Literature Survey on Fundamental Issues of

Voltage and Reactive Power Control

Literature Survey
Deliverable of the MARS Project financially supported by "swisselectric research"

Omid Alizadeh Mousavi

omid.alizadeh@epfl.ch
EPF Lausanne Power System Group

Rachid Cherkaoui
Rachid.Cherkaoui@epfl.ch
EPF Lausanne Power System Group

10. June 2011

Table of Contents

1-

INTRODUCTION

2-

VOLTAGE CONTROL CLASSIFICATION AND DEFINITIONS

2-1- HIERARCHICAL CLASSIFICATION

2-2- CLASSIFICATION BASED ON TSOS (CURRENT PRACTICES IN DIFFERENT TSOS)
2-2-1- ENTSO-E CONTINENTAL EUROPE
2-2-1-1- France
2-2-1-2- Italy
2-2-1-3- Belgium
2-2-1-4- Switzerland
2-2-1-5- Spain
2-2-1-6- Germany
2-2-1-7- NORDEL
2-2-1-8- Netherlands
2-2-2- NERC
2-2-3- CONCLUSION

8
10
11
12
14
15
17
18
19
20
21
21
22

3-

TIME SCALE CLASSIFICATION OF VOLTAGE CONTROL AND PHENOMENA

24

4-

PROVISION OF VOLTAGE CONTROL

27

4-1- REACTIVE POWER RESERVE

4-2- EMERGENCY COUNTERMEASURE
4-3- PROBLEMS OF VOLTAGE CONTROL PROVISION
4-4- ANALYSIS OF THE VOLTAGE CONTROL IN THE SYSTEM
4-5- TWO BUS TEST CASE SIMULATION
4-6- PROVISION OF VOLTAGE CONTROL IN PLANNING
4-7- PROVISION OF VOLTAGE CONTROL IN OPERATIONAL PLANNING AND REAL-TIME
4-7-1- RPR PROVISION
4-7-2- EC PROVISION
4-7-3- PREVENTIVE AND CORRECTIVE CONTROL ACTIONS
4-8- CONCLUSION

27
29
30
30
33
35
36
36
37
38
38

5-

VOLTAGE CONTROL IN MULTI-AREA POWER SYSTEM

38

6-

BIBLIOGRAPHY

43

Abbreviation
AVR: Automatic Voltage Regulator
CSCOPF: Corrective Security Constraint Optimal Power Flow
EC: Emergency Countermeasure
EGRPR: Effective Generator Reactive Power Reserve
FACTS: Flexible Alternating Current Transmission System
GRPR: Generator Reactive Power Reserve
HVDC: High Voltage Direct Current
LRPR: Load Reactive Power Reserve
LTC: Load Tap Changer
MAPS: Multi-Area Power system
OPF: Optimal Power Flow
PVR: Primary Voltage Regulator
RCCOPF: Reactive Reserve based Contingency Constrained Optimal Power Flow
RPR: Reactive Power Reserve
SAPS: Single Area Power System
SCOPF: Security Constrained Optimal Power Flow
SVR: Secondary Voltage Regulator
TGRPR: Technical Generator Reactive Power Reserve
TSO: Transmission System Operator
TVR: Tertiary Voltage Regulator
VSCOPF: Voltage Stability Constrained Optimal Power Flow
VSM: Voltage Stability Margin

1- Introduction
Voltage control service is a critical ancillary service used by all system operators for secure and reliable
operation of the power system. It must be continuously active. On-going researches attempt to welldefine how to measure and commercialize this ancillary service [1], [2].
In order to maintain the voltage for the system reliability, both active and reactive power consumptions
must be controlled. However, a direct link between the voltage and the reactive power makes it possible
to control the voltage to desired values by the control of the reactive power [3].
The voltage control can be achieved by providing sufficient reactive power resources to keep the voltage
level at a desired nominal value regardless how much reactive power it takes. On the other hand,
controlling the amount of reactive power injection at each node can be accomplished through the
regulation of the voltage at the node. This brings up the issue of difference between the voltage control
and the reactive power control. Each one of the aforementioned control methods contains limitation.
The control of the voltage by the reactive power is restricted to the limitations of reactive power
resources and the control of the reactive power through the voltage is restricted to the feasible limits of
voltage at each node. Whenever the concern of the control is the reactive power resources, the aim
could be either voltage or reactive power control, but not both of them at the same time. In the case of
the transmission system, the control would be implemented on the system voltage.
In normal operation state, the reactive power balance must be kept in such a way that the voltages are
within acceptable limits. In fact inequality between reactive power generation and consumption does
not exist and the reactive power generated and consumed is always equal. Therefore, an improper
reactive power generation and consumption level in the system will result in inappropriate voltage
profile.
Unlike the active power ancillary services (frequency control reserves), the reactive power cannot be
transmitted efficiently through long distances because it leads to additional active and reactive power
losses. Reactive power losses are due to the large reactive impedance of the high voltage transmission
system1. As a result, the voltage has to be controlled by using special devices dispersed throughout the
system. In other words, reactive power generation and consumption have to be as close as possible to
each other to avoid excessive reactive power transmission.
The operator of the power system is responsible to control the transmission system voltage which
means enough reactive power available to prevent or mitigate voltage violation conditions. The system
operator could respond to the voltage problem conditions asking for all available reactive support from
its area and also from the neighboring systems. The system operators usually provide the voltage
control services from generators and consumers within their own controlled area. It is due to the fact
that reactive power transmission is a highly localized service. Principally this ancillary service is provided
by the generators [1]. Moreover, the regulation establishes some services to be supplied also by
transmission and distribution systems.
Power system equipments provide a variety of actions for the system operator which could be
undertaken to control the voltage and to schedule the production of reactive power. Synchronous
generators are the backbone of the voltage control in the network2. They are already available over
entire the system and their voltage support are low-cost and simple to control. However, they are not
the only ones and other reactive power resources in the power system are automatic transformer tap
changer, synchronous condenser, capacitor banks, capacitance of overhead lines and cables, static VAR
compensators and FACTS devices.

As an example for a 345 kV line, the reactive impedance is approximately ten times the real portion of the lines impedance.
The reactive power output of synchronous machine can, for a given active power level, be adjusted within the limits of the
capability curve by the excitation system. These limitations are field winding thermal limit, stator winding thermal limit and
thermal limit of the end-turn area of the stator core. Hydro units are water cooled and they are not subjected to end-turn
thermal limitations. So their leading reactive capability is much greater than that of a thermal unit.
2

The voltage control from generation resources is a necessary supplement to static reactive devices to
prevent voltage problem because:

Generation supplied reactive resources do not lose effectiveness at low voltage as do static
reactive devices.
The response of a generator to an emergency reactive requirement is much faster and more
accurate than the static reactive sources (except power electronic based devices).

The voltage control capability of synchronous generators is limited by saturation of both: field current
and armature current. The generators under heavy real power loading require high amount of field
current to maintain the desired terminal voltage which pushes the generator and exciter to the
saturation region. When armature current limitation is in effect, a large reduction in the reactive power
output is needed if the active power output is to remain constant [4].
Among different types of generating units, hydro power plants have less limitations and so higher
capabilities in voltage and reactive power control. Pumped storage power plants, as a specific type of
hydro power plants, not only can improve the frequency control but also can participate in reactive
power control. New technologies like variable speed pumped storage power plants with higher
capabilities than conventional ones, like frequency control during night time (at low loading) and
independent active and reactive power control, bring more flexibility for the system control. However,
the provided support by these generating units is usually affected by their far geographical location from
load centers.
The transmission customers can also supply reactive power to the system or can reduce the use of
reactive resources by power factor correction. Note that even with a unity power factor, reactive supply
and voltage control from generation sources is still required for dynamic voltage control, supplying
reactive losses of the transmission system, and maintaining reactive reserves for security. Recently,
provision of ancillary services by dispersed generation and demand side response became important.
However, TSOs cannot effectively manage and operate the provided ancillary service by thousands of
DG units. Therefore, their participation in the ancillary services is confronted with barriers at this time
[5].
These voltage regulators can be operated in automatic or manual mode. From the system operation
perspective, all voltage regulators should remain in automatic mode. Power plant operators for a short
period of time may need to place voltage regulators in the manual mode because of maintenance,
testing, or any problem in the generating units voltage regulator. These automatic controllers are set by
the control area operators in order to maintain a scheduled voltage in response to system changes due
to a disturbance or an unusual increase of power demand.
The control of voltage could be accomplished with passive (shunt and series capacitors and reactors)
and/or active (synchronous generators, synchronous condenser, and FACTS) devices. The former devices
contribute to the voltage control by modifying the network characteristics, while the latters
automatically adjust the absorbed or supplied reactive power to maintain the voltages of buses at
specific points in the system [6].
Another classification divides the voltage control devices into static and dynamic types [7]. Dynamic
reactive power resources refer to equipment that can respond within cycle of a disturbance where static
devices are not capable of reacting fast enough. Appropriate balance between static and dynamic
reactive power resources in an area should be provided to obtain a feasible operating point after a
reactive power deficit in the area [8].
A well-planned and coordinated application of these devices is essential for the economical design and
operation of a reliable system [9]. The proper selection and coordination of equipment for controlling
reactive power and voltage are among the major challenges of the power system engineering [9].
For efficient and reliable operation of the power system, the control of voltage should a) maintain the
voltages of all terminals in the system within acceptable limits, b) enhance the system stability to
5

maximize utilization of the transmission system, and c) minimize the reactive power flow so as to reduce
active (RI2)and reactive (XI2) losses [9].
A power system at a given operating state and subjected to a given disturbance is voltage instable if the
voltages could not approach post-disturbance equilibrium values. Basically, voltage instability has two
origins: first, gradual increases of power demand without sufficient reactive power support, and second,
a sudden change in the network topology which redirect the power flow in such a way that the required
reactive power cannot be delivered to some buses.
Overvoltage instability could be excluded because the over-excitation of machines is not permitted. The
risk of overvoltage in the system during low loading conditions is normally more of an equipment
problem rather than a power system stability problem (Page 525) [10]. In order to avoid such
overvoltage problems reactive power sources and transmission equipments should be managed
appropriately. One possible approaches for solving this problem could be disconnection of low loading
transmission lines1 which doesnt seriously affect the thermal limit margins or other constraints of the
other paths in parallel [11].
Voltage instability is commonly analyzed by employing two techniques, namely time-domain (dynamic)
simulation and steady-state analysis. Depending on the phenomena under investigation, one or both of
these techniques may be applied [8].
The process by which the sequence of events accompanying the voltage instability leads to the loss of
voltage in a significant part of the system is called voltage collapse. It means that, a power system
undergoes the voltage collapse if the post disturbance equilibrium voltages are below acceptable limits.
Voltage instability commonly occurs as a result of reactive power deficiency. The voltage collapse may
be total (blackout) or partial [12], [13].
The term voltage security refers to the ability of the system to maintain the voltages within some limits
following any credible contingency. In other words, there should be a considerable margin from an
operating point to the voltage instability point (or to the maximum power transfer point) after a
contingency [12]. System security can be distinguished from stability in terms of the resulting
consequences. For example, two systems with equal stability margins, but one may be relatively more
secure because the consequences of instability are less severe. During the disturbances, sufficient
capabilities to supply static and dynamic reactive power are required to prevent the collapse and have
to be mobilized on request even if this enforces a reduction of active power supply [14]. Inadequate
voltage support can result in equipment damage and in the extreme case it can lead to voltage collapse
and system instability.
Voltage instability could be considered as important as thermal overloads and the associated risk of
cascading outages. In recent two decades power system has revealed with widespread blackouts which
insufficient reactive power support was an origin or a factor in major power outages worldwide. Lack of
reactive power reserves response to the increased reactive power demand in contingencies, can lead to
operation of protection system and also limit the generators reactive power support. As a consequence,
both of active and reactive power deficient participate in the separation of the system and the spread of
cascading events over the entire system and finally make a large blackout. Therefore, insufficient
reactive power reserves in one area can increase the propagation of disturbance even in neighboring
areas.
Voltage collapse was a causal factor in the blackouts of August 4, 1982, Belgian; August 22, 1987, West
Tennessee; July 2, 1996, in WSCC; August 10, 1996, in West Coast; July 12, 2004, in Greece. Voltage
collapse also factored in the blackouts of December 19, 1978, in France; March 2, 1979, at Zealand in
Denmark; July 1979, Canada in B. C Hydros north coast region; December 27, 1983, Sweden; May 17,
1985, South Florida; July 1985, Czechoslovakia; July 23, 1987, in Tokyo; January 12, 1987, in Western
France; March 13, 1989, in Qubec; August 1992, Southern Finland; August 14, 2003, North America;
August 28, 2003, in London; September 23, 2003, in Sweden and Denmark; and September 28, 2003, in
1

It reduces the reactive power circulation and so its losses.

Italy. The following given examples which ended in voltage collapse and blackouts can demonstrate
some aspects of the voltage control problems and their consequences.

Greece
The Hellenic system was prone to voltage instability on July 12th 2004. This phenomenon is related
to the maximum power transfer from the generating areas in the North and West of Greece to the
main load center in the Athens metropolitan area. The Hellenic interconnected system (Greece
network) blackout was a sever voltage collapse. At that time two generating units in Peloponnese
and Northern Greece were out of service which was further stressing the Athens grid.
The sequence of events leading to the blackout was started with the failure of 300 MW generating
unit in Athens area. This unit was reconnected to the network but it was lost again due to high drum
level.
A manual load shedding is implemented by the transmission system operator which was not enough
to stop the voltage decline. So a further load shedding command was requested which didnt have
time to be executed, because of a generating unit trip was occurred at central Greece automatically.
Another unit was manually tripped and the voltage was collapsed. The system was split into two
parts. In one part, the remaining generators were disconnected by under-voltage protection leading
to the blackout. The other part, North and Western of the Hellenic system, was saved due to the
split of the system. This part was interconnected to the 2nd UCTE synchronous zone. The resulting
surplus of power in this part created a severe disturbance in the neighboring systems of the UCTE
network. This excess generation changed the flow in the northern interconnections. As a result,
interconnection with FYROM was overloaded and tripped, the Bulgarian interconnection was
received huge surplus power, and the frequency increased to 50.75 Hz.
During the incident, the power stations in the affected area lost their voltage control due to the
over-excitation. Therefore, they lowered their pre-disturbance active generation in an attempt to
increase their reactive capability and controlling their terminal voltage. This, however, had an
adverse effect, as it increased the import of power into the affected area, thus creating further
voltage drop despite the increased reactive generation [15], [16].

Sweden and Denmark

The system of southern Sweden and eastern Denmark were experienced blackout on 23th
September 2003. The operating conditions were stable within the Nordic security requirements.
Initial disturbance was outage of nuclear power plant due to mechanical problem and lose of 1175
MW generation. This contingency managed through operational reserves and the supply-demand
balance was restored. Within 15 minutes time to restore the system into N-1 secure state, a double
bus bar failure is occurred which disconnected four 400kV transmission lines.
Power flow increased on the remaining transmission link between central and southern Sweden. At
this stage, the level of reactive power support for voltage control was reduced because no major
generators were left connected to the transmission system in southern Sweden. As a result, voltage
levels on the lines dropped to critical levels and consequently a voltage collapse was developed in a
section of the transmission network. Mal-function of distance protection tripped some transmission
lines and severed all remaining transmission connections between north and south of Sweden.
An electrical island containing southern Sweden and eastern Denmark was formed. However, the
large generation deficiency led to collapse of frequency and voltage in the islanded system [17],
[16].

