudes of all odd harmonics. Fig. 8 shows therelative amplitudes of the fundamental frequency and harmonics up to the 13th as thepeak magnitude of the current pulse increases. The third harmonic componentpredominates, but the amplitudes of the higher harmonics diminish quickly as theapplied voltage and flux levels increase.

Fig. 7 Magnetic constant voltage transformer

362 D. J. Young

3.2 Harmonics in Three-Phase Self-Saturated Reactors

The pattern of harmonics flowing in a three-phase saturated reactor (or a set of threesingle-phase reactors) depends on the type of connection, star or delta, earthed, orisolated neutral. With a simple set of 3 single-phase reactors, there will be 6 currentpulses per cycle of the fundamental frequency; with more advanced reactorsconsisting of two or 3 sets of 3 reactors, there may be 12 or 18 pulses per cycle.By analogy with rectifier terminology, the dominant residual harmonics in the linecurrents can be categorized as 6n � 1, where n represents the number of sets ofreactors and has the values 1, 2, 3, etc. See, for example, ▶Chap. 6, “TechnicalDescription of Static Var Compensators (SVC)” in this Green Book.

3.2.1 Earthed Star ConnectionWhen the windings on three single-phase reactors are connected in star and the starpoint is connected to the system neutral, as in Fig. 9a, the reactors saturate one at a time,and the line current contains the fundamental and all odd harmonic components. Whenthe other reactors are unsaturated, they have a very high impedance, and the saturationcurrent return path is from the star point into the system neutral. This pattern is repeatedas each phase conducts in sequence. Fig. 9b shows the phase and neutral currents; therepetition rate of the neutral current pulses is 3 times the system frequency and thereforecontains all the triplen harmonics (but not the fundamental and non-triplen harmonics).In the example shown, the neutral current, In, is clearly much larger than each of the linecurrents. The reactors could be combined on to a three-limb core, but the earthed starconfiguration is generally undesirable because of the very large third harmonic currents

1

Hn/H1

0.8

0.6

0.4

0.2

0

–0.2

0 0.1 0.2 0.3 0.4(Bcrest - Bsat)/Bsat

13th

11th

9th

7th

5th

3rd

Fig. 8 Amplitudes of harmonics relative to the fundamental, for increasing crest flux


https://doi.org/10.1007/978-3-030-35386-5_7

https://doi.org/10.1007/978-3-030-35386-5_7

flowing in the neutral connection to the supply system. Third harmonic overloading ofneutral conductors sometimes occurs in distribution systems.

3.2.2 Unearthed Star ConnectionWhen the windings on the single-phase reactors are connected in star but with afloating star point, as in Fig. 10a, the reactors are forced to go into saturation in pairs.When none of the reactors is saturated, the star point is aligned with the systemneutral, but each time a pair of reactors goes into saturation, the star point moves to themidpoint of the associated line-line voltage, and this develops a third harmonicvoltage (which is approximately a square wave) between the star point and the systemneutral. Consequently, it is the flux wave in each limb which contains a thirdharmonic component instead of the current wave. If the three reactors are combined

Ia Ib

Va

a

b

Vb Vc

Vn

N

0 90 180 270 360

Ibc

Ica

Ia

Ib

Ic

Vn

Iab

Vca

Vbc

Vab

Ic

Fig. 10 Unearthed star reactor currents and voltages

Ia Ib

Va

a

b

Vb Vc N

0 90 180 270 360In

Ic

Ib

Ia

Vc

Vb

Va

Ic In

Fig. 9 Earthed star reactor currents and voltages

364 D. J. Young

on to a multi-limb core, one or two unwound limbs are required to provide a returnpath for the third harmonic fluxes. The current in the line, Fig. 10b, has a double pulsewaveshape instead of the single pulse of Fig. 9b; it does not include any 3rd-orderharmonic components.

3.2.3 Unearthed Star Connection with Mesh WindingFig. 11a shows an unearthed star-connected arrangement of the same set of single-phase reactors, but with an extra “mesh” winding (also called a delta-connectedwinding) on each phase; these mesh windings are connected to form a closed loop. Inthis case, each reactor saturates individually, because the mesh winding allows eachof the unsaturated phases to conduct half of the saturated current by transformeraction. The mesh current also flows in the saturated phase; the polarity of the meshcurrent is the same as that in the main winding and thus reinforces the magnetizingforce on the saturated core. Because each reactor saturates individually, the phase-to-phase saturation voltage is reduced to √3/2 compared with the simple unearthed starof Fig. 10a in which reactors saturate in pairs. Fig. 11b shows the line and meshcurrents. The current in the mesh winding contains the triplen harmonics; the linecurrents (with a triple-pulse waveshape) consist of the fundamental and the other oddharmonics. The presence of the mesh winding eliminates the third harmonic voltageand flux between line and star-point terminals and allows a three-limb core to beused. Harmonic analysis shows that the phase angles of the 5th and 7th harmonicshave reversed polarity compared with those harmonics when there is no meshwinding as in Fig. 10 (or when the mesh winding of Fig. 11 is left open-circuited).

3.2.4 Delta ConnectionWhen the windings on the three single-phase reactors are connected in delta, as inFig. 12a, each reactor goes into saturation independently, with delta phase currents,I1, I2, and I3. The line currents are given by the differences of the phase currents,

0

In

90

ab

180 270 360

In

Ic

Ib

Ia

Vc

Vb

Va

Ia Ib

Va Vb Vc

Ic

Fig. 11 Unearthed star reactor, currents, and voltages with mesh winding


Ia = I1 – I2, Ib = I2 – I3, and Ic = I3 – I1 and thus comprise two equal pulses in eachhalf cycle, separated by 60� (π/3), Fig. 12b. The line current waveshapes are thesame as those in Fig. 10b. The triplen, zero sequence, 3rd, 9th, 15th, etc., harmonicsin the phase currents are all in-phase and circulate round the delta-connectedwindings; therefore the triplen harmonics do not appear in the line currents underbalanced conditions. The line currents contain the fundamental with the remainingodd harmonics, 5, 7, 11, 13, etc. The voltages and fluxes sum to zero round the delta-connected units, which can therefore be combined on to a three-limb core.

3.3 Reduction of Harmonics

3.3.1 Phase Displacement of FluxesThe waveshapes of the line currents in Figs. 11b and 12b are very different althoughthey both include the same proportions of harmonic currents to fundamental. Thereis a 30� (or 150�)-phase displacement of the fundamental frequency fluxes andcurrents in the delta-connected reactors compared with the star-connected units.This fundamental frequency phase displacement causes the fifth and seventh har-monic components to have a relative phase displacement of 180� with reference toeach other. Thus, when a star-connected reactor with a mesh winding, Fig. 11a, isconnected in parallel with a delta-connected reactor, Fig. 12a (designed to have thesame saturation voltage), the individual and combined currents for one phase areillustrated in Fig. 13. The star- and delta-connected reactors form a 12-pulse set, andthe fifth- and seventh-order harmonics circulate between the reactors and are there-fore absent from the line currents. The eleventh and thirteenth harmonics (12 � 1)are not cancelled and are the lowest-order harmonics in the line currents.

In practical applications, the combination of star- and delta-connected reactors isnot normally used for harmonic cancellation. For convenience of design and pro-duction, two identical reactors are manufactured, using zigzag (inter-star) intercon-nections to give 30� relative phase displacement, Fig. 14a. The interconnections ofthe phase-shifting windings are arranged so that the flux in one set of limbs is phase

0

ab

90 180 270 360 Ic

Ib

Ia

Ica

Ibc

Iab

Vca

Vbc

Vab

Ia Ib

Va Vb Vc

IcI1 I2 I3

Fig. 12 Delta-connected reactor currents and voltages

366 D. J. Young

displaced by 15� positive and in the other by 15� negative, resulting in the cancel-lation of the 5th and 7th harmonics, as shown in Fig. 14b. The fifth and seventhharmonics are also eliminated from the line currents if the reactors are connected inseries: the fifth and seventh harmonic distortions then appear (and are cancelled) inthe fluxes of the reactors.

