21, rue d’Artois, F-75008 PARIS A2-301 CIGRE 2008

http : //www.cigre.org

CORE-FORM VERSUS SHELL-FORM SHUNT REACTORS,

UTILITY AND MANUFACTURER POSITION

A. CANCINO*, R. OCÓN, G. ENRÍQUEZ, R. MALEWSKI

Industrias IEM Comisión Federal de Electricidad Instytut Elektrotechniki
México México Poland

SUMMARY

Extra high voltage lines have a large ground capacitance, and the line-capacitance charging current has
to be compensated by shunt reactors connected at the line end to avoid dangerously-high operating
voltages at terminal equipments. Also, a fast growing load in densely populated areas and difficulties
in getting permit to construct new lines force utilities to operate the existing lines at their maximum
capacity. This can be achieved by compensation of the line reactive impedance with shunt reactors.
Utilities often contemplate acquisition of shell or core type reactors that have different design and
operational characteristics. To compare these two design options, technological issues and operational
characteristics such as the core configuration, loss and ability to withstand over-voltages are reviewed.
A choice between these two types is dictated by several factors, the most important is the cost of
acquisition, but this may be partly counter balanced by the lower cumulative cost of loss dissipated
over the reactor technical life. Technical issues have to be taken into account: higher mechanical
rigidity of shell form winding and lack of segmented part of the core column are considered an asset.
Environmental considerations, such as noise, are of importance for reactors installed in the downtown
area of large agglomerations. A choice of three phase unit versus a bank of three single phase reactors,
usually with the common reactor at the ground side, is dictated by size and weight of a large three
phase reactor that have to meet the road or rail transport limits. Transient over voltages induced in the
windings by frequent switching operations call for a high dielectric-strength of the reactor insulation,
and a trade off is to be determined between the higher costs of HV circuit breakers with pre-insertion
resistors, or shunt reactors with reinforced series insulation. In the latter case, the switching transients
induce oscillatory over voltages in the reactor windings. The utility has to determine the reactor
operating conditions in the technical specification to enable the design engineer to account for them
while calculating the insulation safety margin. When reactors are connected to the tertiary winding of
a large network transformer the possible overvoltage conditions on the transformer shall be consider at
transformer design stage.
The reactor acceptance test involves application of the voltage 50% higher than the rated, at a higher
frequency, usually 180 or 200 Hz. Such voltage is generated by a rotating machine of much smaller
capacity than the reactive power of the reactor under test. To compensate this power a large capacitor
bank has to be tuned to resonance at the test frequency. The reactor loss measurement is rather
difficult as the loss-current to reactive-current ratio is often lower than 1:500, but a high measuring
accuracy is required, since a penalty for excessive loss attains thousands dollars per kW.
Relative merits, operational characteristics, technological issues and problems of HV shunt-reactors of
a different design are reviewed and discussed in the paper. The paper also presents main aspects of the
content of technical specification addressed by the major Mexican utility.

* Email : [email protected]
KEYWORDS

Shunt reactor, shell form, core form, switching, over voltage, resonant frequency.

1. INTRODUCTION

Capacitive reactance of transmission lines increases with their voltage, and above 220 kV has to be
compensated by shunt reactors. Initially, shunt reactors have been designed for connection to tertiary
winding of the transmission-network transformer [1]. Their three-phase reactive-power attained 75
MVAr, and rated voltage ranged between 10 kV and 20 kV. Both air-core and oil-immersed reactors
has been implemented and installed in substations since 1950-ies.

Fig. 1. Air-core shunt reactors of the older helix design A, and the modern
“spider type” B, connected to the tertiary winding of a large transmission-
network transformer.
With the expansion of 400 kV network and introduction of 750 kV lines at the end of 1960-ties the
demand for shunt reactors that compensate the line capacitive reactance has increased, and new
designs were introduced for direct connection to the station bus bars without an intermediate
transformer [2, 3].
Shell-form reactors have been designed as a stake of pancake coils enclosed in a laminated steel core
adjacent to the oil-tank walls. The winding is usually composed of two coils connected at the HV
terminal located at the winding mid-length, and the neutral is made from both ends near the grounded
magnetic shield circuit. The magnetic flux lines pass mainly inside the oil filled winding window, and
return via the laminated steel core (see Fig 2).

