Professional Documents
Culture Documents
Doctor of Philosophy
in
ELECTRICAL ENGINEERING
2009
i
Author
Engr. Aamir Hanif
(04-UET/PhD-EE-16)
Approved by:
2009
ii
ABSTRACT
Aamir Hanif
(04-UET/PhD-EE-16)
Prof. Dr. Mohammad Ahmad Choudhry, Advisor
Electrical Engineering Department
Due to increased proliferation of sensitive loads in the distribution network, voltage
sags, swells and phase jumps mitigation has become the focus of power quality (PQ)
research. This reduces its adverse economical impact. Equipment sensitivity curve shows
that small duration outage of one-cycle is tolerable for sensitive loads. Therefore, the
control system should be capable for voltage sag detection and mitigation in less than 20
ms at 50 Hz.
There are various custom power devices for protecting industrial processes against
voltage sags. Due to simple control, fast response and fewer transients; dynamic voltage
restorer (DVR) is investigated in this research work. Several voltage-sourced converter
(VSC) control strategies for the DVR are proposed to maintain constant load voltage for
smooth and clean power delivery to the load. A new leading series voltage injection
(LSVI) technique is presented as a first strategy for the mitigation of sag. A mathematical
model of LSVI by a DVR for voltage sag mitigation has been developed. Simulation
results have been obtained to validate the developed model. Conventionally, distributed
generation (DG) source is connected in parallel with utility to share load and export its
excess power to utility. A new concept of power export through series connection of DG
supported DVR is presented, whose power export capability is managed by magnitude of
injected voltage as well as its phase-angle instead of the current. However, the approach
for mitigating voltage sag using LSVI scheme is suitable for those locations where the
phase-shift in the voltage will not cause any problem.
A feed-forward VSC control strategy for the DVR is then presented, based on
unipolar SPWM switching strategy. This is a novel but a simple control strategy in which
the VSC not only converts DC to AC but also mitigates voltage sags, swells and phase
jumps in the utility voltage, acting as power flow controller. Comparison of instantaneous
utility voltage and generated three-phase reference voltage is carried out to detect sags,
swells and phase jumps in the utility voltage. It then generates appropriate switching
signals for the VSC of DVR to mitigate them.With the help of proposed control scheme
of DVR, the dynamic capability of DVR increases the sensitive load penetration in the
distribution system. LSVI control uses leading series voltage injection technique whereas
a feed-forward method of compensation uses in-phase injection to regulate the load
voltage. The magnitude of injected DVR voltage is minimal in in-phase injection
compensation strategy. However, the presented power export concept is valid only for the
LSVI scheme.
The simulation for the presented control strategies have been performed using
SimPowerSystems tool box of MATLAB. The proposed control strategies detect
utility disturbances and mitigate them within half-cycle against a target of one-cycle.
These strategies succeed in keeping load voltage THD limit within 5%.
iii
ACKNOWLEDGMENTS
The first acknowledgement goes to my supervisor, Prof. Dr. Muhammad Ahmad
Choudhry for his kind guidance, valuable suggestions, technical assistance and
encouragement during the course of this research that enabled me to carryout this
research.
I am thankful to members of the research committee for their important discussions and
helpful comments in achieving the objectives of my research work. They always
encouraged me and their guidance was the driving force for me during my dissertation
work.
Special thanks to anonymous reviewers of Journal of Zhejiang University Science A,
co-published with Springer-Verlag whom valuable comments helped me to improve
further my research work.
The author greatly acknowledges the facilitation provided by the Directorate Advanced
Studies Research and Technological Development, University of Engineering and
Technology Taxila with the financial support from Higher Education Commission
Pakistan.
I am also grateful to all of my colleagues especially Prof. Dr. Ahmed Khalil Khan and
friends for their assistance and encouragement.
I cant forget at this moment to pay gratitude to my late father Prof. Muhammaad Hanif
Janjua, Principal Government Hashmat Ali Islamia College, Rawalpindi for his
inspiration, encouragement and moral support to initiate the performed research work.
I would like to appreciate the encouragement and prayers of my mother, wife and
children who never let me alone throughout this work.
iv
TABLE OF CONTENTS
ABSTRACT
ACKNOWLEDGEMENT
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
NOMENCLATURE
INTRODUCTION
DISTRIBUTION SYSTEM POWER FLOW AND DG
2.1
Introduction
2.2
Issues with Distribution System Power Flow
2.3
Power Flow Control Techniques
2.4
DG and Distribution System Power Flow
2.4.1
Advantages of Distributed Generation
2.4.2
Interconnection of DGs with Electric Power Systems
2.5
Summary
iii
iv
v
vii
viii
xiii
1
8
8
16
24
29
35
36
36
39
39
41
42
43
62
91
97
98
99
103
105
107
107
108
109
110
113
114
115
116
118
120
120
121
121
123
124
124
125
129
130
130
137
140
143
173
175
178
186
187
187
190
192
196
196
198
201
205
205
207
210
217
220
222
224
PERTINENT PUBLICATIONS
VITA
253
254
vi
LIST OF TABLES
Table ....................................................................................................................Page
Table 2.1
Table 2.2
Table 2.3
Table 2.4
Table 2.5
Table 2.6
Table 3.1
Table 4.1
Table 5.1
Table 5.2
Table 5.3
Table 5.4
Table 5.5
Table 5.6
Table 5.7
Table 5.8
Table 5.9
Table 5.10
Table 5.11
Table 5.12
Table 5.13
Table 5.14
Table 5.15
vii
LIST OF FIGURES
Figure .......................................................................................................................Page
Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 2.7
Figure 2.8
Figure 2.9
Figure 2.10
Figure 2.11
Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 3.6
Figure 3.7
Figure 3.8
Figure 3.9
Figure 3.10
Figure 3.11
Figure 3.12
Figure 3.13
Figure 3.14
Figure 3.15
Figure 3.16
Figure 3.17
Figure 3.18
Figure 3.19
Figure 3.20
Figure 3.21
Figure 3.22
Figure 3.23
Figure 3.24
Figure 3.25
Figure 3.26
Figure 3.27
Figure 3.28
Figure 3.29
Figure 3.30
Figure 3.31
Figure 3.32
Figure 3.33
Figure 3.34
Figure 4.1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8
Figure 4.9
Figure 4.10
Figure 4.11
Figure 4.12
Figure 4.13
Figure 4.14
Figure 4.15
Figure 4.16
Figure 5.1
Figure 5.2
Figure 5.3
Figure 5.4
Figure 5.5
Figure 5.6
Figure 5.7
Figure 5.8
Figure 5.9
Figure 5.10
Figure 5.11
Figure 5.12
Figure 5.13
Figure 5.14
Figure 5.15
Figure 5.16
Figure 5.17
Figure 5.18
Figure 5.19
Figure 5.20
Figure 5.21
Figure 5.22
Figure 5.23
Figure 5.24
Figure 5.25
Figure 5.26
Figure 5.27
Figure 5.28
Figure 5.29
Figure 5.30
Figure 5.31
Figure 5.32
Figure 5.33
Figure 5.34
Figure 5.35
Figure 5.36
Figure 5.37
Figure 5.38
Figure 5.39
Figure 5.40
Figure 5.41
Figure 5.42
Figure 5.43
Figure 5.44
Figure 5.45
Figure 5.46
Figure 5.47
Figure 5.48
Dynamic voltage variation with sag with =130 (Case II)........................... 167
o
Dynamic active power variation with sag with =130 (Case II) .................. 167
o
Dynamic reactive power variation with sag with =130 (Case II) ............... 168
o
Dynamic voltage variation with sag with =150 (Case II)........................... 168
o
Dynamic active power variation with sag with =150 (Case II) .................. 169
o
Dynamic reactive power variation with sag with =150 (Case II) ............... 169
o
Dynamic voltage variation with sag with =165 (Case I) ............................ 170
o
Dynamic active power variation with sag with =165 (Case I) ................... 171
o
Dynamic reactive power variation with sag with =165 (Case I) ................ 171
o
Dynamic active power variation without sag with =180
When loas is 500 kW (Case II) ...................................................................... 172
Figure 5.49 Dynamic reactive power variation without sag with =180o
When load is 1000 kW (Case II).................................................................... 172
Figure 5.50
Figure 5.51
Figure 5.52
Figure 5.53
Figure 5.54
Figure 5.55
Figure 5.56
Figure 5.57
Figure 5.58
Figure 5.59
Figure 5.60
Figure 5.61
Figure 5.62
Figure 5.63
Figure 5.64
Figure 5.65
Figure 5.66
Figure 5.67
Figure 5.68
Figure 5.69
Figure 5.70
Figure 5.71
Figure 5.72
Figure 5.73
Figure 5.74
Figure 5.75
Figure 5.76
Figure 5.77
Figure 5.78 Dynamic active powers for 45% swell in utility voltage......................204
Figure 5.79 Dynamic reactive powers for 45% swell in utility voltage...................204
Figure 5.80 Dynamic voltages for +25 phase jump in phase-A of
utility voltage ........................................................................................205
Figure 5.81 Dynamic active powers for +25 phase jump in phase-A of
utility voltage ........................................................................................206
Figure 5.82 Dynamic reactive powers for +25 phase jump in phase-A of
utility voltage ........................................................................................207
Figure 5.83 Dynamic voltages for -25 phase jump in phase-A of
utility voltage ........................................................................................208
Figure 5.84 Dynamic active powers for -25 phase jump in phase-A of
utility voltage ........................................................................................209
Figure 5.85 Dynamic reactive powers for -25 phase jump in phase-A of
utility voltage ........................................................................................209
Figure 5.86 Mitigation devices cost solution ...........................................................216
xii
NOMENCLATURE
ASVC
ANSI
APF
AF
BSES
CPD
CP
CHP
Cdc
CPP
CVT
CPM
CEIDS
DVR
D-STATCOM
DER
DG
DCS
DWT
DCMLI
D-SMES
DE
EMS
EPRI
EMTDC
ESKOM
ETC
Estorage
e(t)
FACTS
FCMLI
FES
GTO
GWh
HSAF
HISP
I i*
ITIC
IEC
IEEE
IGBT
IDVR
IFL
IGCT
IPC
IR
IX
kVA
kW
LSVI
LTC
LPF
LQR
MATLAB
MG set
MOSFET
NO
NC
PCC
PQ
PWM
PSCAD
PLL
PAC
PV source
PSC
PAC
PF
PI controller
pf
QRA
qf
RMS value
SEMI
SVR
SSSC
STATCOM
SVC
Si
SSC
SPWM
STS
SRFT
SHC
SEC
SABER
SSCL
SSCB
SSTS
SCR
SCH
TCSC
T&D
TSC
TCR
TBESS
TCSR
THD
s
UPFC
UPQC
UPS
VSC
VAR
Vdc
Vdrop
VD/VDVR/Vinj
Vpcc
Vsag
Vmissing
V pre sag
VL,pre-dip
Vs
Vu
VL
VD*
VL*
VU
vRa(t)
vua(t)
Vmod
Vcarr
ZS
Load voltage
Shunt converter voltage
Series converter voltage
Reference voltage for DVR to track
Desired load voltage
Utility/Supply voltage available at the point of common connection
Instantaneous reference signal for phase-a
Instantaneous utility voltage signal for phase-a
Modulating signal
Carrier signal
Supply impedance
Zf
vshunt
vseries
xvi
CHAPTER I
INTRODUCTION
Disturbance free supply of electricity with regulated voltage at permissible level is the
key system requirement in the utility network. Conventionally mechanical switched
capacitor and series voltage regulator (SVR) utilizing transformer taps are used to control
voltage in distribution system. Mechanical switched capacitor operates in the order of
300 ms; hence, it has very slow response. So the voltage sag in the distribution network
may not be corrected in this time frame. Even SVR is quite slow in response to voltage
sag and its operations are in steps. Reactive power supplied by the capacitor is
proportional to the square of system voltage and this reduces its effectiveness in high and
low-voltages. Even combined usage of switched and fixed capacitor does not ensure to
generate continuously variable reactive power in case of changing load conditions.
Whereas, combined application of SVR and switched capacitor to regulate load bus
voltage suffers from the voltage swing that may result in power quality (PQ) problems
for the consumers. An electronic solution for the voltage regulation problem includes
custom power devices (CPDs) to maintain PQ in distribution system. The advantage of
electronic solutions is speed, as rapid fluctuation in load current or supply voltage can be
compensated within one-cycle as required by Information Technology Industry Council
(ITIC) curve and Semiconductor Equipment and Materials International (SEMI-F47)
standard. Supply voltage is referred by some authors as utility voltage. Therefore, these
two terms will be used interchangeably in this research work.
For the deregulating electric power industry, the demand of flexible power flow
control is becoming an important goal achievable through innovative power electronics
technology. . Energy management system (EMS) requires quick on-line decision making
ability whereas; the equipment that is operated mechanically has slow response. The
rapid development of power electronics has made it possible to design power electronic
equipment of high rating for high voltage systems. The voltage stability problem and
cheap power transfer can be improved through flexible alternating current transmission
system (FACTS) controllers. The objective of FACTS devices mainly static synchronous
compensators (STATCOM), static synchronous series compensators (SSSC), static VAR
compensators (SVC) and unified power flow controller (UPFC), etc., is to bring a system
under control and to transfer power in required network configuration. Thus FACTS
technology has enabled the engineers to use different controllers for efficient series and
shunt compensation.
The FACTS technology offers an enhanced control of power flow. Since these
devices offer very quick power swing damping, the power transmission lines can be used
to transfer electrical energy upto their thermal capacity. Similarly, the application of
power electronic devices is also helpful for the power distribution system as it enhances
the quality and the continuity of power supplied to the consumers. The concept of
applying the power electronics based technology to power distribution system for the
benefit of a customer or group of customers is called Custom Power. Custom power (CP)
technology includes switching equipment and controllers that can be utilized to offer
disturbance/interruption free service to the customers.
The present PQ issues that are seen by the utilities and power distributors are sudden
and short duration deviations, e.g. impulsive and oscillatory transients, voltage dips,
voltage swells, phase jumps, short interruptions, as well as steady-state deviations, such
as harmonics and flicker. A compact solution to these issues may be offered by custom
power technology. The reliability of the power supplied using this technology can be
enhanced in the form of less interruptions and small voltage variations. Therefore, the use
of this technology may benefit all the industrial, commercial and domestic customers
through power quality and reliable service. However, it has been reported in various
survey results that more than 90% of the interruptions in the industrial installations are
due to voltage sag.
In the last decade, there is growing interest in equipment for mitigation of PQ
disturbances, especially that are based on power electronics called Custom Power
Devices. These devices provide cost effective solutions to PQ issues. In this way, load
voltage can be regulated which results in smooth and clean power flow to the load. The
most common CPDs are dynamic voltage restorer (DVR), distribution static compensator
(D-STATCOM), unified power quality conditioner (UPQC) and uninterruptible power
supply (UPS). Dynamic voltage restorer is a series compensator that provides technically
and economically advanced solution to the most PQ issues. It provides a cost effective
solution as compared to other CPDs. Hence, this device is investigated in this research
for the mitigation of voltage sag, swell and phase jumps issues in utility voltage. New
control strategies for the voltage-sourced converter (VSC) of DVR have been developed.
The new developments in the area of power electronics, generation methods and
energy storage devices provide an opportunity to utilize distributed energy resources
3
(DERs) in the distribution system. Moreover, laws have been introduced to encourage
more decentralized approach to power distribution. DERs or DGs can be used with or
without CPDs.