It is important to note that complete harmonic compensation can only occur withbalanced system conditions. If the three phases of the system lose AC symmetry, thenegative sequence fundamental component will reintroduce small amounts of oddharmonics into the line currents; if there is a zero sequence component present (DCor third harmonic), even harmonic currents will be generated. The presence of evenharmonic distortion in the supply voltages will similarly cause a slight disruption ofthe harmonic balance in a saturated reactor.

0

a b

90 180 270 360sum

plus 15°

minus 15°

Fig. 14 Cancellation of 5th and 7th harmonics with �15� displacement of fluxes

Voltage

I delta

I star

star fundamental

I line

line fundamental360270180900

delta fundamental

Fig. 13 Cancellation of 5th and 7th harmonics between star- and delta-connected reactors


3.3.2 Mesh LoadingIf the mesh winding of an unearthed star-connected reactor, Fig. 11a, is suddenlyopen-circuited, the third harmonic current becomes zero, and the star point developsa third harmonic voltage; the current waveshapes change from Fig. 11b to Fig. 10b.The difference in waveshapes is because of the reversal of polarity of the dominantfifth and seventh harmonics with respect to the fundamental. If the transition frommesh short-circuit to open-circuit is made gradual, by introducing reactive imped-ance into the mesh loop, the amplitudes of the fifth and seventh harmonics steadilyreduce, passing through zero, with a line current waveshape similar to Fig. 13, beforethey increase again, with reversed polarity, to the open-circuit condition. However,for a particular value of mesh impedance, the fifth and seventh harmonic currents areonly minimized at one value of fundamental frequency current. When the funda-mental current changes, the mesh impedance must also be changed to maintain anoptimum waveshape. A small saturated reactor, operating at third harmonic fre-quency, inherently provides a self-adjusting impedance which closely approximatesto the optimum value throughout the current range.

3.4 Magnetic Frequency Multipliers

In the metal processing industries, induction furnaces provide a clean and efficientmethod of heating metal for refining, alloying, and/or raising temperature prepara-tory to teeming. Mains frequency furnaces are frequently used for this purpose butare inefficient for melting metal. Higher-frequency induction furnaces are oftenpreferred, supplied from motor generators or static frequency multipliers. Themagnetic frequency tripler utilizes a three-phase, unearthed star-connected saturatedreactor with a mesh winding (as in Fig. 11) connected to the terminals of aninduction coil surrounding a crucible containing the metallic material. Shunt har-monic filters are used both to improve the input power factor and also to improve thethree-phase current waveshape by absorbing the predominant 5th and 7th harmoniccurrents from the saturated reactors. The third harmonic output voltage and power isadjusted by means of contactor switching of shunt capacitors in parallel with theinduction coil, Fig. 15. The effect of the load impedance during the different stages

Harmonic Filters/pdf

Inductioncoil

Contactor-switchedcapacitors

Fig. 15 Frequency tripler arrangement

368 D. J. Young

of the heating and melting process is to reduce the harmonics in the supply current,as indicated in the previous section.

The general principles for magnetic frequency multipliers at higher frequencieswere described by Friedlander. For example, a quintupler uses an iron core with fiveactive limbs and two flux return limbs. Four of the active limbs use zigzag windingsto provide saturation flux displacements at 36� intervals, Fig. 16 (Friedlander andYoung 1966). The mesh winding has an output voltage at the fifth harmonicfrequency, and the voltage is adjustable in the same way as for a frequency tripler.The septupler, for seventh harmonic generation, takes a corresponding format withseven active limbs (Friedlander 1958).

4 The Magnetic Amplifier or Transductor

For many years magnetic amplifiers provided a very convenient and effective way ofcontrolling a high level of either AC or DC output power flow, using only alow-power control input. Fig. 17a represents a closed iron core with an AC powerwinding and a DC control winding. A sinusoidal voltage which does not causesaturation of the iron core is applied to the AC winding. When there is no current inthe DC winding, the core remains unsaturated so that the impedance of the ACwinding is very high and only a very small magnetizing current will flow in it. Whenthe DC winding is energized with direct current, this provides a magnetizing forcewhich drives the core into saturation; the alternating flux wave becomes offset, sothat part of the flux wave projects into the saturated part of the B-H characteristic,Fig. 17b. While the core is saturated, the AC winding impedance is very low (equalto an air-cored winding) and a pulse of current flows in the winding. At the end of theconduction period, the alternating flux wave drives the core out of saturation, and

R

R

Y

Y

III

III

IIL3 a

IV

b

IVII

V

V

B

I

I

B

Fig. 16 Frequency quintupler


only a low value of magnetizing current can flow in the AC winding for the rest ofthe cycle. During conduction, the mean value of the ampere-turns in the AC windingbalances the DC ampere-turns in the control winding. When the DC control currentis increased, the time during which the flux wave projects into saturation is pro-longed (to maintain ampere-turn balance), which allows the current in the ACwinding to increase; this current includes both direct and alternating components.For practical applications, the direct component is eliminated from the line currentby using a pair of identical iron cores, Fig. 17c, with opposite relative polarity of theDC windings so that one core permits current during the positive half cycle, and theother permits an equal current during the negative half cycle. The current in the AC

Fig. 17 Operating principle of a transductor

370 D. J. Young

windings is approximately proportional to the DC control current for a wide range ofapplied voltages, as indicated in Fig. 17d.

In order to distinguish them from self-saturated reactors, transductors are some-times called “saturable reactors.” In comparison with a self-saturated reactor, atransductor requires twice as much active magnetic material, a greater number ofwindings, auxiliary equipment for control of its output, and it has a relatively slowresponse. Consequently, after a few early applications of transductors, as describedbelow, Friedlander concentrated on the development of self-saturated reactors as thekey elements of static compensators for power systems.

4.1 100 MVA Transductor for Alternator Testing

Factory tests on very large machines such as turbine alternators cannot normally bedone at the actual output load for which the machines are designed and rated becausethe required input power is not available and the output could not be containedwithin the factory. Very occasionally a suitable machine might be available for aback-to-back test at a high load level. Large machines are normally subjected to twomain electrical tests in the factory, short-circuit and open-circuit tests. The impact ofthe complex heat flows and stray losses occurring under rated service conditions aredifficult to determine from these factory tests. However, testing at full kVA load, butat almost zero power factor, can provide a great deal of information on stray lossesand heat flows.

In the 1950s, Friedlander designed two large transductors to enable generators tobe tested in this way (Easton et al. 1958). These provided an easy and stepless meansof controlling the reactive load current in the machine under test at the voltage set bythe machine excitation. The nominal rating of each transductor was 100 Mvarthroughout the range of voltages between 6.6 and 22 kV.

In order to incorporate harmonic compensation, two six-limb transformer typecores were needed. As shown in Fig. 18, the winding arrangement used zigzagconnections giving �15� phase displacement of fluxes to eliminate 5th and 7thharmonic currents. Delta-connected tertiary windings allow triplen harmonic cur-rents to circulate and eliminate the need for any unwound flux return limbs. Theresidual 11th and 13th harmonic currents have maximum amplitudes of about 1.5%(causing a similar magnitude of voltage at the machine and transductor terminals)but in most operating conditions are much lower. These residual harmonics had aninsignificant effect on the performance of the machine under test.

The control current for the transductor was drawn from a grid-controlled 500 kWrectifier. Each DC control winding embraces three limbs of the core, so that theinduced voltage at fundamental frequency is zero. These control windings wereconnected in series. Fig. 19 shows the wound core of one of the two units of thetransductor before being mounted in a conventional oil-filled tank.


The primary windings could be connected either in star or in delta to cover a widerange of operating conditions, shown as hatched areas in Fig. 20. The voltage/currentcharacteristics with the units in star connection are shown in Fig. 21.