North America
In 14 August 2003, several hours before the start of events, there was large volume of power
transmissions through Ohio to the other areas. It led to high reactive power demand and
consequently a severe shortage of reactive power in northern Ohio. But the supply of reactive
power was low because some power plants were out of service and others were not producing
enough reactive power. The sign of insufficient reactive power increased VAR production at nine of
power plants in this area. It caused the generators to operate near limits with reduced reactive
power reserves for contingencies [18].
Then a transmission line went out in southwest Ohio due to a contingency, and two hours after, low
voltage situation shut down the Eastlake plant. This state redirected active power and consequently
changed the need for reactive power. It made transmission lines between Cleveland and southern
Ohio and also links between northern Ohio and southern Ohio tripped. Ultimately, this situation
caused the power plant and transmission line failures and participated in the spread of blackout.
Although this blackout was not due to a voltage collapse but the Task Force Final Report said that
insufficient reactive power was an issue in the blackout.The report also cites overestimation of
dynamic reactive output of system generators as a common factor among major outages in the
United States [19].
As it is demonstrated above, although reactive power control is primarily a local problem, it may involve
several TSOs in the interconnected systems and increase the scale of blackouts and even affect on the
intact areas.
Some difficulties associated with the voltage control are a) the possible spread of a local but
uncontrolled voltage collapse whereas a relatively small remedial action at the right place and time may
stop the system degradation, and b) a sufficiently detailed representation of major components is
needed even if they are located geographically far from the disturbed area [20].
This report intends to study different aspects of the power system voltage control. The main concern is
whether the voltage control should be considered in the scope of multi-area power system. For this
purpose, the current practices of different system operators and their different voltage control
classifications are investigated in section 2. The different time scale classifications in the power system
voltage control are described in section 3. Section 4 studies the role of providing reactive power reserve
in the power system security and surveys different proposed methods by literatures in provision of
reactive power reserve for voltage control. Possible studies for the voltage control in multi-area power
system are intended in section 5.

2- Voltage control classification and definitions

The organization of the voltage control can be decomposed in different levels. The voltage control is
separated into various classifications depending on the time response of the controllers and strategy of
each TSO. The classification based on the time response is called hierarchical voltage control in the
literatures. The latter classification concerning the specific strategies of TSOs highlights the complexity
of the interactions between the different system operations.

2-1- Hierarchical classification

The hierarchical scheme in the voltage control is a common classification which is usually implemented
in three levels; primary, secondary and tertiary voltage control [6]. The primary voltage control referrers
to the local response of the generators. The voltage control at a zonal level is related to as secondary
control and the tertiary voltage control is on a global system level. The interactions between these
8

control levels can be managed based on different objectives, time responses, and geographical
implications.
The primary voltage control or Primary Voltage Regulation (PVR) is performed by automatic and rapid
voltage regulators which control reactive power output so that the output voltage magnitudes are kept
at specified values. PVR has only local scope targeting the control in the particular bus assigned to the
controllable device. Generally, the primary voltage control is performed by the generators Automatic
Voltage Regulator (AVR). The AVR regulates the voltage by controlling the excitation system. The other
controllable devices like synchronous condenser and SVC can also be used for PVR.
The PVR set values are selected so that the desired voltage profile of the system is obtained. The
coordination and the supervision of the PVR set point values within a given geographical zone are the
tasks of the secondary voltage control also referred to as secondary voltage regulation (SVR). The main
idea behind SVR is to coordinate the various regional reactive power resources in such a manner that
they control the voltage at given pilot nodes. Pilot nodes are selected such that the voltage magnitude
at the pilot node represents the voltage profile over the associated zone. Usually the pilot nodes are the
ones with the highest short-circuit power in a given zone. The other method for pilot node selection is
based on electrical distance concept between nodes [21].
The SVRs should not be implemented in a centralized manner because the system operators should not
be involved in local voltage matter, if it is not necessary. Therefore, the SVRs zones should be selected
to have the minimum interaction between neighboring zones. In order to decentralize the SVRs, control
generators can be grouped into homogenous zones according to the electrical distance between them
or their capability to affect the pilot bus voltage. Hence, decentralized SVR zones should be selected to
minimize the effects of the generators of each area on the pilot bus voltages of the others. For the
extensively coupled zones the generators that produce the coupling effect can be taken out of the SVR
or each control center should utilize additional measurements to offset the effect of neighboring SVR
zones.
This hierarchical, zonal voltage control approach can be further enhanced using a Tertiary Voltage
Regulation (TVR) scheme. The basic idea of the TVR is to increase the operating security and efficiency of
the system through a centralized coordination of the zonal SVR structure [22]. The TVR considers the
counteract coupling between controls at the SVR levels. In fact, TVR defines the optimal voltage setpoints for the SVR pilot nodes. Different objectives like minimizing the grid losses or maximizing the
reactive reserve can be taken into account when selecting these set-points. Normally SVR and TVR are
implemented with a delay and they involve both automatic and manual actions.
The definition and the implementation of the SVR and the TVR vary from one TSO to another [23]. Some
TSOs consider secondary and tertiary voltage controls together. In this case, the voltage control is
divided only into two classes: primary and centralized voltage controls [24], [25]. From the perspective
of providers of voltage control services, the production of reactive power can be divided into a basic and
an enhanced reactive power service. The basic or compulsory service includes the generating units
requirements that must be fulfilled to be connected to the network. The enhanced reactive power
service is a non-compulsory service that is provided as supplement to the basic requirements [26], [27].
The hierarchical structure of the control system consists of numerous loops. Generally, the loops on
lower levels are characterized by smaller time constants than the loops on higher levels [6]. For
example, AVR typically responds in a timescale of seconds, while SVR operates within ten times slower
than AVR. The TVRs response time depends on the presumed operation time horizon of the operator.
These different timescales in operation result in a decoupling between the control loops and minimize
the interactions between different control levels. The requirement for communication devices on each
control level are given such that the delays must be lower than time constants of the controllers on
these control levels.
A generic hierarchical voltage control scheme is illustrated in Figure 2-1. In the depicted control scheme,
the system control center, which act as TVR, determines the optimal voltage set-points for the pilot
buses based on a given optimization criterion applied to the whole system. These set-points are then fed
9

to the SVRs and would be used by local voltage/reactive power regulators, which are PVRs, to control
their voltage/reactive power output with respect to their own reserves.

SVR
Ctrl.

Figure 2-1: PVR, SVR, and TVR overall structure.

The operation of the system under the hierarchical scheme increases the transmission capacity
associated with improvements in voltage stability characteristics of the interconnected grid [6].
Note that the voltage control areas are affected by both system topology and loading condition, and
hence these areas change dynamically during system operation, which is an issue with the current SVR
approaches, which assume that these areas remain fixed. This particular problem is addressed in [28] by
a combined Secondary and Tertiary Voltage Regulation (SVR+TVR) methodology based on real-time
optimal power flows (OPFs) to periodically update the generators AVR set points. Since the method is
mainly software-based, the voltage control areas boundaries can be readily redefined to better reflect
changes in the system operating and/or topological conditions. However, this method corresponds to
centralized OPF models, where in practice, this is likely to be an issue due to the large size and the
complexity of real systems.
In the case of hierarchical voltage control in power systems, [29] proposes wide area voltage protection
system, whenever the operating limits are reached and control efforts are saturated, including active
power rescheduling and load shedding on the area which is the first cause of the voltage instability. The
objective is the removal of the risk of voltage instability within the saturated voltage control area.

2-2- Classification based on TSOs (Current practices in different TSOs)

The growing interest in creating a reactive reserve market indicates the development of reactive control
as a specific ancillary service [1]. Several utilities developed special scheme for the control of the
network voltages and the reactive power. It is likely that the preferred methods differ from network to
network, and depend on the network structure and the reactive power compensation practices [23].
Some TSOs like Italy and France implemented some kind of automatic SVR and TVR. However, in many
countries the adjustments of the AVR set-points are performed manually from a control center [30]. In
this survey, recommendations and practices of voltage control in different TSOs in continental Europe
and NERC are studied.
10

2-2-1- ENTSO-E Continental Europe

Several hierarchical controls based on the network area subdivision and the automatic coordination of
the reactive power resources were first studied in Europe for achieving the network voltage control. The
ENTSO-E operation handbook (former UCTE) [25] provides a procedure to keep the network voltage
within predefined ranges according to the N-1 security principle by different facilities. Load shedding can
also be initiated when voltages have declined to abnormal levels. Voltage must be maintained within a
range of values in order to be compatible with the equipments size, to maintain the supply voltage
within the contractual range, to guarantee the system reliability and the static stability and to avoid
occurrence of voltage collapse.
Furthermore, ENTSO-E released a draft for the grid connection requirements of power generating
facilities [31]. The generating units are obliged to meet the requirements and to provide the technical
capabilities with relevance to the system security. According to [31] each generating unit shall be
capable of providing reactive power automatically by either voltage control mode, reactive power
control mode, or power factor control mode in coordination with the relevant TSO. The control scheme
characteristics, parameters and settings of the voltage control system components shall be coordinated
in agreement with the relevant TSO. Each generating unit shall be equipped with over and under
excitation limiters and stator current limiter and shall inform its network operator about its capabilities
to provide reactive power. The relevant network operator shall have the right at any time to change the
reactive power target value within the agreed reactive power range.
Each TSO have the right to define voltage-against-time-profile at the connection point for fault
conditions which describes the conditions in which the generating unit stay connected to the network
and continue stable operation after the power system has been disturbed. Also generating units shall be
capable to fulfill the relevant TSO requirements for automatic disconnection in case of voltage deviation
at the connection point for a specified range and a minimum time period, as shown in table 2-1 for
Continental Europe. The terms shall be agreed with the relevant TSO in the conditions set forth by
national legislation, connection agreement or any other bilateral contracts or by the TSO.
Table 2-1: The minimum time periods each generating unit has to operate for voltages deviating from the nominal value at the
connection point without disconnecting from the network [31].
Generators in Continental Europe

The voltage base is between 110 kV and 300 kV

Voltage Range

Time Period for Operation

0.80 pu 0.85 pu

30 minutes

0.85 pu 0.90 pu

180 min

0.90 pu 1.115 pu

Unlimited

1.115 pu 1.15 pu

60 minutes

0.80 pu 0.85 pu

30 minutes

0.85 pu 0.90 pu

180 minutes

0.90 pu 1.0875 pu

Unlimited

1.0875 pu 1.10 pu

60 minutes

The voltage base is between 300 kV and 400 kV

According to the [25], the voltage control has been divided into primary, secondary, and tertiary levels.
Various TSOs employ different voltage control methods based on their policy. In most cases, a single
transmission system operator (TSO) is responsible for the primary voltage control, whereas the other
control modes might involve several TSOs.
Each TSO continuously and coordinately support the voltage in its own network. The TSOs must have
information of the available reactive resources and their restrictions. Besides, they have to exchange
data for real time operation and network security analysis.
In order to ensure a safe operation of the synchronous area, adjacent TSOs should agree on common
voltage ranges on each side of the borders. In addition, they (adjacent TSOs) should provide coordinated
11

voltage control near the boundaries preventing that individual actions have opposite effects to the
security of the neighbors in normal operation and in case of disturbances.
TSO can have contract with the reactive power providers to get proper, adequate, and rapid reactive
power resources for normal and emergency operation. It is declared that, if the reactive power can be
produced in the adjacent TSOs, specific bilateral contracts should be made to transfer reactive power
through the tie-lines.
Moreover, [25] states that the TSOs are committed to have available a sufficient reserve of fast reactive
power resources participating to the PVR in order to ensure normal operation condition with a
continuous evolving of load and transits, and to prevent voltage collapse after any contingency of the
contingency list. TSOs have to keep available a sufficient number of reactive power resources connected
to the grid, which contribute to reactive power generation or absorption, in order to maintain or get
back the voltage in normal ranges after any contingency.
Different European grid operators, depending on their hierarchical level, developed and implemented
specific voltage control schemes. Here, the current practices in France, Italy, Belgium, Switzerland,
Spain, Germany, Nordic, and Netherlands TSOs are studied in depth.
2-2-1-1- France
France TSO (RTE1) has organized a three level voltage control which concern distinct geographical zones.
The zones are mutually decoupled and a decentralized secondary voltage control (DSVC) coordinate the
action of different generating sets at zonal level. The DSVC acts on all PVRs of regulating units within the
zone to control the zone pilot bus voltage and to maintain their uniform reactive loadings.
RTE has designed a coordinated secondary voltage control (CSVC) which it can be considered as the first
industrial implementation to improve the SVR. The CSVC is a closed loop centralized voltage control
scheme with a dynamic of a few minutes. This coordinated control remains on the regional level and it is
formed of several strongly coupled zones. In fact the CSVC take into account the interactions between
voltage regulation zones.
In each region one control center (CSVC) gathers the information of the pilot nodes voltage and critical
nodes voltage and also generators participating in the CSVC. This information is used to determine the
pilot nodes voltage and the set point of all PVRs in a region. The aim of the CSVC system is the
controlling of the voltages at the pilot nodes and generator terminals to set point values while
maximizing the reactive power reserves and improving the system voltage stability within a region.
Actually the CSVC continuously employs optimization for computing the voltage set points of the
generators in the supervised zone2. The CSVC is afforded to use the existing reactive resources and to
avoid installation of new devices for the voltage control [23], [32], [33], [34]. The CSVC is installed only in
the Western region which is particularly sensitive to voltage problems [35]. The described hierarchical
voltage control of French system is depicted in figure 2-2.

1
2

RTE (Rseau de transport d'lectricit) is the France Transmission System Operator.

For this purpose, each 10 seconds, measurements of "pilot node" voltages and generator reactive outputs are collected, from
which new AVR voltage set-points are computed and sent to generators at the next time step. This computation consists in
minimizing the sum of squared pilot node voltage deviations and machine reactive productions, with inequality constraints on
controls, pilot node voltages, generator reactive outputs and sensitive bus voltages

12

Figure 2-2: Organization of DSVC and CSVC in the system of France

In France, RTE ensure the provision of the primary and secondary voltage control reserves through the
ancillary services contract with the power plants. Long-term contract is signed with all the producers to
ensure the availability of the ancillary services. The voltage control remuneration depends on the unit's
geographic zone. RTE defines reactive power sensitive zones. In these zones, the generators are
remunerated to provide the voltage control whereas in the other zones, this remuneration is not
payable. The sensitive zones cover roughly a third of French territory. These areas are colored in brown
on the map given in figure 2-3.

Figure 2-3: voltage control sensitive zones in France [1].

However, regardless of the zone, the energy generator's operating expenses are remunerated at a fixed
rate per MVar/hr when the unit is run. This remuneration is increased by 50% if the unit participates in
the secondary automatic voltage control. These voltage controls are procured through the bilateral
contract with generators.
Some specific regulations related to the reactive energy are considered at the interface between the
transmission and the distribution networks. A coordinated policy and incentives is provided for
distribution companies to maintain their power factors near unity [1].

13

2-2-1-2- Italy
The Italian system operator (TERNA1) employs a hierarchical voltage control scheme for controlling the
network voltages and the generator reactive outputs. All generating units connected to the transmission
and sub-transmission grids shall contribute to the PVR. The PVR is a mandatory service without any
financial compensation. The most significant levels of this hierarchical control are the SVR and the TVR.
The hierarchical structure of the Italian transmission system voltage control is shown in figure 2-4.

Figure 2-4: Hierarchical structure for transmission network voltage control in Italy [22].