Power electronic interface is used to connect renewable energy sources to common
A.C. load bus at 50 or 60 Hz. In this regard the use of power electronics based VSCs
help to control power flow as required from DER, working with utility supply at the A.C
load bus. The benefits of VSC connected DG on feeder performance can be significant in
reducing the voltage variations on the feeder. It can also be used for peak load shaving to
facilitate line to meet a rising peak demand. This mode of operation requires the DG to
operate by delivering real power at times of peak loading for that particular line. For rest
of the time, reactive control of the VSC can be used to control voltage and reduce voltage
dips on the line.
In this research work, different DVR control strategies are proposed to tackle PQ
issues such as sag, swell and phase jumps in the utility voltage so that customer have an
interruption free service from the utility. Moreover, the proposed strategies enhance the
quality/reliability/availability (QRA) of utility supply to customers. Industrial and
commercial customers require QRA of utility supply in order to avoid interruption in
their production and work environment. This will provide them opportunity to meet their
planed goals and can earn profit. Otherwise, sudden tripping of electronic and digital
equipment due to sag, swell or phase jumps will result in shutdown of plant processes.
This may cause huge loss in term of money and components damage.
Traditionally, DG is connected in parallel with utility to supply load demand or
export power to grid to release the utility from stress due to growing demand of
4
electricity and shortage of generation. However, in this research work, a new DVR
control scheme is presented that not only compensates sag in the utility voltage to
provide disturbance free power to load but also exports excess power to grid (utility) as
well. The power is exported to grid through series connection of DVR which is fed from
a DG.
Chapter II describes the distribution system power flow and the role of distributed
generation (DG) on it. Firstly, structure of power system along with its subsystems is
described. Different primary distribution system configurations are mentioned and
advantage of radial distribution system over various distribution circuit configurations is
presented. Secondly, issues of power flow in distribution systems and conventional
power flow control techniques are discussed. Finally, DG and distribution system power
flow along with interconnection of DGs with electric power system is explained.
Summary of chapter recommends the use of power electronics based solutions especially
CPDs (with or without DGs) with utility to deal with voltage regulation issues at load
bus.
Chapter III provides literature review for CPDs applications in distribution system.
Firstly, the application of power electronics to power transmission system as well as to
power distribution system is highlighted. Then the concept of Custom Power is
introduced and its advantages are described. Next, the PQ issues are briefly described and
need of equipment for mitigation of PQ disturbances is highlighted. Afterward, concept
of CPDs utilizing power electronics technology is introduced and different types of CPDs
are mentioned. Applications of these CPDs in literature are discussed in detail that
utilizes different control strategies to tackle PQ issues. At the end, summary of chapter is
5
and phase jumps. The results of simulations show its effectiveness in handling these PQ
issues, so that smooth and clean power flow to the load.
CHAPTER II
DISTRIBUTION SYSTEM POWER FLOW AND DG
2.1
Introduction
Nowadays power systems are complicated networks. They have hundreds of
generating stations and load centers that are interconnected through power transmission
lines [1]. Electricity is generated and supplied to consumers via transmission and
distribution networks and contributes a major share in the consumer markets of the world
[2]. Power systems of the modern era are more reliable and serve customer load without
any interruption in utility voltage. Generation facilities should have the capacity to
produce required power to meet the customer demand. Bulk power generated must be
transported through transmission systems over a long distance without overheating or
jeopardizing system stability. It is responsibility of the distribution system to deliver
electricity to each customers service entrance. In the context of reliability; generation,
transmission, and distribution systems are made up of several subsystems as listed in
Table 2.1 [3].
Table 2.1
Electric power systems and their subsystems
Sr. No.
Systems
Generation
Transmission
Distribution
Subsystems
Generation plants
Generation substations
Transmission lines
Transmission substations/ switching stations
Sub-transmission systems
Distribution substations
Primary distribution systems
Dispersed distribution transformers
Secondary distribution systems
A simplified description of an overall power system along with its subsystems is shown
in Fig. 2.1[2].
Generation Plants are required to convert fuels (coal, gas, oil, fossil, nuclear,
etc.) and other energy resources (water, wind, solar radiation, etc.) into electrical
power [4]. Typically, a prime-mover turns an alternator that generates voltage between
substation step down sub-transmission voltage levels to primary distribution levels. Fig.
2.2 shows a distribution substation layout [5].
11
service entrances. Generally, the secondary distribution systems are designed in singlephase for areas of residential customers and in three-phase for areas of industrial or
commercial customers with high-load densities [6]. Voltage levels are typically
120/240V single-phase, 120/208V three phase or 277/480V three phase in U.S system,
whereas, 220/380V or 230/400V or 240/416V three phase in European System.
A distribution system consists of feeders, distributors and service mains. Feeders are
large current carrying conductors that have the capacity to carry the current in bulk to the
consumer premises whereas current is tapped off from distributors for supply to the
consumer [7]. The service mains are the small cables between the distributors and the
consumers premises.
Fig. 2.3 shows single line diagram indicating elements in a modern distribution
system. The essential elements in the Fig. 2.3 include the following [8]:
(1) Sub-transmission system (220kV, 132kV, 66kV, 33kV)
(2) Distribution substations (step down voltage)
(3) Primary distribution system (33kV, 22 kV, 15kV, 11kV, 6.6kV, 3.3kV)
(4) Distribution transformers
(5) Secondary distribution system (400V, 3-phase 4-wire)
(6) Service connections (400V, 3-phase and 230V, single-phase)
Fig. 2.4 shows a simple one line diagram of the basic structure of a typical power system
[1], [9]. Transformer T3 connects sub-transmission network/primary distribution network
with secondary distribution network. It step downs the voltage from 66 kV to 11 kV.
Load in the distribution system is either commercial type or residential type. Generally,
12
commercial customers get power supply at 11 kV, whereas, the domestic consumers are
supplied at 400-440 V at 50 Hz. Since domestic customers are fed through single-phase
supplies, hence they normally receive 230-250 V at their inlet points at 50 Hz.
It is to be noted that the voltage levels quoted above are not standard and vary from
country to country. For example, in Pakistan the generation voltage level varies from
13
13kV to 15 kV. The domestic customers get supply at 400 V 10% (line-to-line) or 220
V 10 V (line-to-neutral). The primary distribution is at 11 kV. The sub-transmission
voltages are either 66 kV or 132 kV and the transmission voltages are 220 kV or 500 kV.
Detailed reference values of nominal voltage levels and nominal system voltages as
recommended by IEC for a power system at 50 Hz are shown in Table 2.2 [8].
Fig. 2.4 One line diagram of the basic structure of a typical power system [1, 9].
There are various stages between the points of power generation to the stage when
electric power is delivered to the end users. The correct operation of all components of a
power system is absolutely critical for a reliable power delivery.
Distribution system design varies in philosophy from country to country [10]. Many
books are published that cover all the aspects on the design and planning of electric
distribution networks [11]-[13]. Lengths of distribution circuits and their configuration
varies however, they share similar characteristics. There are several configurations of
distribution systems that increase customer reliability and five of them are shown in Fig.
14
2.5. Each configuration is discussed in detail in [2], [14]. These distribution circuit
configurations have their own merits and demerits. However, most distribution circuits
have radial configuration. Following are the advantages of radial distribution circuits
over other distribution circuit configurations.
Lowest cost
Table 2.2
Reference values of nominal voltage levels and nominal system voltages [8]
Nominal voltage
of 3, AC
system at 50 Hz,
phase-to-phase kV, rms
3, AC, 50 Hz
phase-to-phase, and
phase-to-neutral
EHV = Extra high voltage
HV = High voltage
765 kV
EHV
500 kV
EHV
380 - 400 kV EHV
345 kV
EHV
275 kV
EHV
220 kV
HV
132 kV
HV
66 kV
HV
220 kV
HV
132 kV
HV
66 kV
HV
33 kV
HV
33 kV
MV
22 kV
MV
15 kV
MV
11 kV
MV
6.6 kV
MV
3.3 kV
MV
400V
LV
or 415 V
(ph. to ph.)
230 V or
240 V
(ph. to n.)
MV = Medium voltage
LV = Low voltage
15
Transmission
Subtransmission
Distribution
primary
circuits and
direct feeders
to large consumers
Distribution
secondary circuits and
service to
consumers
2.2
Voltage drop
Power loss
Power quality
Voltage drop in the distribution circuits during peak loads and voltage rise during low
loads is the main problem, in addition to the problem of overloading of distribution
16
transformers and secondary feeders. The total voltage drop in the secondary circuit of
distribution system is highest at the remote end. To minimize the secondary-circuit
length, distribution engineers locate the distribution transformers close to the load centers
and try to have the secondary service drops to the individual customers as close as
possible.
Table 2.3
Loads and corresponding secondary distribution supply voltage [8]
Connected load
Upto 1.5 kW
2 kW to 50 kW
50 kW to 100 kW
100 kW to 1.5 MW
1.5 MW to 5 MW
5 MW and above
Supply voltage
1, 230 V AC
(phase-to-neutral)
3, 4-wire AC
(phase-to-phase)
3, 3.3 kV or
3, 400 V AC
(phase-to-phase)
3, 11 kV
(phase-to-phase)
3, 33 kV or
3, 66 kV AC
(phase-to-phase)
3, 132 kV or
3, 220 kV AC
(phase-to-phase)
Remarks
Residential LT
Secondary distribution supply
(3-phase 4-wire AC system)
Secondary distribution LT supply
(3-phase 4-wire AC system)
From primary or secondary of distribution
system
(3-phase 4-wire AC system)
From primary distribution system
(3-phase 4-wire AC system)
From primary or secondary of distribution
system or Sub-transmission system
(3-phase 4-wire AC system)
From receiving substation bus of adequate
fault level
Table 2.3 gives reference values of various MW loads and corresponding preferred
supply voltage [8]. Higher the load, higher is the required supply voltage for achieving
reduced I2Rt losses and size of feeding transformer and distribution feeder. It is core
responsibility of the utility to maintain specified voltage at consumers premises within a
suitable range, so utilities must regulate the voltage. Impedance of line causes voltage
drop in it. Hence voltage drop have been allocated as necessary along the primary and
secondary distribution system circuits to provide end users with suitable voltage.
Voltage decreases in radial distribution systems when one moves away from the
substation to the end of a distribution feeder. This decrease in voltage is also known as
17
the line drop and is mathematically represented by Ohm's law, V = IZ, where
I is line current , A
According to ohms law voltage drop is affected by variation in load. As line impedance
is constant hence large load requires more power. This leads to higher current drawn by
the load and higher voltage drop. As it is impossible to obtain exactly flat voltage level at
the end of the distribution feeder, utilities have allowed 5% tolerance for the nominal
voltage in the United States according to ANSI C84.1 standard [15]. This implies that a
customer can have voltage in between 114 and 126 V at his service drop which
nominally receives 120 V at 60 Hz.
The problem of voltage drop along a feeder is evident from the Fig. 2.6. This figure
shows the effect of feeder length on the voltage and hence tolerance limit. Therefore,
voltage regulation measures have to be adopted to maintain permissible voltage level
along the whole length of a feeder. As voltage at the end of feeder changes with load
variation hence voltage compensation is required at different times.
Voltage drop along a circuit can be approximated as [5]
|V drop| = | Vs | - | Vr | IR . R + IX . X
(1.1)
Where
V drop = Feeder Voltage drop, V
R = Resistance of line,
X = Reactance of line,
IR = Current flowing in the line due to real power flow (in phase with the voltage), A
18
IX = Current flowing in the line due to reactive power flow (90 out of phase with the
voltage), A
Considering the power factor (pf), equations for the real and reactive line currents can be
written as:
IR = I . pf = I cos
(1.2)
(1.3)
Where
I = Line current magnitude, A
pf = Load power factor
qf = Load reactive power factor /reactive part of load power factor = sin (cos 1 (pf ))
= Phase-angle difference between voltage and current
While just an approximation: voltage drop given by equation IR R + IX X yields good
results for most distribution feeders [16]. This equation summarizes two important
aspects about voltage drop:
19
The reasons behind the higher voltage drop include lower distribution system voltage
levels, poor power factor, single-phase circuits, and unbalanced circuits. Reduction of
voltage drop in a distribution system can be achieved by:
Balance circuits
Reducing overload
Voltage drop can be tolerated as long as compensation means are provided by voltage
regulation equipment that adjust the voltage drop on the circuit.
Voltage ranges in ANSI C84.1 standard are mostly opted by the utilities for the
service entrance [14]. However the methods for voltage regulation and voltage drop
allocation vary from case to case. Fig. 2.7 shows the voltage variation along the circuit.
Usually load tap changer (LTC) or bus regulator at the substation regulates the voltage at
the source. Voltage drops may occur in primary distribution line, distribution
transformer, and secondary distribution line. Therefore, in order to maintain voltage, one
cannot ignore consumers connected at the starting point as well as at the terminating
point of the distribution network. At the end of line, voltage level is usually dependent on
the transformer loading as well as length of secondary distribution line. However, the
voltage drop allocation for the secondary line also relies on utilitys design practice [16].
20
Fig. 2.7 Voltage drop along a radial circuit with no capacitors or line regulators [14].
At the starting point, voltage of the line is kept slightly above the rated value. It is
responsibility of the utility to monitor and control first customer voltage so that one does
not have voltage level higher than 126 V. It is common practice of utilities to assume
voltage drop of 1 V at the consumer premises, when it is lightly loaded. With this value
of drop, the upper voltage limit at the start of line is 127 V. Use of low impedance
distribution transformers and larger size transformers also help to keep the voltage drop
beyond the limits. Larger size conductors also help to reduce secondary voltage drop and
hence power loss.
Besides voltage drop and power loss issues that occur during power flow in a
distribution network, the power quality (PQ) of supply to consumers under the severe
conditions of starting currents, frequent switching of connected loads, thyristor switched
loads, industrial welding sets, industrial arc furnace loads etc. is a challenge for the
utilities [8]. Conventionally, load current was calculated without considering harmonic
21
content in the load, but nowadays there are standards for harmonic content present in the
load current. Moreover, sensitive load was fed through separate feeder; hence, issue of
PQ was not prominent and, distribution systems were designed without PQ concerns. But
today sensitive load is scattered and it almost exists with every consumer. Hence,
distribution system has to tackle PQ issues as well to provide customers with electricity
of acceptable quality in terms of:
As the sensitive load and complexity of control systems in industrial processes have
grown, the quality of power supplied to consumers has a vital role especially, in recent
years. Nowadays, industries are more concerned about the power quality issues because
they affect them financially. Voltage sag is one of the severe PQ problems that affect
them [17]. It is a momentary reduction in available supply for a short duration below 90%
of the rated value [18]. Faults in the power system or starting of large load such as
induction motor, induction furnace, etc are the prime source of voltage sag [19]. Main
characteristics of such voltage sag are its duration and magnitude of available utility
voltage [20]. Table 2.4 shows categories of PQ variations [21], [22].
22
Table 2.4
Categories of PQ variations [21], [22]
Categories
Transient
Impulsive
Oscillatory
Short-duration
variations
Instantaneous
Momentary
Temporary
Long-duration
variations
Voltage
unbalance
Waveform
distortion
Nanosecond
Microsecond
Millisecond
Low Frequency
Medium
Frequency
High Frequency
Sag (dip)
Swell
Interruption
Sag
Swell
Interruption
Sag
Swell
Typical
spectral
content
5 ns rise
1 s rise
0.1 ms rise
< 5 kHz
5-500 kHz
0.5-5 MHz
Interruption
Sustained
Under
voltages
Over voltages
D.C. offset
Harmonics
Interharmonics
Notching
Noise
0-100th H
0-6 kHz
Broadband
< 25 Hz
Voltage
fluctuations
Power
frequency
variations
Typical
duration
Typical
voltage
magnitude
< 50 ns
50 ns to 1ms
> 1 ms
0.3-50 ms
20 s
5 s
0-4 p.u.