The two identical 100 Mvar transductors were manufactured and entered servicein 1953 and, as alternator ratings increased, a third identical unit was added a fewyears later. The three transductors had an overload capability which enabled them tobe operated with a combined controlled output of 360 Mvar and, using theuncontrolled natural saturation characteristic, at up to 460 Mvar.

Fig. 19 Wound core of 100 MVA transductor

Fig. 18 Transductor in parallel inter-star connection

372 D. J. Young

Fig. 21 Voltage/current characteristics in star connection

25ZERO d.c EXCITATION

ZERO d.c EXCITATION

kv

20

STAR CONNECTION

MESH CONNECTION

MAX. d.e. E

XCITATION

MAX. d.e. EXCITATION

15

10

5

0 20 40 60 80 100 120 MVA

Fig. 20 Operating ranges of100 MVA transductor


4.2 Tertiary-Connected Transductor for Dynamic Var Balancingin a 132/275/400 kV Network

As its grid network developed in the United Kingdom, the Central ElectricityGenerating Board (CEGB) had embraced the use of shunt reactors and shuntcapacitors to control the network voltage profile by balancing steady-state vardemands at its substations. For dynamic compensation of sudden system changesfollowing faults or equipment outages, tertiary-connected synchronous compensatorswere installed at important substations. The usual dynamic range for these compen-sators was 90 Mvar (30 Mvar inductive absorption to 60 Mvar capacitive generationat a nominal 13 kV). They were usually installed in pairs so that one should always beavailable when the other was taken out of service from time to time for routinemaintenance or factory refurbishment.

Recognizing that the controllability of a transductor was similar to that of asynchronous machine and that regular factory refurbishment would not be needed,CEGB decided to install a +60/�30 Mvar static compensator at one of its 275/132 kVsubstations, at Exeter, to provide a direct comparison of the behavior and performanceof the static and the rotating compensators. In steady state operation, the synchronouscompensators were normally brought back to a “float” condition at 0Mvar, so that theirdynamic range, with a slope of 5% on 60 Mvar, was available for sudden changes. Avoltage range of �10% to +15% was allowed on the tertiary winding.

The transductor design used �15� flux phase displacement for harmonic com-pensation similar to the earlier 100 Mvar transductors, but the control windings werewound round individual limbs to help achieve the required response time.

It was agreed that the 60 Mvar capacitive range of the static compensatorshould be obtained by means of three switched capacitor banks each of 20 Mvarso that a smaller transductor could be used and could operate with minimum losses atthe 0 Mvar float condition. Air-blast circuit breakers were used for the capacitorswitching. The transductor was designed to have a dynamic range of 34.5 Mvar (at13 kV), sufficient to give an overlap at each capacitor switching point so that the totalMvar output range could be covered smoothly without any discontinuities.

With forced excitation, the synchronous compensators had a response timeof about 5 cycles (at 50 Hz). The control system of the transductor providedthe same response time for sudden disturbances. Capacitor switching was insti-gated when the transductor reached the end of its dynamic range. It was acceptedthat, even though the full-range dynamic response included the short delaysrequired for switchgear operation, the static compensator would provide a satis-factory Mvar contribution toward severe network disturbances. Fig. 22 shows thecontrollable range of the voltage/current characteristic and a simple single-linediagram of the installation at the CEGB substation. The static compensator wascommissioned in 1967. When the substation was upgraded to 400/132 kV opera-tion, both the compensators were transferred to the 13 kV tertiaries on the newtransformers.

The static compensator remained in service for some 30 years without signif-icant downtime until the substation underwent a further major upgrade and new

374 D. J. Young

thyristor controlled SVCs were installed. The successful and reliable operation ofthis original static var compensator was a factor in support of the later decision byNational Grid (the successor to CEGB) to install thyristor-based SVCs fordynamic compensation of the 275 kV and 400 kV networks at many substationsin England and Wales.

4.3 Magnetically Controlled Shunt Reactors (MCSR)

Magnetically controlled shunt reactors (MCSRs) are being used to provide variablereactive absorption for power grids and industrial plant in Russia, Kazakhstan, andsome other countries. An MCSR is a transductor which has its primary windingconnected directly to the high voltage system. Although a transductor requires twomagnetic cores, one for each direction of magnetic saturation, this duplication ofactive core material is offset by the direct connection to the HV system, which eliminatesthe need for a stepdown transformer; consequently, the cost of anMCSR is relatively low.

Typically, MCSRs are being used to compensate the net charging current of high-voltage transmission lines, to give better control of the voltage along the line and toincrease their total power transfer capability. MCSRs are also used at substationswhich are distribution nodes of the power grid in order to contain voltage variationswithin a narrow range close to the nominal voltage and also to reduce the Mvarcontrol requirements of nearby synchronous generators.

Details of the MCSR design and its application are given in ▶Chap. 11, “Devel-opment of Magnetically Controlled Shunt Reactors in Russia” in this Green Book.

Fig. 22 Operating range and diagram of +60/�30 Mvar transductor-type SVC


https://doi.org/10.1007/978-3-030-35386-5_28

https://doi.org/10.1007/978-3-030-35386-5_28

5 Development of Effective Compensation for Arc Furnaces

5.1 Characteristic Features of Arc Furnaces

A three-phase electric arc furnace is a valuable tool for melting scrap metal toproduce high-quality steel for use in steelworks and foundries. The output voltageof the arc furnace transformer is a few hundred volts and is supplied to three, large-diameter, graphite electrodes. Arcs are developed between the electrodes, and thescrap metal in the furnace and the electrodes are controlled to move vertically to tryto maintain the arc current in each phase at a chosen target value. It is the heat fromthe arcs that is responsible for melting the scrap and not conduction of current withinthe scrap. In the early stages of the melting cycle, the arcs are very unstable andunbalanced, and the currents can change by a large amount from one half cycle to thenext, in the worst case from open-circuit to short-circuit and vice versa. In addition,the scrap metal settles as it melts, sometimes collapsing around the electrodes tocause a short-circuit which may persist for several tens of cycles. In the later stagesof melting and refining, when the arcs are established between the electrodes andmolten metal, they become much more stable and balanced.

The fluctuating currents pass through the supply system and cause correspondingvoltage fluctuations which can disturb other consumers. The most commonlyreported disturbance has been the flicker effect on the light output of filamentlamps; fluctuations of picture size on early television sets also caused complaints.It is not always possible to mitigate lamp flicker by connecting the arc furnace to astronger supply point. When a series capacitor installation was used to offset for themains frequency inductive reactance of the supply system, it had a very limitedability to reduce disturbances.

Synchronous compensators, designed to have low transient reactance, absorbedsome of the reactive current fluctuations of arc furnace loads when they were connectedin parallel with the primary of the furnace transformer. To improve the sharing effect ofthese compensators, the supply impedance was sometimes increased by means of aseries buffer reactance. Good results were reported for this system, and fast excitationcontrol was claimed to give some further improvement, but the results were misleadingbecause the transducers used to measure voltage fluctuations had a slow response.Caution was needed in selecting the impedance of the buffer reactor – too large a valuerisked a loss of synchronism because of the significant rotor oscillations which occurredas the machine attempted to respond to the fluctuations of furnace power as well as thereactive fluctuations (Concordia et al. 1957).

5.2 Experimental Arc Furnace Compensation by Transductor

When a large arc furnace complex was being planned for the Sheffield area inEngland, using six 40 MVA arc furnaces, it was clear that the existing supply networkwould be inadequate to supply the load because other customers would experienceunacceptably large voltage fluctuations. The use of synchronous compensators was

376 D. J. Young

being considered, but, because transductors offered a static equivalent to synchronouscompensators, Friedlander was approached for his views and to offer a design. Asynchronous compensator does not compensate the unbalanced components as effec-tively as the balanced components of furnace reactive demand, because its negativephase sequence reactances are higher than its positive-sequence reactances. In con-trast, transductors can be arranged as single-phase units, so that they can compensatebalanced and unbalanced components equally; in addition, transductors do not haveinertia and thus cannot lose synchronism with the supply voltage.