SART2 regulates the units' reactive power or local nodes voltage by directly controlling the AVR setpoints and sharing out total generated reactive power among power plant units in a balanced way. The
Regional Voltage Regulators (RVRs) within each area provide a specific reactive power level which
controls the SARTs. The RVR also controls capacitor banks, shunt reactors, OLTCs, and SVCs to avoid
saturation of area generators. The combination of the SART and the RVR implements the SVR. The SVR is
installed on all generators and is coordinated by the Italian National Control Center with the objective of
controlling voltages in some selected pilot nodes. The SVR is now a voluntary service. The pilot nodes
and the control power plants of the SVR in the Italian system are selected such that the network is
subdivided into 18 automatic coordinated control areas [36].
At the highest hierarchical control level, a TVR automatically coordinates the RVRs in a real-time closed
loop with a time constant of about 5 minutes. The TVR aims to both minimize network losses and
improve the operation voltage security [37], [38]. An Optimal Reactive Power Flow (ORPF) for Losses
Minimization Control (LMC) computes the forecasted optimal voltages and reactive levels. On the basis
of a forecasted state estimation, LMC computes in advance (i. e. the day ahead) the provisional optimal
voltage and reactive power plan, which is stored and used by the TVR. The TVR minimizes the
differences between the actual field measurements and the optimal forecasted references. The
combination of the TVR and the LMC forms the National Voltage Regulator (NVR), which links the ORPF
forecasting with real-time optimization of the SVR set-points [36]. The schematic diagram of the
hierarchical voltage control in Italian transmission system is depicted in figure 2-5 with more details.

TERNA (Rete Elettrica Nazionale) is the France Transmission System Operator

Automatic system for the regulation of voltage of power stations. (Sistema Automatico per la Regolazione della Tensione di
centrale)

14

Figure 2-5: Schematic diagram of the Hierarchical voltage control for the Italian transmission system [36].

The main achieved operational benefits of the hierarchical voltage control implementation are: the
reduction of the real losses, the increase of the reactive reserves for facing large perturbations, the
increase of the active power transfer capability, and the reduction of the risk of voltage collapse [22].
In addition to the described hierarchical voltage control, TERNA introduced a mandatory framework of
payments (/MVar/hr) for consumers and Distribution System Operators (DSOs) with excess reactive
energy withdrawals [36].
2-2-1-3- Belgium
The coordinated voltage control has been employed in Belgium since 1998, as a tool to support
decisions made by the system operators. Elia, the TSO of the Belgium network, ensure sufficient
absorption or generation of the reactive power to stabilize the system voltage through making contract
with the producers. In the network of Belgium, the secondary and tertiary hierarchical levels, as defined
by the French and Italian propositions, is not utilized. In this application, the voltage control exercised
both as primary control and centralized control. The primary voltage control automatically adjusts the
voltage variation within a given band defined by the producer, while the centralized control is activated
by the producer upon request of Elia depending on the contracted band. The main goal of alignment
objective function is to spread and maximize the reactive power reserves on the different generators
taking part to the voltage control of the system. The proper operation of this objective requires that the
import and the export of reactive power from the neighboring system tend to zero. In the present
implementation generating units, shunt capacitors, and UHV-HV transformers tap changers are
considered as controllers [39].
15

The reactive power generating units are divided into regulating units and non-regulating units. The
regulating units are capable to participate in both of the primary and centralized controls while the nonregulating units are only involved in the centralized control. The generating units with capacity over 25
MW are required to participate in the primary voltage control of the Belgium network. As additional
mean to solve a problem, Elia can ask the regulating units to activate the reactive power beyond the
bands, if this activation didnt jeopardize the security of the producing unit. Figures 2-6 and 2-7
demonstrate the aforementioned characteristics of the regulating and the non-regulating units,
respectively [24].

Figure 2-6: Utilization of the reactive band for the regulating unit in Belgium [24].

Figure 2-7: Utilization of the reactive band for the non-regulated unit in Belgium [24].

Elia launches a tender for providing the voltage control, and chooses the providers based on the price of
the received bids and the location of the generating units in the grid. The producers are paid for the
actual consumed or generated volumes of the reactive power (/MVar/hr). In addition, the required
reactive power reserve is provided through adjusting the set of the generating units [24].
In order to improve the voltage control of the network of Belgium, a hierarchical voltage control scheme
with SVR and TVR is studied in [40]. In this scheme, SVR calculates the voltage of the pilot nodes and
sends the reactive power set-points to generators. The objective of TVR is defined as minimization of
generators reactive production, capacitor switching, reactive power exchange with neighbor grids, and
voltage deviation. General structure of the proposed control system is shown in figure 2-8.

16

Figure 2-8: General Diagram of the proposed hierarchical voltage control for the network of Belgium.

2-2-1-4- Switzerland
There is no formal SVR and TVR in the Swiss grid. The TSO of Switzerland (Swissgrid) is responsible for
ensuring voltage support in coordination with the prequalified ancillary service provider (ASP), power
plant operators (PPO), distribution system operators (DSO) and TSOs in other countries (FTSO). Since
2009, Swissgrid has implemented a central voltage/reactive power control, which coordinates the
generators AVR and the transformers tap changers through a Day-Ahead Reactive Planning (DARP). The
DARP process is shown schematically in figure 2-9.

Figure 2-9: Overview of the DARP procedure in Switzerland [41].

The main input data of the DARP process is the 24 DACF (Day-Ahead Congestion Forecast) snapshots
that contain the 24 hours day-ahead power flow forecast. Moreover, it is necessary to add reactive
power limits of power plants and tap changer transformers model. The minimum available reactive
power of each plant is derived from its total active power production of the DACF model (Qlim=f(P)).
The optimal set-point for the power plants and transformer tap changers are determined such that they
minimize the cost of active power losses throughout the transmission system plus the cost of reactive
energy payment to the generators. The optimization has to ensure a number of technical and
operational constraints like voltage limits of generations and nodes, transformer tap position limits, and
reactive power flow branch group limits at the borders and for Switzerland [42], [41]. The Optimal
Power Flow (OPF) is performed in a consecutive manner for all 24 snapshots. The day-ahead voltage
17

schedule contains the individual 24 hourly set-point values for all transmission system production
nodes.
The power plant operators must regulate their units reactive power in such a way that the magnitudes
of voltages at nodes are within the reference of 3 kV. All generators in operation must support the
voltage within the obligatory reactive power margin which doesnt cause any opportunity cost for
generators. In addition, generators can provide enhanced reactive power support beyond the obligatory
service by concluding bilateral agreements with Swissgrid.
It can be seen that there is no closed-loop automatic voltage control and the voltage set-points are
determined off-line. Therefore, Swissgrid provided an operational emergency plan for the operators,
including a number of procedures to keep the voltages within operational limits in the event of
violations [43]. Based on the severity of the situations the following countermeasures can be taken:
completely utilize the obligatory reactive power, order for enhanced reactive power services, call for
synchronizing all available units as the last measure on a national level, and international redispatch
procedure with the surrounding TSOs.
The generators provided reactive energy for the voltage control is compensated by a constant default
(CHF/MVar/hr) payment rate. The precondition for this remuneration is supporting the hourly defined
voltage set-points that it is checked every 15 minutes.
At the moment, DSOs and end consumers directly connected to the transmission system have limited
ability to control the system voltage [41]. However, they have the opportunity to choose between active
or passive participation in the voltage control. The active participants are reimbursed and the passive
participants are charged. The goal of this elaborated concept is to increase the reactive power reserves
through increasing of the active participants voltage support, and improving load factor of the passive
participants [44].
Possible future enhancements of voltage control in the Swiss network could be an active participation of
DSOs and end consumers directly connected to the transmission system, better coordination with
surrounding networks, and realization of a central closed-loop voltage control with direct online control
of generators reactive power [41].
2-2-1-5- Spain
REE (Red Elctrica de Espana) is in charge of the unified operation of the Peninsular Power System and is
the owner of the High voltage Power System. REE performs its basic functions and 5 regional control
centers perform complementary functions to those of the REE.
To improve reactive power management, off-line OPF studies are performed using the worst real time
cases whose solution was obtained by the state estimator and saved in the applications computer.
These voltage optimization studies were performed in a regional level. In one region, an expert system,
called SEGRE (Reactive Power Management Expert System), to assist the operator has been developed
and it is currently in operation.
The Spain system operator adopted a two layers reactive service. The first layer is a minimum reactive
service required of all generators, which is mandatory, and the second layer the reactive services
exceeding the minimum requirement [1]. The service providers are the generators connected to the
transmission system, the available means in the network like capacitors and transformers with
regulation for the reactive management, the qualified consumers with contracted power greater than
15 MW, and the distribution networks. The providers may offer the availability of an additional band of
generation and/or absorption of reactive power that exceeds the corresponding required resources.
Voltage control and reactive power support in the Spanish transmission network is regulated by two
market mechanisms: 1) the network constraint management and 2) the voltage control ancillary service
(VCAS). Constraint management is conducted by the system operator, who asks generators to modify
their outputs to solve power system voltage problems. REE fulfill security criteria while minimizing the
cost of generation redispatch based on the generators submitted bids to the day-ahead energy market
18

[45]. VCAS has been implemented under two different time scopes: 1) annual and 2) daily. In the annual
time scope, the service suppliers submit optional reactive offers exceeding the minimum mandatory
one. According to the offers of different providers, the required resources are selected through a
tendering process and by analyzing a number of scenarios that represent the future operation of the
power system. In the daily time scope, an OPF determines the optimal operation of the power system
based on the available reactive power offers: mandatory and optional assigned in the annual time scope
[46]. Nowadays, only the first mechanism is fully implemented, while the VCAS is partially established.
The voltage set-points and the limits of reactive power output in the boundary points are determined in
a day-ahead basis. The system operator undertakes a sampling of the voltage values in the controlled
nodes every five minutes. The operator set a permissible band of 2.5 kV around the set point voltage
value.
The provision of the voltage control ensure the optimal safety and the quality of supply while minimizing
transmission losses and keeping the system away from voltage instability. The latter target can be
achieved through by: 1) maintaining an adequate voltage profile in normal operating condition and 2)
assuring that generators exhibit enough reactive margins that guarantee that the system voltages will
remain under acceptable values in case of contingencies.
The voltage control is paid for the various service providers on a monthly basis. The remuneration not
only contains the generated and absorbed reactive powers (/MVar/hr), but also considers availability
of additional band for reactive power generation and absorption (/MVar) [47].
Besides, REE has launched a research project on analysis, design and implementation of a hierarchical
and automatic voltage control, in which the current primary generator AVR would be completed by a
regional SVR and, eventually, by an automatic or semi-automatic TVR. Figure 2 shows the complete
voltage control system. The reference values for the SVR will be provided by an expert system. [48]

Figure 2-10: Proposed reactive power management and voltage control in Spanish electric system [48].

2-2-1-6- Germany
The TSOs in the network of Germany are responsible for the voltage control in the system as a part of a
secure supply. Each TSO must ensure voltage stability in its own controlled area, which involves the
power grid (transmission and distribution networks), generating units, power stations, consumers, and
the boundary areas of the adjacent networks. The following parties are involved in the voltage control
under the coordination of the responsible system operator:

the system operators own network,

the synchronously interconnected transmission systems,
the generating units connected to the system operators network,
19

the distribution networks connected to the system operators network,

the consumers connected to the system operators network.

Generating facilities, reactive compensation installations, transformer tap changing, and modification of
network topology are taken into consideration by the TSOs to ensure enough reactive power generation
and demand in the system. The conditions for supply and purchase of reactive power are specified in
bilateral agreements between the concerned parties.
Each generating unit must meet the defined minimum requirements regarding to the specified power
factor in the transmission codes. The generating unit under operation should provide the requested
reactive power as specified by the TSO.
According to the contractual agreements, if the suppliers notified a restriction in reactive power
generation, the TSO should be immediately notified. In addition, if a TSO during daily operational
planning cannot be ascertain of reactive power management by the available means (its own passive
facilities and contractually guaranteed ancillary services), it should ask for supplementary generating
units to supply reactive power. Financial compensation for that is settled on a bilateral basis [49].
2-2-1-7- NORDEL
The network of NORDEL is composed of four countries, including Norway, Sweden, Finland, and
Denmark. Each one of the system operator is responsible for voltage regulation in its own grid.
Deregulated markets in the Nordic countries do not have any provision for payments towards reactive
power services [50]. For example, Sweden follows a policy wherein reactive power is supplied by
generators on a mandatory basis and without any financial compensation. Some large generators are
rarely used for voltage control and are operated at a constant reactive power output. Also in Norway
there is mandatory reactive power supply, within power factor range of 0.93 lagging to 0.98 leading,
without any financial compensation. Additional reactive power supply could be individually imposed to
generators which it would be remunerated yearly by negotiation between system operator and
producers.
Moreover, the interaction between the system operators is considered such as communication between
the Norwegian and the Swedish system operators. The voltage of the Norwegian system is monitored by
the National Centre and also Regional Centers. If the Regional Centers do not have sufficient resources
to maintain the voltage within the given limits, the National Centre will be contacted. Two operation
centers in the Swedish system are responsible for voltage regulation in the northern and southern parts
of the grid. If the operations centers do not have sufficient resources to maintain the voltage within the
given limits, they should contact each others. In normal operation, the goal is the higher voltage within
the normal operation range. In conjunction with operational disturbances and switching, the respective
operations centers in Sweden and Norway can agree on actions to maintain the voltage within the given
intervals.
The margin for the PVR is set by each system operator for its own system and bilaterally between the
system operators in borders between the systems. Voltage regulation in each system should be
conducted in such a way that the operational security standards1 are upheld and the reactive flow
between the systems does not entail operational problems. The Parties rights and liabilities regarding
reactive power flows on the AC interconnections are limited to what corresponds to zero exchange (no
reactive exchange) at the national border, based on values measured at the terminals of the links.
NORDEL operational security standard states that there must be a reserve of reactive power within each
subsystem. It must be constituted with regard to the size, the regulation capability and the localization
to prevent the system collapse [51].

Operational security standards are criteria which the system operators use when conducting operational planning in order to
uphold the reliable operation of the power system.

20

2-2-1-8- Netherlands
In the Netherlands, individual network companies have to provide for their own reactive power, usually
through bilateral contracts with local generators, who are only paid for the reactive capacity but not for
reactive energy [36].
The operating point for the reactive power exchange at the active power output is defined by one of the
following three possibilities: power factor (cos ), reactive power level, voltage level, if necessary with
tolerance band. The operating points are defined by agreement of a value or online set-point
specification. The generators which active power is taken from must maintain a power factor of cos =
0.95 (inductive) to 1. Further exchange of reactive power is permissible and has to be agreed separately
[52].
2-2-2- NERC
In the North America power system, the enhanced voltage control is not utilized. Power plants are the
primary resources used to control the transmission system voltage. The effectiveness of the existing
reactive power and voltage control standards and how they are being implemented in practice has been
reevaluated in the ten NERC regions [53]. New generators should have an over and under-excited power
factor capability of 0.95 or less. If a generator could not meet this requirement, it should make alternate
arrangements for supplying an equivalent dynamic reactive power capability. The provision of the basic
voltage controls is compulsory in NERC. The generators are remunerated based on a regulated price.
This price incurs the fixed and opportunity cost of the generators [27].
Generators must declare their reactive power capabilities for the system operator such as
characteristics of the unit automatic voltage regulator, maximum and minimum reactive power output
capabilities, and speed of response. The generators accept and confirm the scheduled voltage or the
scheduled reactive output requests from the system operator within two minutes. These generators
must modify MVAR output to keep the voltage or the reactive output error less than the specified band
around the scheduled voltage. The generator must meet either the voltage or reactive output
requirements, but not both of them at the same time [54].
However, the performance requirements for voltage controls are not dealt with in great detail by NERC.
Thereby targets of network voltage schedules are left for the regional coordinating and operating
entities to define these requirements more specifically. NERC voltage and reactive control requirements
states that each transmission operator should acquire adequate dynamic and static reactive resources
within its area to protect the voltage levels of interconnected system under normal and contingency
conditions. Reactive resources should be dispersed so that they can be applied effectively and quickly
when contingencies occur [55]. The adjacent transmission operators are responsible for facilitating the
resolution of any potential conflicts in the applicable voltage limits.
In addition to NERC standard, the Western Electricity Coordinating Council (WECC) defines more
detailed performance requirements for automatic voltage regulators. WECC-specific standards
addressed the required active and reactive power margin in the system for both transfer paths and load
areas. The established standards of WECC assess the required static stability margin and reactive power
margin through conducting PV and PQ analysis, respectively. It defines the active power margin
requirement in such a way that the path flow transfer (the area load) should be kept 5% below the path
flow transfer (the area load) of the nose-point on the PV curve for normal operation and worst single
contingency. Also the active power margin should be kept 2.5% below the path flow transfer (the area
load) of the nose-point on the PV curve for worst multiple contingency. Similarly, the reactive power
margin requirement at the critical node under consideration is equal to the change in the reactive
power margin between

100% and 105% of forecast loading (or path transfer) for single contingencies.
100% and 102.5% of forecast loading (or path transfer) for multiple contingencies.