Core-form design of HV shunt reactor was developed and introduced at the end of 1970 decade.
Essentially, the reactor resembled a single winding transformer with the wound core leg composed of
cylindrical segments separated by gaps with spacers. At a lower voltage level the core segments were
manufactured as stacks of transformer steel strips, kept together by bolts. Then radial laminated core
segments bonded by epoxy resin were developed and gained acceptance of both reactor manufacturers
and users (see Fig. 3).

The most recent product has been smoothing reactor for high-voltage direct-current (HVDC) stations.
The main function of HVDC reactor is filtering of harmonics, but limiting the steepness of ground
fault currents and attenuation of lightning surges are also valuable features for HVDC transmission
system.

2
 Fig. 2. Cross section of single phase, shell-form shunt reactor (left) and view
of 400 kV, 50 MVA single phase, shell form reactor under test (right).

A C
Fig. 3. Three-phase core-form reactor core (A), composed of segments made
of radial laminated grain oriented steel (B), separated by ceramic spacers (C).

2. REACTORS CONNECTED TO TERTIARY

WINDING OF NETWORK TRANSFORMERS

Capacitive impedance of HV transmission-line can be compensated by a reactor connected to the

tertiary winding of a large network transformer. An obvious advantage of this solution lies in a
significantly lower cost of the reactor and associated breaker that operate at voltage lower than 20 kV.
Both oil immersed and air core reactors are used. The latter is generally less expensive, but requires a
larger distance to adjacent installations, since its stray magnetic field induces a non negligible current
in nearby steel structures. Initially, the air core reactors have been wound using a large cross section
cable as a simple helix supported by a massive structure to ensure sufficient mechanical rigidity. An
invention by Anthony Trench of the “spider design” that consists of individual strands of the reactor
winding wire connected to radial arms of the “spider” welded of aluminium bars has been a milestone
in the air core reactor design [4]. This design reduces copper loss in the winding, since the current is
spread between parallel wires connected to corresponding arms of the “spider”. In consequence,
proximity as well as skin effects is reduced and current density distribution in the wire cross section is
more uniform.

Utilities have to consider two factors when such reactors are specified:
* An increased risk of failure of the big network transformer, since such failure can be initiated
by a high dielectric stress and dynamic forces acting on the tertiary winding when the reactor
is frequently switched on and off.
* An inherently higher loss dissipated by the network transformer.

3
The possible overvoltage conditions on the transformer shall be consider at transformer design stage
but it should be mentioned that delta connected tertiary winding of older transformers was designed
with the insulation and mechanical strength sufficient to carry the imbalance current of the main three-
phase windings; no additional safety margin was applied by the designer to account for the dielectric
and mechanical stress resulting from frequent switching on and off the shunt reactor connected to the
tertiary winding.
These factors convinced many power transmission companies to accept the higher cost of shunt
reactors directly connected at the HV line terminals, since the risk of compromising reliability of the
large network transformer is considered unacceptable.

3. SWITCHING OF SHUNT REACTORS AT THE TRANSMISSION LINE VOLTAGE

For utilities reliability of HV shunt reactors, and specifically in-service failures attributed to switching
transients is an important issue [5].
An interruption of relatively small inductive current at very high voltage has always been considered a
challenge for circuit breakers. Multiple re-ignitions that often occur before the final breaking, affect
the breaker technical life, and result in a dangerous dielectric stress of the reactor insulation.

(kV)

Fig. 4. Simulated transient over voltage (left) at the reactor terminals after switching on and
off near current-zero. A low-damping oscillatory transient overvoltage across the breaker
(Ubreaker), and at the reactor input terminal (Uinp) are plotted together with the current trace. A
switching off voltage recorded at the reactor input terminal (right graph) reveals several re-
ignitions at SF6 breaker poles, and very steep-front transients after each re-ignition.