0-8 p.u.
0-4 p.u.
0.5-30 cycles
0.5-30 cycles
0.5 cycles -3s
30 cycles - 3s
30 cycles - 3s
3s to 1min
3s to 1min
3s to 1min
> 1 min.
0.1-0.9 p.u.
1.1-1.8 p.u.
< 0.1 p.u.
0.1-0.9 p.u.
1.1-1.4 p.u.
< 0.1 p.u.
0.1-0.9 p.u.
1.1-1.2 p.u.
0.0 p.u.
> 1 min
> 1 min
0.8-0.9 p.u.
1.1-1.2 p.u.
Steady-state
0.5-2%
Steady-state
1-0.1%
Steady-state
Steady-state
Steady-state
Steady-state
Intermittent
0-20%
0-2%
0-1%
0.1-7%
10s
Voltage quality problems in the form of voltage sags, voltage harmonics, and voltage
swells may cause severe disruptions to critical processes, resulting in substantial
economic and/or data losses [23], [24]. The process-control constitutes brain or
Central Nervous System of the modern-day automated manufacturing and process
industry. Failure of control part of any process may lead to failure of that process and
therefore, loss of raw material and production time and even risk to human life.
23
2.3
in power systems with large number of generators [25]-[35]. In this method, the generators share the system load by drooping the frequency of every generator with active
power (P) supplied by the generator [34]. This allows each generator to take up changes
in total load in a manner determined by its frequency droop characteristic, and essentially
utilizes the system frequency as a communication link between the generator control
systems [25]-[29]. In the same way, reactive power sharing may be achieved by
involving droop in the voltage amplitude (Vmax) [35].
Conventionally, load sharing by power system generators is based on the power flow
theory in an AC system. According to this theory, the flow of the active power (P) and
reactive power (Q) between two sources can be controlled by varying the power angle
and the voltage magnitude of each system i.e., the active power flow (P) is mainly
controlled by the power angle, while the reactive power (Q) is mostly controlled by the
voltage magnitude. Fig. 2.8 shows the typical variables involved for load-sharing control
of paralleled connected power generators. It shows two generators (sources) represented
24
by two voltage sources connected to a load through line impedance represented by pure
inductances L1 and L2 for simplified analysis purpose.
The complex power at the load due to ith voltage source is given by
S i = P i + jQi = V .I i *
(1.4)
where i = 1, 2..n., and Ii* is the complex conjugate of the current for the ith voltage
source and is given by
E cos i + jE i sin i V
I i* = i
jwLi
(1.5)
E cos i + jE i sin i V
Si = V i
jwLi
(1.6)
This equation involves active and reactive power flowing from the ith source that can be
written as
Pi =
VEi
sin i
wLi
(1.7)
25
VEi cos i V 2
Qi =
wLi
(1.8)
Equations (1.4)-(1.8) show that, when values of 1 and 2 are very small, then the
flow of active power is dominantly realized by the power angles 1 and 2, whereas the
flow of reactive power mostly relies on the source output voltages E1 and E2. Hence, flow
of active and reactive power can be achieved independently to some extent. As
controlling the frequencies, dynamically controls the power angles, therefore the real
power flow control can be equivalently realized by controlling the frequencies of the
voltages generated by the sources. Consequently, the power angle and the source output
voltage magnitude are main variables that can directly control the real and reactive power
flow for proper load sharing of sources connected in parallel. Effect on power flow for
increasing modulation index value by keeping constant and vice versa for a DC type
DER connected in parallel with utility to supply load have been investigated in [36].
Distribution utilities have several options to control steady-state load bus voltage and
in-turn power flow to load. The most popular regulation methods include the following
traditional equipment:
Most utilities use LTCs to regulate the substation bus voltage. However, supplementary
feeder regulators and /or switched capacitor banks may be used where needed. Hence, in
general, voltage in the transmission and distribution system can be controlled or
26
supported through transformer taps and reactive power injection. Changing taps on
distribution transformers is another tool to supply proper voltage to customers.
Distribution transformers are available with standard taps of +2.5 and +5%. Utilities may
use this option to provide a fixed boost for customers on a circuit with low primary
voltage. By moving the tap up or down, the effective number of turns of the transformer
winding is changed. An LTC typically has some number of discrete settings to be
adjusted by distribution operators according to conditions on the circuit.
The same basic device, when installed midway on an individual feeder rather than at
the substation transformer, is called a line voltage regulator. A traditional step-voltage
regulator is a regulating transformer in which the voltage of the regulated circuit is
automatically controlled in steps by means of taps, without interrupting the load. This
transformer can boost or buck the voltage supplied to a load with a delay on the order of
seconds. Voltage regulators look like tall transformers on distribution poles, often with
large fins for heat dissipation, but with no secondary lines going out to customers; they
are simply transformers between two segments of the same line. Their turns ratio is
adjusted to boost the voltage just enough to compensate for line voltage drop. Raising the
nominal voltage by using LTCs, regulators, and switched capacitors helps the voltage sag
ride through capability of many types of customer equipment. Programmable logic
controllers, computers, adjustable-speed-drives, and other equipment with capacitors
benefit if the voltage is regulated near the upper end of voltage range A of ANSI C84.1
standard, or at least avoid the lower end.
Reactive power affects voltage in a very different manner. When the consumption of
reactive power is more than supplied, voltage decreases and vice versa. Its deficiency is
27
28
2.4
The electric power system is traditionally designed and operated to transport large
bulk of the electric power from the generation units to the electric energy consumers
through the transmission and distribution networks. Distribution systems are traditionally
29
passive and designed to operate with only such unidirectional energy flow, in contrast to
the transmission system, which is designed for two-way power flow. In operation and
planning, it is assumed that electric power always flows from the secondary winding of
the transformer in the substation to the end of feeder.
In recent years Distributed Generation (DG) appeared as an important identity in
the power system. It is predicted to perform a dominant role in the future power system.
The expansion of distributed generation has the potential to considerably modify the
nature of the distribution system and the coupled PQ problems. With the introduction of
DG [34]-[44], the distribution system becomes an active system with both energy
generation and consumption at the formerly exclusive load nodes. Now bidirectional
power flows should be incorporated in the hierarchical network design and its operating
criteria [44].
Conventionally, distribution systems were designed without considering PQ issues
because sensitive load was isolated. Nowadays sensitive load is situated in every house,
commercial place etc and has significant portion of total load and is not supplied in
isolation. But instead custom power devices (CPDs) with or without DG as well as DGs
independently are used to deal with sensitive loads at distribution side.
Traditionally DGs are connected in parallel with utility to serve increasing load
demand or to meet grid power requirements. Various schemes have been proposed in
literature that shows how various types of DGs share power in parallel mode with utility
to serve load demand. Besides acting as source of power, DG has been increasingly used
for voltage regulation at the load bus/point of common coupling (PCC) in distribution
system. The connection between distributed generation source and distribution network
30
alters the direction of flow of power in the distribution network. In addition, voltage at
various nodes is also affected in power system. Also, the power flow contributed by
distributed generation source may become bidirectional though the distribution networks
were originally configured for unidirectional flow of power.
Distributed Generation is scattered in the service area of utility. Different terms are
preferred in different parts of the world to refer it, for example:
However, literature studies show that the word Distributed Generation is most
commonly used through out the world. When also considering storage systems, the name
distributed energy resource (DER) is also used. It has the ability to inject various
proportions of active and reactive power to utility and load.
The distributed generation is either directly connected to the utility distribution
system or the consumer load bus [37]. The idea of using DG in electric power system is
not a new one. Various consumers although small have been installing it for decades.
However, the introduction of competitive retail electric markets along with innovative
generation technologies has sparked new and broader interest in DG. One of the most
important benefits of DG is that it may be used to defer or avoid transmission and
distribution expenditures [45]. Electric utilities spend significant amounts of money to
expand transmission and distribution networks. Although such expansions are necessary
but usually the existing transmission and distribution networks are not fully-utilized
while they are built for to serve large peak loads. Usual methods to serve an expansion of
31
peak load includes: line reconductoring, more transformer installed capacity and new
feeders or lines. While utilizing DG, peak load demand can be met by operating the local
generating units. Proponents of DG claim that financial savings offered by its application
in distribution system are large and significant. However, some utilities have reported
that these savings are insignificant, if not bringing net costs to the electric power delivery
system. Fig. 2.11 shows the categories that are generally made for DGs capacities.
Fossil-fuel based DG: diesel or natural gas engines, gas turbines, stirling engines,
fuel cells, etc. These technologies are often used in combined heat and power
(CHP) systems.
Table 2.5 shows typical available size per module of different DG technologies.
Table 2.5
Technologies for distributed generation [37]
Technology
Combined cycle gas Turbine
Internal combustion engines
Combustion turbine
Micro-Turbines
Renewable
Small hydro
Micro hydro
Wind turbine
Photovoltaic arrays
Solar thermal, central receiver
Solar thermal, Lutz system
Biomass. e.g. based on gasification
Fuel cells, phosphoric acid
Fuel cells, molten carbonate
Fuel cells, proton exchange
Fuel cells, solid oxide
Geothermal
Ocean energy
Stirling engine
Battery storage
According to new technology, the electric power generation trend uses distributed
generator sized from kW to MW at load sites instead of using traditional centralized
generation units sized from 100 MW to GW and located far from the loads where the
natural recourses are available. With the advent of new technologies, the distribution
34
companys planners continuously endeavor to develop new planning strategies for their
distribution systems in order to serve the load growth and provide their customers with a
reliable electric power. In the new era of competitive electricity market, distribution
engineers investigate economical and technical feasibility of new capacity expansion
alternatives. DG is one out of the available options.
2.4.1
Utilization of DG in the distribution network has lot of benefits. Some of them are
summarized in Table 2.6.
Table 2.6
Economical and operational benefits of DG [39]
Sr. No.
1
2
3
4
5
6
7
Economical
Serve load increase without building new T&D lines and
upgrade the existing power system.
Can be assembled easily anywhere and installed in a
very short time [37].
Capacity can be increased or decreased by modules.
Not restricted to particular location, although, renewable
DGs require certain geographical conditions.
Can be installed in small increments to provide the exact
customer load demand.
Reduce wholesale price by supplying power to the grid that
leads to reduction in demand [41].
Defers construction schedules of power plants.
Type of energy resource and fuel used are diversified and
there is no need of certain type of fuel more than other.
Operational
Reduce
distribution
system
losses.
Help in peak load shaving
[40], [41].
Improve voltage of load bus.
Help system continuity and
reliability.
Used as standby or emergency
reserve for system outages.
Provides transmission capacity
release by locally serving load.
Reduce the output process
emission.
Acts as spinning reserve to
improve system stability.
Despite all the benefits provided in this table, the connection of DG into a distribution
system has raised many issues. The foremost issue that arose in the past was of PQ
problems due to line commutated VSCs. However, with the self commutated power
electronics converters employing pulse-width modulated (PWM) switching or other
advanced switching controllers, the harmonic problems have been diminished. Also,
voltage regulation in a power distribution system may be improved or degraded with the
35
2.5
Summary
Reactive power injection at the point of common coupling (PCC) may be required to
regulate load bus voltage. Conventionally, mechanically switched shunt capacitors are
installed in the distribution primary to support load bus voltage. Installation of
36
mechanically switched shunt capacitors is the simplest and under certain conditions is a
very useful way to compensate the load reactive power demand. However, this method is
reported with some major disadvantages, such as fixed compensation, possible
occurrence of resonance condition with nearby loads, switching transient, bulky size,
aging factor etc. Mechanical switching operates in the order of 300 ms or (much) more
and some sags are not corrected in this time frame. Traditional series voltage regulators
(SVR) utilizing transformer taps may be used for voltage regulation in distribution
system. These devices do not have the ability to produce reactive power and by its
operation only compel the source to generate requested reactive power. Furthermore,
they have quite slow response and these operations are in steps and tap changing under
load has its disadvantages (e.g. cost). SVR may also be coupled with capacitors that are
switched on and off to regulate load voltage. However, this arrangement will result in
voltage swings that can cause PQ problems for customers. An electronic solution to the
voltage regulation problem is use of CPDs for providing reliable distribution power
quality. Applications of CPDs in distribution system are discussed in detail in chapter III.
The advantage of electronic solutions is speed: rapid fluctuation in load current or supply
voltage can be corrected or compensated within the one-cycle.
Modern technological advancements in the area of power electronics, electricity
generation methods and energy storage devices have offered an innovative opportunity
for distributed energy resources (DER) at the distribution level. Moreover, the driving
force provided by the laws to use distributed energy resources has also encouraged a
more decentralized approach to power distribution. DERs or DGs can be used with or
without CPDs.
37
38
CHAPTER III
CUSTOM POWER DEVICES APPLICATIONS IN
DISTRIBUTION SYSTEM
3.1
Introduction
Nowadays, electric power systems are under growing stress, more complex to
operate, and more unstable with unscheduled/uncontrolled power flows and higher losses
[49]. The reason behind is higher demand and less generation as well as constraints on
the construction of new lines. On the other hand, a number of high-voltage power
systems are operating below their thermal ratings because of voltage and stability limits.
Conventional electric power delivery systems, in most cases, are not designed to handle
the control requirements of complex, highly interconnected power systems [50]. This
situation necessitates the review of traditional power delivery methods and practices and
the creation of new concepts. These steps would allow the use of existing lines up to their
full capabilities without reduction in system stability and security [51], [52]. Another
reason that is forcing the review of conventional power transfer methods is the tendency
of modern power systems to follow the changing global economy [53]. Leading trend of
deregulation of electrical power markets due to the changing global economy is
stimulating competition between utilities [54].
Deregulation of electric power industry requires flexible control of power flow that
may be feasibly attained by using innovative power electronics technology.
Developments in power electronics technology have lead to design power electronics
equipment that may be used in high voltage systems [55]. The power electronics
39
involvement in power system has a long tradition. It has been in use for high voltage
direct current (HVDC) power transmission where, electronically controlled power
converters allow the power flows as per operator actions. Hence HVDC line can be used
to its thermal capacity. However, HVDC is expensive for general use, and is only
economical when power systems in different regions separated by long distances are
interconnected.
The parameters and configuration of network are approximately fixed and the
network is slow in response to unforeseen events and is difficult to deal with system load
flow control required in terms of speed and versatility [56]. Energy Management Systems
require fast on-line decision making ability from power system equipment. However, the
mechanical switched equipment in power system can not perform this job. In this
scenario, concept of flexible ac transmission system (FACTS) was presented for the long
term development of power systems of the future [57]. This power electronics based
concept involving switching equipment and controllers was introduced by the electric
power research institute (EPRI) in the late 1980 [58], [59]. The main purpose of FACTS
is to bring a system under control and to transfer power in required network configuration
[60], [61]. It proposes that power in a transmission line can be controlled by varying
three parameters namely impedance, voltage magnitude and voltage angle difference
between the ends of the line [55], [62]. Maximum power flow can then be limited to
transmission line rated thermal limit capacity if less power flows on it or under
contingency conditions when this line is expected to carry more power due to the loss of
parallel line.
40
Just like the use of FACTS devices to solve voltage stability problem and attain cheap
power transfer, the application of power electronic devices can also be effective for the
power distribution systems to enhance the quality and the reliability of power delivered
to the consumers. A compact solution to the present issues that are seen by the utilities
and power distributors can be offered by custom power technology that offers fast
response for the benefit of a customer or group of customers. Custom power (CP)
technology utilizes switching equipment and controllers in order to enhance the
reliability of electrical energy supplied to consumers in the form of less interruptions and
small voltage variations [52]. Hence if this technology is appropriately applied then all
the industrial, commercial and domestic customers may benefit from it.
