THE SPECIFICATION AND CONTROL OF THE PHASE SHIFTING TRANSFORMERS [607637]

THE SPECIFICATION AND CONTROL OF THE PHASE SHIFTING TRANSFORMERS

FOR THE ENHANCED INTERCONNECTION BETWEEN NORTHERN IRELAND AND

THE REPUBLIC OF IRELAND

Authors: R Sweeney* NIE Powerteam, Northern Ireland
G Stewart NIE Powerteam, Northern Ireland P O'Donoghue ESB International, Ireland P Gaffney PCAS Ireland

PSTPST110kV Line
110kV LineSTRABANE
ENNISKILLEN
TANDRAGEELETTERKENNY
CORRACLASSY
LOUTH275kV Double Circuit Line
NIE Network ESB Network
Figure 1 AC Interconnection Between NIE and ESB
1. INTRODUCTION
NIE is the electricity transmission and distribution company for Northern Ireland and the ESB is the electricity generating, transmission and distribution company for the Republic of Ireland. The interconnection of both electricity networks brings benefits to each company in the form of sharing the generation reserves, facilitating the liberalising of the electricity market in the island of Ireland and generally improving the quality and continuity of electricity to their respective customers. At present there is a 275kV AC double circuit interconnection and two standby 110kV AC interconnections between the NIE and ESB networks as shown in Figure 1. The 110kV AC interconnections are run normally open and only closed in the event of a system emergency on either network. They need to run normally open because it is currently not possible to control the power flows through them and under certain network conditions it is possible that these interconnections could be overloaded. The installation of phase shifting transformers on these lines will control power flows and will enable the 110kV interconnections to run normally closed and thus enable both NIE and ESB to accrue the benefits of full interconnection.
* Northern Ireland Electricity, 120 Malone Road, Belfast, N. Ireland, BT9 5HT 21, rue d'Artois, F-75008 Paris
http://www.cigre.orgSession 2002
© CIGRÉ 14-118

2. PHASE SHIFTING TRANSFORMER
FUNDAMENTALS

Interconnected Power Systems
Power system theory shows that when power flows
between two networks there is a voltage change and a phase angle shift between the sending end and the receiving end voltages. If the systems are connected
together in two or more parallel paths so that a loop
exists, any difference in impedance will cause unbalanced line loading. For two interconnected networks it can be shown that:
• MW flow is proportional to voltage phase angle
difference.

• MVAR flow is proportional to scalar voltage
difference.
If a PST is inserted into one of the lines it is possible to control MW and MVAR flow with a respective phase shift tap changer and voltage tap changer.

Types of PST
PSTs can be either non-symmetric type or symmetric
type. Non-symmetric devices add a quadrature voltage to the input voltage. The output voltage is the vector vector sum of these two perpendicular voltages (therefore the output voltage is boosted by a small
amount and this “boost” reaches a maximum on
extreme taps). Symmetric PSTs are different in that they have constant or nearly constant input and output voltages across the phase shift range (the input and output
voltage vectors would describe circular arcs through
the phase shifting range). This is achieved by adding a symmetrical quadrature voltage to the input and output phase angle, as shown in Figure 2.
Figure 2 Symmetrical PST Depending on the direction of the power flow two
different operation conditions are defined: Advance
Operation – being the phase angle α that results when
the load (L) leads the source (S) voltage. Retard
Operation – being the phase angle α that results when
the load lags the source voltage.

PST Under Load
A PST will in practice, have an internal impedance and
this will modify the characteristics of the device. Take for example the equivalent circuit for a PST in Figure3
and the associated phasor diagram in Figure 4.

V
L* = Load Voltage (No load)
VL = Load Voltage (Loaded)
Cos ϕL = Load power factor
ZT = Transformer Impedance
IL = Load current
VS(a) = Source Voltage (advanced)
VS(r) = Source Voltage (retarded)
β = Transformer Load Angle
α = Phase Shift Angle (No load)
+ Advanced (Leading)
– Retard (Lagging)
α∗ = Phase Shift Angle (loaded)

Figure 4 PST Under Load
Figure 3 PST Equivalent Circuit

From Figure 4 it can be shown that:
Advanced phase shift angle (loaded) α*
(a) = α − β ..(1)
Retard phase shift angle (loaded) α∗
(r) = -(α−β)……. (2)

At unity power factor load it can be proven that the
transformer load angle β can be described in terms of
percentage impedance:
β = tan-1 {z%/100} ……………………………………(3)

These three formulae are important for the design of a PST because under loaded conditions the impedance of the device will reduce the advanced phase angle range of operation.