From reports of the operation of large arc furnaces in other parts of the world, itappeared that the most severe effects of voltage fluctuations causing lamp flickerwould be the very large, low-frequency voltage dips during the early stages ofmelting cold scrap iron. Disturbances lasting less than 3 cycles (60 ms) were notconsidered important and could be ignored. Consequently, the specification requiredthe transductor compensator to respond completely to large disturbances within5 cycles at 50 Hz (100 ms). Friedlander arranged a laboratory-scale model demon-stration to show that this speed of response could be achieved with transductors. Itwas agreed that there should be a larger trial using an actual arc furnace. There was asuitable 500 kVA arc furnace, supplied at 11 kV, at the British Iron and SteelResearch Association Laboratory in Sheffield. Recordings were made of the opera-tion of this furnace to guide the design of the compensator.

The disturbing currents of the arc furnace lie within the range from open-circuit toshort-circuit. In order to aim at complete compensation of these currents, thetransductors would need to be large enough to perpetuate the short-circuit currentof the arc furnace. Small arc furnaces generally have a lower per-unit circuitreactance than large furnaces. For the 500 kVA furnace, the short-circuit currentwas more than 3 times the rated current, and the specification required the correctionto be 75% of this maximum value.

The arc furnace current includes odd harmonic currents because the arc has anonlinear impedance characteristic; there is often asymmetry between successivehalf cycles of the furnace current, resulting in even harmonics and brief durationdirect currents. The odd harmonics generated by the single-phase transductors wouldbe attenuated by the inclusion of an inductor in series with the control windings. Abank of shunt harmonic filters tuned to 2nd, 3rd, 4th, 5th, 6th, and 7th harmonics wasconnected in parallel with the arc furnace transformer and the transductors tominimize voltage distortion on the supply busbars. The capacitive output of thesefilters was approximately equal to the maximum transductor load.

A six-phase grid-controlled rectifier fed the transductor DC winding, with aclosed-loop control system to regulate the alternating current. 2nd, 4th, and 6thharmonic filters were connected across the DC supply to minimize interactionbetween the transductor harmonics and the rectifier. The forcing power of therectifiers was great enough to enable the transductors to change between minimumand maximum current in 50 ms, i.e., only half the specified time. Fig. 23 is a single-line representation of the compensator arrangement.

The normal short-circuit level at the 11 kV substation was 93 MVA, and thevoltage fluctuations at this “reference busbar” were barely noticeable. In order to


produce more severe voltage fluctuations for the experiment, a large buffer reactorwas introduced to reduce the short-circuit level at the furnace transformer to only13 MVA. The visual effect of the disturbances on this “flicker busbar” was moni-tored by a 110 volt, 60 watt lamp fed via a voltage transformer, and this could becompared to the lamp flicker on the reference busbar, which was at a completelyacceptable low level.

High-speed recordings showed an impressive reduction of the large,low-frequency, voltage fluctuations, and the transductor compensator fully met thetarget performance criteria in this respect. Nevertheless, the results of the tests weredisappointing with reference to the reduction of visual flicker effects on the 60 wattlamp. The visual perception of the higher-frequency components of lamp flicker wasstill disturbingly strong with the compensator in service.

It became evident that the sensitivity of the human eye and brain to the smaller, butmore frequent, step-function changes in illumination caused by an arc furnace wasmuch greater than had previously been considered and allowed for; it was concludedthat sudden voltage changes would need to be cancelled within about one half cycle ofthe AC frequency in order to be strongly attenuated (Dixon et al. 1964).

5.3 Experimental Arc Furnace Compensation by Self-SaturatedReactor

Within a few days of the evaluation of the transductor test results, Friedlander proposedan alternative concept which promised to meet the required compensating performance.He realized that the magnetic constant voltage transformer, which is described in sect.2.2 and is used to protect sensitive loads from supply voltage fluctuations, could be used

Fig. 23 Circuit of experimental transductor-type SVC for arc furnace compensation

378 D. J. Young

in the reverse mode to protect a supply system from a disturbing load, Fig. 24a; it alsohas the required very fast speed of response, as described in sect. 2.1.1. For arc furnacecompensation, when the slope reactance of the self-saturated reactor, Xs, matches thenegative reactance,� nX1, of the tapped branch in the equivalent circuit of Fig. 24b, thevoltage at the star point is a constant value equal to the saturation voltage Vs. Thecurrent, I1, drawn from the supply must therefore also be constant so that the supplyvoltage, Vp, becomes immune to the variations of furnace reactive current, I3. Vs alsobecomes the effective supply voltage for the furnace. After a laboratory demonstration ofthe principle, further tests using self-saturated reactors were agreed.

For the full-scale trial, again using the 500 kVA arc furnace, the compensation wasarranged as a three-phase, mesh-connected version of the basic single-phase circuitshown in Fig. 25. The auxiliary transformer in parallel with the arc furnace transformeris used to magnify the voltage change across the saturated reactor in a manner equivalentto the tapped reactor of the constant voltage transformer. The ratio of the auxiliarytransformer was adjusted to match the slope reactance of the saturated reactor to thereactance of the buffer reactor, Xo, which was connected in series with the furnacetransformer. The shunt harmonic filters were tuned to 3rd, 5th, and 7th harmonics. Themonitoring arrangements were similar to those for the earlier experiments.

The results were highly successful, giving a reduction of perceptible lamp flickerby a factor of approximately 7 to 1, and all observers found the residual flickereffects on the 60 watt lamp at the flicker busbar to be barely visible or not visible(Dixon et al. 1964). Phase sequence filter measurements showed that the phase-by-phase compensation reduced the negative phase sequence current drawn from thesupply system to an insignificant level, even when the furnace was operating in itsmost unbalanced conditions.

Because the trial was intended to prove the principle of effective flicker compen-sation, it was accepted that the compensated voltage supplied to the furnace (the

Fig. 24 Tapped reactor SVC for arc furnace compensation


saturation voltage of the saturated reactor) was lower than the normal supply voltageand that this would reduce the melting power of the furnace and increase the meltingtime. Fig. 26 shows how the rated power of the furnace can be restored byincorporating an autotransformer into the saturated reactor in order to restore thenormal rated supply voltage to the arc furnace transformer.

Fig. 26 Tapped reactor SVC with boosting of furnace voltage

Fig. 25 SVC using auxiliary transformer arrangement for arc furnace compensation

380 D. J. Young

5.4 Commercial Applications of Saturated Reactors for ArcFurnace Compensation

Shortly after the publication of the very positive results of the trials referred to insect. 5.3, there was a successful commercial application in Ethiopia (Friedlanderet al. 1965). At the time, the main source of power for Addis Ababa was a remotehydroelectric power station consisting of three 18 MW alternators. An arc furnacerated at 1.7 MVA caused such severe disturbances throughout the system (they wereeven clearly visible at the generating station) that the furnace was only allowed tooperate for 6–8 hours a day from about midnight.

GEC installed a compensator to control the voltage fluctuations caused bythe arc furnace based on the successful tapped reactor scheme, using three -single-phase saturated reactors with a three-phase rating of 7 Mvar and a bankof shunt harmonic filters for 3rd, 5th, 7th, 9th, 11th, and 13th harmonicsalso totaling nominally 7 Mvar (Fig. 27). The arc furnace transformer couldbe connected either in star or in delta to cover the required range ofoutput voltages, and the compensation equipment was connected in star; inorder to suppress third harmonic phase voltages and currents, a mesh windingwas applied to the single-phase saturated reactors and also looped over the series

Fig. 27 Circuit of SVC for arc furnace compensation in Ethiopia


buffer reactors. The saturated reactors included auto transformer windingsto restore the arc furnace primary voltage to its rated value as-illustratedabove in Fig. 26.