3.2
Recent Developments
The term power quality involves all possible circumstances in which the waveform of
the supply voltage (or load current) differs from the sinusoidal waveform at rated
frequency with amplitude corresponding to the rated rms value for all three-phases of a
three-phase system [63]. The PQ disturbances involve sudden, short duration deviations,
e.g. impulsive and oscillatory transients, voltage dips, voltage swells, short interruptions,
as well as steady-state deviations, such as harmonics and flicker [18]. However, survey
results pointed out that interruptions of more than 90% in the industries are due to
voltage sags [64]. Generally, PQ issues are either related to the quality of supply voltage
and those related to the quality of the current drawn by the load.
Besides the other causes, voltage dips and interruptions are the main reasons related
to supply voltage issue. These are usually caused by faults in the power system and may
cause tripping of sensitive electronic equipment with disastrous consequences in
41
industrial plants, where tripping of critical equipment can bear the stoppage of the whole
production with associated high costs [65]. One can claim that in this situation, it is the
source that disturbs the load. The second issue is related with low quality of the current
drawn by the load. In this situation, it is the load that disturbs the source. Due to these
two reasons, there is an increasing trend to utilize the devices that serves to compensates
PQ events, particularly that are based on power electronics called Custom Power
Devices [66], [67]. These devices have the ability to provide remedial measures to PQ
problems. Hence, load voltage can be regulated which results in smooth and clean power
flow to the load.
Fig. 3.1 shows basic types and commonly used configurations of custom power
devices (CPDs). Main categories of CPDs are compensating type and network
reconfiguring type.
42
compensate for the variations in the supply voltage. As DVR is series connected device,
hence it may serve as an active series filter [69], [70]. The basic function of a controller
of a series compensator is to detect any abnormalities in the utility voltage such as
sag/swell etc; computation of the correcting voltage; generation of trigger pulses to the
switching element of a DCAC voltage-sourced converter (VSC); correction of any
anomalies in the series voltage injection; and termination of the trigger pulses when the
system event is over [71].
Even though this device has same structure as that of static synchronous series
compensator (SSSC), the operating principles of the two devices differ significantly.
While the SSSC injects a balanced voltage in series, the DVR may have to inject
unbalanced voltage to maintain the load voltage at the rated value in case of an
unbalanced sag in the supply voltage. Moreover, when there is a distortion in the source
voltage, the DVR may also have to inject a distorted voltage to offset the harmonic
voltage.
A. V. Jouanne and his colleagues pointed out that short-duration shallow voltage sag
can be mitigated by improving equipment tolerance characteristics [72]. Whereas, longduration, deep sag can be avoided by changing structure and/or operation of the power
system [73]. Usually, industrial customers do not normally have access to system or
equipment improvement, therefore, the installation of supplementary mitigation
equipment is in general the only solution left to attain the desired quality of supply
voltage at the system-load interface [63], [74].
J. G. Nielson and his colleagues have reported experimental test results that were
obtained on a medium voltage (10 kV) level using a series compensator (DVR) at a
44
distribution test facility [75]. A feed forward and feed back approach has been combined
in the control circuit of the DVR to obtain good responses in case of transient and steadystate disturbances in the supply voltage, for different types of loads. Fig. 3.3 shows the
proposed control structure of the combined feed-forward/feed-back medium voltage
(MV) DVR. On the experimental results, authors proposed implementation of the DVR at
both a low-voltage (LV) level, as well as a medium voltage (MV) level for the protection
of sensitive load from disturbances in the voltage. Control structure of the DVR shows
that d-q technique is used for the detection of sag. However, in this technique, the
selection of the parameters for the low pass filter (LPF) and order of the filter play a vital
role in the filter performance. Hence filtration process gets better with the increase of
filter order, but the response time gets worse and the phase-shift of the output signal
becomes bigger. Also the d-q technique takes at least one-cycle time to track the sagged
voltage [76]. Moreover, this technique indicates some differences in voltage magnitudes
and phase-shifts in the non-faulted phases. In addition, this technique does not provide
acceptable results for unbalanced voltage sags.
45
S.S. Choi and his colleagues have proposed a voltage sag/swell mitigation strategy in
source side voltage [23]. Authors have shown that, in case of sag in voltage energy
injection from the DVR to external system (load) takes place so that load voltage can be
restored to its rated value. However, energy is absorbed by the DVR during voltage swell
from the external system that increases the DC-link voltage of inverter. With the
proposed strategy, voltage sag/swell in load voltage can be quickly compensated
whereas; DC-link voltage can also be controlled, simultaneously. Performance of the
strategy was illustrated using numerical examples. Simulation results have shown that
constant voltage across DC-link can be maintained only by using the zero-power
injection strategy. But, this limits the DVR capability to mitigate deep and long sags.
Moreover, dip (sag) mitigation demands injection of active power as well when load has
high power factor.
Fig. 3.4 Block diagram of the DVR system with proposed controller [77].
scheme used in the DVR to regulate load voltage. Fig. 3.4 shows block diagram of the
proposed control system of the DVR. Authors have shown through simulation studies
and experimental tests that the suggested control technique is better than the existing
technique as improved damping is observed in it. However, improved damping is
obtained at the cost of increased complexity of proposed control scheme. Proposed
closed-loop load voltage and current-mode control has been also applied by the authors
to IDVR (Interline DVR). Mostly, the voltage-restoration process by the DVR requires
active power injection into the distribution system. The ability of a specific DVR
topology, particularly for mitigating long-duration voltage sag, depends on the energy
storage capacity of the DVR. The interline DVR (IDVR) considered in the work restock
DC-link energy storage dynamically. Schematic diagram of an IDVR in a two-feeder
system is shown in Fig. 3.5.
47
The IDVR consists of numerous DVRs that are connected to different distribution
feeders in the power distribution system. The DVRs in the IDVR system have common
DC-link energy storage. When one of the DVR mitigates voltage sag appearing in that
feeder, the other DVRs restock the DC-link energy dynamically. The control strategy for
the IDVR consists of a multiloop feedback control system that uses closed-loop load
voltage and current-mode-control techniques in the two modes of operation. Hence, one
DVR in the IDVR system operates in voltage-sag compensation mode while the other
DVR in the IDVR system work in power-flow control mode. Proposed control system for
DVR-1 in voltage sag compensation mode is similar to that as shown in Fig. 3.4 except
addition of one gain block after voltage error signal. The proposed IDVR system work
efficiently when several DVRs in the system are electrically far apart. The analysis and
experimental results obtained for a laboratory prototype of the IDVR have been presented
to show performance of a two-line IDVR system that can compensate about 40% voltage
sag with long duration.
M.R. Banaei and his colleagues have also utilized IDVR for their proposed control
strategy to restore load voltage with minimum energy injection to mitigate sag [79]. Fig.
3.6 shows the proposed control strategy. However, the presented IDVR systems by M. R.
Banaei et al. and M. Vilathgamuwa et al. require at least one parallel healthy feeder that
limits the application of this system to a typical situation due to increased cost.
Chi-Seng Lam and his colleagues have addressed voltage swell and over voltage
compensation problems in a diode-bridge rectifier supported transformerless-coupled
DVR [24]. Existing DVR voltage boosting method results in a rapid rise in DC-link
voltage in case voltage swell or over voltage occurs. This may damage the storage
48
capacitors and switching devices, and raise the switching loss. This paper shows that the
minimum energy injection strategy cannot handle these problems during unbalanced
situation. A unidirectional power flow control algorithm was proposed with DVR
maximum compensation limit consideration to restrict the DC-link voltage from
exceeding its permissible operation limit.
Fig. 3.6 The suggested control strategy for the DVR [79].
the proposed algorithm over existing minimum energy scheme. However, results
obtained from experimental prototype show that the swell or overvoltage is mitigated
after 50 ms by using the suggested control algorithm. This slow response by the series
device for the proposed control algorithm against swell or overvoltage phenomenon may
damage rectifier unit of rectifier supported DVR. Moreover, it affects sensitive load as
well.
Fig. 3.7 Block diagram of proposed controller for DVR system [80].
H. Kim and S. K. Sul have suggested a method to control DVR compensation voltage
[80]. In order to define tasks for the control of compensation voltage, power circuit of the
DVR has been analyzed. A control structure for the DVR has been proposed by authors
utilizing a feed-forward and state feedback based scheme. Digital control systems (DCSs)
have time delay concern. This issue in DCS of a DVR has been also discussed in the
paper. Usually, control delay in DCSs results from the sampling period, transmission
time of the sensor and the inverter switching frequency, etc. This paper has analyzed and
presented some important design guidelines for the control gains and the inverter
switching frequency of DVRs proposed controller related with the control delay and filter
cut off frequency. Fig. 3.7 shows control system block diagram for the DVR with control
50
delay. An experimental setup with complete DCS has been utilized to validate the
analysis and simulation results for the suggested theory. Simulations for the proposed
controller for DVR system show that output voltage after compensation becomes stable
and shows good dynamics only at the higher switching frequencies. The reason for is that
ripples and the settling time increases at lower switching frequencies.
M. J. Newman and his colleagues have proposed feedback control strategy based on
selective harmonics elimination. It has been suggested by authors that this strategy can be
incorporated with great ease to MV DVR systems to counter voltage harmonics with
negligible effect on the dip mitigation ability of the DVR [81]. In the suggested control
scheme, harmonics in load voltage are eliminated using resonant feedback controller. The
controller is designed by adding resonant control filters to the existing PI control scheme.
Experimental results with a prototype of MV three-phase DVR have been obtained to
verify the proposed controller for various situations, including distortion in supply
voltage, nonlinear load, and operation during voltage sag. These results show that
selected harmonics are effectively regulated without jeopardizing the performance of
existing sag compensator. However, the main problem with the proposed DVR structure
in above mentioned study is that it requires separate filter to eliminate each harmonic, if
system is unbalanced. In case of balanced system this number reduces to half.
P. R. Snchez and his colleagues have presented two control systems for the DVR to
mitigate various PQ problems such as sag, voltage harmonics, and imbalance in voltage
[82], [83]. The proposed control system uses repetitive controller. Only one controller
has been used to deal with all three disturbances, simultaneously. Proposed closed loop
control scheme is shown in Fig. 3.8. Proposed repetitive controller has been claimed to
51
have a fast transient response and zero steady-state error for any sinusoidal reference
input and for any sinusoidal disturbance whose frequencies are integral multiple of the
50/60 Hz. The PSCAD/EMTDC environment has been utilized to perform the modeling
aspects required for the repetitive controller and test system. Results of simulation have
been included only to show that the DVR control approach yield excellent voltage
regulation. Still, hardware implementation is necessary to verify the simulation results
presented for the proposed repetitive controller.
The other system proposed by P. R. Sanchez is based on the repetitive control for a
five-level flying-capacitor DVR. Multilevel converters are suitable in high power
medium-voltage (MV) applications and offer advantages, such as the flexibility in the
choice of switching combinations. The control scheme developed for multilevel
converters find its application in DVR to ameliorate voltage sags as well as other PQ
phenomena, such as voltage imbalances and voltage harmonics. These applications of
DVR require more versatile control system than the classical controller, such as the PI
regulator. The developed control system consists of three subsystems: the first one
eliminates the resonance peak of the filter used in the converter output voltage; while the
second one is the repetitive control, which improves the transient response and zero52
tracking error in steady-state for any sinusoidal reference and for any sinusoidal
disturbance whose frequencies are integral multiples of the 50/60 Hz.. The third
subsystem maintains constant, balanced voltages in the five level flying capacitors.
PSCAD/EMTDC environment has been used to implement the control system, together
with the DVR. Detailed simulation results with an MV test system shows that the control
scheme developed in this paper performs very effectively and provides excellent voltage
regulation. However, the proposed control scheme is complex and its hardware
implementation is difficult.
B. J. Quirl and his colleagues have addressed DVR control strategies that eliminate
propagation of voltage sag or phase-shift to the load [84]. Voltage sag with phase jump is
in some situations more apt to load tripping. Two open loop control methods that protect
the load from voltage sag with phase jump have been explored through simulation in
SIMULINK. Space vector control has been used in the control scheme of the DVR. Fig.
3.9 outlines the control methodology for the DVR. The first method is described as presag to in-phase compensation and monitors the supply such that initial pre-sag conditions
are used, but it adjusts approximately after 25 ms to inject a voltage that is in phase with
the new supply voltage. This method includes compensation during the sag and after the
sag has left the system. The second method makes use of the condition before sag.
Ideally, with this approach, the load voltage remains same as before voltage sag. These
compensation methods are very effective in keeping the load voltage at the correct level
and phase-shift free. The benefit of the pre-sag to in-phase compensation is that the
magnitude of injected DVR voltage is finally minimized for constant load voltage
53
magnitude. This can be beneficial; however, in case of deep sag, the voltage rating of the
DVR is a concern.
In case of the simulated sag, both methods need active power injection, but the presag to in-phase compensation involve the development of less instantaneous active power
towards the end of the sag than that of the pre-sag compensation. However, the pre-sag to
in-phase compensation also includes active power injection after the sag, while the presag method only requires compensation during the sag. The important advantage of the
pre-sag method is that phase-shift is not seen by the load. However, this would be the
best method if the sag duration is relatively small and power drain on the DVR is not a
concern.
Fig. 3.9 Open loop control methodology for the DVR [84].
54
decrease the power factor during sags. In this way, operating limit of DVR and its sag
mitigation duration is increased. This inturn increases the DVR operating limit and
mitigation interval. Fig. 3.10 shows the proposed control structure of DVR. The proposed
DVR control system has the ability to reduce the active power injection and, thus,
increases the sag mitigation limit in situations where sag mitigation also requires active
power injection. Operation of the new device for the DVR proposed strategy has been
investigated using PSCAD/EMTDC simulation program as well as by experiments
performed on a laboratory hardware prototype. Although results of investigation show
the feasibility and the practicality of the suggested DVR control system, however,
presented simulation results show that load voltage after sag mitigation takes 3-5 cycles
to reach steady-state value. Hence settling time for the load voltage after sag mitigation
needs to be improved for the proposed DVR control system.
E. K. Kenneth Sng and his colleagues have presented the control of a transformerless
and self-charging DVR using a capacitor bank for energy storage to mitigate voltage sag
[87]. The proposed DVR works without an isolation transformer for series injection.
Moreover, it is capable of self charging under normal supply conditions, without the need
for a separate charging circuit and transformer. The voltage recovery and self-charging
processes of the new restorer have been explained in detail and validated through
simulation. The study of the DC-link voltage regulation dynamics has revealed the
existence of a nonlinear element in the form of magnetic energy of the filter inductor.
This affects stability for high feedback gains. Verification of this aspect on regulator
performance and closed loop stability problem has been carried out and a cancellation
technique has been proposed. The proposed controller is simple to implement using
56
either digital or analog devices. This allows flexibility of choice in the face of device cost
or robustness constraints. Simulation and experimental results have been provided for a 1
kVA prototype to show the usefulness of the restorer design on series compensation and
the self-charging process, the effects of the nonlinear element in regulator dynamics, and
the effectiveness of the proposed cancellation method. However, this technique is
effective in enhancing stability when higher feedback gains are used. But, if uncertain
parameters are present in the nonlinear term then higher feedback gains in proposed
controller may lead to instability [88].
57
58
A new reference voltage generation algorithm for a DVR has been presented by A.
Ghosh and A. Joshi. It uses the method of instantaneous symmetrical components [90].