3. PST DESIGN AND SPECIFICATION

The choice of impedance of the PST is dependent on a
number of issues:

% Impedance
• Transient stability requirements –The impedance
of the PSTs have an influence on the ability of NIE and ESB to transfer power between their networks. This could be significant during network emergencies, to ensure that networks remain
within transient stability limits.

Network studies showed that the networks would be stable for PSTs with impedances below 20%. It was recommended to minimise the impedance to
provide an additional margin of stability.

• Losses – It is envisaged that for normal system operation power flows between NIE and ESB would be via the main 275kV interconnection and the normal function of the PSTs would be to
inhibit MW and MVAR flow that would normally
occur on an interconnected network if the PSTs were absent. The enquiry document for the PSTs therefore included a value for reactive losses.

• Short Circuit – Studies indicated that fault levels
would be maintained within equipment ratings
even for a device with zero impedance.

• Sensitivity of PSTs to impedance – A number of load flow studies were undertaken to ensure that
the chosen impedance would not result in large
changes in power flows under circuit outage conditions. The results showed that for a circuit outage, the power flows were relatively insensitive to PST impedance. Reducing the impedance from 10% to 3% increased the power flow through the
PST in the limiting case by only 3MW

Further system analysis showed that with the PSTs initially operating at a OMW transfer set-point they would be capable of restoring the flow to
within their maximum continuous rating for any single
simultaneous circuit outage on the combined NIE and
ESB networks.
On the basis of the above it was considered that the impedance of the PSTs should be specified as less than 10% on all tap combinations.
No-Load phase shift advance and retard angles
Load flow studies were performed for operating
conditions associated with the extreme ranges of power transfer between NIE and ESB to determine the no-load phase shift angles. In these studies the PSTs were blocking power flow in the 110kV interconnectors. This resulted in a no-load advance angle of +45
o and a
no-load retard angle of –38o. Manufacturers of PSTs
require this information to design the exciter winding and to assist with the tap-changer selection.
Load Phase shift advance and retard angles
Similarly, load flow studies were undertaken for those
operating conditions associated with the extreme
ranges of power transfer between NIE and ESB to determine the load phase shift angles. With a PST impedance of 10% the studies required a load advance angle of +27
o and a load retard angle of –17o, with
reference to controlling the power on the line side of
the PST. Manufacturers of PSTs require this information to ensure that under full load conditions the voltage drop across the PSTs will still permit the specified load advance and load retard angle range.

Figure 5 Load and No-Load Phase Angles

Figure 5 compares the specified advance and retard angles with the actual angles of the designed PST. The actual PSTs will have symmetrical no-load advance and retard angles of 45
o. This symmetry was due to the
symmetrical design of the tap-changer. The actual load
advance and retard angles are asymmetrical. The internal impedance of the PST has the effect of reducing the advance angle with increasing load and increasing the retard angle with increasing load.

ndn2
n3n1N2
N3N1L3 S3 S2
b caIaLOAD (L) SOURCE (S)
Active part of exciting transformer Active part of series transformer
a
b cA1
B1C1
Ja
bI'1
Jc
a
Stabilizing deltaI1 A
B
CA'
B'
C'
VaVA1A'1
A'1
EXCITATION
WINDINGS
EXCITED
WINDINGSSERIES
WINDINGS
REGULATING
WINDINGS
Jbc
EXCITING UNIT SERIES UNITK1 K2
b &c phases as per a phase
(omitted for clarity)K1 and K2 change over
simultaneously to provide
Advance or Retard
operation.
PSTCVTCS1 L2 L1
Tap 33
Tap 1 & 65

Figure 6 Winding Connections for Phase Shifting Transformer

Power flow step size
A quadrature tap-changer has a discrete number of
steps to control the flow of power over the no-load
advance and retard angle range. It is necessary for the client to advise the PST manufacturer of the number of steps that are required. A balance needs to be determined between a practicable number of steps and
the magnitude of the MW change per step, as this has
an operational impact on the two networks.