Because of the sensitivity of the higher-frequency harmonic filters toany drifting of the system frequency (which had been manually controlled), afrequency controller was supplied for the generators to ensure that the frequencywas stable to 50 Hz � 1%. A target reduction of disturbances to not more than15% of the uncompensated fluctuations was agreed. This target was succes-sfully achieved so that all the constraints on the timing of furnace operationwere removed. The residual fluctuations were barely perceptible and were wellbelow the level which would cause irritation to other users of the system(Friedlander et al. 1965).

There were several other successful commercial applications of the tapped reactorscheme to single arc furnaces for which a large reduction of disturbances wasneeded, down to 15 to 20% of the uncompensated level. These levels of lamp flickerreduction do not yet appear to have been matched by power electronic devices.

Because of the way in which the tapped reactor scheme must be integratedinto the circuit of an individual arc furnace, it cannot conveniently bearranged to compensate multiple furnaces which operate simultaneously in themelting mode. Nevertheless, harmonic compensated three-phase saturatedreactors have been applied to compensate multiple furnace loads. They havethe same fast response as single-phase reactors, but their ability tocompensate unbalanced conditions is less effective; their negative phasesequence reactance is greater than their positive-sequence reactance, but they havebeen used to reduce disturbances caused by single and multiple arc furnaces to about30% of the uncompensated level (Kennedy et al. 1974).

5.5 Compensation by Decoupling Transformer-Reactor

Another interesting and very effective method of compensation also uses a tappedreactor. In this application the negative reactance at the tapping point of the tappedreactor is matched to the system impedance, to provide one output to the disturbingload and a second output to the protected load with minimal interaction from oneload to the other. This principle of compensation had previously found applicationfor protection against welding loads but was originally patented for use in earlytelevision transmitters.

This compensating system was installed in Scotland at a substation where onestepdown transformer supplied two arc furnaces and the other supplied the localcommunity. The point of common coupling at high voltage was strong enough toavoid any interaction or disturbance between the two loads, but the enforcedsegregation of supplies at 11 kV resulted in a lack of back-up for maintenance oremergency conditions.

Two tapped reactor-transformers were provided, one for each stepdowntransformer; each was rated to supply the combined load to the furnaces

382 D. J. Young

and the community, but they could also be operated in parallel, Fig. 28. Asanticipated, the reduction of disturbances was extremely strong, such thatthe residual disturbances on the protected busbar were imperceptible to almost allobservers.

6 Three-Phase Self-Saturated Reactors with HarmonicCompensation

In a balanced three-phase system, the third harmonic currents of a three-phase set ofsingle-phase saturated reactors would circulate between the reactors and would notflow in the line currents, but the other odd harmonic currents, especially the fifth andseventh harmonics, would be unacceptably high for general applications in powersystems. These harmonics are reduced to negligible levels in three-phase saturatedreactors by using the phase-shifting techniques that are described in sect. 3.2.

6.1 The Twin Tripler Saturated Reactor

The twin tripler reactor was the first practical design of a harmonic compensatedsaturated reactor devised by Friedlander. It was used for ratings from a few kvar up toabout 50 Mvar. Its name is derived from the idea that a frequency tripler is a three-phase saturated reactor with a mesh winding arranged to feed a load atthird harmonic frequency, see sect. 3.3. The twin tripler is made up of two such

Fig. 28 Decoupling transformer-reactor for multiple arc furnace compensation


triplers combined to operate in such a way that their low-order harmonic currentscancel out.

The core is constructed with 6 active limbs; each group of 3 limbs has zigzagwindings to provide flux displacements of plus or minus 15� and also has its ownmesh winding for third harmonic currents. The net 30� flux displacement at mainsfrequency is multiplied at fifth and seventh harmonic frequencies such that theyare phase displaced by 180�; therefore, when the windings of the reactor groupsare connected in parallel, these currents circulate between the groups and areeliminated from the line currents as was shown in Fig. 14. If the windings of thereactor groups are connected in series, the fifth and seventh harmonics develop inthe fluxes and again cancel out. As described in sect. 3.2.2, when the thirdharmonic currents in the mesh (or delta) windings are controlled to optimumvalues, fifth and seventh harmonic currents in the primary windings of each“tripler” are greatly reduced, as are the losses that they would cause. In addition,the predominant 11th and 13th residual harmonics become very small as can beseen in Fig. 29.

In order to control the third harmonic currents in proportion to the primary(mains frequency) currents, saturated reactors are used as the mesh loading reac-tors. The third harmonic currents in the mesh windings are displaced by three timesthe mains frequency fluxes, at �45�, i.e., at 90� with respect to each other. Themesh loading reactor is therefore arranged as a harmonically compensatedtwo-phase saturated reactor with 4 active limbs, again using zigzag windingswhich provide �22.5� flux displacement on each pair of limbs. This arrangementalso helps to improve the waveshape of the third harmonic currents in the mainmesh windings.

Fig. 30 illustrates the core and winding arrangement. Conventionaltransformer design and manufacturing techniques are used for building the coresand windings. Because of saturation, the core losses are greater than in a compa-rable transformer, and additional cooling ducts are included in the coreconstruction.

Fig. 29 Twin tripler currents without and with mesh loading reactor. (a) Mesh winding open (b)Mesh winding with optimum loading

384 D. J. Young

6.2 The Treble Tripler Saturated Reactor

Similarly to the twin tripler, the name of the treble tripler is derived from the conceptof three frequency triplers, in this case arranged with 20� flux phase displacements.The core has nine active limbs. There are simple windings on the central group of3 limbs, which give no phase displacement of the fluxes; the groups of limbs oneither side have zigzag windings which give, respectively, � 20� and + 20� phasedisplacements. In normal operation, this design of saturated reactor cancels allharmonics below 17th and 19th harmonics (6n � 1), but even these are reduced toinsignificant levels by the mesh loading reactors.

In the treble tripler reactor, the third harmonic currents are phase displaced by120� and thus form a three-phase system at triple the mains frequency. In this case,the mesh loading reactor is a three-phase saturated reactor, with three active limbs,and is itself fitted with a mesh winding which operates at 9th harmonic frequency.Fig. 31 illustrates the winding arrangements of a treble tripler. Fig. 32 shows acompleted core and winding assembly before tanking.

Fig. 30 Winding arrangement for a twin tripler reactor


Fig. 31 Winding arrangement of a treble tripler reactor

Fig. 32 Treble tripler core and winding assembly

386 D. J. Young

Treble tripler reactors have been installed with continuous ratings up to 170 Mvar,but substantially higher ratings are possible. The voltage-current characteristic has asharp knee point, and the slope is linear to within about 1% from 10% up to morethan 300% of rated current.

6.3 Slope Correction for Saturated Reactors

The natural slope reactance of self-saturated reactors is typically 8 to 15% overits normal range of operation. This is uncomfortably large for most applicationsin power systems, for which only a small change of voltage is desirable. Seriescapacitors offer an obvious way to provide the negative reactive impedanceneeded to reduce the effective slope reactance to practicable values. However,the connection of capacitors in series with saturated reactors is a potentiallyunstable arrangement and, in the absence of preventive measures, would causesubharmonic oscillations. A bypass filter, which includes a damping resistor asshown in Fig. 33, is always used in conjunction with slope correcting seriescapacitors to ensure that this circuit never presents a capacitive reactance to anysubharmonic frequency (one half, one third, one fourth, etc., of the supplyfrequency).

Fig. 33 Series capacitorbypass filter


7 Applications of Self-Saturated Reactors

7.1 Disturbances Caused by Industrial Loads

When the electrification of industrial processes gathered pace, converter-fed DCdrive motors became increasingly popular because of their easy and accurate con-trollability. The drawback of using this type of drive was that the associated reactivedemand was often relatively large and would cause unacceptable voltage fluctuationsfor other users of the supply system. Reinforcement of the supply system was usuallyexpensive and not always feasible. The availability of a reliable and robust design ofstatic compensator, which could reduce the voltage fluctuations to acceptable levels,provided a lifeline in such cases.