Series injection of reference voltage in distribution feeder is made by a VSC. It has the
ability to regulate load voltage at its rated value against harmonics or imbalance in the
supply voltage. Fundamental positive-sequence voltage components have been extracted
to facilitate the DVR to produce balance steady-state voltage within a half-cycle. The
proposed algorithm does not require injection of active power from the series
compensator during its steady-state operation. The reference voltage generation
algorithm has been investigated and verified using computer simulation studies. In it,
dynamic voltage restorer has been assumed as an ideal voltage source which tracks the
reference voltage. The DVR algorithms discussed in this work are suitable only for
systems with balanced and linear loads. However, the authors in another work have
presented the operating principles, structure, and control of a DVR that is supplied by a
DC capacitor to protect sensitive but unbalanced and/or distorted loads [91]. Harmonics
components in the source and load currents for the desired task have been reduced by
allowing them to pass through a low impedance path. This path is provided by a shunt
capacitor filter in the proposed hybrid structure of an ideal DVR. Three different DVR
configurations have been examined for performing the desired task. However, best
performance is obtained when series device is connected with the shunt capacitor.
Performance of the DVR with shunt capacitor configuration is verified through
simulation results. However, it has revealed that DVR success depends on the choice of
shunt capacitor and DC-link capacitor. Their values must be chosen carefully. Also, the
proposed algorithm for the reference voltage generation allows variation in phase-angle
59
of load voltage; hence above DVR system cannot be used for the loads that are sensitive
to phase-angle shifts.
Il. Y. Chung and his colleagues have proposed a new method of voltage restoration
that is based on load voltage tolerance to inject minimum energy during DVR
compensation period [92]. Generally, voltage tolerance range of almost all the
distribution system load varies. However, capability of DVR to inject real power to
mitigate voltage sag is limited. Usually, phase advance technique is used for the DVR to
supply minimum energy during sag compensation. In this technique, load voltage phaseangle serves as a control parameter in phase advance technique. However, this technique
has limited capability to optimize energy injection to mitigate deep voltage sag. The
proposed method utilizes phase advance angle as well as magnitude of load voltage in the
voltage tolerance area in order to optimize injected active power of DVR. In addition,
authors have also proposed objective function, equality constraints, and inequality
constraints in order to have control parameter. Simulation for the proposed method for
several cases is performed using EMTDC/PSCAD package to verify that the proposed
technique has a merit to optimizing injected real power compared with existing methods,
especially, during deep sag. However, load voltage phase still varies as is the case in
phase advance method which restricts its applications to locations where load is not
sensitive to phase-angle variations.
N. Athanasiadis and D. V. Bandekas has described a technique for the development
of a DVR model for use within distribution system to provide ride-through capability for
an AC variable speed drive under voltage sag conditions [71]. It has been suggested that
presented closed loop control system can be utilized as a potentially promising design
60
method for efficient regulation and control of the DVR. Fig. 3.11 shows the presented
control scheme for the DVR. However, only simulations for the proposed DVR model
and its control system have been presented using electromagnetic transients simulation
program (EMTDC). It does not guarantee that integrated power electronics environment
can be used for successful implementation of the DVR. Experimental investigations are
required to validate the proposed DVR model and its control system.
Literature review reveals that there are numerous methods to inject DVR
compensating voltage to distribution system to serve a typical or multiple PQ issues.
These methods are: pre-sag compensation, in-phase compensation, and phase advance
method [93]-[96]. A brief description of these methods is also provided in chapter IV.
The worlds first DVR was installed in August 1996 at a 12.47 kV substation in
Anderson, South Carolina to provide protection to an automated rug manufacturing plant.
61
However, it was first installed at the Waltz test facility near Pittsburgh for full power
tests [97]. Florida Power Corporations 2 MVA VSC-based DVR provides protection to
one of the six 12.47 kV feeders at the Econ substation (230/12.5 kV) of Orlando, Florida.
This DVR is placed in service in 1996 in a high density residential and commercial area
[98]. The next commissioning of a DVR was in February 1997 at a 22 kV distribution
system at Stanhope, Victoria, Australia [99]. This project was carried out to protect the
diary milk processing plant. Another installation of a DVR was done in April 1997 at the
Sappi Limited, Stanger Mill in South Africa that provides pulp to Sappis paper making
process. The DVR is supplied by superconducting magnetic energy storage (SMES) [98].
Two DVRs, each having rating of 6 MVA, 12.47 kV have been installed in July 1998 at
an important industrial place on the Salt River Project system at Phoenix metropolitan
area in Arizona. Each of these DVRs can boost a 20 MVA load as much as 30% [98]. In
April 1998 a 4 MVA series compensator is installed by Scottish Power at the Caledonian
Paper Mill at Irvine, Scotland [98]. The power to the mill is supplied by a South African
electricity public utility ESKOM. Platform mounted DVR was first installed in 1999 to
provide protection to Northern Lights Community College and several other loads in
Dawson Creek, British Columbia, Canada [100].
3.2.1.2 Reactive Power and Harmonic Compensation Devices
Fig. 3.1 shows that second type of compensating CPDs are called Reactive
power and harmonic compensation devices. Generally, these devices are connected
in shunt at the load bus, with the objective of injecting a current to the load.
Reactive power and harmonic compensation CPDs include static VAR
62
PQ problems may arise from nonlinear customer loads such as, arc furnaces, welding
operations. Voltage flicker and harmonic problems can affect the entire distribution
feeder. Various devices have been considered to reduce the impact of these variations.
The main concept is to use dynamic capacitance and reactance to have steady-state
operation of the power system. This is particularly achieved by using static switching
devices to control the capacitance and reactance to stabilize the power system. SVC and
the D-STATCOM are the mostly used flicker mitigation devices [63]. However, series
capacitor and series line reactors are also used for this purpose. Moreover, literature
studies show that reactive power and harmonic devices have been also investigated for
voltage sag, swell and phase jump mitigation as well.
Series Capacitors and Series Line Reactors
installed in distribution networks. Although they were initially intended for installation in
transmission grids but only a few of them has been applied. They increase the voltage
stability limit through reduction in the line reactance and maintaining constant voltage
irrespective of quick load changes. Hence, they can decrease power losses and line
voltage drop in a distribution feeder. However, constant voltage can be maintained only
in a certain range of power change, depending upon the equivalent reactance of series
capacitors controlled by thyristors [102].
By inserting a linear reactor in series with the arc furnace, the short-circuit current of
the furnace is reduced owing to the higher total impedance. Moreover, increased
reactance in the circuit results in larger phase-shift between voltage and current and
hence a more stable electric arc in the furnace. But, the selection of series reactor must be
done carefully as high value of the series impedance will decrease the furnace power and
thereby the steel production rate. Saturatable reactors in series have also been used to cut
current peak due to for instance short-circuit in the arc furnace [103]. It has been
observed that independent usage of series capacitors and series line reactors without
thyristor control have no effect on voltage sag, swell or phase jumps mitigation.
However, Thyristor controlled series capacitors and reactors can be used for voltage sag
mitigation as discussed in section 3.2.2.1 under heading Solid State Current Limiter and
Solid State Circuit Breaker.
Static VAR Compensators
The SVC can be used for AC voltage control by generation and absorption of reactive
power through passive elements. It can also be used for balancing unsymmetrical loads.
Usually, an SVC is implemented with a thyristor controllable reactor (TCR) and a
64
number of thyristor switched capacitor (TSC) branches as shown in Fig. 3.12 [52]. In
addition to TCR and TSCs branches, VSC circuit also contains tuned filters to suppress
harmonic current from flowing into the AC system. Moreover, there are firing and
control circuits which are not shown in the Fig. 3.12.
The SVC is connected in shunt to an AC line through a step down transformer. Due
to this arrangement, the SVC has the ability to generate continuously variable reactive
power in a specific range, and the size of the TCR is limited to the rating of one TSC
branch. Evidently, the reactor size limits the power that can be absorbed in the inductive
range. SVCs placed closed to load can be very effective in providing voltage support,
thereby avoiding voltage instability [5]. Although, electrically switched shunt capacitors
(TSCs) in SVC branch can supply variable reactive power to the system, however,
65
reactive power output of a capacitor is proportional to the square of the system voltage
which may reduce its effectiveness in high and low-voltages.
The SVC has been found in applications such as power line compensation [104],
compensation of traction system [105], reducing disturbance from rolling mills [106] and
arc furnace compensation [107]. The potential to absorb changes in reactive power makes
to some extent the SVC appropriate for flicker reduction. In this situation, the SVC
usually consists of a TCR branch with a filter. An SVC installed together with an arc
furnace not only reduces the flicker, but also, increases the steel production and its
quality due to stabilized AC voltage [108]. However, SVC has a low control update rate.
Therefore, the capability of the SVC to mitigate flicker is limited. Also, it has been
reported to have slow response time to compensate sag.
P. Wang and his colleagues experimentally investigated voltage sag mitigation by an
advanced static VAR compensator (ASVC) [109]. The results of this investigation have
shown that the ASVC can completely compensate shallow sag, and considerably lift the
sagged voltage during deep sag. However, due to opted sag mitigation strategy, post sag
voltage overshoots. Response time of ASVC for sag mitigation has been measured about
two and half cycles of the power frequency. This shows that response of the device is
quite slow and some of the sags cant be mitigated in this time frame according to
Semiconductor Equipment and Material International, SEMI- F47 standard. Moreover, it
was further investigated that phase-angle jumps can disturb the working of the ASVC,
and offset its mitigation outcome on voltage sags.
SVC sag mitigation capability is strongly dependent on its ability to provide variable
reactive power to the system. However, in case of a sag mitigation for high power factor
66
loads where active power contribution is also required to compensate sag then SVC lacks
in its operation to achieve the desired sag mitigation task. Hence, due to the pitfalls
associated with SVCs, the D-STATCOM was introduced into the market as an alternate
for it.
Distribution Static Compensator
used in the context of FACTS at the transmission level [110], custom power controllers
at the distribution level and in end users electrical installations [111]. A typical
configuration of a VSC based D-STATCOM is shown in Fig. 3.13 [112]. The DSTATCOM has emerged as a promising CPD to provide not only for voltage sag
mitigation but a host of other PQ solutions. Important applications of it include
voltage regulation, load balancing, power factor correction, harmonic filtering, and
flicker mitigation [113].
C. K. Sao and his colleagues have provided a bench mark D-STATCOM that
has digitally controlled VSC for studying the sinusoidal pulse-width-modulated
(SPWM) D-STACOM [114]. Authors have also described principle of operation of
D-STATCOM. The proposed multifunctional controller has been investigated and
67
it has offered DC-bus voltage control, power factor correction, and AC voltage
control. Block diagram of complete D-STATCOM control system is shown in Fig.
3.14. Performance of D-STACOM in various modes, including power factor
correction mode and AC voltage regulation mode has been analyzed through
simulations with PSCAD/EMTDC software. Simulated results have been also
experimentally verified on a 10-kVA laboratory D-STATCOM that have shown
excellent match between the experimental and simulated results. This provide a
validated benchmark model for simulation studies of the pulse width-modulated
D-STATCOM. Block diagram of the D-SATCOM shows that phase-locked loop
(PLL) technique is used for voltage sag detection and mitigation. However, this
technique provides good results only if voltage sag is not coupled with phase-angle
jump.
Fig. 3.13 Schematic representation of the D-STATCOM as a custom power Device [112].
68
G. Ledwich and A. Ghosh have discussed various topologies and controls of a DSTATCOM for the voltage or current control mode, load compensation at a weak AC
bus, and voltage regulation for a mini custom power park [113], [115]-[116]. In the
voltage control mode D-STATCOM has forced the voltage of a distribution bus to be
balanced sinusoids. Moreover, it has cancelled distortion caused by the load, such that
current drawn by the compensated load is pure balanced sinusoid in the current control
mode. Three single-phase VSCs working in parallel to a filter-capacitor have been
connected in chosen D-STACOM topology. A switching control scheme has been
adopted and its suitability has been verified through simulations for the voltage and
current control modes of D-STATCOM. However, operation of a D-STATCOM for
69
weak AC buses will result in distortions in the line current or the voltage at the point of
common coupling. The line current distortion has been eliminated using the fundamental
voltage of the PCC, but the distortion in voltage cannot be eliminated without adding a
filter capacitor in parallel with the D-STATCOM. However, the addition of the filter
capacitor generates control issues and complicates the tracking problem, as standard
controls are not appropriate to be used in this situation. A switching control scheme was
then proposed that depends on the extraction of the reference signals to solve the tracking
problem. The reference current generation scheme along with extraction of fundamental
component of PCC voltage constitutes the reference of this controller. Computer
simulation studies have been carried out to validate the performance of the proposed
scheme. However, the employed reference signal generation method also suffers from the
problem of harmonic contamination due to VSC of D-STATCOM. Although authors
have used only capacitors to filter the harmonics produced due to VSC switching,
however, LC filters can be used for this purpose as well.
A configuration of the mini custom power park in which the voltage inside the park is
power park, which shows satisfactory results. The above discussion clearly demonstrates
that the D-STATCOM can hold the custom power park bus voltage during moderate
voltage sag. However, for a very deep sag or fault in the preferred feeder, the supply has
to be switched to the AF through the static transfer switch (STS). Moreover simulation
results show that CPP voltage bus is regulated after 20 ms. This shows that controller
response time needs to be improved in order to respond disturbances in supply or load
side.
control loop. Block diagram of closed loop voltage control is shown in Fig. 3.16. The
load bus voltage magnitude is chosen as nominal value, i.e., 1.0 p.u., while its phaseangle is attained through a feedback loop that regulates the voltage across the DC storage
capacitors. It has been shown through detailed simulation and experimental results, that
the D-STATCOM is capable of regulating the PCC voltage against any unbalance and
distortion in either on the load or supply side. However, the employed dead-beat control
scheme for the VSC to maintain the AC bus voltage is very sensitive to system
parameters and it must be carefully used even if it has a very fast convergence property.
W. Freitas and his colleagues have presented a dynamic study about the simultaneous
usage of AC generators and D-STATCOM devices on the dynamic behavior of
distribution networks [119]. The performance of a D-STATCOM as a power factor
controller or a voltage controller has been analyzed. D-STATCOM voltage controller is
shown in Fig. 3.17. The controllers impact on the stability performance and protection
system of distribution networks with distributed generators has been determined.
Simulation results have shown that a D-STATCOM voltage controller can improve the
stability performance of induction generators significantly. Whereas, a D-STATCOM
power factor controller may adversely affect the stability performance of synchronous
72
generators. Moreover, simulation results shows that although load bus is maintained at
constant value but D-STATCOM response to stabilize load voltage is poor and it took 35
cycles to attain steady-state value. This shows that D-STATCOM controller performance
needs to be improved.
Elnady and his colleagues have introduced a unified approach for the mitigation of
the voltage sag and voltage flicker by the D-STATCOM using Kalman filter [KF] and its
derivatives [120]. Also, the authors of the paper have shown the advantage of using the
Kalman filter instead of the existing tools for tracking and extracting voltage
disturbances. The Kalman filter is an algorithm used as a tool to extract both the
instantaneous envelope of the voltage sags, and to extract the instantaneous flicker level
(IFL) of the voltage flicker. The proposed modular D-STATCOM has been investigated
73
to show its ability to mitigate voltage sags and voltage flicker, even if the voltage
disturbances are unbalanced. The presented mitigating approach that depends on KF is
fast and accurate for the tracking and extraction of the voltage disturbance. Therefore,
unbalanced voltage sags can be mitigated.