Load flow studies based on a peak demand of 490MW export were performed for varying power transfers through the PST and showed that, on average a 20MW
change in power flow would require a 5
o change in
phase angle across the PST. To provide the power system operators with an appropriate level of control i.e. 5MW, a 1.25% step tap changer was selected.

With a no-load range of 90
o a 72 step tap changer was
needed. It is not practicable to purchase a single tap-
changer with this number of steps. The number of connections, leads and contacts would impact on the reliability of the PST. There would also be the issues of cost and potential problems with transport due to the weight of the PST.
Taking these factors into consideration, a 35 position
tap-changer with an advance retard switch (ARS) was selected. The ARS switch reverses the polarity of the
quadrature voltage in the PST and almost doubles the
tap-changer range. The combination of tap changer and ARS switch resulted in a 65 step device that would provide approximately 5.6 MW/step power flow control. This option also helped to minimise the
number of connections, leads and tap changer contacts.
The final PST design was of shell type design having two cores, in two separate tanks; the series and the exciting unit, as shown in Figure 6.

Voltage step size (MVAR )
As the voltages on each network may vary by +
10%
and with the need to synchronise the two networks, a tapping range of +
20% is required. With the tap-
changers that are commercially available and by selecting from a regular step fraction a 35 step tap
changer was specified, to give a step size of 1.33
o and a
range of + 22.67%.
With the PST in service the voltage tap-changer will be used to control MVAR flow between the two networks. NIE and ESB require the ability to control reactive

δV2
ESB
δV3
δV1
NIE
+ P MW
– P MW+ Q MVArA
B
D C- Q MVAr
Figure 8 – PST Power Flow Controls
power flows between the respective networks in the
range + 30 MVAr. System studies were performed to
investigate
what the expected MVAR flow would be per step. The analysis produced an average figure of 6.0 MVAr/step which is similar in magnitude to the MW/step for the phase shifting tap changer.

MVA Rating
The Letterkenny–Strabane and Enniskillen –
Corraclassey 110kV overhead lines have winter ratings of 125 MVA. The ratings of the PSTs were thus matched to the overhead lines.
Voltage Rating
The 110kV voltage on the NIE and ESB systems may
vary by + 10% and the maximum voltage could rise to 121kV. The PSTs were therefore rated to the IEC standard of 123kV.
Figure 7 Tapchanger Interactions

4. TAPCHANGER INTERACTIONS

PSTs which control both MW and MVAR must be given special consideration as the two tap changers are not totally independent of each other. The phasor diagram in Figure 7 illustrates that
increasing the voltage tap position will have the effect
of slightly reducing the phase angle α. The converse
will occur with the phase angle α when decreasing the
voltage tap position.

5. CONTROL OF THE PSTs

Various PST control modes were investigated to
determine the best mode of operation of the tap-change control to achieve optimal convergence of the operating point and the set point (MW, MVAR, V). To
compare the various control modes it was first
necessary to model the network to establish the
expected operating steps of the phase and voltage tap-changers. In each mode the active and reactive power step displacements are determined by the following: 1. The system network impedances behind the PST
terminals in both directions.
2. The actual phase shift between the ESB and NIE
systems.
3. The voltages driving the power flow. Three combinations of parameter control were
identified for the PST under normal operation as
follows:

1. P, jQ controls both active and reactive power
to a given set point.
2. P, V_s controls both active power and source
side voltage to a given set point.
3. P, V_l. controls both active power and load
side voltage to a given set point.

For the above parameter control combinations four
possible control modes were identified for
investigation. The object was to find if there was an
acceptable single mode which would yield an acceptable solution for each of the three parameter control combinations. The solution was to be implemented using a programmable logic controller. The four modes considered were as follows:

P: This gives priority to the power control,
changing the phase until it lies inside the band-width of the set point before switching to off and handing control to the voltage tap-changer.
V: This mode is the reverse of P above.
Er: This mode continuously calculates the largest
error of MW and MVAR and changes that tap accordingly.
PV: In this case the Phase and Voltage operate in
tandem until one has reached the bandwidth setting
when the other take control alone.
Figure 8 PST Power Flow Controls