7.1.1 Compensation for Mine Winders, Hoists, and Rolling MillsIn the United Kingdom, efficient AC induction motor drives were often used toreplace old steam-driven winding engines in coal mines, but converter-fed DC drivesbecame increasingly popular because of their greater controllability. The var demandof these drives, typically in the range of 3–5 Mvar, would often cause voltagedisturbances which were larger than the local 11 kV distribution system could acceptwithout reinforcement. The harmonic distortion due to 6-pulse converters was alsopotentially a problem. For several mine winder drives, a standardized SVC was usedto reduce the voltage and harmonic disturbances to acceptable levels.

The SVC used slope-corrected twin tripler reactors, with a continuous inductiverating of approximately 5Mvar, together with a 5 Mvar shunt filter capacitor bank. Inorder to adapt the fixed compensation voltage of the SVC to the value of distributionvoltage chosen by the supply utility, a small transformer with tap changer wasinterposed between the distribution busbar and the SVC, so that the saturated reactorcould always operate within the necessary compensation range.

The fluctuating Mvar demand due to thyristor converter-fed drives for rollingmills, particularly reversing mills, is much larger than that for mining applications.Larger drives may use 12- or 24-pulse converters, but often some 6 pulse auxiliariesrequire attenuation for 5th and 7th harmonics. There are usually several large drives,including reversing mills, which are the chief sources of fluctuating var demand. It isnecessary to make a statistical assessment of the probability of coincidence ofsimultaneous high peaks of demand from all the drives and to allow for this insetting the operational range of static compensating equipment. Twin and trebletriplers with slope correction and shunt harmonic filters were used for such SVCs(Clegg et al. 1974).

7.2 Compensation for Long Transmission Lines

As referred to in ▶Chap. 3, “AC Network Control Using Conventional Means,”Baum had described how transmission distances could be increased by using asynchronous machine to provide voltage support at the midpoint of a long

388 D. J. Young

https://doi.org/10.1007/978-3-030-35386-5_3

transmission line (Baum 1921). Subsequently, Griscom had illustrated this stabiliz-ing action by means of his mechanical model of a transmission line (Griscom 1926).Rudenberg and Friedlander had identified that a self-saturated reactor has a naturalsaturation characteristic which is similar to the control characteristic of a synchro-nous machine and it should therefore be able to fulfill an equivalent stabilizing action(Friedlander 1930). Furthermore, a saturated reactor should have a much lower costthan a synchronous machine. The concept of saturated reactor stabilization waspatented by Friedlander (DRP 592510 1931), and he described it in a contributionto a book edited by Rudenberg (1932), but there were no practical applications formany years.

7.2.1 Model StudiesFriedlander continued to look for opportunities to convert principle into practice andarranged for the construction of a model transmission line to demonstrate to consul-tants and potential customers the stabilizing action of self-saturated reactors(Friedlander and Jones 1969; Ainsworth et al. 1974). A three-phase 415 volt labo-ratory supply represented the strong network and a 200 kW motor/generator set,driven by a synchronous machine, represented a weaker, remote generation or loadsystem. Transmission lines were represented by multiple lumped components; somesections were equivalent to individual lines, while others represented two (or more)lines in parallel. Voltage stabilizers consisting of 100 kvar treble triplers, with seriesslope-correction circuits and shunt capacitors, provided the intermediate voltagesupport Fig. 34. The model demonstrated the possibility of steady and stableoperation with angles between the equivalent machine voltages well in excess of180�. Even at such a large angle, it was shown to be possible to ride through a faulton, for example, one of two lines in parallel. Fig. 35 shows the layout in thelaboratory, with two of the line sections in the foreground and two voltage stabilizersin the center of the picture.

11kVsupply

Regulator

Double-circuit

line unit

Single-circuit

line unit

Single-circuit

line unit

Voltagestabilizer

Voltagestabilizer

Voltagestabilizer

Load

M-Gset d.c.

Double-circuit

line unit

Fig. 34 Model line arranged as three line sections


7.3 Commercial Applications

In 1978 the first commercial application of saturated reactor SVCs was for a 132 kVtransmission line in Nigeria, which was about 750 km long; 8 Mvar slope-correctedsaturated reactors were connected to the distribution busbars at two intermediatesubstations.

This system was followed in 1984 by a larger, more complex arrangement inWestern Australia, which is applied to the 700 km long, 220 kV single-circuittransmission line from Muja power station, near Perth, to the extensive EasternGoldfields area in the Kalgoorlie region, Fig. 36 (Lowe 1989). About halfway alongthe line, there is a substation at Merredin which interconnects with the 132 kVnetwork around Perth. At Kalgoorlie, the 220/132 kV substation supplies commu-nities and industries in the Eastern Goldfields mining area, which includes a rela-tively small amount of local generation. The stepdown transformers at eachsubstation have on-load tap changers and a 29.5 kV tertiary winding. Each SVChas a rated output of 44/�32 Mvar, and one is connected to each tertiary. Slopecorrection series capacitors provide a constant voltage characteristic at the tertiaryterminals.

Fig. 35 GEC’s model transmission line

390 D. J. Young

The SVCs permit stable power transfer in excess of the SIL of about 125 MW. Inthe event of an outage of one SVC or its transformer, the remaining SVCs ensurecontinuing voltage control on the 220 kV transmission line but with a reduced rangeof compensating vars. The total requirement for balancing vars is obtained byadjusting the voltage profile along the line.

7.3.1 Studies for the 735 kV James Bay Transmission SchemeIn 1971, Hydro-Quebec announced its intention to develop the hydroelectric poten-tial of the James Bay basin in northwest Quebec, to supply about 8 GW into itsnetwork. James Bay is about 1000 km north of Montreal. Hydro-Quebec haddeveloped and adopted 735 kV transmission systems, but even at this voltage, itappeared that as many as 13 parallel circuits might be needed for stable transmissionif conventional line-connected shunt reactors were used for each line section, tocontrol the Ferranti effect at the receiving end (see▶Chap. 3, “AC Network ControlUsing Conventional Means” in this Green Book). Dynamic shunt compensation hadthe potential to allow a reduction of the number of transmission lines needed andhence a major reduction of the costs of the scheme. Friedlander’s work encouragedHydro-Quebec engineers to consider the possible use of static shunt dynamiccompensation at intermediate substations. To support their calculations, they under-took studies using GEC’s model transmission line, compensated by saturated reac-tors, and concluded that the number of circuits could be reduced by about half, withconsequent major cost savings. The final arrangement uses 6 AC transmission linesat 735 kV (and one 450 kV HVDC transmission circuit). The shunt SVCs connectedat intermediate substations use TCRs rather than saturated reactors, and seriescapacitors (TCSCs) have been added subsequently to give a further enhancementof the transmission capacity of the lines.

Fig. 36 SVC compensation of a long transmission line


https://doi.org/10.1007/978-3-030-35386-5_3

https://doi.org/10.1007/978-3-030-35386-5_3

7.4 Static Var Compensation for the 2000 MW HVDC Cross-Channel Link

7.4.1 Features of the HVDC SchemeA 2000MWHVDC scheme links France and England by submarine cable across theEnglish Channel, from Les Mandarins Converter Station near Calais, in NorthernFrance, to the Sellindge Converter Station, near Folkestone in the southeast ofEngland. Full-rated power can be transmitted in either direction, depending onsystem requirements in each country.