Obviously, the reactor insulation can be designed to withstand such frequent steep-front, repetitive
overvoltages, but the increased reactor price will reflect the cost of reinforced insulation design and
manufacturing [6]. In the past special circuit breakers with additional set of poles and pre-insertion
switching resistors was used, to provide an effective damping of the oscillatory transients. This design
increased complexity of circuit breaker mechanisms and its cost. At present, some utilities employ the
synchronized switching of SF6 breakers [7], and apparently its operational experience is satisfactory.
Switching three-phase reactor by air-blast or SF6 circuit breaker results in a certain time delay
between closing the breaker poles. During this time, a current circulates in one phase winding of the
three-phase reactor, and induces a transient overvoltage in the two remaining phase windings.

An oscillatory transient of approximately 1 kHz frequency follows the reactor switching off, and
damping of this oscillation depends on the type of circuit breaker. Magnitude of such oscillation
usually does not exceed the peak value of power frequency voltage, but the high rate of rise results in a
dangerous stress on insulation of bushings and instrument transformers installed in vicinity of the
breaker. Several failures of the reactor bushings [9] have been attributed to these transients, and their
insulating system contained electrostatic screens embedded in paper layers to control voltage
distribution along the bushing core surface.

4
 ∆t ∆t

Fig. 5. From [8], phase to ground voltage recorded at the shunt reactor switched off
by SF6 breaker (left) and by air-blast breaker (right). The time lag ∆t between
closing of breaker poles is of the order of milliseconds.

A technical specification for the HV shunt reactor has to specify the type and switching mode of
circuit breaker that will be installed at the reactor. In such a way an optimum cost distribution between
the enhanced reactor insulation and parameters of the circuit breaker can be achieved.

4. SHELL VERSUS CORE-FORM HV SHUNT REACTORS

To compare these two design options, technological issues and operational characteristics such as the
core configuration, loss and ability to withstand overvoltages are reviewed.
Three phase and single phase design has been specified by the utilities. The former is less expensive
than the bank of three single phase reactors, but a physical size and weight of a large three phase
reactor may exceed the road or rail transport limits. In general, single phase reactors are preferred by
American utilities, whereas in Europe three phase design is more common.
In the case of core form reactors, five limb core is often employed, with the winding wound around
each of three central limbs composed of core segments, and two outer limbs without gaps and
winding. These provide an independent return path for the flux induced by one phase winding. Such
design results in a larger and heavier core but prevents the magnetic flux from flowing through the
tank walls during switch on, and reduces the ensued dynamic forces.
From the manufacturer point of view, five limb core designs has an advantage, since it enables use of
single phase excitation during the acceptance tests and the same apply for three phase shell-form
reactors (see Fig. 6). On these cases the ratio between zero and positive-sequence reactance is 100%.

φ
φ

(A) (B) (C)

Fig. 6. Simplified sketch of the magnetic flux Φ in the core or tank of three-phase core-form
reactor with five limbs (A), three limbs (B), and in the core of three-phase shell-form reactor
when only one phase winding is excited (C).

The energy density is highest in the core gaps or inside the shell form reactor winding, where the
magnetic permeability µr=1 [2], in order to design a reactor of high rated power, a high flux density B

5
has to be maintained in this critical volume. On core-type reactors this energy is divided and
concentrated on each gap, allowing the use of relative high flux density along the winding axis. In
other hand, the bigger mean diameter of coils on shell-form reactors, maintains the required inductive
reactance at a reduced flux density level. However, on both designs the magnetic field acting on the
winding results in a non-uniform current density in the wires due to skin and proximity effect.
On shell-form reactors to minimize this copper-loss, the rectangular cross-section wires have to be
oriented along the field lines or continuously transposed wires are used, especially when low loss
designs are needed.
Core-form reactor with the segmented core is less affected by the proximity effect since the magnetic
flux attains high density in the gaps between adjacent core segments. The flux lines fringe at the
segment edge, but the winding should be kept from the segments at a distance that reduces proximity
effect on current density distribution in the wires. Iron loss in the solid part of the core adjacent to the
highest and lowest segment contributes significantly to the overall core-form reactor loss.
The higher mechanical rigidity of shell form winding and lack of segmented part of the core column
are considered as an asset.