21

The described active and reactive power margins requirements are depicted in figures 2-11 and 2-12,
respectively [8].
2-2-3- Conclusion
Study of current practices in voltage and reactive power control demonstrates the intention of different
system operators toward implementation of more sophisticated schemes like centralized and
hierarchical voltage controls. Table 2-2 and 2-3 summarize the current practices of different TSOs in
ENTSO-E and NERC for voltage and reactive power control and its remuneration, respectively, which
explicitly described in this chapter.
Moreover, a combined SVR and TVR (SVR+TVR) methodology based on real-time centralized optimal
power flows (OPFs) is proposed in [28] to periodically update the generators voltage regulator set
points. Minimum active power losses (MAPL) and maximum loadability (ML) OPF approaches are used
for the proposed SVR+TVR control. However, in practice implementation of centralized OPF is
problematic and regionalization of the OPFs for the proposed SVR+TVR needs to be studied.

Figure 2-11: active power margins requirements [8].

Figure 2-12: reactive power margins requirements [8].

22

Table2-2: Summary of different practices in voltage and reactive power control

Hierarchical voltage control

Centralized
System

TSOs
Voltage
Control

SVR

Reactive
Power
Control

Voltage
Control

France

RTE

Italy

ENEL

Practice

Belgium

Proposed
Swissgrid
Practice

Spain

Reactive
Power
Control

Voltage
Control

Reactive
Power
Control

Elia

Switzerland

TVR

REE

Proposed
Vattenfal, EON,
RWE, EnBW

Germany
NORDEL
Netherlands
PJM

NERC

Table2-3: Summary of different provision and remuneration methods in voltage and reactive power control
System

Provision

Remuneration

France

Long term bilateral contract with generators

It is only specified for sensitive geographical zones. Energy

generators at fix rate (/MVar/hr). 50% increase if
generator participates in SVR.

Italy

Hierarchical voltage control

Excess reactive energy withdrawal should pay (/MVar/hr).

Belgium

Tender for providing voltage control resources

based on price and location of generating units

The producers are paid for the actual consumed or

generated reactive power (/MVar/hr)

Switzerland

Day ahead reactive power planning, and bilateral

agreement for enhanced reactive power support

The provided reactive energy is compensated by

(CHF/MVar/hr)

Spain

For reactive power generation day ahead reactive

planning, and voltage control ancillary service in
annual and daily time scopes. Both through
tendering process.

Monthly payment for both production and absorption

considering; 1) utilized reactive power (/MVar/hr),
2)availability of additional band (/MVar)

1- Bilateral agreements between concerned parties

Germany

NORDEL

2- Supplementary reactive power support in daily

operational planning
1- generators compensatory reactive power supply

1- Opportunity cost has to be included

2- Financial compensation based on bilateral agreements
1- without financial compensation

2- additional reactive power supply

2- yearly negotiation between system operator and

producer

Netherlands

bilateral contracts with local generators

They are only paid for the reactive capacity not for reactive
energy

PJM

Compulsory basic voltage control

The generators remunerated based on a regulated price

including fixed ($/Month) plus opportunity cost

23

3- Time scale classification of voltage control and phenomena

The scope of the voltage control studies is important in order to select the appropriate countermeasures
to enhance stability and to avoid voltage collapse. The countermeasures can be taken based on different
design stages. Various design stages in voltage control studies are given in table 3-1.
Table 3-1: Different design stages in voltage control studies [30].

Design Stage

Power System Planning

System Protection Design
Operational
Planning

Procurement
Scheduling

Real-Time
In the power system planning stage, the system operator has to ensure the viability of voltage controls
requirement of the future system. The most perceptible countermeasures actions in this stage are:

Transmission reinforcement by adding new VAR resources such as series and shunt
compensations, FACTS devices, and etc.
Construction of generation units with more capability of reactive power control.
Improving reactive power management through implementing automatic SVR and TVR schemes.

The system protection against voltage collapse consists of automatic control actions based on local or
wide area measurements. Reactive compensation switching, load shedding and load tap changers
blocking can be implemented in this stage. In these two stages the aforementioned countermeasures,
except the load shedding, deal with long-term stability.
The operational planning and the real-time control typically involve different generator responses
(including PVR, SVR and TVR) and reactive device switching. In emergency states, load shedding and load
tap changer blocking also can be taken into account. These actions aim in maintaining voltage profile
and reactive power reserves. They can be implemented either manually or automatically.
In the context of deregulated electricity markets, there are two classes of problems when analyzing the
reactive power provision (operational planning), namely, reactive power procurement and reactive
power dispatch. Reactive power procurement is essentially a long-term issue, i.e., a seasonal problem.
The system operator looks for optimal and secure reactive power allocations from possible suppliers in
the given time period. Reactive power dispatch, on the other hand, corresponds to the short-term
allocation of reactive power to suppliers based on current operating conditions. In this stage the system
operator determines the optimal reactive power schedule for all providers [56].
The system operator in the various voltage control studies deals with different phenomena in different
time-scales. The voltage stability can be classified into two categories based on the size of the
disturbance. Small-disturbance voltage stability concerns the systems ability to control voltages
following small perturbations, such as gradual change in load. This form of stability can be effectively
studied by steady-state approaches based on load flow. A lot of methods are developed for this purpose
such PV and VQ curves, Jacobian matrix, continuation power flow, and quasy steady state simulation.
Largedisturbance voltage stability concerns the systems ability to control voltages following large
disturbances such as system faults, loss of load, or loss of generation. It can be studied by using nonlinear time domain simulations in the short-term timeframe and steady state analysis in the long-term
time frame [57].
In addition, according to the different time-scale of phenomena, the voltage stability can be classified
into short, mid and long-term stability [12], [13]. In some reviews, the mid-term voltage stability with
24

time-scale within a few minutes is treated as a separate class [4], [8], [58], [59]. Nevertheless, as
discussed in [9] (p.1078) distinction between mid-term and long-term stability appears less and less
justified. Thus, in recent literatures the mid and long-term voltage instabilities are considered in the
same category [13], [60], [61], [62]. In this report, the expression long-term voltage instability concerns
all studies beyond the short-term.
The short-term voltage stability is characterized by fast acting dynamics of the power system and its
components following a disturbance. The time frame is from less than one second to several seconds.
The response of the PVR is in this time scale. Time-domain or dynamic simulation considering different
control actions are commonly used for the short-term studies.
The long-term voltage stability involves slow phenomena and slower acting equipments. Its time frame
may extend from several minutes to hours. It contains automatic or manual actions of higher level
controls like the SVR and the TVR. The investigation in this time period is done through static analysis
methods based on power flow models while considering fast dynamics stable.
The mechanisms that make the system instable in short-term and long-term dynamics are a) loss of
post-disturbance equilibrium (ST1 and LT1)1, b) lack of attraction toward stable equilibrium (ST2 and
LT2), and c) post-disturbance oscillatory instability (ST3 and LT3). Usually the evolution of the long-term
voltage instability, leads to a short-term instability. Similarly, this type of instability (S-LT1, S-LT2, and SLT3) can be distinguished according to the three aforementioned mechanisms. To face the problem, the
generators and the synchronous condensers can be asked to provide reactive power in excess of their
current limits for a limited time. But it transforms a short-term voltage problem to a long-term one.
This time scale decomposition perspective can be utilized to indicate time horizon of various
phenomena and system components actions taking part in the voltage stability. The fast acting
automatically controlled equipments participate in the short-term stability dynamics such as: generators
automatic control devices (excitation system), synchronous condensers, automatic switched shunt
capacitors, SVC, induction motor, voltage dependent loads, FACTS, HVDC links, etc. In the long-term
stability dynamics, SVR, TVR, transformer tap changers, generator limiters, switched shunt
compensation, and in the last resort load shedding could be enumerated. They typically act over several
minutes. The response of the components in the long-term voltage stability is designed in such a
manner that it has no interaction with the short-term dynamics.
This time decoupling allows to categorize the voltage controls and to perform more precise analysis.
Figure 3-1 depicts the dynamic of voltage control response in comparison with the response of other
power system controllers timescale.
Operator LFC
Long-term
voltage control
AGC
Short-term
voltage control
Protection
Power electronic
controllers
0.01

0.1

10

100

1000

time (s)

Figure 3-1: Short and long-term voltage controls in comparison with different time scales of power system controls [6], [63].
1

ST is for the short-term and LT is for the long-term phenomena

25

It can be derived that fast response of automatic voltage controls, which are commonly available over
the entire power system, is even faster than rapid active power control. For the first few seconds or
even tens of seconds after a system disturbance, there is no active power control from the generators.
The controllability of the reactive power in the generators is fast since it involves electronic control of
excitation current and does not need any mechanical power control [64]. Therefore, when there is a
sudden change in load, the voltage profile at the load buses can be controlled with rapid reactive power
regulation of generators and then the generators governor restores the active power balance at their
low speed [64]. All of the aforementioned controllers could be applied in both preventive and corrective
strategies.
The preventive and the corrective controls are two main defenses against instability incidents. These
control actions must be taken appropriately to provide sufficient margin for security. The objective of
the voltage security assessment in operational planning and real-time environments is to ensure the
system security through taking into account both types of the remedial actions. Usually the secure
operation point can be obtained with applying different countermeasure. Such decisions are taken in
accordance with each actions cost as a trade-off between reliability and economy.
In the case of the short-term voltage problem, there is not always enough time to implement the
corrective actions. Therefore, sufficient reactive power margin should be provided for the short-term
voltage instability prior the disturbance by the automatic support of the control devices. The
countermeasures for the long-term voltage instability contain both preventive and corrective actions,
because in the long-term voltage instability usually there is time for operator actions. The SVR control
actions are basically in the time scale of the long-term voltage studies. The various remedial actions for
different time-scales of the voltage instability are shown in table 3-2.
Table 3-2: Various preventive and corrective countermeasures for different time-scales of voltage instability.

Preventive action

Corrective action

short-term

PVR

long-term

SVR
TVR
Capacitor switching
Load tap changer

Load shedding
Generation redispatch
Capacitor switching
Load tap changer blocking

Automatic or manual preventive controls include the optimization of the amount and location of
reactive reserves. The PVR in the short-term voltage control and the SVR and the TVR schemes in the
long-term voltage control can be used in this respect. Note that SVR is also in charge of the shunt
compensation switching and the transformer tap changing with the objective of maintaining reactive
reserves on generators to face incidents. Automatic or manual corrective controls in the long-term
instability include shunt compensation switching, load tap changer blocking, load shedding and
generation redispatch.
In order to avoid voltage instability, three characteristics of countermeasures including amount,
location, and execution time should be appropriately adjusted. The location of corrective actions should
be selected such that the minimum amount could be achieved. The farther the countermeasure from
the location with voltage instability, the more countermeasures is needed to save the system.
Moreover, execution of corrective actions can restore the long-term equilibrium when they are
performed before the time limit. If corrective actions were realized after the time limit, the system
would be prone to LT2 instability. Otherwise, more corrective actions are required to restore the stable
post-contingency equilibrium.

26

The provided emergency controls to protect the system against the voltage collapse are divided into two
categories. The first group has no impact on consumers. This group could include the topology change,
the modification of cross border flow, the reduction of exchange, the fast generation rescheduling, and
the load tap changers. If they were available, they would be the first controls to be utilized. Some of
these actions such as generation rescheduling may involve additional cost to the utilities. In this case, if
the generators reactive power production affects on their active power dispatch, they receive an
opportunity cost payment. The second group of emergency controls, like load shedding, have a direct
impact on consumer and usually are used as the ultimate remedial action. In serious contingencies, the
system cannot be efficiently restored without some form of load shedding. It is shown that among
different emergency controls only load shedding is able to restore the long-term system equilibrium in
the presence of load self-restoration1 [65]. However, the efficient load shedding scheme should be
designed so that appropriate amounts of loads are disconnected within a delay to protect the system
against the voltage collapse. For instance, the Hydro-Quebec operator has implemented an under
voltage load shedding scheme -TDST- to have an extensive defense plan against major disturbances [66],
[67].
Usually power systems are operated with sufficient preventive controls in such a way that they can
survive credible contingencies. For more severe incidents the TSOs relies on corrective actions. When
the preventive actions are insufficient or cannot be implemented fast enough, corrective measures
should be adopted. The remedies for higher order contingencies could be covered either by its own TSO
or with neighbors. During such situation, the TSO which is on alert provide the necessary information to
the neighbors and also looks for convenient remedial actions with them [25]. Moreover to the various
mentioned countermeasures, modification of cross boarder flow and reduction of exchanged power
also can be included in the remedial actions.
The reactive power reserve in the system should be managed to improve the voltage stability and to
avoid the voltage control problems in case of disturbances. It requires adequate response of the
equipments and coordination of the control and the protection equipments. Next chapter studies the
concepts of reactive power reserve and emergency countermeasures in depth.

4- Provision of Voltage Control

The proper provision of the voltage and reactive power control resources is required to maintain the
security of the bulk power system against the short- and long-term instabilities. These resources are
anticipated to support reactive capacity and reactive energy through automatic and manual actions of
the controllers. Also these resources should be managed to keep the system voltages within established
limits, under both pre- and post-contingency condition [64]. For this purpose, the TSOs should
continuously acquire, deploy and maintain adequate amount of control actions from their resources to
meet contingencies.
These control actions comprise Reactive Power Reserves (RPR) and emergency countermeasures. The
RPR and emergency countermeasures, respectively, can be considered as the preventive and corrective
controls for the security of the system voltage. In the following sections both preventive and corrective
actions are studied more in depth.

4-1- Reactive Power Reserve

In addition to reactive power requirements to support the transmission system under normal
conditions, RPRs should be maintained for contingency conditions to enable the secure system
operation against the voltage instability and collapse. The RPR is spare reactive capability available in
1

Following a disturbance in the supply voltage, the active and reactive powers drawn by the load are restored by internal
controllers (like thermostats).