At Sellindge, there are four 500 MW 12-pulse converters combined into two1000 MW bipoles, which can operate independently, with an output voltage of�275 kV. At the rated power transfer of 2000 MW, either importing or exporting,the reactive demand of the converters is about 1200 Mvar lagging. This demand islargely compensated by eight-switched capacitor banks, each rated at 130 Mvar. Thecapacitor banks are configured as harmonic filters. Four of the filters (two per bipole)are broadly tuned to attenuate the predominant 11th and 13th characteristic har-monics of the converters. The other four filters (again two per bipole) provideattenuation for low-order non-characteristic harmonics, including background dis-tortion on the 400 kV supply network.

7.4.2 Features of the Supply NetworkThere are several nuclear power stations in Northern France which ensure a strong400 kV system at Les Mandarins; no additional dynamic reactive support orharmonic filtering was considered to be necessary to support the operation of theFrench Converter Station. In contrast, the features of the 400 kV network in SouthEast England are less conducive to the connection of a large HVDC terminal.Dungeness is the only generating station near to the Sellindge Converter Station.When in export mode from England, most of the power needs to be drawn from amore remote group of power stations on the Thames Estuary, Fig. 37.

Fig. 37 400 kV supplies toSellindge Converter Station

392 D. J. Young

The short-circuit level at the Sellindge 400 kV terminal is normally not less than9 GVA, but under summer light load conditions, it is about 6 GVA; it might even fallto 4 GVA under severe outage conditions. Operation of the Converter Station shouldnot cause disturbance to other customers. When the shunt capacitor banks areswitched in or out, the voltage step change should not exceed 1.5% at a short-circuit level of 9 GVA. Harmonic distortion on the 400 kV network should belimited to 1% for an individual harmonic and 1.5% total rss (root sum of squares)distortion. The rate of change of the transmitted power should be slow enough toavoid interference with normal frequency and voltage control strategies for thenetwork. A scheduled change from no load to full load and vice versa takesapproximately 30 min. Unscheduled changes of transmitted power (for example,due to faults) are much more problematic and need special precautions to mitigatetheir effects.

7.4.3 The Need for Dynamic Mvar SupportA fault in either the English or the French network can lead to blocking of theconverters, with total loss of power flow and reactive demand. If the converters aredisconnected and the capacitor banks tripped immediately, it will take 30 min for fullpower to be restored. The alternative is to keep the capacitor banks connected and totry to recover converter control and pre-fault power flow as soon as the fault(in England or in France) has been cleared. This causes a potentially severe voltageswell on the 400 kV system. The overvoltage is greater for the power exportcondition, and, at the short-circuit level of 6 GVA in summer, the overvoltagecould reach 1.4 per unit. A temporary overvoltage of this magnitude would becompletely unacceptable for consumers at the distribution level, which has a normaltolerance of �6%.

In order to reduce short-time overvoltages to a more acceptable level ofabout 15%, fast-acting equipment capable of absorbing at least 900 Mvar wasneeded. The capabilities of synchronous compensators and thyristor-controlledSVCs were evaluated, but it was found difficult to achieve designs whichreconciled normal continuous operation and the severe overload duties (Allonet al. 1982).

Because of their transformer-like construction, SVCs using saturated reactorshave a very large inherent short-time overload capability and are well able to satisfyboth normal and overload compensating duties.

7.4.4 SVC Performance RequirementsThe 400 kV network is normally operated within the voltage range of �5%, i.e.,between 380 kV and 420 kV. The compensating equipment was specified to be ableto operate continuously between a capacitive generation of 300 Mvar at 380 kVandan inductive absorption of 300 Mvar at 420 kV. The nominal dynamic characteristichas a slope of 5%: the nominal “float” condition is 0 Mvar at 400 kV, but the floatcondition must be adjustable between 380 kV and 420 kV. The dynamic character-istic gives a change of 600 Mvar in response to a voltage change of 10% and thus


makes a contribution of 6000 MVA to the effective short-circuit capacity of the400 kV network at Sellindge and therefore reduces its sensitivity to switching of theHVDC capacitor banks.

With regard to the short-time overload duty, many network faults are cleared veryquickly, within two to three cycles, and the preset power flow can be quicklyrestored, but allowance must be made for some faults to persist for longer periods;the HVDC shunt capacitor banks will remain connected to allow for power flow tobe restored after this longer delay. If, however, converter control has not beenrestored within 300 ms, circuit breaker tripping is initiated; the HVDC convertersand capacitor circuits are interrupted within a further 55 ms. The overload absorptioncapability of the SVC installation was specified as 3.3 pu (990 Mvar), for 0.5 s, at1.165 pu voltage (corresponding to the design slope of 5% on 300 Mvar). Fig. 38illustrates the specified requirements, with the envelope for the required continuousrange shown shaded.

Because the HVDC capacitor/filters had already been specified, it was necessaryto arrange that the compensation equipment would not compromise the harmonicperformance and ratings of these 400 kV capacitor banks.

7.4.5 The SVC EquipmentThe compensation equipment is subdivided into two identical SVCs, each with adynamic range of 150 Mvar within a total rated Mvar range of �150 Mvar. TheSVCs are based on treble tripler saturated reactors; they include series slope-correction capacitors and fixed and switchable shunt capacitors; they are connectedto the 400 kV system by stepdown transformers.

The saturation voltage of each treble tripler (the knee of the voltage-currentcharacteristic) is 56.6 kV, and its slope characteristic is 11% on rated current. The

Fig. 38 Voltage-Mvar characteristic for SVC at Sellindge

394 D. J. Young

slope-correcting series capacitor circuit is connected to the neutral terminals ofthe saturated reactor, so that it is not vulnerable to through-fault currents. It hasthe same nominal impedance as the saturated reactor; consequently, the netimpedance of the saturated reactor, seen from its line terminals, is zero, andthis creates a constant voltage busbar at 56.6 kV. The slope-correcting capacitorsinclude a damping bypass circuit, which prevents the generation of subharmoniccurrents, and a discharge voltage transformer to ensure readiness forre-energizing within a few cycles. Although the slope-correction circuit is ratedto withstand the specified half-second overload duty, some transient conditionscould impose even higher voltage stresses; therefore nonlinear resistors areconnected in parallel with the capacitors to ensure that all overvoltages arelimited to safe values. The capacitor units in the slope-correction circuits arefitted with internal fuses.

The capacitive output range requires each SVC to include shunt capacitor bankswhich can be switched out when the SVC inductive range is needed. The 56.6 kVvoltage does not change when the SVC filters are switched in or out, so that SVCcapacitor switching causes no net Mvar change on the 400 kV network.

The overall 5% slope of the SVCs, seen from the 400 kV busbars, isprovided by the reactance of the 150 MVA, 400/56.6 kV stepdown trans-formers. The transformers have on-load tap changers so that the SVC floatvoltage can be adjusted through the range 400 kV � 5%, if required by systemconditions.

The harmonic studies showed a particular need to avoid amplification of thenon-characteristic second and third harmonics, which can always be present aslow-amplitude background distortion of the network voltage. Third harmonic cur-rents can be generated by inherent phase unbalances in the converters as well asunbalances in the 400 kV voltage. Any unbalance in the French network willgenerate a second harmonic modulation of the direct current, which is convertedinto third harmonic currents by the converters at Sellindge.

In order to coordinate the filtering characteristics of the SVC shunt capacitorswith the 400 kV shunt capacitor banks, two small filters are permanentlyconnected in parallel with the saturated reactors; one is tuned to just belowthird harmonic, and the other is a “C” Filter,1 which is broadly tuned to dampany incipient second harmonic distortion. Two larger filter capacitor banks, whichare also tuned to just below third harmonic, are switched when needed to satisfythe requirements for leading vars. The shunt capacitor units are fitted with

1Friedlander needed to identify a filter design which would provide damping for low order harmonicfrequencies without introducing significant losses at fundamental frequency. He compared abouteight possibilities and decided to use the third item in his alphabetical list, “C”. Thus, “C-Filter”became the shorthand reference for this filter arrangement.


external, expulsion-type fuses. All the shunt filters include discharge voltagetransformers. All the linear reactors for the filter and the bypass circuits are ofthe air-cored, air-cooled design. The rapid discharge of switched capacitor volt-age can be seen in Fig. 39.