Regarding with overvoltage insulation design: Shell form reactors has large pancake coils given a
much larger series capacitance between coils than the edge of the coils has to ground, which minimize
transient overvoltages, and high voltage core-form reactors normally use layer type or interleaved disc
windings in order to control transient overvoltages.
Besides, the winding insulation of reactors rated at 750 kV and 400 kV level, have to be designed to
withstand the electric field in oil gaps, and along the paper-oil creep path in critically stressed areas.

Fig. 9. An example of the dielectric stress distribution calculated in the critically stressed volume
between the input coils of 400 kV shell-form reactors and its core. Electric field intensity is colour-
coded at the original plot, but the higher stress is shown by the lighter area, in grey scale.

5. SHUNT REACTOR ACOUSTIC NOISE

From the utility point of view, the requirement of a low acoustic noise is gaining importance, as the
compensation of line reactive impedance, and in particular HV cables, is needed in large
agglomerations, in vicinity of residential area [12].
A force F acting on the magnetic material surface S, to which the induction vector B is normal,
amounts to:
dF B2 (1)
=
dS 2µ r µ 0
where: dF/dS is expressed in [N/m2], B in [T] and µo=4π·10-7 [H/m].

For instance, energy density in the gap where the peak value of induction B=1.6 [T] and µ=1 is given
dF 1.6 2 x107
by: = = 10.2 x105 , and the maximum force F acting on 33x33 cm core-segment surface S
dS 8π
≈ 0.1 m2 attains 11.1 x 104 N or 11.3 tons.

6
To withstand the very high dynamic force acting on the segments their height has to be precisely
uniform to ensure a uniform force distribution on all the ceramic spacers. Usually the spacers are first
glued to core segment and then grinded [11]. Then the segments are stacked up and pressed using
threaded rods that compress the upper and lower yoke. Such manufacturing procedures are essential to
ensure the required mechanical strength of the gapped core reactor, and to reduce the vibrations as
well as the acoustic noise. Such force oscillates at 120 Hz (or 100 Hz) and contributes to vibrations
and acoustic noise produced by the reactor [10]. This is an important consideration for the design
engineer, and for the utility that may receive complaints from people residing in vicinity of a noisy
reactor.

In principle, the gapped core with distributed gaps produces a higher noise than the magnetic-shield
type or shell-form reactor with one large gap. Radial lamination of core segments helps to reduce the
vibrations, and to control the resulting acoustic noise [13].
One way to reduce the noise is to design the reactor with a lower flux density, but then the reactor
volume, weight and cost will be increased. A certain reduction of acoustic noise can be achieved by
special tank designs. For instance, on figure 10, mechanical de-couplers on the tank can reduce the
noise level by a few decibels.
Active noise absorbers have been also developed and often implemented when the requirements of
reduced noise were imposed on existing reactors [14].
In HV shunt reactors the magnetic circuit is the predominant source of noise, but reactors used to
suppress harmonics in AC/DC conversion stations may generate a stronger noise by vibration of
heavy-current windings.

Noise level is very important for Mexican utilities, and the proposal for the next revision of CFE
shunt-reactors technical specifications is to limit of noise level at 81 dB for shunt-reactors larger that
20 MVA.

Fig. 9. 15kV, 30 MVA three phase, core-form shunt reactor tank(A), finite element model for dynamic
response calculation under 120 Hz pressure excitation (B) and internal view showing the special
mechanical de-couplers to control noise level (C).