27

the system to assist the voltage control. This capability should be held in reserve to respond to
unforeseen events which lead to a sudden change of required reactive power. Contingencies like outage
of transmission line, capacitor, SVC, and large generator supplying reactive power can increase the
required reactive power substantially and immediately.
The RPR has to be composed of both reactive supply and absorption capabilities [54]. It can be activated
either automatically or manually during the momentary operational situation. The generators are the
main source of RPR which also referred as spinning RPR. Generator supplied reactive power is especially
an effective resource of RPRs because of a) its superior performance at low voltage as compared to
static reactive devices, b) fast response of excitation system, and c) large reactive range. The other
equipments which can maintain the RPR are synchronous condensers, spare shunt capacitors and shunt
reactors, and SVCs.
The RPR can be viewed from the loads and generators perspective. The two bus test system, shown in
Figure 4-1-a, is used to illustrate the various viewpoints of the RPR. A generator and a load are
connected to bus 1 and bus 2, respectively. The QVcurve method, which more details are given in [68],
is used to obtain the reactive power margin to voltage collapse point. For this purpose a fictitious
reactive power support Qf is connected to the load bus (pilot node). The QVcurve, shown in Figure 4-1c, expresses the relationship between the reactive power support (Qf) at the given bus and the amount
of the voltage at that bus (V) [30]. The minimum of the QVcurve distinguishes the reactive power
margin to loss of the current operating point. This point is called voltage collapse point and it is shown
by the white circle. The current operating point without compensation (Qf = 0) is shown with black
circle. The generators reactive power outputs of the current operating point and of the voltage collapse
point are shown on the generator capability curve in Figure 1-b. In this paper, the optimal power flow is
used as an alternative method to calculate the reactive power margin to the voltage collapse point [4].
The Load RPR (LRPR), shown in Figure 4-1-c, is the difference between the reactive power at the current
operating point and the reactive power at the voltage collapse point. It is also called reactive power
margin and usually the literatures concentrate on this point of view. The Generator RPR (GRPR) focuses
is on the quantity and the value of RPR provided by each generator. The simplest definition, which is
called Technical Generator RPR (TGRPR), is the difference between the maximum reactive power
capability of the generator and its reactive power generation at the current operating point. However,
V<

E<0

PL , QL

PPig,,QQi g

Bus 1

Bus 2

QQcf

a) Single line diagram of the two bus test system

Current Operating Point
Voltage Collapse Point
Max Reactive Capability

Qi

Qf

Qc

Qi

TGRPR

Qi

Current Operating Point

Voltage Collapse Point

EGRPR
Qf

LRPR
Pi
P
i

b) Generator capability curve at bus #1

c) QV-curve for bus #2

Figure 4-2: LRPR, TGRPR, and EGRPR for the two bus test system.

28

this quantity may not represent the useful quantity of the RPR since at the collapse point all amount of
the TGRPR cannot be utilized. Effective Generator RPR (EGRPR), as a more accurate representative of
the GRPR, is defined as the difference between the generators reactive power output at the voltage
collapse point and the generators reactive power output at the current operating point. The TGRPR is
an upper bound for the EGRPR. The LRPR, the TGRPR, and the EGRPR for the two bus test system are
shown in Figure 4-1-c and 4-1-b. Fewer studies are performed around the GRPRs rather than the LRPR.
Both perspectives need to be considered for system operation and planning to meet the system
reliability criteria [69].
Furthermore, the RPR can be classified into static and dynamic RPR based on the resources. In case of a
contingency, both of static and dynamic reactive resources are necessary for the system to survive the
transitions and settle to new operating condition. Appropriate balance between them and their location
should be well determined [8]. The dynamic RPR respond to the system voltage deviations within a few
cycles. Synchronous generators, synchronous condensers, and SVCs are equipments which fit the
definition of dynamic RPRs. Manually controlled elements are static RPRs. Note that automatic
controlled shunt capacitors do not qualify as dynamic RPRs since their control systems limit their
response.
In order to respond to the contingencies and to support the voltage during extreme system operating
conditions, the system operator needs to carry sufficient RPRs according to the best response capability
of the resources. Thus it would be a wise practice to control the system in such a way to keep maximum
amount of RPRs on the generators. In other words, generators should be operated at low reactive
power production to ensure sufficient RPR response to the system voltage changes. For this purpose,
the system operator may have to switch shunt reactors or shunt capacitors to relieve the MVar supply of
generators and allow an increase to their RPRs while maintaining the desired voltage profile [64].
However, beyond a certain level of compensation with shunt capacitors the voltage regulation tends to
be poor and stable operation is unattainable. The reason is that the generated reactive power by shunt
capacitors is proportional to square of voltage and during low voltage conditions their VAR support
drops and thus exacerbating the problem.
An effective way for the system operator to manage the RPRs in the system is the application of SVR and
TVR. After a disturbance, the operating points of the generators change according to PVR action. It leads
to poor distribution of production, consumption and total reserve of reactive power. Thus, a nonoptimum solution is attained for the overall network. As a result, the network voltage control plan
degrades due to major reactive power flows, increase of line losses, and generator overloads. SVR can
act on the set point levels of PVRs to restore the operation point to a more optimal situation. SVR
performs this modification in a coordinated manner in a control zone. This action is performed through
scheduling the voltage of the pilot points. SVR scheme increases RPR because all the generators tend to
have their RPRs exhausted at the same time. Moreover, the system operator has to respond the
scheduled and unscheduled changes of the system operating conditions. Thus the application of the TVR
would be necessary to adapt and harmonize the various set point voltages to the pilot points in all the
networks so as to economically optimize the operation of the system and ensure its optimum safety
[70]. However, it should be noted that SVR has not been devised to face emergency situations, where a
faster and coordinated control of generator is needed. If this control is performed over a large enough
area, it could preserve generator RPR by reducing network reactive losses and by increasing the
production of shunt compensation [60].

4-2- Emergency Countermeasure

When the provided RPR could not attain the desired voltage, emergency countermeasures (EC) like Load
Tap Changer (LTC) modified control, generation redispatch, voltage and active power transfer controls,
and at last resort load shedding schemes must be implemented. As it is mentioned in chapter 3, the
amount, the location and the time of implementing are important in the effectiveness of the emergency
countermeasures and should be allocated carefully. The minimization of the amount of corrective
29

actions subject to the system equalities and inequalities is often referred to as solvability restoration.
The system operators must be able to recognize voltage stability related symptoms and take
appropriate remedial actions.
Generation redispatch can be one of the emergency countermeasures, since the available RPR of a
generator varies depending on its loading condition. The generator RPR is determined by its capability
curve. Note that for a given real power output, the reactive power generation is limited by both
armature and field heating limits.
Usually emergency action on load is the ultimate countermeasure. This can be implemented indirectly
through a modified control of LTCs or directly as load shedding. Emergency control of LTCs can be
achieved by LTC blocking or LTC voltage reduction. This emergency action has to be coordinated
between different LTC levels in EHV transmission, HV sub-transmission and MV distribution levels. LTC
emergency control slows down the system degradation, but its response is affected by counteraction of
load power restoration mechanism and complex implementation due to large number of distribution
transformer to control [60]. Appropriate load shedding is the ultimate way of stopping voltage
instability. Under-voltage load shedding scheme can restore a long-term equilibrium by increasing the
active power margin of the system.

4-3- Problems of Voltage Control Provision

Adequate stability margin should be ensured by proper provision of voltage control including both the
RPR and the EC. However, at present there are no widely accepted guidelines for the selection of the
degree of reactive power margin. For the EC, each TSO utilizes different schemes to obtain sufficient
active and/or reactive power margin to the voltage instable point. The specific EC can be taken based on
the requirements of each network and TSO.
The margins to keep the voltage secure depend on provided RPR by different reactive resources which
should be managed by each TSO. In one hand, TSOs can define an acceptable voltage level for normal
operation and contingencies and must guarantee that the voltage level is not near the critical voltage in
these situations [25]. On the other hand, TSOs can determine appropriate RPR with respect to operating
constraints and voltage stability criteria. In this case, the RPRs are taken in such a way to ensure the
secure operation limit. The appropriateness of the provided RPR should be tested through contingency
analysis.
The RPR provision is affected by several problems and concerns regarding the current procurement
practices and pricing policies for reactive power. These comprise a lack of transparent planning
standards, noncompetitive procurement, discriminatory compensation policies, rigid interconnection
standards that may not meet local needs, and poor real-time incentives for production, consumption
and dispatch [53]. Existing standards are not specific for RPR requirements because the TSOs may not
bear the full reliability costs of inadequate RPRs. So the TSOs may not consider all available alternatives
in the procurement of reactive power capacity.

4-4- Analysis of the Voltage Control in the System

Voltage instability imposes important limitations on the power systems operation. The system should be
operated with an adequate voltage stability margin by the appropriate scheduling of reactive power
resources and voltage profile. If the required reactive power margin cannot be met by using available
reactive power resources and voltage control facilities, it may be necessary to limit power transfers
and/or to ask additional reactive power resources to provide voltage support at critical areas. The
knowledge of the reactive power reserve condition is of paramount importance in the system operation
and strongly affects the reliability of the power systems [71].
Voltage instability scenarios and correspondingly system security should be analyzed at various decision
stages from planning to real-time. T. Van Cutsem in [60] classifies these analysis methods in four
30

categories namely: contingency analysis, loadability limit determination, determination of security

limits, and preventive and corrective control.
Contingency analysis aims at analyzing the system response on a particular operating point to credible
contingencies that may lead to instability or voltage collapse. The system should be operated such that
to survive the credible contingencies by providing appropriate pre- and post-contingency controls. The
analysis can be accomplished by static methods based on load flow, modified load flow, multi-time scale
simulation, Quasy Steady State (QSS) simulation, and time-domain methods.
Loadability limit determines how far a system can move away from its operating point and still remain
in a stable state. This type of analysis typically utilizes load increase and/or generation rescheduling
which stress the system by increasing power transfer or by drawing RPR. The singularity of Jacobian
matrix and continuation load flow has been widely used in literature. In some applications loadability
limits can be obtained as the solution of an optimal power flow. Traditional PV and VQ curves are the
most well-known methods to distinguish the margin of the system operation point to instability. These
curves provide the results with acceptable accuracy and little computational effort since these analyses
are based on static approaches.
The PV curve plot the relationship between the active power transfer (P) and load bus voltages (V). It
demonstrates the active power stability margin. The relationship between Q and V can be used to show
the sensitivity and variation of bus voltages (V) with respect to reactive power injections or absorptions
(Q). It demonstrates the stability margin of the reactive power. The advantages of the latter method are
as follow [9]:

It could be more readily derived for non-radial system.

Better suited for examining the requirements for reactive power compensation.
Not only it identifies the stability limit, but also defines the minimum reactive power
requirement for stable operation.

The PV and the VQ curves are utilized to obtain the active and the reactive power margins for two-bus
test system as shown in figure 4-2.
E<0

V<

X
X
X

P L , QL

Bus 1

Bus 2

Qc

Figure 4-2: Two-bus test system.

In the given system the synchronous generator at bus 1 with voltage E<0 (E=1.1) feeds the load at bus 2
with active power PL (PL=2) and reactive power QL (QL=0.4). Three parallel transmission lines, each one
with inductance X (X=0.3), connect the generation bus to the consumption bus.
The PV and the VQ curves are calculated for this system in pre-contingency, post-contingency #1
(outage of one transmission line), and post-contingency #2 (outage of two transmission lines) and are
shown in figures 4-3 and 4-4, respectively.
The point on the nose-curve where the maximum power occurs is called the critical point and in
literature is often considered to be the voltage stability limit.
In figure 4-3, the active power margin to the voltage collapse is the distance of the operating point
(black circle) to the nose-point on the PV curve. This margin for the post-contingency #1 (1) decreases
comparing with the pre-contingency (0) due to the loss of one transmission line (1<0). The
contingency #2 has a negative active power margin (2) with respect to the current operating point
which means voltage instability. The system operator can restore the system to voltage stable area by
applying fast enough corrective countermeasures like shedding more than 2 MW of the loads.

31

As shown in figure 4-2, a fictitious reactive power injection (QC) is added to bus 2 to obtain the VQ
curves. The VQ curves for the three aforementioned scenarios and the corresponding reactive power
margin are depicted in figure 4-4. The black circles show the operating points of the system where the
fictitious injections are equal to zero (QC=0). The difference between the minimum of the VQ curve and
the operating point is defined as the reactive power margin at the bus, which is equal to the negative
value of the fictitious injected reactive power. The positive margins of the pre-contingency and postcontingency #1 are, Q0 and Q1, respectively. For the post-contingency #2, the reactive power margin (Q2)
became negative. This value (Q2) is the reactive power margin to operability. The reactive power margin
can be managed to keep the voltage secure by using different reactive resources.
Note that at the minimum of the VQ curve, the RPRs of depleted generators are the effective reserves
for the area and, thereby, determine the reactive margin in the area. The amount of effective RPR is a
key index in voltage stability assessment [72]. Developing of VQ curves is recommended as an
alternative method for time-domain or dynamic simulations to identify the appropriate RPR.
1

V/E

tan()=0.2
0

Secure
Voltage Zone

Margin from the

critical voltage

pre-contingency

post-contingency
#1
post-contingency
#2

0
0

PL
Figure 4-3: The PV curves of the two-bus test system.

(QCX)/E2

Q is the Margin to Critical Point

i
Q > 0 : Positive Margin
i
Q < 0 : Negative Margin
i

post-contingency
#2

0
Q

post-contingency
#1

pre-contingency

1
Figure 4-4: The VQ curves of the two-bus test system.

32

V/E

The MW and MVAr margin to the critical point in the PV and the VQ curves can be used as stability
indices. Many voltage stability indices are derived based on the information obtained from the proximity
to voltage collapse. These indices can be classified into state-based indices and large-deviation indices
[30]. The state-based indices compare the current state of the system with a theoretically calculated
critical value. The calculation of these indices doesnt need to obtain the real critical point. For instance,
voltage drops, eigenvalues and singular values can be enumerated. The large deviation-based indices
are determined by tracking a certain parameter from the operating point to the voltage instability point.
These indices, like active and reactive power margins, are widely used in voltage stability analysis [73].
Practical indices should be observable and controllable parameters for operator and anticipate the
effects of contingencies. Moreover, composite system reliability can be evaluated by incorporating
voltage stability indices to OPF [74].
As mentioned before, OPF can be used to obtain the critical information of the PV and VQ curves, such
as active and reactive power margin to instability.
The security limit is defined as the maximum stress that the system can accept, taking into account
contingencies. There are two types of security limits for a given direction of the system stress and a list
of contingencies:
1) Post-contingency loadability limits (PCLL) indicate how far the system can be stressed after the
occurrence of each contingency. It provides a measure of the security margin left after a
contingency.
2) Secure operation limits (SOL) indicate how far the system can be stressed before any contingency
such that it will remain stable after the contingencies. Interpretation of SOL is easier since it
separately refers to pre-contingency parameters that operators can either observe or control.
Also, SOL makes a clear separation between the pre- and post-contingency actions.
A set of contingencies should be selected for these studies to avoid the huge computational effort of
analyzing all contingencies. Contingency filtering is a key process for the success of voltage security
studies [75].
A combination of SOL determination and contingency filtering procedure, namely binary search, is
widely used to determine the secure operation limit of the system [30]. Binary search determines the
system stress limit for a single contingency. Similarly, simultaneous binary search can be used when the
study is for several contingencies. The presented methods in [76] and [77] extrapolates this limit
through extracting more information from the simulated responses.
Preventive control actions are taken in pre-contingency situation to increase the security margin with
respect to a set of postulated contingencies. Corrective control actions are taken in a given postdisturbance configuration in order to restore system stability. Each one of the preventive and the
corrective controls can be identified from eigenvector or based on Optimal Power Flow (OPF).
Different analysis methods described above can be used in appropriate provision of voltage control for
planning, operational planning, and real-time time scales.