A detailed single-line diagram for one SVC is presented in Fig. 40.

Fig. 39 Oscillogram of switching a 56.6 kV filter

396 D. J. Young

SDT Stepdown transformer

ET Earthing transformer

SR Saturated reactor

(C ) (L ) STA (Capacitor) (inductor) switched tuned arm

(C ) (L ) UTA (Capacitor) (inductor) unswitched tuned arm

(CM) (CA) (L) (R) CF (Main capacitor) (auxiliary capacitor) (inductor) (resistor) “C” filter

SC Series capacitor

B (C ) (L ) (R ) Bypass (capacitor) (inductor) (resistor)

DVT Discharge voltage transformer

NLR Non-linear resistor

SA (SC) (L) Surge arrester (static compensator) (reactor)

CB Circuit breaker

N Star point of filter arm (isolated)

Fig. 41 is a photograph of one of the SVCs at Sellindge. The treble tripler is in anoise enclosure in the center of the picture. The switched filters are in the foregroundand the slope correction circuits in the background.

While the SVC design was in progress, CEGB (the Central Electricity GeneratingBoard, now named National Grid) identified that dynamic Mvar support was alsorequired, for system operational reasons, on the 400 kV network to the west of theDungeness nuclear power station. CEGB decided to locate this at Ninfield, about

400kVX

XX

X X

CB

56 6kV

CB

LSTA LSTA LUTA

CUTACSTACSTA

DVT

STA STA UTA CHCFN

NCF Slope Correction Circuit

BRNLR

SC BL BC

DVTDVT

DVT

RCF

CACF

LCF

NN

DVT DVT

DVT

CBET

SRSAL

SVC 2

SDT

SASC

Fig. 40 Single-line diagram of one SVC. See the component identifications below the figure


30 miles from Sellindge, and to use an SVC identical to those at Sellindge. Thisdecision did not affect the design of the main SVC components but resulted in smallchanges to the harmonic filter specifications.

The commissioning tests were completed smoothly in only 8 days, and the SVCs atSellindge were taken into service in 1984 in advance of the start of commissioning ofthe HVDC equipment. The SVC at Ninfield was taken into service a few months later(Brewer et al. 1986). During rare system and converter disturbances, the SVCs haveresponded fully in accordance with the design specification and performance objec-tives and remain in service as essential components of the Sellindge Converter Station.

References

Ainsworth, J.D., Cooper, C.B., Friedlander, E., Thanawala H.L.: Long distance AC transmissionusing static voltage stabilisers and switched linear reactors. CIGRE, 31–01 (1974)

Allon, H., Gardner, G.E., Harris, L.A., Thanawala, H.L., Welch, I.M., Young, D.J.: Dynamiccompensation for the England-France 2000 MW Link. CIGRE, 14–04 (1982)

Baum, F.G.: Voltage regulation and insulation for large power long distance transmission systems.J. AIEE. 40, 1017–1077 (1921)

Brewer, G.L., Horwill, C., Thanawala, H.L., Welch, I.M., Young, D.J.: The application of static varcompensators to the English terminal of the 2000 MW HVDC cross channel link. CIGRE,14–07 (1986)

CIGRE TB 25: Static var compensators; WG 38-01, Task Force 2, (1986)

Fig. 41 View of one SVC at Sellindge

398 D. J. Young

Clegg, E., Heath, A.J., Young, D.J.: The static compensator for the British Steel Corporation –anchor project. In: IEE International Conference on Sources and Effects of Power SystemDisturbances, IEE Conference Publication 110, (1974)

Concordia, C., Levoy, L.G., Thomas, C.H.: Selection of buffer reactors and synchronous con-densers on power systems supplying arc furnace loads. AIEE Trans. 76(part 2), 170–183 (1957)

Dixon, G.F.L., Friedlander, E., Seddon, F., Young, D.J.: Static shunt compensation for voltage-flickersuppression. In: IEE Symposium on Transient, Fluctuating and Distorting Loads and their Effectson Power Systems and Communications; paper no 7, February 1963. IEE Conference ReportSeries No 8, Abnormal loads on power systems, p. 49. (1964)

DRP 592510, Friedlander, (1931)Easton, V., Fisher, F.J., Friedlander, E.: A 100 MVA Transductor for Testing Alternators; paper

117, CIGRE (1958)Friedlander, E.: Uber Kippschwingungen, insbesondere bei Elektronenrohren; Doctoral thesis,

Berlin 1926, also Archiv fur Elektrotechnik, vol 16, p 273 and vol 17, p 1. (1926)Friedlander, E.: Selbstattige Blindstromkompensation auf langen Hochspannungsleitungen; Sie-

mens Zeitschrift, p. 494. (1930)Friedlander, E.: Der Spannungsgleichhalten, ein verzögerungsarmes, statisches Regelgerät zum

Ausgleich von Wechselspannungschwankungen; Siemens Zeitschrift 15, 177–181 (1935)Friedlander, E.: Grundlagen der Ausnutzung hochster Eisensattigungen fur die starkstrom technik;

ETZ, Ausgabe A, 11 Feb 1958Friedlander, E.: Static network stabilization: recent progress in reactive power control. GEC J. Sci.

Technol. 33(2), 58–65 (1966)Friedlander, E, Jones, K.M.: Saturated reactors for long distance bulk transmission lines. Electr.

Rev., 27 June 1969Friedlander, E., Young, D.J.: TheQuin-reactor for Voltage Stabilisation. Electr. Rev. 126–9, 22 July 1966Friedlander, E., Telahun, A., Young, D.J.: Arc-furnace flicker compensation in Ethiopia. GEC

J. Sci. Technol. 32(1), 2–10 (1965)Griscom, S.B.: A mechanical analogy to the problem of transmission stability. Pittsburgh, Electr J.

23, 230–5 (1926)Kennedy, M.W., Loughran, J., Young, D.J.: Application of a static suppressor to reduce voltage

fluctuations caused by a multiple arc furnace installation. In: IEE Conference on Sources andEffects of Power System Disturbances, IEE Conference Publication No 110, (1974)

Lowe, S.K.: Static var compensators and their applications in Australia. IEE Power Eng. J. 3(5),247–256 (1989)

Rudenberg, R.: Elektrische Hochleistungsubertragung auf weite Entfernung; pp. 182–239.Springer, Berlin (1932)

David J Young, David Young was educated at King Edward’sSchool, Birmingham, and read Mechanical Sciences at CambridgeUniversity. After joining the General Electric Company (GEC), hewas appointed as Assistant to the Company’s Consultant,Dr. Erich Friedlander, at Witton, Birmingham. He was immedi-ately involved in the early development of static var compensators(SVC) for flicker correction and then for their wider application intransmission and distribution systems. He became the Chief Engi-neer responsible for SVC and FACTS projects using saturatedreactors and power electronic devices, initially at Trafford Park,Manchester, and then at Stafford where he was also responsible forharmonic filter design, including filters for HVDC projects. Hewas appointed as a Consultant after the company became part ofAlstom and worked as an independent consultant after retiring. Hewas a member of the Disturbances Study Committee of UIE


(International Union for Electro-heat, now the International Unionfor Electricity Applications) which specified and produced theUIE/IEC Flickermeter. He also served on the IEE (Institution ofElectrical Engineers) Panel P9 and was a member of severalCIGRE Working Groups reporting on the application of SVCsand on reactive compensation and harmonic filtering for HVDC.In 1996 he was awarded GEC’s Nelson Gold Medal and hereceived the IEEE PES FACTS Award in 2000.

400 D. J. Young

doi: 10.1007/978-3-030-35386-5 29abstract before the advent of facts controllers, breaker-switched...

Documents