Control block diagram of proposed algorithm is shown in Fig. 3.18. Different levels
of voltage sags and a voltage swell has been compensated simultaneously. Also, the
cyclic voltage flicker is mitigated efficiently and the IFL is reduced by approximately
60% of its value before the compensation. Consequently, voltage flicker that occurs due
to an arc furnace is transferred from the irritation region to lower than the observable
region, based on IEEE Std. 141-1993, and does not disturb the other customers.
Simulation results have been shown to demonstrate the mitigation of unbalanced voltage
sags and the compensation of the voltage flicker by employing the proposed algorithm.
74
However, the presented mitigation approach is complex and its hardware implementation
is difficult.
N. Marium and his colleagues have presented the design of a prototype D-
using multipulse VSCs. The algorithm has been tested by carrying out detailed
simulation using PSCAD. Simulation results have shown good dynamic performance
especially in the case of unbalanced load compensation. However, as the d-q method is
used for sag detection in the control algorithm, it provides good results for balanced
voltage sag. Also use of d-q approach takes 2 cycles to detect the magnitude or envelop
of sag with a little oscillations. This shows that if voltage sag magnitude is more than
50% of nominal load voltage then according to SEMI-F47 standard, sensitive load will be
tripped as voltage magnitude should be 50% of rated load voltage from 1st cycle to 10th
cycle of supply voltage. Hence sag detection and mitigation time for the proposed control
algorithm needs to be improved according to SEMI-F47 standard to avoid tripping of
sensitive load for the deep voltage sag.
A. Shukla and his colleagues have presented the implementation of D-STATCOM
using multilevel inverter [123]-[125]. Five-level flying capacitor multilevel inverter
(FCMLI) is a multiple voltage level inverter scheme that has been designed for high
voltage and power operations with low distortion. Capacitors, called flying capacitors
have been used in FCMLI for clamping the voltage across the power semiconductor
devices. A current control technique has been derived for controlling the injected current
by the FCMLI-based D-STATCOM. A new scheme has been proposed for controlling
the flying capacitor voltage such that their voltage remains constant and at the same time
produces the required output line currents under the current control operation.
PSCAD/EMTDC environment has been used to perform simulation studies to validate
the effectiveness of the proposed control scheme and the FCMLI-based DSTATCOM.
The simulation results show that the proposed D-STATCOM provides acceptable
76
Fig. 3.19 State feedback control with a nonlinear element in the forward path [124].
simulations
using
MATLAB
software
with
its
Simulink
and
SimpowerSystemsTM toolbox. Fig. 3.20 shows that d-q technique is used for driving the
reference control signals, but this technique does not provide good results for unbalanced
voltage sags. Also three phase PLL is used to synchronize load current in dqo frame with
the source voltage. However, PLL approach provides good results if voltage sag is not
accompanied with phase-angle jump.
Comparison of D-STATCOM and DVR shows that shunt compensator (DSTATCOM) protects the utility system from the ill effects of customer loads with the
79
Fig. 3.20 Synchronous reference frame theory (SRFT) based control of D-STATCOM [126].
Current harmonics from the grid are removed by installing, passive or active shuntfilters. Passive filters for harmonic minimization provide low impedance paths for current
harmonics. Hence, the current harmonics are fed into the shunt filters and they do not
return to the supply. Construction wise passive filter consists of series LC filters tuned
for specific harmonics, usually associated with a high pass filter used to wipe out the rest
of the higher-order current harmonics. Passive filters have the drawbacks that they are
strongly dependent on the system impedance, which depends on the distribution network
80
configuration and the loads [63]. Consequently, the system impedance, which changes
continuously, strongly influences the filtering characteristics. An unwanted resonance
can take place between the filter and the system in the worst case. This may compel the
passive filter to act as a sink for harmonic currents from other sources in the grid.
Hence, the passive filter can be overloaded by a current higher than the rated value.
Ultimately, the capacitors of the passive filter generate reactive power that may not
necessarily be needed for power factor correction. Literature studies show that when
passive filters are used independently then they are unable to mitigate voltage sag.
Active filters or active power filters (APFs) can have connection configuration either
in series or shunt. But, a shunt active power filter configuration is preferred over series
because of greater ease of protection. It is represented by a controllable voltage source
behind a reactance acting as a current source. The VSC based shunt active filter is mostly
used, due to its well-known topology and straightforward installation procedure. It
includes a DC-link capacitor, power electronic switches and filter inductors between the
VSC and the grid, shown in Fig. 3.21.
81
active shunt and series filters have not been investigated to compensate voltage swell and
phase jumps.
Unified Power Quality Conditioner
The literature survey shows that the D-STATCOM is proposed for compensating PQ
problems in the current whereas DVR is used for mitigating PQ problems in the voltage.
However, when D-STATCOM and DVR are combined to provide a multi purpose
compensator for solving current and voltage PQ problems then this concept has been
given the name of unified power-quality conditioner (UPQC) [63]. Although, literature
studies reveal that APLCs were rebranded as UPQC. Schematic block diagram of singlephase UPQC is shown in Fig. 3.22. A UPQC is a device that is similar in construction to
a unified power flow controller (UPFC) [133]. The concept of UPFC, as the combination
of STATCOM and a SSSC, was first introduced in 1991 by Gyugi [134]. The UPQC like
UPFC utilizes two voltage-sourced converters (VSCs) which are connected to a DC
energy storage capacitor. Out of these two VSCs, one is connected in series with AC line
while the other is connected in shunt with the AC system. A UPFC is used in a power
transmission system to provide shunt and series compensation at the same time. In the
similar analogy, a UPQC can also be utilized to provide shunt and series compensation in
a power distribution system. But, operating principles resemblance of these two devices
comes to an end at this point. Operation of a power transmission line is characterized by
balanced, distortion (harmonic) free environment. Hence, a UPFC must only offer
balanced shunt or series compensation. On the other hand, a power distribution system
may have unbalance, distortion and even DC components in the load bus voltage. Hence
83
UPQC have to function with all these aspects so as to provide shunt or series
compensation [135]. However, due to its high cost, it is rarely used in its combined form.
V. Khadkikar and A. Chandra have explored a new idea to balance the load-reactive
power demand using a three-phase UPQC [136]. Usually shunt converter of the UPQC is
employed for load-reactive power compensation, whereas the series converter handles
voltage-related problems. Authors of this paper have presented a new application of
UPQC in which both the series and shunt APFs provide the load-reactive power demand.
This property facilitates the UPQC APFs not only to share the load-reactive power
demand, but also helps to reduce the shunt APF rating, and hence, the overall cost of
UPQC. In this way series converter is effectively utilized during normal operating
condition. Power angle control (PAC) concept has been utilized to achieve the desired
objectives. Fig. 3.23 shows the block diagram that has been used to determine
84
A. Ghosh and G. Ledwich have discussed the operation and control of a UPQC that
can be used simultaneously in voltage or current control mode in a power distribution
system [137]. The series component of the UPQC called DVR forces the voltage of a
distribution bus to be balanced sinusoids and free of distortion in the voltage control
mode. At the same time shunt component of UPQC called D-STATCOM performs load
compensation resulting in the drawing of balanced sinusoidal currents from the
distribution system bus in the current control mode. Both these objectives have been
85
have verified the simulation results. The proposed UPQC system with DG has the ability
to enhance the power quality at the point of installation on power distribution system or
industrial power system. However, if the diesel generator is replaced with PV source and
battery storage system then it would be economical and environment friendly.
Moreover, as cost of UPQC is already quite higher as compared to DVR or DSTATCOM and its cost would be further increased if Diesel generator is connected as
DG with DC-link through rectifier. Hence, cost versus benefit analysis should be carried
out first to opt the proposed UPQC configuration with DG for the required application.
M. Basu and his colleagues have explained two control scheme models for UPQC, for
improving PQ of sensitive nonlinear loads [140]. Designed control schemes are based on
87
two different kinds of voltage compensation strategy for series converter (SEC) of
UPQC. These schemes are termed as UPQC-Q and UPQC-P. UPQC-Q has the benefit of
VAR sharing between the two compensators. The SEC, while injecting voltage to
mitigate the supply voltage sag, shares a part of VAR of the load and does not consume
any active power. A disadvantage of UPQC-Q is that it cannot mitigate unbalanced
voltage sag. But for situations where reactive power requirement is very high UPQC-Q
could be a potential control scheme for action. Besides voltage sag compensation, the
SEC control scheme of UPQC-P has the ability to mitigate the supply voltage unbalance
problem as well. But, this control scheme utilizes d-q technique to generate reference
signals for unbalanced sag mitigation that suffers from the drawback that it does not
provide satisfactory results for unbalanced voltage sag mitigation. The reason is that the
unbalanced voltages in synchronous reference frame (SRF) develop two components (d
and q).
frequency [141]. These oscillations in d-q components are not suitable for control
purpose. It has been observed that under balanced voltage sag condition, the load voltage
angle after compensation is not altered. Hence, the shunt converter (SHC) of the UPQC-P
does not require compensating any additional VAR created due to SEC control action. A
comparative analysis of VA loading of SEC, SHC and total loading of UPQC has been
carried out. The performance of the two control schemes has been shown through
extensive simulation using the software SABER. Since the power circuit design of UPQC
remains same for both the model, the usage of UPQC can be optimized depending upon
the application requirement.
88
Y. Y. Kolhatkar and S. P. Das have presented an optimized UPQC, which aims at the
integration of series active and shunt active power filters with minimum volt-ampere
(VA) loading of the UPQC [142]. The series active filter is a DVR, which regulates the
voltage at the load end with minimum VA loading of the overall UPQC by injecting the
voltage at an optimum angle. Closed loop block diagrams of DVR as well as DSTATCOM in the proposed scheme are shown in Fig. 3.25 and Fig. 3.26 respectively.
The D-STATCOM has been activated in the current control mode to stabilize the DClink voltage, supply the load reactive power, and eliminate harmonics from the supply
current.
The control scheme flexibility for the sudden load change during the sag
condition has also been investigated with the help of simulation as well as through singlephase experimental setup in the laboratory at 120 V, 0.6 kVA. It has been shown that
voltage can be restored in 12 cycles after the occurrence of the sag. However, angle
closed loop may take another 23 cycles to match with calculated voltage value.
Consequently, the optimization process starts a little later.
89
A. K. Jindal and his colleagues have illustrated the operation and control of a new
connection for a UPQC to enhance the PQ of two feeders in a distribution system [143].
Connection of this device between two independent feeders originating from different
substations has been explained in this paper. As UPQC is connected between two
different lines, hence the device has been given the name interline UPQC (IUPQC) as
shown in Fig. 3.27. The proposed configuration of UPQC serves to control the bus
voltage of one of the feeders while regulating the voltage across a sensitive load in the
other feeder. Different case studies have been carried out to verify the efficacy of the
proposed configuration through simulation using PSCAD/EMTDC. On the other hand,
proposed configuration of UPQC always requires two parallel feeders to be present. This
restricts its application to specific situations that have two, side by side running parallel
feeders.
3.2.1.3 Backup Stored Energy Devices
Fig. 3.1 shows that third type of compensating CPDs are backup stored energy
devices. Usually, a D-STATCOM is used for flicker mitigation and active filtering
90
Sensitive load is isolated from the power system by a static switch on sensing a
disturbance and is fed by the VSC. As mentioned in chapter II, there are several devices
for storing the necessary electrical energy e.g. DC storage capacitors, batteries
(transportable battery energy storage system, TBESS [144]), supercapacitors, flywheels,
and superconducting coils [145]. These backup energy storage devices can manage to
supply up to seconds or even minutes of backup power. This permits sensitive loads to
continue operation during complete power interruptions. TBESS have the capability to
compensate short term PQ variations and enhance reliability provided they are configured
with longer-term backup generation. The above mentioned storage devices are shortly
described in a Table 3.1 along with their merits and demerits [146]:
91
92
Sr.
No.
1
Table 3.1
Backup storage devices and their merits and demerits [146].
Type of backup stored
Merits and demerits
energy devices
DC storage capacitors
Store energy in their capacitance.
Useful for short ride through times.
Require DC/DC converter between the constant voltage
bus and the capacitance.
Cost increases with the increase in ride through time.
Batteries
Most common method of storing energy.
Do not require DC/DC converter as they are directly
connected to VSC.
Capacitor can compete with batteries but only for short
ride through times.
Utilize environmentally unfriendly materials.
Have limited life time.
Require regular maintenance.
Some new types of batteries do not have above mentioned
limitations but have higher cost.
Supercapacitors
Energy densities comparable to batteries.
Improve equipment voltage tolerance.
Have much longer lifetime than batteries.
Require much less maintenance than batteries.
Discharge time is not less than 1 minute.
Faster than batteries but much slower than capacitors.
Only available for voltages of a few volts.
Flywheels
Store energy in fast-spinning flywheels.
Stored energy cannot be extracted fully.
Require an additional DC/DC converter.
Superconducting coils
Energy is stored in superconducting magnetic energy
storage (SMES) coils.
Most cost attractive solution for high power short time ride
through applications.
Fast extraction of energy as compared to batteries.
Have reduced size and lower maintenance cost as
compared to batteries.
Can be quickly and easily installed with short lead times.
Have modular design to meet future load growth and are
portable [145], [148], [149].
Require an additional DC/DC converter between SMES
and constant voltage bus (Fig. 3.29).
M. Moschakis and his colleagues have presented some of the operational and design
characteristics of distributed superconducting magnetic energy storage (D-SMES) device
[146]. The main parts that a D-SMES includes: VSC, D-SMES system, DC-DC
converter, coupling transformer, and harmonic filter. Numbers of configurations and
operational techniques for D-SMES have been discussed in the paper. The merits and
93
demerits of SMES compared to other solutions have been provided. The issues that a DSMES can deal with have been listed and discussed. However, authors in the present
work have focused only on the ability of a D-SMES to compensate voltage fluctuations
and sags (dips) and support a sensitive load during power system outages or
interruptions. Simulations results have been shown using the EMTDC simulation
package that demonstrates the ability of D-SMES to mitigate voltage fluctuations caused
by load variations. Moreover, it has been shown that D-SMES can safeguard a sensitive
load from faults on the distribution or the transmission system when it is accompanied
with a static switch. Hence, the load faces no voltage sag during the fault. However,
without the uses of static switch upstream the D-SMES system, the D-SMES would
require enormous amount of energy storage to adequately support the voltage of a
sensitive load [18]. But, the use of static switch requires an alternate feeder as well.
Moreover simulations in the paper have been performed with the assumption that DC
side of the SMES system is replaced by a DC voltage source. Hence amount of energy
that is available in SMES is not taken into consideration and the response in the demand
for real power is instantaneous. However, this response in a real D-SMES unit would not
be instantaneous but it will have delay of few milli seconds.
Two possible functions of flywheel energy storage systems (FESS) are voltage
support and the uninterruptible power supply. These applications are helpful to protect
sensitive loads on distribution feeders. Robert S. and his colleagues have looked at the
possibility of combining voltage support and the uninterruptible power supply functions
by considering a flywheel energy storage system (FESS) [147]. Authors have proposed
FESS system that performs the above mentioned functions by adopting a new control
94
Normally, Backup energy stored devices available in the market can be considered
modification of UPS. These devices can be classified on the basis of three different
factors [150]:
1. Type of storage used
2. Duration of load support (or back-up time)
3. Mode of connection (online, offline, line interactive)
95
In a conventional online energy storage system, the protection unit supplies the load.