-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120-120-100-80-60-40-20020406080100120
MWMV
ArP
V
ErP-V
set lineMode Comparisons D – B
-100 -50 0 50 1000102030405060
MWMVAR
-100 -50 0 50 1000102030405060
MWMVAR
The performance of each control mode was
investigated for the P,jQ parameter control. Figure 8
showed how it was proposed to look at the most
extreme movements of the OLTC to find out how each of the modes of operation behaved, by permitting the largest possible magnitudes of A-B, A-C A-D and D-B within the thermal limit of the transformer as indicated in Figure 8. The effects of large changes in P (and Q)
combined with small changes in Q (and P) were also
investigated. The analysis for the D-B case is shown in Figure 9:

Figure 9 D-B Control Mode Comparisons

P-V Control Mode
From the analysis of each control mode the following
conclusions were drawn as shown in Table 1:

Shift Unacceptable Acceptable
A-B
A-C A-D D-B P or Er
V
Er or V
P V or P-V
P, Er or P-V
P or P-V
Er, V or P-V

Table 1 Control Mode Comparisons

In each case the P-V mode was found to be acceptable.
Before confirming that this was to be the recommended mode the speed of convergence from the initial operating point to the set point for each mode was compared i.e. the total number of P and V taps. The Er
mode appeared to be the fastest but caused problems in
certain cases. This is due to the fact that when the start
and final set points of Q for example are close together, the transition error is only controlled by the magnitude of the P error during transition and can become quite large. The other three modes used a similar number of tap-change combinations to get to the set points.

Finally, a means to control the transition error by controlling the number of intermediate tap positions Figure 10 Intermediate set points

This consisted of generating a series of intermediate set points along the line between the start and finish points
and is shown in Figure 10. As can be seen the
reduction in number of intermediate taps reduces the transition error but can slow the convergence of the final set point as overall it can require slightly more tap-changes. It was therefore decided not to include this feature into the P-V control mode.

6. SYSTEM EMERGENCY CONTROL

The PSTs are series devices used to interconnect both
the NIE and ESB networks. Abnormal power flows
through these devices will occur during system
emergencies on either of these networks. A feature of the control system was to ensure that the controls would not exceed PST permissible thermal overload limits and over-voltage limits during emergencies.
Figure 11 Detection of Abnormal Power Flows

The control system was designed to be capable of distinguishing between normal and abnormal load variances by monitoring the rate of change of MW and MVAR demand. When an abnormal increase or
decrease in demand is detected the control system will
lock and hold the current tap positions continuing to deliver an emergency block of load until the remote operator despatches an instruction to change state. There are programmable maximum levels for the magnitude of MWs and MVARs which can be TimeMW, P
MVAr, QA
B
CP or Q
TPQdPdQ
Previous
Average Value
see Note eA, B & C are
instantaneous
values

supported during an emergency period and the system
automatically controls it’s own actions to ensure that
the programmed levels are not exceeded during that
period as shown in Figure 11 where: a. dP, and dQ are user settable levels. b. T
PQ is user settable.
c. T PQ is the time period over which the actual
change of P or Q remains greater than or equal to
dP and dQ.
d. The time period over which the Previous Average
Value of P or Q is obtained will be user settable.
e. Waveform A will activate Emergency Response
(ER) as the change in instantaneous value exceeds
the previous average by ≥ dP or dQ for a time
period ≥ TPQ.
f. Waveform B will not activate ER as the change in
instantaneous value does not exceed the previous
average by ≥ dP or dQ.W
g. Waveform C will not activate ER as the change in
instantaneous value exceeds the previous average
by ≥ dP or dQ, but for a time period < TPQ.

7. CONCLUSIONS

• PSTs are generally designed to be site specific and
many system studies need to be performed to adequately specify these devices.

• Utilities need to specify an adequate number of
steps per tap changer to allow tolerable network
MW and MVAR flows per step change.

• PST control schemes are site specific and may be
very simple or complex, and need to be
specifically developed to meet the requirements of the network operators.

8. REFERENCES

1. ANSI/IEEE PC 57.13S – “Guide for the
Application, Specification and Testing of Phase Shifting Transformers”.
2. Seitlinger, W. – Phase shifting Transformers
Discussion of specific Characteristics, 1998,
CIGRE Session 12-306.
3. CIGRE Working Group WG12.05 report of 1983.