4-5- Two bus test case simulation

The system operator defines voltage and reactive power controllers set-points according to different
objectives such as minimization of reactive power injection, voltage profile deviation, transmission
losses and etc. However, it is necessary to keep the set points after the contingencies through
appropriate provision of voltage and reactive power controls. The different objectives of the system
operators would result to different amount of RPRs and consequently different security margins. Here,
the simple two bus test case, shown in figure 4-2, is considered to demonstrate the effect of the
different objectives on the amount of the RPR. In addition to the given data in figure 4-2, in this test case
for each line the unavailability is equal to 0.01, the resistance is equal to 0.03, and the shunt impedance
is 0.3.
33

The simulation is accomplished in two steps as described below:

Step 1 Solve an OPF based on the different objectives (Min. Qg , Min. (vi-1), Min Ploss). If there is a
solution for the OPF, the maximum amount of reactive power load up to voltage collapse point
is calculated in step 2. Otherwise, the minimum amount of required reactive power injection to
obtain a solution for the system is calculated in step 2. Moreover, in latter case appropriate EC,
such as active and reactive generation rescheduling and load shedding, is calculated by utilizing
an OPF to obtain operable point with minimum amount of corrective actions.
Step 2 Find the reactive power margin of the system according to the bottom of the VQ curve as
described in Fig 4-3. Here, an OPF is utilized to obtain amount of fictitious injected or absorbed
reactive power (Qc).
The obtained results are given in table 4-1. Technical GRPR, effective GRPR, and LRPR are calculated as
defined in section 4-1.
Table 4-1: The results of the two bus test case
Objective
VQ-curve info
Min. Qg

Contingency #0

Min. (vi-1)

Min Ploss

Pg0= 1.2144

Pg0=1.2154

Pg0= 1.2144

Qg0= 0.5301

Qg0= 0.5458

Qg0= 0.5301

V10= 1.0500 < 0

V10= 1.0206 < 0

V10= 1.0500 < 0

V20= 1.0120 < -0.1113

V20= 0.9807 < -0.1181

V20= 1.0120 < -0.1113

=2.81360.5301= 2.2835

= 2.81330.5458=2.2675

=2.81360.5301= 2.2835

=1.17810.5301= 0.6480

=1.1781 0.5458=0.6323

=1.17810.5301= 0.6480

Pg1= 1.2225

Pg1= 1.2234

Pg1= 1.2225

Qg1= 0.5612

Qg1= 0.5756

Qg1= 0.5612

V11= 1.0500 < 0

V11= 1.0332

V11= 1.0500 < 0

V21= 0.9896 < -0.1715

V21= 0.9707< -0.1777

V21= 0.9896 < -0.1715

Qgc= 1.1781
V1c= 1.0500 < 0

Prob=0.9697

Contingency #1

Pgc= 1.2225

V2c= 0.9500 < -0.1129

Qc= QLRPR= 0.5617

Pgc= 1.2267
Qgc= 0.8384
V1c= 1.0500 < 0

Prob =0.03
=2.81100.5612=2.2499

2.8107-0.5756=2.2351

=2.81100.5612=2.2499

=0.8384 0.5612=0.2772

=0.83840.5756=0.2628

=0.8384 0.5612=0.2772

V2c= 0.9500 < -0.1752

Qc= QLRPR = 0.2293

Pgc= 1.2479
Contingency #2
Prob =0.0003

Qgc= 0.5270
No convergence

No convergence

No convergence

V1c= 1.0500 < 0

V2c= 0.9500 < -0.3702
Qc= QLRPR = 0.1432

In the presented results, the following points should be noted:

a) Minimization of transmission losses (Min Ploss) is defined as minimization of Pg.

34

b) Maximum reactive power capability of generator

is defined based on the generators

capability curve by using this formula

. The generators capability

1, where Pg and Qg are active and reactive

curve is estimated with an ellipse with

power of generator, and PM and QM are maximum active and reactive power of generator
which are equal to
3.5,
3. The effect of generators voltage on its capability
curve is neglected in this formulation.
c) In this test case the minimization of active power losses (Min Ploss) is in the direction of line
reactive power flow minimization which means minimization of reactive power generation
(Min. Qg). It is the reason that the results of objectives (Min Ploss) and (Min. Qg) are exactly
the same.
In contingency #2, there is not sufficient RPR to obstacle against the voltage collapse. In this case the
obtained results in the last column of table 4-1 can be considered as an index which measures the
severity of instability. In order to obtain an operable point, emergency countermeasures (ECs) such as
active and reactive generation rescheduling and load shedding need to be utilized. Here, since there is
only one generator, the generation rescheduling is not possible and load shedding should be
implemented. For this purpose, an OPF is developed to obtain minimum amount of load shedding with
respect to the system voltage collapse. The results are given in table 4-2.
Table 4-2: The results of EC for the two bus test case.
Pgc = 0.9022
Contingency #2
Prob =0.0003

Qgc = 0.4508
V1c = 1.0500 < 0
V2c = 0.9500 < -0.2631
PLS = 0.3238

4-6- Provision of Voltage Control in Planning

In the planning perspective, the system operator has to evaluate reactive power margin requirement of
the future system and ensure the viability of voltage controls, including RPR and EC.
The RPR can be provisioned by the system reinforcements through construction of new generating
units, transmission lines, series and shunt compensation. A suitable placement of shunt capacitors can
free up spinning RPR in generators. Application of some devices and controllers can contribute to
voltage control such as the line drop compensation in AVR, control of generator step-up transformer,
and automatic shunt compensation switching. Moreover, the voltage control scheme and reactive
power management can be enhanced through implementing SVR and TVR.
The studies under the title of VAr planning are also included in the RPR provision. It is aimed at
minimizing the installation cost of additional reactive support necessary to maintain the system in a
secure manner. The planning priority is to minimize cost and also to minimize future operations costs
(Page-349) [78].
The increasing participation of the variable distributed generation resources in the power system
exacerbates the necessity of voltage and reactive power control. This condition heightens the need to
pay more attention to the issue of voltage and reactive power control provision to maintain the
reliability of the system.
In order to face the severe disturbances with low probability of occurrence, automatic curative actions
aimed at avoiding instability should be established. For this purpose, one TSO can implement system
protection schemes as EC, such as LTC emergency control and emergency load shedding.
Decisions are taken based on the cost of installation and the improved value of security.
35

4-7- Provision of Voltage Control in Operational Planning and Real-Time

The provision of the voltage control in operational planning and real-time operation is dealt with
different aspects in literatures. In this time horizon the main purpose is to best utilize the available
system components for the voltage control.
In order to define the set-points of the voltage controllers, different system operators utilize various
objectives such as maximization of generator RPR (whether technical or effective), minimization of
transmission losses, and minimization of voltage deviation.
The system voltage should be kept in the secure region after a contingency by using RPR and EC. The
available reactive resources and countermeasures accompany the system operator for the RPR. The SVR
and the TVR schemes can be utilized to enhance the RPR in the system. For the EC, the system operator
can count on all the available reactive resources and also the active power redispatch and as the last
resort, load shedding. According to the earlier mentioned classification, here the provision of the
voltage controls is separately investigated in the provision of the RPR and the EC.
4-7-1- RPR provision
The provision of the RPR can be investigated in the perspective of the load RPR (LRPR) or the generation
RPR (GRPR). More studies are performed around the former (reactive power support) rather than the
latter (efforts in the RPR assessment). In both of LRPR and GRPR, the RPR requirements have been
investigated in the context of voltage stability for enhancement of its margin. Most of the approaches
use static analysis but some references confronted with this topic with regard to the stability analysis
[72], [75].
In addition, the provision of RPR based on security constraints is widely proposed in literature to
improve voltage stability. Various formulation of Security Constraint Optimal Power Flow (SCOPF) is
employed to assess the RPRs with constraints on operation [75] and contingencies [79]. For this purpose
some references specifically use the term Voltage Stability Constrained Optimal Power Flow (VSCOPF)
[73]. The VSCOPF are divided in two classes. The first class aims at maintaining steady-state voltage
stability which can be used for corrective control in severe emergency states. The second one aims at
determining preventive control strategies in the normal state considering voltage stability. Some
references utilize time-domain simulation to check the voltage stability in the preventive controls [80].
Small Perturbation Stability Constrained OPF (SSC-OPF) proposed in [81] include a stability index in the
OPF algorithm.
In the following section, different literatures around the RPR provision are surveyed based on the
perspective of the load RPR and the generation RPR.
a) Load RPR
Reference [82] defines a reactive reserve basin for each zone as the sum of the exhausted reactive
reserves at the minimum of the VQ curve. After a disturbance, the remained percentage of the basin
reactive reserve is used as a measure of the proximity to voltage instability.
The reactive reserve based Contingency Constrained Optimal Power Flow (RCCOPF) presented in [79]
aims at enhancing the voltage stability margin (VSM) of the interface flow. This method solves
preventive control in the normal state concerning VSM of post-contingency state by using a
decomposition method. Active power margin of post-contingent states are determined with modified
continuation power flow (MCPF) and an OPF is performed for preventive control. In fact, the RCCOPF
uses reactive power dispatch as a control mean to relieve reactive power generation and increase the
effective RPRs of the chosen generators.
The proposed RPR management in [72], manages RPRs in critical areas based on the OPF and as a result
it improves the voltage stability. A two-level benders decomposition is used for base case and a set of
stressed cases sub-problems. In the base case, the RPR is maximized while the transmission losses is
minimized. The optimization in the stressed cases deals with minimization of the fictitious reactive
36

power injections. The optimization procedure in this paper focuses on reactive reserve margin
optimization instead of reactive power rescheduling. In the management scheme, the participation
factors of the involved generators are determined based on the VQ curves.
b) Generation RPR
As described in section 4-1, the generators RPR (GRPR) can be classified into technical GRPR and
effective GRPR. Many studies in this area utilize the technical GRPR since it can be calculated easily
regardless stability analysis [83], [84], [85], [86], [87], and [88]. However, in [89] it is shown that the
effective RPR not only depends on the generators capability curve, but also on the network
characteristics that play an important role in generators RPRs. That means maximization of generators
RPR could not demonstrate the effective RPR all the times. Moreover, active and reactive power limits
of generators are linked together according to their capability curve. This dependency entails that the
RPR of a generator depends on the active power reserves provided by the generator. The effective RPR
for a bus or an area is determined in [90] as the weighted some of the individual RPRs of generators at
the minimum of the VQ curve. The weights are calculated based on sensitivities of generator reactive
outputs to reactive loads.
In [91], correlative relationship between GRPR and system voltage stability margins (VSM) is
investigated and a method for on-line voltage stability monitoring is proposed. Nonlinearity relationship
between GRPR and both VSM and voltage violations is investigated in [92].
The TSOs is responsible to provide and to coordinate the RPR requirements. In one hand, inappropriate
RPR provision treats the security of the system, and on the other hand, devoting large amount of RPRs
increase the operating cost of the system. In order to maximize the efficient use of assets, the minimum
amount and optimal location of required RPR should be well determined.
The proposed approach in [93], determines the minimal RPR to face a contingency, while stressing the
system in its pre-contingency state, until reaching an unacceptable post-contingency response.
In [75] a two-step approach is proposed to assess the required RPR with respect to operating constraints
and voltage stability for a set of assumed operating scenarios. At the first step a Security Constraint
Optimal Power Flow (SCOPF) determines the minimum overall needed RPR of generators such that the
system withstands any postulated scenario. In the second step additional RPR is determined to ensure
voltage stability of scenarios, whenever the obtained RPRs by SCOPF are insufficient to confront with
dynamic system behavior.

Application of the aforementioned reactive power rescheduling and RPR management methods can be
proposed as the objective of SVR and TVR, to increase voltage stability, and active and reactive power
margins.
4-7-2- EC provision
Given the cost of corrective countermeasures and the low probability of contingency occurrences, it
would be desirable to resort to post-disturbance controls. However, an essential characteristic in these
actions is the implementation time needed since the speed of response is an issue for long-term voltage
stability.
The system operator must be able to recognize voltage instability and take appropriate remedial actions
such as voltage and power transfer controls, generation rescheduling, and as a last resort, load
curtailment. One may try to find suitable remedial actions to restore the system to the voltage secure
zone. Branch and generator participation factors are among the proposed methods to determine the
appropriate remedial actions. The branch participation factor indicates which branches consume the
most reactive power in response to an incremental change in reactive load. It would be useful for
identifying remedial measures to alleviate the voltage stability problems and also for contingency
selection. Similarly, the generator participation factor shows which generators supply the most reactive
37

power for a given reactive power variation. It provides important information regarding proper
distribution of RPRs among all the machines in order to maintain an adequate voltage stability margin.
Moreover, the proposed method in [58] utilizes the analytical methods to determine the corrective
actions through a linear approximation of the feasible set boundary. The analytical methods use
continuous models of the system components, and describe the system with a set of differentialalgebraic equations.
In addition, specific OPFs can be developed to find appropriate remedial action. For this purpose,
literatures use the term Corrective Security Constraint Optimal Power Flow (CSCOPF) [94], [95]. The
constraints of the CSCOPF formulation is improved in [94] by incorporation of the system dynamics to
ensure existence and viability of the post-contingency short-term equilibrium of the system until
corrective actions can start. The proposed optimal load shedding in [96] utilizes time-domain simulation
to check voltage stability in corrective controls.
4-7-3- Preventive and Corrective control actions
Some references defined a coupling security-constraint optimal power flow to determine an optimal
combination of preventive and corrective controls ensuring long-term voltage stability.
An optimal combination of preventive and corrective control actions is determined in [97] while
ensuring dynamic transition of the system. The result of quasi steady state simulation is used iteratively
to modify the combination of both control actions. Post-disturbance generations rescheduling is applied
as corrective countermeasure against long-term voltage instability.

4-8- Conclusion
Although the main concern of this chapter is provision of preventive and corrective voltage and reactive
power control in Single Area Power System (SAPS), this concern remains important in Multi-Area Power
system (MAPS). A coordinated provision of reactive power supply and voltage control with system wide
voltage scheduling and emergency response capabilities is required for the reliable system operation.
The TSO dispatches the reactive resources and provide adequate reactive supply and voltage control.
The required RPR and EC must be identified and maintained by each TSO within its own voltage control
area. However, the interaction of different control levels and actions between neighboring TSOs should
be investigated for the security of the interconnected system.

5- Voltage Control in Multi-Area Power System

Reactive power support and reserve with emergency countermeasures for voltage control are typically
provided locally since reactive power cannot be transmitted over long distances efficiently. Thus, there
are fewer opportunities in exchanging these services regionally. As a result, each TSO has defined a
number of zones within its own control area for a locally voltage control. In order to obtain an optimum
and secure operation, the operator of each zonal voltage control may need to have access to some
information of the other controlled zones. Hence, a supervision of the TSO, as a regional control, is
required for coordination among the controlled zones.
In recent two decades power system were subjected to widespread blackouts where insufficient
reactive power support was a major factor responsible of such outages. The lack of reactive power
reserves response to the increased reactive power demand in contingencies, can lead to the operation
of the protection system and also cascading overloads of generators. Also inappropriate emergency
countermeasures can propagate the disturbances in the system and increase the scale of problem.
Therefore, the idea of centrally coordinated voltage control and hierarchical voltage control is
introduced to obstacle the voltage control concerns. As a consequence, different TSOs designed and
38

applied various coordinated voltage control schemes to improve the system voltage control against
voltage collapse.
In addition, evolution of the power systems toward the continental interconnection, obviously,
demonstrates the intention of TSOs to be operated as a member of a Multi-Area Power System (MAPS).
The interconnected networks in continental Europe (ENTSOE) and North America (NERC) which involve
multi independent TSOs are examples of such MAPSs. Improvement of system security and economic
efficiency within the entire system are the main motivations for operation as a member of MAPS. Higher
security margins can be obtained because of shared active power reserves within MAPS. However,
interconnections may also have some drawbacks when the system is operated by non-coordinated
TSOs. The non-coordinated operation situation in MAPS can be the result of each TSOs intention to not
disclose its control action to other TSOs. Moreover, operation of MAPS under supervision of a super TSO
with higher level of control would be more expensive and requires more communications.
Since voltage control primarily is of regional concern, the ENTSO-E operational handbook [25]
recommends that interconnected TSOs should coordinate their actions and agree on an acceptable
voltage range at each interconnection link, which can be roughly formulated as a zero reactive power
flow at every interconnection link [85]. However, No reactive power flow at the interconnection links is
difficult to apply. Observations show that reactive power flows are rarely negligible at the
interconnection links [98].
As described above, although voltage control is mainly a local issue, voltage control problem can spread
in the interconnected systems and increase the scale of blackouts and even affect on the intact areas.
For this purpose, each TSO utilizes specific hierarchical voltage control or centralized one, within its own
control area. Since the voltage control approaches and practices are different from one TSO to another,
it is necessary to study the interaction of different control levels of neighboring TSOs with each other.
As a result, as shown in figure 5-1, in MAPS each TSO requires to take into account interactions of its
own centralized voltage control or TVR with the ones in the neighboring TSOs (type I). In addition, if the
neighboring TSOs utilize a hierarchical voltage control, the counteractions of the corresponding SVRs at
the borders need to be investigated (type II). These higher levels of coordination are necessary in MAPS
since the TSOs accessibility to the neighboring TSOs information is limited. There are little relevant
works for required additional coordination between TSOs [98], and so voltage and reactive power
control in MAPS need more attention by using distributed or decentralized control schemes.