Protection unit consists of a diode rectifier followed by a VSC as shown in Fig. 3.30. The
incoming AC power is converted into DC power that is applied to charge a bank of
batteries or some of the other energy storage mediums. This DC power is then converted
into AC power to feed the load. If the incoming AC power fails, the VSC is fed from the
batteries and continues to supply the load. Generally, UPS systems are designed to
provide 5 to 15 minutes of backup. In addition to providing ride-through for power
outages, an online UPS system provides isolation of the critical load from all power-line
disturbances.
An offline energy-storage system allows the utility to power the protected loads until
a disturbance is detected and a static or a mechanical switch transfers the load to the
battery or other energy-storage backed VSC. Since there is a short-duration interruption
during the time it takes to detect a mains failure, start the VSC and transfer the load to
battery power. A load with some inherent ride-through capability is required for the
interruption to go unnoticed. A standby UPS using a static switch rather than a
mechanical switch can provide nearly seamless transfer from utility power to battery
power during utility interruption.
The line interactive energy-storage system is a variation of the offline system in
which the design includes some means of load voltage regulation. This can be
accomplished by integrating some of the technologies such as an injection transformer or
static voltage regulator [150]. In theory, installing an uninterruptible power supply (UPS)
is the easiest way to protect sensitive processes against all sags. However, these batterydependent devices are often expensive and energy insufficient and they require
96
maintenance that may exceed the owners available in-house resources. Hence, due to its
considerable purchase and maintenance costs, UPS equipment is installed on a structural
basis only in places where the damage resulting from power supply problems is very
high, such as in hospitals etc. Amongst the CPDs, UPS and DVR have the ability to inject
a voltage waveform to the distribution line. However, DVR costs less as compared to
UPS. While evaluating the UPS and DVR; the UPS is constantly providing the full
voltage to the load irrespective of whether the waveform is distorted or not. As a result,
the UPS is always working at its full power. In contrast, the DVR injects only the
difference between the pre-sag and the sagged voltage and that too is only during the
sagged period. Hence operating losses of the DVR and the power rating are very small as
compared to the UPS. Hence DVR is deemed as a power efficient device as compared to
the UPS.
that network reconfiguring type equipment are gate turn-off (GTO) based or thyristor
based switches [1]. Network reconfiguring devices include solid state current limiter
(SSCL), solid state circuit breaker (SSCB) and solid state transfer switch (SSTS).
However, most popular network reconfiguring device is SSTS. The solid state transfer
switch or static source transfer switch (SSTS) uses solid-state switches to provide an
almost seamless load transfer to an alternate feeder/source to protect a sensitive load
from momentary interruptions and voltage sag/swell or fault in the supplying feeder. The
transfer time can be as low as quarter of the rated frequency cycle [22].
97
SSCL used to limit fault current in power system have two types.
GTO-Switched
Thyristor- Controlled
In the GTO based switched device, fault current-limiting impedance is shunted across a
pair of anti-parallel GTO switches. It comes in series with a faulted circuit as soon as the
fault is sensed. In this way large fault current is interrupted by turning off the GTO
switches during fault. However, this leads to high transient overvoltage and nuisance
tripping of small PWM adjustable-speed-drives [151]. Moreover, the transient
oscillations initiated by the operation of GTO based switched SSCL excite LC circuits
formed by the customer step down transformers and the low-voltage power factor
correction capacitors at the customer vicinities. This phenomenon gives rise to voltage
magnification at the customer vicinities [152]. Hence it can be concluded that GTO based
switched SSCL is not suitable for use on power system if voltage quality is a critical
issue.
Thyristor-Controlled SSCL utilizes thyristors to incorporate limiting impedance to the
system as soon as fault occurs. In contrast to GTO switched SSCL the switches of this
device turn on during fault. In this way, fault current interruption during fault is avoided
and system does not suffer from transient complication. There are two types of thyristorcontrolled SSCL devices. Thyristor controlled series capacitors (TCSCs) and thyristor
controlled series reactors (TCSRs).
TCSC is a combination of conventional series capacitor with thyristor controlled
reactor (TCR). This allows continuous control of the TCSC reactance. When TCSC is
98
used to limit fault current, it is consequently improving the quality of the supply by
mitigating sag severity [153]. However these devices are not able to completely mitigate
sag as desired and hence have limited capacity to compensate voltage sag.
TCSR have been also investigated for voltage sag mitigation by the researches but
they were found to have limited capacity to ride through AC power line dips [154].
Moreover, value of inductance in TCSC and TCSR design should not be kept large in
order to gain higher voltage level, smaller phase-angle jump and lower fault current;
otherwise it will give in rise to prolonged voltage sags [155].
SSCL has been initially proposed for the limitation of large fault current. However,
literature studies reveal that, through the correct selection of circuit configuration, SSCL
can also be operated for the mitigation of voltage sag. In contrast to UPS and DVR that
are installed on the feeder supplying sensitive loads, a SSCL requires to be installed on
every potentially faulty line. Hence, due to the large number of SSCLs required, this
method of solution is expensive.
In contrast to SSCL, SSCB is based on a combination of GTO and thyristor switches.
It has the ability to intrude a fault current very rapidly and can also perform autoreclosing function. It is a lot faster than its mechanical equivalent and is therefore an
ideal device for custom power application. However, according to above discussion,
SSCL and SSCB are not commonly used for voltage sag, swell and phase-angle jump
mitigation.
3.2.2.2 Static Source Transfer Switch
Fig. 3.31 shows a device that is obtained by series combination of STS and static
series compensator (SSC) [22]. This arrangement can manage total protection against
99
interruptions and voltage dips, with the STS taking care of interruptions and dips
originated by faults in the distribution system, which are long and deep and would
deplete the energy storage of the SSC. In this case, the responsibility of the SSC would
be to compensate for the voltage dips originated by faults in the transmission systems,
which the STS cannot handle. It is reported by Bollen that transmission system dips are
normally short and shallow [63]. The combined series configuration of SSC and STS will
reduce the size of the energy storage of the SSC, with a consequent reduction of the cost
of the device. However, this arrangement is bounded with availability of alternate feeder.
A. Bagginii pointed out that most commonly used solid-state device in SSTS is the
thyristor or silicon-controlled rectifier (SCR) [22]. In case SCR is used in SSTS
configuration then transfer time may be half-cycle as SCR cannot turn off until the
current crosses zero. Fig. 3.32 shows a common static switch configuration.
100
T.A. Short mentioned another fast transfer technology that can be used in the same
applications as static switches [15]. High-speed mechanical source transfer switches use
high speed vacuum switches and a sophisticated microprocessor based control to
provide "break-before-make" transfers in approximately 25ms or 1.5 cycles. The
two sources are not parallel during the transfer; therefore, the load experiences an
interruption of approximately 1.5 cycles. The authors also pointed out that this
level of protection may be acceptable for some equipment. However, l.5-cycle of
interruption may trip some sensitive equipment. These switches have the
advantage of being very efficient (99%), inexpensive (one-fifth to one-tenth the
cost of a static switch), and small. Both pole-mounted and pad-mounted versions
are available. For most loads, the fast transfer switch provides significant benefit.
Relays and contactors, though, can drop out for an l.5-cycle interruption.
Normally, fast transfer switches have been applied at individual customers in a
primary-selective scheme. The technology could nicely apply to feeder level
application as shown in Fig. 3.33. This configuration offers better performance for
101
voltage sags, momentary interruptions, and long duration interruptions for the
customers at the end of the circuit that normally get the worst power quality [15].
The SSTS performance concerning transfer time has been analyzed by [99, 151].
Simulation results conclude that in general the speed of SSTS operation is high enough to
attain a seamless transfer for sensitive loads. However, the transfer time increases even
more in case of regenerative load, e.g. induction motors [152], [153].
Fig. 3.33 Fast transfers switch to enhance power quality to a downstream section
of circuit [15].
every half-cycle. This checks whether or not load voltage is within a prescribed range. In
case it is found outside the allowed limit, an abnormal condition is sensed and the firing
signals to the SSTS thyristors are changed to shift the load to the healthy feeder. The
main reason for the success of the SSTS is its rather low cost compared with other
solutions. Although, the SSTS has been proved to be a suitable CPD for screening
selected load points against faulted conditions, but it does require an alternate feeder
being available that has high cost. Therefore, this solution is particularly attractive for
installations that already have mechanical transfer systems, where upgrading to a static
system does not require major changes in the layout of the distribution system. In
addition, SSTS cannot protect against dips originating in the transmission system that
also affects the alternative supply. However, a significant improvement can be attained in
the performance of the industrial system against faults at distribution level, that usually
are responsible for long duration sags and short interruptions.
phase transformers [67]. Generally, this type of device is limited in design to provide up
to 50% boost [150]. The rating of an ETC needs to be the same as the full rating of the
load that it will protect, because the ETC will carry the entire load during sag or swell
events. An ETC is not effective during voltage interruptions because there is no voltage
to transform due to absence of energy storage element.
3.3
Summary
CPDs are used in the power distribution systems to improve the quality and reliability
of power delivered to the consumers. Compensating type CPDs include DVR, reactive
power and harmonic compensation devices and backup stored energy devices. SSTS is
most commonly used for network reconfiguring. An overview of each CPD has been
included in this chapter for voltage sag, swell and phase jump mitigation. Specific
applications of CPDs have been also explored.
There are various CPDs for protecting industrial processes against voltage sags like
static UPS, flywheel, shunt-connected synchronous motor, SSTS, D-STATCOM, DVR
and UPQC, etc. Although these CPDs work efficiently with some constraints to protect
sensitive load from voltage sags, but the most efficient and robust voltage restoration is
possible by the use of DVR. The main reason for extensive use of DVR is its simple
control, fast response, and fewer transients. The DVR is a CPD that provides technically
105
advanced and economical solution to voltage sags, swells and phase jumps as compared
to other CPDs.
DVRs are now a matured option in industry to reduce the impact of voltage sags to
sensitive loads. The use of DVR in PQ-related applications is increasing. Although, the
most popular application of DVR is to control voltage sags (swells) but the harmonics
and power factor correction may also be achieved through robust control schemes. DVR
application offers a number of benefits to industrial, commercial and residential
customers. Decrease in shut down time of process industries, small losses in the
production process, reduction of insulation damage on transformers and motors due to
power irregularities, and smooth operation of sophisticated electronic equipments are few
of them. Proliferation of sensitive load has opened the venues for DVR application in all
consumer categories.
106
CHAPTER IV
DVR INTEGRATION WITH
DISTRIBUTED GENERATION
4.1
Introduction
Due to deregulation, ever growing competition in the market and declining profit
have compelled the industry to recognize the impact of high-power quality. This can be
managed by ensuring uninterruptible flow of power at the correct voltage levels. Electric
utilities are finding ways to ensure premium power supply to their customers. The
emergence of DVR technology provides a good opportunity to have disturbance free
supply to load. DVR is primarily responsible for rapidly restoring the load-side voltage to
its rated value in the event of a disturbance in order to avoid any power interruption to
load [161], [162]. DVR usages facilitate consumers to shield and isolate themselves
from disturbances and transients caused by sags and swells on the transmission
lines or distribution networks. Besides compensating voltage sags, swells, and phase
jumps, DVR can also provide other features such as: line voltage harmonics
compensation, fault current protection and reduction of transients in voltage [163]. The
operation of DVR is not specific to any type of fault or any event that occurs in the
system, but the whole system should remain connected to the supply grid, i.e. the line
breaker does not trip [164].
The idea of integration of DG with utility system is now well established. The
purpose of connecting DG in parallel with utility is two fold. One is to either supply
power to load to share utility burden or support grid by exporting power to it. During this
107
mode of operation, it also regulates load bus voltage. Second is to tackle PQ issues. DG
may help to eliminate the sag in voltage at load bus. It depends on the type of technology
used, location and configuration of DG in the distribution network. From operational
point of view, DG has two basic types.
DC source type: fuel cells, photovoltaic, and battery storage, etc.
High frequency AC source type: micro-turbine.
In each case, the source is required to be interfaced with utility AC network, using
voltage-sourced converter (VSC). The VSC can be configured like an adaptive VAR
compensator to balance and boost supply voltage. However, literature review reveals that
the idea of power export to grid through series connection of DG supported DVR is not
yet explored. In this research work, DG supported DVR is extensively investigated to
compensate PQ issues (voltage sag, swell and phase jumps in the utility voltage) to
regulate load bus voltage for the smooth and clean power delivery to the load. However,
the idea of power export using DG supported DVR and its working theory is also
explored for one of the proposed control strategies. The theory of power export concept
and its validation through simulations using series generator (DVR) has been discussed in
chapter V.
This chapter discusses DVR components, its different modes of operations, DVR
compensation strategies and their limitations, and different system topologies for DVRs.
At the end of this chapter a brief description of DG supported DVR has been provided.
4.2
DVR Components
The DVR structure mainly involves a power circuit and a control circuit. Derivation
of the parameters (magnitude, frequency, phase-shift, etc) of the control signal injected
108
by the DVR is linked to control circuit. Depending upon the control signal, the voltage to
be injected is generated by the switches in the power circuit [165], [166]. Moreover, the
basic structure of the DVR is described by power circuit and is explained in this section.
Power circuit principally consists of five components as shown in Fig. 4.1. The role and
requirements of each component is discussed below.
109
angle. To serve this purpose, VSC employs insulated gate bipolar transistors (IGBTs). In
the DVR application, the VSC is used to temporarily replace the supply voltage or to
generate the part of the supply voltage which is missing [18]. In the DVR power circuit
step up voltage injection transformer is used. Thus a VSC with a low-voltage rating is
sufficient [171]. Numerous circuit topologies are available for the VSC. The VSC
connection methods that are usually used for three phase DVRs are:
However, H-bridge VSC is used for single-phase DVRs [168]. This VSC configuration is
preferred due to its simplicity as shown in Fig. 4.2. It can be also used in a three phase
system by connecting it with each individual phase as DVRs are usually meant to provide
individual phase compensation independently.
Fig. 4.2 shows a 2-level VSC configuration as the output voltage can have values of
+V and -V. But there exists a category of VSCs that can have more than two values
(levels) and are referred to as multi level VSCs [1]. For example a 3-level VSC has the
111
ability to generate voltage levels such as + V, 0, and V where V is the voltage of the
DC supply. Likewise, four unique voltage levels can be realized from a 4-level VSC and
so on. It is to be noted that multilevel VSCs are used for high-power high voltage
applications. But the efficiency and reliability of the converter is decreased owing to an
increasing number of devices [172]. Common multilevel VSC topologies/configurations
include:
There are four important types of switching devices that can be used in VSC design
[164]:
Each type has its own merits and demerits. The MOSFET demands a high on-state
resistance and has sharp switching times [173]. It has the ability of working beyond the
20 kHz frequency [174]. On the other hand, increasing on-state resistance with increasing
voltage, limits the device to applications with just a few hundred volts. The GTO is
termed as a latching device that can be switched off by a negative pulse of current to its
gate [175]. The GTO is mostly suitable for high voltage applications [176]. The
drawback of the GTO based devices is that they do not have the ability to fulfill the
dynamic requirements of a DVR [177]. The IGBT is relatively an advanced device as
112
compared to the MOSFET and GTO. It showed the presence in the market in early 1980s
and because of superior characteristics; it become a popular device [173]. It is a
controllable medium power range three terminal switch that includes the fast switching
times of the MOSFET with the high voltage characteristics of the GTO. Moreover, IGBT
can handle large current and has low loss as compared to GTO or MOSFET [1]. The
need of expensive and bulky snubber circuits as well as the complex gate drive is the
basis that GTOs are being substituted by IGCTs [178]. The IGCT is a highly
sophisticated converter switching device with improved performance and reliability that
facilitates designing VSC with very large power ratings. This enables DVR to mitigate
sags that are beyond the capability of the past DVRs using existing devices. However, in
the proposed DVR control strategies, VSC model uses IGBTs for the circuit simulations
as they meet the design requirements in terms of voltage and power rating.
the injection transformer. This necessitates the use of higher rating of the transformer
[167], [181]. However, the transformer leakage reactance may be used as a part of the
filter, which will be supportive in tuning the filter [168], [180], [181].