6. REGULATED SHUNT REACTORS

An optimal compensation of transmission line reactive-impedance requires changing the shunt reactor
inductance to follow daily load cycles. The most flexible control of the reactive power can be achieved
using the Static VAR Compensator (SVC), which contains a step-down transformer, stack of thyristors
and shunt reactor, as well as a capacitor bank. Continuous control of the thyristors firing-angle keeps
the reactor inductance at the level required to compensate the transmission-line capacitive-impedance.
However, SVC’s are significantly more expensive than the shunt reactors, and very often the
compensation consists in switching the reactor on and off twice a day. Another approach is based on
changing the reactor inductance by the conventional on-load tap changer [15]. The reactor can also be

7
installed inside the network-transformer tank, with the on-load tap changer in the transformer HV
winding [16]. In addition to compensation of the transmission line reactance, shunt reactors are used to
control load flow, particularly in the network connected to large wind-farm that have to be integrated
into power system [17].

7. SHUNT REACTOR TESTING

Dielectric tests of shunt reactor involve application of a higher frequency (usually around 200 Hz) test
voltage at 150% of the rated level [18]. Several tens or even hundreds of MVAr reactive power have
to be supplied to excite the reactor to the required test voltage at such frequency. No HV test
laboratory of the world is equipped with the generator with such parameters, but most of the test
stations operate a rotating machine that generates a constant frequency of say 180 Hz, or a variable
frequency, with the generator capacity of a few MVA [19]. The reactive power drawn by the reactor
under test can be compensated by a huge capacitor-bank tuned to resonance with the reactor at the test
frequency. Theoretically, such compensation can be easily achieved by adjusting the number of
capacitors in the bank. However, the tuning performed at a low voltage level need to be adjusted at the
test voltage level, due to a slight non-linearity of the reactor inductance, and a slightly voltage-
dependent dielectric constant of the capacitors. An inexperienced test floor operator may over-
compensate the reactor and force the huge capacitive load on the relatively small rotating generator.
Such condition may result in a rapid increase of the rotor speed that is too fast to switch the generator
off and prevent destruction of the rotating machine. A few such accidents have been reported with the
rotor expelled from the stator and flying over several tens of meters.
Test transformer represents a major investment to many test stations, and some of them test three-
phase reactors using a single-phase excitation.
The reactor loss specified in the manufacturer bid has to be confirmed by measurements during
acceptance test. A penalty of a few thousand dollars is applied for every kilowatt of loss power that
exceeds the value promised by the manufacturer in his bid. Such penalty may be contested on the
grounds that an excessive measuring-uncertainty of loss measurement masks the actual loss value.
Effectively, a very high measuring accuracy is required to assess the reactor loss. Despite the very low
power-factor, the loss power shall be measured with uncertainty of ±1%, but the measuring
instruments see the reactive power that may be 500 higher than the measured loss. Then the overall
uncertainty of loss measuring system shall not exceed 20 ppm. In an industrial laboratory that tests
equipment at several hundred kilovolts such measurement represents a major challenge, and the
measuring system has to be calibrated by a national or international reference institution [20].
However, bridge technique that has been mastered for single-phase reactor loss measurement [23],
does not lend itself to simultaneous three-phase loss measurements. A very high precision wattmeter
equipped with “zero-flux” current transformer and HV compressed gas, standard capacitor can meet
the requirement of 20-ppm loss-measuring uncertainty on reactors with 0.2% power factor. Such
three-phase loss measuring systems are commercially available, although rather expensive if designed
to test reactors rated at 400 kV.
An additional test of HV shunt reactor linearity is also required by some utilities and is included as
special test on IEC 60076-6 [26]. As the reactors have to operate under a higher than rated voltage,
their magnetic core shall not saturate within a specified voltage range, usually up to 130% of Un.
Both core-form and shell-form reactors are designed in accordance with this linearity requirement, but
it is well known that shell-form reactors have a more linear impedance characteristics.
To demonstrate the reactor linear behaviour at power frequency the regulated test voltage source shall
have very small internal impedance that prevents deformation of the test voltage sine. The sine
waveform of the test voltage can also be maintained on non-linear loads, such saturated magnetic core,
by insertion of a large HV capacitor-bank [24]. A simple, synthetic test has been proposed [25] to
demonstrate the shunt reactor linearity range using a regulated supply of direct current.