Figure 5-1. Possible interactions between different levels of voltage control in MAPS.

39

In fact, when a contingency happens in an area, an automatic and non-coordinated response of voltage
controller (PVR) by the generators electrically closer to the disturbance1 may lead to unacceptable
reactive power flow or voltage level in its own control area or even in neighboring areas. That means
some TSOs dont provide sufficient MVar support. Due to locally provision characteristic of voltage
control, normally in real time, each TSO is aware of this situation in its own control area. The added
value of a wide coordinated control is to propose a global optimum remedial action to restore the
system to a secure state. Otherwise, the system would be operated in a non-optimal state which means
less security margin. This situation requires a higher level control to achieve the global optimum
operation point. The experience of CORESO2 in ENTSO-E demonstrated such operating situation may
happen and a coordinated control action can improve the taken remedial action.
In the case of sever contingency in presence of interregional voltage management in the system, the
TSOs can take the advantages of the available voltage controls in the neighboring TSOs to counteract
with their problem or to limit the extent of the problem in the MAPS. Similarly, one TSO may utilize its
own control facilities to help the neighboring TSO or to avoid the effects of external contingency in its
own control area.
The following uncertainties made the wide coordinated control system necessary [8]:

Neighboring system voltage profile for the operating condition.

Variations on neighboring systems generation dispatch.
Large and variable reactive exchanges with neighboring systems.
Restrictive reactive power constraints on neighboring system generators.
Outages not routinely studied on neighboring systems.

Combination of both preventive and corrective control actions including amount and place of RPRs and
emergency controls should be determined to respond to the system sudden changes. The optimization
of a system wide coordination is proposed as important measure for sharing reactive reserves when
some control limits are reached [32]. If voltage constraints begin to be approached, a wide voltage
scheduling regarding the effect of neighboring regions becomes significant.
However, the implementation of a centralized control in MAPS is not possible since not only the TSOs
dont intend to reveal their operational information for the other TSOs but also implementation of a
wide area control scheme would be technically more expensive and requires more communication.
Therefore, distributed [99] or decentralized [98] control manners are needed to be considered for this
purpose, which has got little attention up to now. These control schemes are difficult to be effectively
implemented or might achieve suboptimal performance.
In [99] distributed voltage control and Model Predictive Control (MPC) technique are applied for
emergency voltage control to coordinate the control actions among the various grids while each
operator preserves its own sensitive local system data. The proposed centralized control scheme is
solved in a distributed fashion through Lagrangian decomposition method. Although the control
problem is global, only local information is employed to achieve the overall optimum control. All AVR
references and load shedding at some buses are assumed as available controls. At each iteration of
solution procedure, the information which are sent out to the external control centers are the obtained
local optimal cost and the calculated interface bus voltages. It is shown that in some operation
conditions; when a contingency happens within an area, the control actions also must be taken in other
areas to restore the grid to safe operation state according to the chosen globally optimal criterion. It
should be noted that the proposed method in [99] cannot consider the different TSOs with different
objectives, since a centralized control scheme has been taken.
1

The required reactive power will be produced by the generators electrically closer to the disturbance and hence the remained
reserve may be unevenly distributed.
2
CORESO (Coordination of Electricity System Operators) is a centralized control center to coordinate control actions and
strength operational security in the Central Western Europe.

40

The decentralized voltage control in MAPS is studied in [100], [84], [85], [86], [87] and [88] with different
approaches such as neighboring network equivalent [84], fairness of different TSOs objective [86], [87],
and advantages and disadvantages of centralized and decentralized voltage controls [85], [88]. A new
layer of hierarchical control to coordinate long-term control actions over several control regions in
normal operating conditions is proposed in [98]. The corresponding time horizon of the proposed MAPS
voltage control in this work is shown in figure 5-2, in comparison to the different levels of hierarchical
voltage control (PVR, SVR, and TVR). These works study reactive power scheduling using multi-objective
optimization for minimizing reactive power support and active power losses. The TSOs different
strategies in voltage control are considered with different combination of this multi-objective function.

International Scale
Regional or
National Scale
Zonal Scale
By each generator
or compensator

Figure 5-2: Time-space delineation of a four-layer hierarchical voltage control scheme [98].

The reference [100] compares two strategies for accessing to the information of the neighboring TSOs.
In the first strategy, two TSOs have access to all the information of each other except their controllers
value in the next step (It is called wide observation). In the second strategy each TSO just knows about
its own control area and the access to the information of the neighboring TSO is limited to the borders.
It should be mentioned that the latter strategy is closer to the present situation of power system. It is
shown that the decentralized control of the system with wide observation leads to the worst result since
each area behaves in greedy way and wants to import reactive power from the other area resources.
However, the decentralized voltage control with limited access to the neighboring TSOs information can
be considered in MAPS studies. In [84], [85], [86], [87] and [88] the centralized voltage control which
optimizes a unique objective over the entire system is assumed as the utopian optimum. The result of
the decentralized voltage control is evaluated based on the distance to this utopian optimum. The
difference between global optimization results and decentralized optimization results is the additional
cost that should be paid for decentralized control. In the decentralized manner each TSO solves its own
objective function considering its own network constraints and imposed constraints of the external
networks. Then all TSOs apply the solution to their own systems as a part of interconnected system and
each TSO measures parameters for external network equivalents. If the control values dont comply the
constraints of the entire network, faster voltage control loop will change the operation setting while
slow devices are remained unchanged. The fast voltage control actions use available reactive power
reserves. The dynamics of the proposed method strongly depend on the number of interconnections
and the size of the power system.
In [84] the neighboring TSOs is modeled with a constant PQ injection corresponding to the value of flows
outside of area. [85] compares the different models of neighboring areas such as PV, PQ, Thevenin
equivalent, and more advanced models like REI equivalent and non-reduced power system equivalent.
These equivalents replace the power system beyond an interconnection of a TSO with a single
interconnection. The parameters of different models for neighboring TSOs are fitted by using different
least square based methods according to the past and current observations. This method is only based
on local voltage and current measurements in the interconnections and doesnt need any coordination
between the different TSOs. It is shown that PQ equivalent could achieve near optimal performance. In
41

addition, PQ equivalent provides the possibility to consider the exchange of active and reactive power
between areas according to bilateral contracts.
The concept of fairness is introduced in [86] and [87] to evaluate a compromise between different
objectives of TSOs so that each TSO is less displeased. It should be noted that the system security is not
taken into account by the mentioned works in [84], [85], [86], [87], [88].
In conclusion, in addition to the aforementioned studies in this area, it is still needed to study the
current practices in voltage and reactive power control from security point of view. Possible problems in
current practices which can threat the security of MAPS would be investigated. Furthermore, it is
necessary to enhance the voltage and reactive power control methods in SAPS based on security
criteria. According to the literature some aspects of this topic like considering both of RPR (particularly
generator RPR) and corrective countermeasures, in addition to security based voltage control provision
need more attention. For this purpose, it is important to propose an approach to enhance the voltage
and reactive power control in SAPS by considering both preventive and corrective actions. In order to
enhance the security of voltage control in MAPS the proposed methodology for SAPS could be extended
for MAPS. For this purpose, a decentralized or distributed control scheme is required to manage
appropriate voltage and reactive power control with respect to the security of MAPS.

42

6- Bibliography
[1] Ancillary Services: an overview of International Practices.: Cigre Working Group C5.06, October
2010.
[2] K. Bhattacharya, M. H. J. Bollen, and J. E. Daalder, Operation of Restructured Power Systems.:
Kluwers Power Electronics and Power Systems Series, 2001.
[3] Y. Rebours, "A Comprehensive Assessment of Markets for Frequency and Voltage Control Ancillary
Services," University of Manchester, PhD Thesis 2008.
[4] F. Dong, B. H. Chowdhury, M. L. Crow, and L. Acar, "Mid-term voltage stability study using the
quasi-steady state analysis method," in IEEE PES General Meeting, Toronto, 2003, pp. 2646-2651.
[5] T. Kapetanovic, B. M. Buchholz, B. Buchholz, and V. Buehner, "Provision of ancillary services by
dispersed generation and demand side response needs, barriers and solutions," Elektrotechnik &
Informationstechnik, vol. 125, no. 12, pp. 452459, Decembre 2008.
[6] A. Gomez-Exposito, A. J. Conejo, and C. Canizares, Electric Energy Systems Analysis and Operation.:
Taylor &. Francis, 2009.
[7] (2009) GMS Regional Trading and Environmentally Sustainable Development of Electricity
Infrastructures. [Online]. http://www.gms-powertrade.net/
[8] Reactive Reserve Working Group, "Guide to WECC/NERC Planning Standards I.D: Voltage Support
and Reactive Power," Western Electricity Coordination Council, 2006.
[9] P. Kundur, Power System Stability and Control.: McGraw-Hilll, 1994.
[10] A. Chakrabarti and S. Halder, Power System Analysis: Operation And Control.: Prentice-Hall of India,
2006.
[11] A. Rattanavit, "Field experience in voltage & reactive power based on EGAT's system
configuration," in International Conference on Energy Management and Power Delivery, 1995, pp.
354 - 357.
[12] C. W. Taylor, Power System Voltage Stability.: McGraw-Hill, 1994.
[13] J. Paserba, V. Ajjarapu, G. Andersson, A. Bose, C. Canizares, N. Hatziargyriou, D. Hill, A. Stankovic,
C. Taylor, T. Van Cutsem, V. Vittal P. Kundur, "Definition and Classification of Power System
Stability," IEEE Trans. on Power Syst., vol. 19, no. 2, pp. 1387-1401, May 2004.
[14] "DEFINITION OF A SET OF REQUIREMENTS TO GENERATING UNITS," ENTSOE, TECHNICAL PAPER
2008.
[15] C. Vournas, "Technical Summary on the Athens and Southern Greece Blackout of July 12, 2004,"
National Technical University of Athens, 2004.
[16] Learning from the blackouts, Transmission system security in competitive electricity market.:
International energy agency, 2005.
[17] S. Larsson and A. Danell, "The black-out in southern Sweden and eastern Denmark, September 23,
2003.," in PSCE , 2006, pp. 309-313.
[18] G. Andersson et al., "Causes of the 2003 Major Grid Blackouts in North America and Europe, and
Recommended Means to Improve System Dynamic Performance," IEEE Trans. on Power Syst., vol.
20, no. 4, pp. 1922 - 1928, Nov. 2005.

43

[19] Federal Energy Regulatory Commission, "Principles for Effecient and Reliable Reactive Power
Supply and Consumption," Federal Energy Regulatory Commission , Washington, Staff Report
AD05-1-000, 2005.
[20] S. C. Savulescu, Real-time Stability in Power Systems: Techniques for Early Detection of the Risk of
Blackout.: Springer, 2006.
[21] T. Amraee, A. M. Ranjbar, and R. Feuillet, "Immune-based selection of pilot nodes for secondary
voltage control," Euro. Trans. Electr. Power, vol. 20, pp. 938951, April 2010.
[22] S. Corsi, M. Pozzi, C. Sabelli, and A. Serrani, "The Coordinated Automatic Voltage Control of the
Italian Transmission GridPart I: Reasons of the Choice and Overview of the Consolidated
Hierarchical System," IEEE Trans. on Power Syst., vol. 19, no. 4, pp. 1723-1732, November 2004.
[23] H. Vu, P. Pruvot, C. Launay, and Y. Harmand, "An improved voltage control large-scale power
system," IEEE Transactions on Power Systems, vol. 11, no. 3, pp. 1295- 1303, August 1996.
[24] Elia. (2007) Voltage support - Produce or absorb reactive power to stabilize the voltage. [Online].
http://www.elia.be
[25] (2009, March) ENTSO-E operation hand book. [Online]. https://www.entsoe.eu
[26] Y. G. Rebours, D. S. Kirschen, M. Trotignon, and S. Rossignol, "A Survey of Frequency and Voltage
Control Ancillary ServicesPart I: Technical Features," IEEE Trans. on Power Syst., vol. 22, no. 1,
pp. 350-357, FEBRUARY 2007.
[27] Y. G. Rebours, D. S. Kirschen, M. Trotignon, and S. Rossignol, "A survey of frequency and voltage
control ancillary services - Part II: Economic features," IEEE Trans. on Power Syst., vol. 22, no. 1, pp.
358-366, February 2007.
[28] B. Hu, C. A. Caizares, and M. Liu, "Secondary and Tertiary Voltage Regulation Based on Optimal
Power Flows," in IREP Symposium- Bulk Power System Dynamics and Control - VII, Buzios, RJ, Brazil,
2010.
[29] S. Corsi, "Wide Area Voltage Regulation & Protection," in IEEE Bucharest PowerTech , Bucharest,
2009, pp. 1 - 7.
[30] T. Van Cutsem and C. Vournas, Voltage Stability of Electric Power Systems.: Kluwer Academic
Publisher, 1998.
[31] (2011, March ) ENTSO-E Draft Requirements for Grid Connection Applicable to all Generators.
[Online].
https://www.entsoe.eu/fileadmin/user_upload/_library/news/110322_Pilot_Network_Code_Conn
ections.pdf
[32] M. D. Ilic, X. Liu, G. Leung, and M. Athans, "Improved Secondary and New Tertiary Voltage
Control," IEEE Trans. on Power Syst., vol. 10, no. 4, pp. 1851-1861, November 1995.
[33] D. S. Popovic, "Impact of secondary voltage control on voltage stability," Electric Power Systems
Research, vol. 40, no. 1, pp. 51-62, January 1997.
[34] J.C. Sabonnadire, J.Y. Lost, J.P. Paul P. Lagonotte, "STRUCTURAL ANALYSIS OF THE ELECTRICAL
SYSTEM : APPLICATION TO SECONDARY VOLTAGE CONTROL IN FRANCE," IEEE Trans. on Power
Syst., vol. 4, no. 2, pp. 479-486, May 1989.
[35] Gestlonnaair du Rseau de Transport d'Electricit, "Rfrentiel Technique, Chapitre 4 Contribution des utilisateurs aux performances du RPT," RTE, 2006.