Based on tested units for the Mexican utilities, some technical requirements above international
standards IEE C57.21.1990 [18] and IEC 60076-6 [26] will be added to the CFE Shunt-Reactors
technical specifications:
• Operation altitude from 0 to 2500 masl, for all units.

8
 • Vibration amplitude levels: 60 µm local maximums peek to peek and 20 µm average for single
phase units and 100 µm local maximums peek to peek and 35 µm average for three phase units.
• Average temperature rise of 60°C tested at 110% nominal voltage.

8. CONCLUSIONS

1. Shunt reactors are mainly used to compensate the capacitive reactance of HV transmission lines in
order to avoid an excessive voltage increase after a load drop, or to use the transmission lines up to
their maximum capacity.
2. Utilities have to make a judicious choice of the reactor directly connected to HV transmission line
or to the network transformer tertiary. In the former case, they have to chose shell or core-form
reactor, and consider a possibility of regulation of the reactive power.
3. The Utility selects the reactor by examining the following factors:
3.1 Commercial considerations, specifically the acquisition cost of the reactor and its operational
cost determined mainly by the loss-power dissipated by the reactor during its technical life.
3.2 Reliability of power transmission that depends on:
• Dielectric withstand-strength of the reactor, and in the case of reactor connected to transformer
tertiary also the withstand strength of this transformer.
• Frequent switching the reactor on and off produces transient oscillatory over-voltage that
stresses insulation of the reactor and of adjacent equipment. This over-voltage can be controlled
by the choice of SF6 or air-blast circuit breaker, and such features as pre-insertion resistors or
synchronized switching.
3.3 Transmission system operation considerations impose requirements on the reactor linearity
within a given over-voltage range and on reactor switching that may affect the load-flow control.
3.4 Environmental considerations, such as acoustic noise of the reactor, and stray magnetic field
in vicinity of the reactor.
3.5 Delivery of the reactor to the substation, taking into account the road and rail transport-
limitations. These may impose the choice of three single-phase or one three-phase reactor.
4. In Europe the air-core shunt reactors are often connected to the tertiary winding of large network
transformers, but American practice prefers the oil-immersed HV reactors directly connected at the
transmission-line end.
5. Manufacturers can supply of core-form and shell-form HV reactors. Their design and operational
characteristics, as well as their relative merits have been discussed.
6. Dielectric tests of large shunt reactors are described and the specialized test installations reviewed.
7. Loss measurement at a very low power factor has been addressed, and reactor loss measuring
techniques analyzed. Special tests such as the reactor linearity range have also been reviewed.

BIBLIOGRAPHY

[1] Meyerhaus, A., "A New Design for Shunt Reactors of High Rating", The Brown Boveri
Review, vol. 45, No. 9, Sep., 1958, pp. 407-410.
[2] Christoffel, M., "The Design and Testing of EHV Shunt Reactors", IEEE Transactions, Vol.
PAS-86, No. 6, Jun. 1967, pp. 684-692.
[3] Edlinger, A., "Transformers and Shunt Reactors for 750 kV", The Brown Boveri Review, vol.
51, No. 1/2, Jan./Feb., 1964, pp. 56-58.
[4] “Trench Reactors/Reactors”, Brochure E600, download from www.trenchgroup.com.
[5] Ma, Z., et al., An investigation of transient over-voltage generation when switching high
voltage shunt reactors by SF6 circuit breaker”, IEEE Transactions, Vol. PWRD-13, Nr. 2,
1998, pp. 472-9.
[6] IEEE C37.015-1993, “Application Guide for Shunt Reactor Switching”.
[7] Chang, G.W., Huang, H.M., Lai, J-H., “Modeling SF6 Circuit Breaker for Characterizing
Shunt Reactor Switching Transients”, IEEE Transactions, Vol. PWRD-22, No. 3, July 2007, p.
1533-40.