44

[36] I. Siviero, "The need to coordinate generation and transmission planning and to ensure a secure
and efficient reactive power provision: two key aspects of the restructured electricity industry,"
Universit degli Studi di Pavia, Doctoral Thesis 2009/2010.
[37] C. A. Canizares, C. Cavallo, M. Pozzi, and S. Corsi, "Comparing secondary voltage regulation and
shunt compensation for improving voltage stability and transfer capability in the Italian power
system," Electric Power Systems Research, vol. 73, pp. 6776, 2005.
[38] S. Corsi, M. Pozzi, M. Sforna, and G. DellOlio, "The Coordinated Automatic Voltage Control of the
Italian Transmission GridPart II: Control Apparatuses and Field Performance of the Consolidated
Hierarchical System," IEEE Trans. on Power Syst., vol. 19, no. 4, pp. 1733-1741, November 2004.
[39] J. Van Hecke, N. Janssens, J. Deuse, and F. Promel, "Coordinatedd Voltage Control Experience in
Belgium," in CIGRE, Paris, 2000.
[40] N. Janssens, "Tertiary and Secondary Voltage Control for the Belgian HV System," in Proceedings of
the IEE Colloquium on International Practices in Reactive Power Control, 1993, pp. 81-84.
[41] M. Geidl, "Implementation of Coordinated Voltage Control for the Swiss Transmission System," in
MELECON - 15th IEEE Mediterranean Electrotechnical Conference, Valletta, 2010, pp. 230 - 236.
[42] (2010) swissgrid website. [Online]. www.swissgrid.ch
[43] swissgrid, "Industry Recommendation for the Swiss Electricity Market - Transmission Code 2010,"
swissgrid, TC 2010, 2009.
[44] M. Zerva, "Voltage Stability Assessment of the Swiss Power Transmission System," Swiss Federal
Institute of Technology (ETH), Zurich, Master Thesis 2010.
[45] T. Gomez, D. Soler P. Frias, "Voltage control and reactive power support in the Spanish
transmission network," in IEEE Mediterranean Electrotechnical Conference, MELECON 2006,
Malaga , 2006, pp. 916 - 919.
[46] E. L. Miguelez, F. M. E. Cerezo, and L. R. Rodriguez, "On the Assignment of Voltage Control Ancillary
Service of Generators in Spain," IEEE Trans. on Power Syst., vol. 22, no. 1, pp. 367 - 375, February
2007.
[47] "Resolucin de 30 de juliode 1998, de la Secretaria de Estado de Energa y Recursos Minerales,por
la que se aprueba un conjunto de procedimientos de carcter tcnico e instrumental necesarios
para realizar la adecuada gestin tcnica del sistema elctrico," Ministerio de industria y energa
Spain, 20053, 1998.
[48] J. L. Sancha, J. L. Femhdez, F. Martinez, and C. Sall, "Spanish Practices in Reactive Power
Management and Voltage Control," in IEE Colloquium on International Practices in Reactive Power
Control, London, 1993, pp. 31-34.
[49] H. Berndt et al., "Network and System Rules of the German Transmission System Operators,"
Verband der Netzbetreiber - VDN e.V. beim VDEW, TransmissionCode 2007 2007.
[50] J. Zhong and K. Bhattacharya, "Reactive Power Management in Deregulated Electricity Markets- A
Review," in IEEE Power Engineering Society Winter Meeting, 2002, pp. 1287 - 1292.
[51] "Nordic Grid Code 2007 (Nordic collection of rules)," Nordel, 2007.
[52] Tennet TSO GmbH. (2010, October ) Grid Code - extra high voltage -. [Online].
http://www.tennettso.de
[53] "Evaluation of Reactive Power Planning and Voltage Control Practices in Response to NERC
Blackout Recommendation 7a," Transmission Issues Subcommittee of the NERC Planning
45

Committee, 2005.
[54] "Measurement of Ancillary Services from Power Plants Operating Reserve Spinning and
Supplemental and Reactive Power Supply from Generation Resources," EPRI , Technical Report
2000.
[55] "Standard VAR-001-1a Voltage and Reactive Control," NERC, 2008.
[56] I. El-Samahy, K. Bhattacharya, and C. A. Caizares, "A Unified Framework for Reactive Power
Management in Deregulated Electricity Markets," in PSCE, IEEE PES, Atlanta, 2006, pp. 901-907.
[57] G. Y. Cao and D. J. Hill, "Power system voltage small-disturbance stability studies based on the
power flow equation," IET Gener. Transm. Distrib., vol. 4, no. 7, pp. 873882, January 2010.
[58] T. Van Cutsem, Y. Jacquemart, J. N. Marquet, and P. Pruvot, "A comprehensive analysis of midterm voltage st ability," IEEE Trans. on Power Syst., vol. 10, no. 3, pp. 1137-1182, August 1995.
[59] C. D. Vournas T. Van Cutsem, "Voltage stability analysis in transient and mid-term time scales,"
IEEE Trans. on Power Syst., vol. 11, no. 1, pp. 146-154, February 1996.
[60] T. Van Cutsem, "Voltage Instability: Phenomena, Counermeasures, and Analysis Methods,"
Proceedings of the IEEE, vol. 88, no. 2, pp. 208-227, February 2000.
[61] T. Van Cutsem and R. Mailhot, "Validation of a fast voltage stability analysis method on the HydroQuebec system," IEEE Trans. on Power Syst., vol. 12, no. 1, pp. 282 - 292, February 1997.
[62] A. R. Nunes, D. M. Falcao T. M. L. Assis, "Mid and Long-Term Voltage Stability Assessment using
Neural Networks and Quasi-Steady-State Simulation," in Large Engineering Systems Conference on
Power Engineering, Montreal, 2007 , pp. 213 - 217.
[63] D. P. Kothari and J. S. Dhillon, Power System Optimization., 2006.
[64] "Ancillary Services Measurement Handbook," EPRI, Technical Report 2001.
[65] C. D. Vournas T. Van Cutsem, "Emergency Voltage Stability Controls: an Overview ," in Power
Engineering Society General Meeting IEEE , 2007, pp. 1-10.
[66] D. Lefebvre, C. Moors, and T. Van Cutsem, "Design of an undervoltage load shedding scheme for
the Hydro-Qu6bec system," in Proceedings IEEE PES General Meeting, Toronto, 2003.
[67] D. Lefebvre, S. Bernard, and T. Van Cutsem, "Undervoltage load shedding scheme for the HydroQubec system," in Power Engineering Society General Meeting IEEE, 2004, pp. 1619 - 1624.
[68] B. H. Chowdhury and C. W. Taylor, "Voltage stability analysis: V-Q power flow simulation versus
dynamic simulation," IEEE Trans. on Power Syst., vol. 15, no. 4, pp. 1354-1359, Nov. 2000.
[69] P. W. Sauer and P. A. Ruiz, "Reactive Power Reserve Issues," in Proc. 2006 North American Power
Symposium, Carbondale, 2006, pp. 539-545.
[70] Hydro Power and other Renewable Energies Study Committee, "Voltage control Reactive
energy," The Union of the Electricity Industry - EURELECTRIC, 03005Ren9712, 1997.
[71] N. Flatab, R. Ognedal, and T. Carlsen, "Voltage stability condition in a power transmission system
calculated by sensitivity methods," IEEE Trans. Power Syst., vol. 5, no. 4, pp. 12861293, Nov 1990.
[72] D. Feng, B. H. Chowdhury, M. L. Crow, and L. Acar, "Improving voltage stability by reactive power
reserve management," IEEE Trans. on Power Syst., vol. 20, no. 1, pp. 338 - 345, February 2005.
[73] H. Song, B. Lee, and Y. H. Moon, "Reactive Optimal Power Flow incorporating margin enhancement
constraints with non linear interior point method," IEE Proceedings on Generation Transmission
46

and Distribution, vol. 152, no. 6, pp. 961-968, November 2005.

[74] G. M. Huang and N. C. Nair, "Voltage Stability Constrained Load Curtailment Procedure to Evaluate
Power System Reliability Measures," , 2002, pp. 761-765.
[75] F. Capitanescu, "Assessing reactive power reserves with respect to operating constraints and
voltage stability," to be appeared in IEEE Trans. on Power Syst.
[76] C. D. Vournas et al., "On-Line Voltage Security Assessment of the Hellenic Interconnected System,"
in PowerTech, Bolognia, 2003.
[77] C. Moisse, R. Mailhot T. Van Cutsem, "Determination of secure operating limits with respect to
voltage collapse," IEEE Trans. on Power Syst., vol. 14, no. 1, pp. 327-335, February 1999.
[78] J. A. Momoh, Electric Power System Applications of Optimization. New York, Basel: Marceld
Ekkerin, 2001.
[79] H. Song, B. Lee, S. H. Kwon, and V. Ajjarapu, "Reactive Reserve-Based Contingency Constrained
Optimal Power Flow (RCCOPF) for Enhancement of Voltage Stability Margins," IEEE Trans. on
Power Syst., vol. 18, no. 4, pp. 1538-1546, November 2003.
[80] F. Capitanescu and T. Van Cutsem, "Preventive control of voltage security margins : a multicontingency sensitivity-based approach," IEEE Trans. on Power Syst., vol. 17, no. 2, pp. 358-364,
May 2002.
[81] S. K. M. Kodsi, "Accounting for the Effects of Power System Controllers and Stability on Power
Dispatch and Electricity Market Prices," University of Waterloo, Waterloo, Ontario, Canada, PhD
thesis 2005.
[82] R. A. Schlueter, "A Voltage Stability Security Assessment Method," IEEE Trans. on Power Syst., vol.
13, no. 4, pp. 1423 - 1438, November 1998.
[83] M. Zima and D. Ernst, "On Multi-Area Control in Electric Power Systems," in PSCC, 2005.
[84] Y. Phulpin, M. Begovic, M. Petit, J. B. Heyberger, and D. Ernst, "Evaluation of network equivalents
for voltage optimization in multi-area power systems," IEEE Trans. on Power Syst., vol. 24, no. 2,
pp. 729-743, May 2009.
[85] Y. Phulpin, M. Begovic, and D. Ernst, "Coordination of voltage control in a power system operated
by multiple transmission utilities," in REP Symposium, Buzios, 2010, pp. 1-8.
[86] Y. Phulpin, M. Begovic, M. Petit, and D. Ernst, "A fair method for centralized optimization of multi
TSO power system," Electrical Power and Energy Systems, vol. 31, no. 9, pp. 482-488, October
2009.
[87] Y. Phulpin, M. Begovic, M. Petit, and D. Ernst, "On the fairness of centralized decision making
strategies in multi TSO power systems," in PSCC008, Glasgow, 2008.
[88] Y. Phulpin, M. Begovic, M. Petit, and D. Ernst, "Decentralized reactive power dispatch for a time
varying multi TSO system," in Hawaii International Conference on Systems Science, 2009.
[89] P. W. Sauer P. A. Ruiz, "Reactive Power Reserve Issues," in Proc. 2006 North American Power
Symposium, Carbondale, 2006, pp. 539-545.
[90] C. W. Taylor and R. Ramanathan, "BPA Reactive Power Monitoring and Control following the
August 10, 1996 Power Failure," in Proc. VIth SEPOPE conference, Salvador (Brazil), 1998.
[91] L. Bao, Z. Huang, and W. Xu, "On-line Voltage Stability Monitoring Using Var Reserves," in IEEE
47

Power Engineering Society General Meeting, 2003, pp. 1754 - 1754.

[92] B. Leonardi and V. Ajjarapu, "Investigation of Various Generator Reactive Power Reserve (GRPR)
Definitions for Online Voltage Stability/Security Assessment," in Power and Energy Society General
Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, Pittsburgh, 2008, pp. 1 7.
[93] F. Capitanescu and T. Van Cutsem, "Evaluation of reactive power reserves with respect to
contingencies," in Bulk Power System Dynamics and Control V, Onomichi, 2001.
[94] F. Capitanescu and L. Wehenkel, "Improving the statement of the conrective security constraint
optimal power flow problem," IEEE Trans. on Power Syst., vol. 22, no. 2, pp. 887-889, May 2007.
[95] F. Capitanescu and L. Wehenkel, "A New Iterative Approach to the Corrective Security-Constrained
Optimal Power Flow Problem," IEEE Trans. on Power Syst., vol. 23, no. 4, pp. 1533-1541, November
2008.
[96] C. Moors and T. Van Cutsem, "Determination of optimal load shedding against voltage instability,"
in Proc. 13th PSCC Conf., Trondheim, 1999, pp. 9931000.
[97] F. Capitanescu, T. Van Cutsem, and L. Wehenkel, "Coupling Optimization and Dynamic Simulation
for Preventive-Corrective Control of Voltage Instability," IEEE Trans. on Power Syst., vol. 24, no. 2,
pp. 796-805, May 2009.
[98] Y. D. Phulpin, "Coordination of Reactive Power Scheduling in a Multi-Area Power System Operated
by Independent Utilities," Georgia Institute of Technology, Doctoral Thesis 2009.
[99] A. G. Beccuti, T. H. Demiray, G. Andersson, and M. Morar, "A Lagrangian Decomposition Algorithm
for Optimal Emergency Voltage Control," IEEE Trans. on Power Syst., vol. 25, no. 4, pp. 1769 - 1779,
November 2010.
[100 Y. Phulpin, M. Begovic, and M. Petit, "External Netwrok Modeling for MVAr Scheduling in Multi
] Area Power Systems," in PowerTech, Lausanne, 2007, pp. 1039-1043.
[101 D. Tziouvaras, E. O. Schweitzer, K. Martin A. Guzmn, "Local and Wide-Area Network Protection
] Systems Improve Power System Reliability," , 2004.
[102 L. M. Honorio, G. L. Torres, G. Lambert-Torres A. C. Z. Souza, "Increasing the loadability of power
] system through optimal-local control actions," IEEE Trans. Power Syst., vol. 19, no. 1, pp. 188194,
Feb 2004.
[103 N. Krantz M. Gustafsson, "Voltage Collapse in Power Systems, Analysis of Component Related
] Phenomena using a Power System Model," Chalmers University of Technology, 215L, 1995.
[104 S. Corsi, M. Pozzi, C. Sabelli, A. Serrani, "The Coordinated Automatic Voltage Control of the Italian
] Transmission GridPart I: Reasons of the Choice and Overview of the Consolidated Hierarchical
System," IEEE Trans. on Power Syst., vol. 19, no. 4, pp. 1723-1732, November 2004.
[105 D. J. Hill and G. Y. Cao, "Power system voltage small-disturbance stability studies based on the
] power flow equation," IET Gener. Transm. Distrib., vol. 4, no. 7, pp. 873882, January 2010.
[106 Y. Harmand, H. Vu, P. Pruvot, and C. Launay, "An improved voltage control large-scale power
] system," IEEE Transactions on Power Systems, vol. 11, no. 3, pp. 1295- 1303, August 1996.
[107 V. Ajjarapu and C. Christy, "The continuation power flow: A tool for steady state voltage stability
] analysis," IEEE Trans. Power Syst., vol. 7, no. 1, pp. 416423, Feb 1992.
[108 L. A. L. Zarate, C. A. Castro, J. L. M. Ramos, and E. R. Ramos, "Fast computation of voltage stability
security margins using nonlinear programming techniques," IEEE Trans Power Syst, vol. 21, no. 1,
48

] pp. 1927, Feb 2006.

[109 G. Heffner, C. Goldman, M. Kintner-Meyer, and B. Kirby, "Loads Providing Ancillary Services:
] Review of International Experience," Technical Appendix: Market Descriptions 2007.
[110 F. Graf, "Real-time application of an optimal power flow algorithm for reactive power allocation of
] the RWE energy control center," in Proceedings of the IEE Colloquium on International Practices in
Reactive Power Control, London, 1993, pp. 71-74.
[111 P. Nedwick, A. F. Mistr, and E. B. Croasdale, "Reactive Management a Key to Survival in the 1990s,"
] IEEE Transactions on Power Systems, vol. 10, no. 2, pp. 1036 - 1043, May 1995.

49