9
[8] Prikler, L., “Disconnecting EHV Shunt Reactors by SF6 and Air-blast Breakers”, ATP case-
study book. Electrical Engineering Dept., TU Budapest, Hungary, 1989.
[9] Petersen, A., Russell, G., Austin, P.L., “Australian Experience with Shunt Reactors –
Reliability, Condition Assessment, End of life and Impact on Specifications“, CIGRE,
Transformer Committee Colloquium, 2007, Brugge, paper PS3-06.
[10] Grundmark, B., “Large Reactors”, ASEA Journal, Vol. 57, Nr. 4, 1984, pp. 20-25.
[11] “Shunt Reactors” ABB Brochure, download from http://library.abb.com/…/1ZSE954001-
011EN.pdf.
[12] IEEE C57.136-2000, “Guide for Sound Level Abatement and Determination for Liquid-
Immersed Power Transformers and Shunt Reactors Rated Over 500 kVA”.
[13] Anger, J., Bengtsson, C, Forslin, J., “Reactor Tank Vibrations in Relation to Sound Level and
Mechanical Stress”, CIGRE, Transformer Committee Colloquium, 2007, Brugges, paper PS3-
05.
[14] Anger, J., Bengtsson, C., Forslin, J., Olsson, K., Yamazaki, L., “An innovative Concept for
Noise Control of Shunt Reactors”, CIGRE, Transformer Committee Colloquium, 2007,
Brugges, paper PS3-08.
[15] Bertagnolli, G., Babare, A., Iliceto, F., Gatta, F.M., “Design and application of variable MVAr
output shunt reactors with on load tap-changer. Operation experience in Africa”, CIGRE 1998,
Paris, paper 12-308.
[16] Veens, J., Verhoeven, R., van Riet, M., “Experience in the Netherlands with a New Concept of
Providing Varible Reactive Compensation in HV Grid”, CIGRE, Transformer Committee
Colloquium, 2007, Brugge, paper PS3-04.
[17] Bollen, M., Bengtsson, C., Nielsen L.H., Larsen, H., Mikkelsen, S.D., “Application of
Regulated Shunt Reactors for Off-Shore Wind Farms”, CIGRE, Transformer Committee
Colloquium, 2007, Brugge, paper PS3-03.
[18] IEEE C57.2-1990 (1995), “Standard Requirements, Terminology, and Test Code for Shunt
Reactors Over 500 kVA”.
[19] Erb, W., Kraaij, D.J., “Design and Testing of Reactors for 735 kV”, Brown Boveri Review,
vol. 47, April 1960, p. 268-278.
[20] The IEEE P1715 (draft,) "Electrical Power Apparatus Low Power Factor Power
Measurements".
[21] Malewski, R., Moore, W.J., Anderson, W.E., "Inter-laboratory Comparison of
Instrumentation Used for the Measurement of Losses in EHV Shunt Reactors", IEEE
Transactions, Vol. PAS-99, No. 4, 1980, pp. 1634-1641.
[22] Malewski, R., Douville, J., "Mesures des pertes dans les réactances shunt EHT par un pont C-
tg* insensibilisé aux fluctuations de la fréquence du réseau", IEEE Canadian Conference,
Montréal 1978, Cat. No. 78 CH 1373.0, Reg. 7, pp. 489-490.
[23] Cancino, A., Ocon, R., Malewski, R., “Testing and Loss Measurement of HV Shell-Type
Shunt-Reactors at Very Low Power-Factor”, CIGRE 2004, Paris, paper #D1-303.
[24] Malewski, R., Kosztaluk, R., Train, D., "Measurements of Iron-Loss in Large Power
Transformers", CIGRE, Paper No. 12-01, 1988, Paris.
[25] Carlson, A., “Shunt Reactors. Determination of Linearity and Over-voltage Characteristics”.
Brochure 1SE 451001-2. “ABB Power Technology Products Transformers”
[26] IEC 60076-6, “Power Transformers- Part 6 Reactors.”, Edition 1.0 2007-12.

10