114
4.3
DVR Operation
The basic source of voltage sag is a short circuit current flowing into a fault on a
transmission or distribution line as shown on the simplified grid model of Fig. 4.4 [183].
However, voltage sags can be caused by switching of heavy loads or starting of large size
motors [184]. Studies have revealed that transmission faults, although comparatively
rare, may trigger widespread sags. These sags can cause process interruptions although
the fault point is far away from the consumer end [185]. Whereas, distribution faults are
relatively more common, however, the consequential sags are more restricted within
geographic location [84]. The magnitude and phase of the sag voltage at the PCC during
the fault can be found by using the voltage divider rule as [186]:
Vsag =
ZF
E
ZS + ZF
(4.1)
Where
Vsag = Sag voltage at the PCC
116
E = Supply voltage
Z S = RS + jX S is the supply impedance, and
Z f = R f + jX f is the impedance from the PCC to the fault on the faulted line.
(4.2)
Where
Vmissing = Missing voltage that is to be injected by the compensator
As voltage sag may accomplish magnitude change and phase jump, hence compensating
device should be able to tackle either of them or both whichever is the case.
phase-to- ground fault, two-phase faults, two-phase faults with ground connection, and
single-phase fault. Moreover, voltage sags may be balanced or unbalanced. Although
various sags type affects the DVR basic structure, however the most important problem is
the depth and duration of the voltage sags that influence the rating of DVR for the
various system topologies. This issue is discussed in detail in section 4.5 of this chapter.
As mentioned earlier, the DVR prime function is to observe the utility or load voltage
waveform in real time and if any disturbance such as sag or swell is detected, the balance
voltage is injected to (or absorbed from) the utility voltage to regulate load bus voltage.
In order to realize the above operation of DVR, a reference voltage waveform is
generated that would be identical in magnitude and phase-angle to that of the utility
voltage. Then by comparing reference and actual voltage waveforms, any disturbance in
the actual voltage waveform is detected. After this an appropriate control action may be
taken to mitigate any disturbance in the utility voltage.
4.3.1
Protection mode
Standby mode
Injection mode
In Protection mode, the DVR is guarded from the over current in the load side owing to
short circuit on the load or large inrush currents [188]. DVR can be removed from the
system under abnormal conditions with the help of bypass switches by supplying an
alternate path to short circuit currents as shown in Fig. 4.6 [187].
118
In standby mode, the series compensator can be configured either to enter into short
circuit operation or inject small voltage to balance the voltage drop on transformer
reactance or losses as shown in Fig. 4.7, [188]. In standby mode with VDVR=0, the lowvoltage winding of series transformer is shorted through the converter. Hence, there will
be no switching of semiconductors, as the individual VSC legs are triggered in such a
way as to create a short-circuit path for the transformer connection [189]. Therefore, the
small conduction losses of the semiconductors in this current loop contribute to the
losses. DVR short circuit operation is normally the favorite solution in steady-state as the
small voltage drop does not upset the load requirement if the distribution circuit is not
weak.
119
With VDVR >0, the DVR enters into injection mode through the injection transformer
as soon as the sag is detected [190]. A single-phase voltage or three single-phase AC
voltages are injected in series with desired magnitude, phase and wave shape for
compensation [188]. Usually, the types of voltage sag; power rating of DVR and load
condition decide the possibility of compensating voltage sag. Some loads cannot afford
phase-angle jumps/variations such as induction motors [23] while others are tolerant to it.
Hence, the selection of control strategy relies on the type of load characteristics [164].
However, the DVR should be able to provide constant voltage at the load bus.
4.4
There are three most popular strategies to inject DVR compensating voltage to
mitigate the effects of voltage disturbance in a distribution system. These methods are:
pre-sag compensation, in-phase compensation and phase advanced or optimal energy
compensation [77], [191], [192]. Depending upon DVR application environment, each
technique has its own merits and demerits.
120
[194]. In this method, the injection voltage phasor is made perpendicular to the load
current phasor so that the injected active power is zero as shown in Fig. 4.10. In this way
consumption of energy stored in the DC-link can be reduced by injecting reactive power
instead of active power [92]. This inturn enhances ride-through capability of the series
device when the energy storage capacity is fixed. In contrast, the injection voltage
magnitude of the phase-advance method is larger than those of pre-sag or in-phase
method. This technique is also known as optimal energy compensation. Since PAC
method utilizes only reactive power therefore not all the sags can be mitigated without
real power especially when power factor of load is high. Consequently, this method is
only appropriate for limited range of sags [163].
122
Voltage limit: The DVR design has limited capability to inject voltage up to a
definite range in order to have low cost and small voltage drop across the device
during nominal conditions [84].
Power limit: DC-link is used to store power. This power is mostly converted from
the supply or from a larger DC storage via a converter that is also used to
maintain the DC-link voltage. The converter rating enforces a power limit to the
DVR [96].
123
Energy limit: A DVR design that involves large energy storage is used to regulate
the load voltage to its rated value and size of energy reserve is kept as low as
possible in order to have reduced cost. [84]. However, some sags will drain the
storage quickly but the control strategy can decrease the risk of load tripping
initiated by insufficient energy storage.
4.5
J. G. Nielsen and F. Blaabjerg have investigated four different system topologies for
series compensators (DVR) with special emphases on the methods used to obtain the
required energy during voltage sag [183]. A comparison has been made between two
topologies that are:
It has been already discussed in chapter II that DG technologies may include energy
storage systems such as batteries, flywheels, supercapacitors, superconducting coils, etc.
Merits and demerits of these storage systems have been highlighted in chapter III.
Moreover connection of DC/high frequency AC source type DGs (energy storage
systems) with utility system is discussed in section 4.1. However, integration of energy
storage device (DG) with DVR is described in section 4.2.2. Now each of the above
system topology is further divided into two types depending upon usage of energy
storage system/device as listed below.
2.
2.
Fig. 4.11 DVR without energy storage and supply-side-connected converter [183].
Fig. 4.12 DVR without energy storage and load-side-connected shunt converter [195].
125
Fig. 4.13 DVR having energy storage and with variable DC-link voltage [183].
Fig. 4.14 DVR having energy storage and with constant DC-link voltage [183], [196].
Fig. 4.15 shows comparison of variation of total (series and shunt) converter ratings
for the DVR four topologies versus sag. The comparison result categorizes the above
explained DVR topologies regarding converter size in the following order starting from
the largest converter size and ending at the lowest.
1. Load side connected converter topology
2. Supply side connected converter topology
3. Constant DC-link topology
4. Variable DC-link topology
126
Fig. 4.15 Effect of variation of sag size on total converter rating of DVR [183].
Table 4.1 summarizes the performance of different DVR topologies [180]. Analysis
shown in Table 4.1 has revealed that the no-energy storage concept is feasible, but an
improved performance can be achieved for certain voltage sags using stored energy
topologies. However, an intermittent type of renewable energy resource (PV source) with
battery storage system is assumed as DG in this research work for analysis purpose. It is
connected with DC-link of the DVR to maintain constant DC-link voltage. Fig. 4.16
shows the DVR structure employing DG (PV source with battery storage) to mitigate
deep and long duration sags. Three DVR control strategies to regulate load voltage are
thoroughly discussed in chapter V that use above mentioned assumption.
127
Table 4.1
Comparison of different DVR topologies with the grading [183]:
Very Good (++), Good (+), Poor (), and Very Poor ( )
Stored Energy
Supply side
Load side
Connected
connected
converter
converter
(1)
(2)
++
++
-+
-+
No Stored Energy
Variable DCConstant DClink voltage
link voltage
(3)
-++
(4)
++
++
++
++
+
-+
-10
7
3
--
---
+
+
+
+
+
+
++
7
10
-3
-+
+
+
8
4
4
+
++
8
6
2
128
4.6
Summary
129
CHAPTER V
CONTROL STRATEGIES
FOR VOLTAGE REGULATION
5.1
Introduction
sag is usually associated with system faults, but it can also be caused by switching of
heavy loads or starting of large size motors [205]. Whereas, voltage swell is usually
associated with system faults, capacitor switching or disconnection of heavy loads.
However it is not as important as voltage sag because it is less common in distribution
systems. Main characteristics of voltage sag are its duration and magnitude in available
utility voltage [206]. Wang et al. have given the performance curves with depth and
duration of voltage sag provided by Information Technology Industry Council (ITIC) and
Semiconductor Equipment and Materials International (SEMI) [207]. SEMI-F47 standard
curve that is usually followed by the processing industry against voltage sag protection is
shown in Fig. 5.1 [208].
Fig. 5.1 SEMI-F47 recommended voltage sag ride-through capability curve [208].
131
Industrial consumers that are generally supplied directly from high voltage system
have provision of an alternate feeder(s). So the interruption for these customers is very
rare. In case of a fault on a transmission system these customers will be affected from
voltage sag. Whereas the fault that occurs in low and medium voltage systems is
responsible for the voltage sag that tend to have more variation. Since voltage sag occurs
more frequently than outage, hence, it is of utmost importance to examine its effects and
possible compensation strategies separately from the remedial measures opted for
complete outages. Fig. 5.2 shows the most common quality/reliability/availability (QRA)
problem pointed out by a North American utility [209].
Type of equipment
1
2
3
4
5
6
PLC
PLC input card
5 H.P AC drive
AC control relay
Motor starter
Personal computer
60%
55%
75%
65%
50%
60%
260 ms
40 ms
50 ms
20 ms
50 ms
50 ms
132
Fig. 5.3 Conceptual diagram of one phase of a DVR with its protection components.
Also, the series nature of the DVR demands a dramatically fast and adequate
mechanism to prevent its malfunction or failure against short circuit on the load side [68].
An electromechanical bypass switch, associated with a fast static switch can be used to
remove the DVR from the system by providing alternate path for the current under
abnormal conditions as shown in Fig. 5.3. Static bypass switch is capable of providing a
path to short circuit current until the operation of the electromechanical switch whose
133
typical time of operation is in the order of 80 ms. As long as the source current is within
rated limits; the bypass switches will remain inactive. This arrangement safeguards the
PWM VSC switches from the forced current overload.
A DVR compensates voltage sag by injecting voltage of suitable magnitude and
phase-angle so that the load voltage is kept within the allowable limits. The PWM VSC
has found applications for this task. The PWM technology gives rapid response and
generally low loss. The injected voltage may or may not be in phase with any other
variable of the main circuit i.e. supply voltage, load current, load voltage etc [211]. The
voltage compensation capability of the DVR is controlled with the turns ratio of its
injection transformer and the line current of the PWM device. A part of this consideration
is the amplitude modulation index of the PWM VSC. The harmonic impact of the VSC is
bounded to higher frequencies which are generally less objectionable than lower
frequencies for the inductive load.
Practical devices for compensating the load voltage to its rated value during sag must
operate under transient as well as steady-state conditions. The DVR transient response
characteristics must have the capability for fast resetting of the PWM VSC in the
beginning of voltage sag or at the recovery moment from voltage sag. The injection
transformer voltage obtained from the PWM VSC will contain harmonics clustered
around Mf f o. Where fo is the power frequency and Mf is frequency modulation index.
For these frequencies, the injection transformer winding current will be simply
proportional to the integral of the voltage and the harmonic impact will be low [212].
For an offline DVR, an important issue is the injection transformer inrush current at
the start of sag compensation. The inrush current may operate the over-current protection
134
of the compensator VSC, and reduce the compensation voltage injected by series
transformer. This will eventually lead the DVR system towards failure to compensate the
voltage sag [61]. An accurate sizing of the series injection transformer is essential for
blocking inrush current being drawn when the DVR begins to compensate for sag. To
account for all initial flux conditions, the series transformer should be rated for two
times the normal steady-state flux requirement at maximum injection voltage without
going to saturation [213]. The consequence of this design consideration is a dramatic
increase in the size, weight and cost of the DVR sag compensation system [75]. Through
proper control of the voltage injection of the sag compensator, the transformer inrush
current at the start of the sag can also be mitigated while preserving the output voltage for
effective sag compensation [64].
The objective of LC filter circuit in the DVR circuitry is to eliminate the high
frequency VSC switching noise. On the other hand DVR filter circuit makes the DVR
system to be lightly damped and introduces transient oscillations at the LC filter cutoff
frequency. The nonlinear load with higher order harmonic currents around the LC filter
cutoff frequency would excite the LC circuit at resonant frequency. Passive or active
damping methods may be used to suppress these harmonics to avoid resonance
phenomenon [214].
Small voltage sag can usually be restored through reactive power but for larger
voltage sag, it is necessary to inject active power into the system by the DVR to mitigate
the voltage sag [215]. There are two classifications of DVR for active power
contribution. In first, only reactive power compensation is provided by injecting a voltage
in quadrature with the load current that means no active power contribution. In second, an
135
active power available from any distributed energy resource is utilized to inject a voltage
that compensates the load voltage to required level. To track the reference voltage for the
DVR following relationship is used [216]
VD* = VL* VU
(5.1)
Where
VD* = Reference voltage for DVR to track
VL* = Desired load voltage
VU = Utility/Supply voltage available at the point of common connection
Theoretically, the phase-angle of desired load voltage in equation (5.1) may have any
value. Ghosh and Ledwich suggested that the phase-angle of desired load voltage should
lag the phase of the available utility voltage [1]. Otherwise, reverse power flow through
the rectifier-supported DVR may damage the rectifier unit. The work presented in this
chapter evaluates different control strategies for DVR to regulate load bus voltage in case
of voltage sag/swell and/or phase jump in the utility voltage.
As a first control strategy in this work, a mathematical model of leading series
voltage injection (LSVI) by a DVR for voltage sag mitigation is developed. The
developed
model
is
then
validated
with
the
results
obtained
from
phase voltage injection scheme is introduced with mathematical modeling and the same
have
been
validated
by
results
obtained
from
MATLAB/SIMULINK
SimPowerSystems simulations.
A unified approach of mitigating voltage sags, swells and phase jumps in three phase
utility voltage using VSC based feed-forward control strategy for DVR with in-phase
voltage injection scheme is next proposed. The proposed strategy for load voltage
regulation is evaluated through simulations to show its performance for mitigating
balanced and unbalanced voltage sags, swells and phase jumps in the utility voltage.
5.2
V L = VU + V D
(5.2)
Where
V L = Load voltage phasor with phase-angle after injection
137
The phase-angle between the utility voltage and load current is thus (-) and phaseangle between DVR injected voltage and load current is (-+).
(5.3)
VL Sin = VD Sin
(5.4)
Equations (5.3) and (5.4) result into the following equations for magnitude of injection
voltage from DVR and the phase-angle of load voltage:
2
(5.5)
VD = VU Cos + VL VU Sin2
138
Negative sign with second term in equation (5.5) is discarded as a magnitude should
remain positive.
VD
Sin
VL
= Sin1
(5.6)
As sine is a dual value function in the range of 0 to 180, equation (5.6) may produce
invalid results. Another relationship for from Fig. 5.4 can be obtained as:
VD Sin
+
V
V
Cos
U
D
= Tan1
(5.7)
It is important to note that inverse tangent algorithms catering for sign of numerator and
denominator values should be used; otherwise, equation (5.7) may also produce incorrect
results.
Active and reactive power flow equations at utility bus (PU, QU), load bus (PL, QL) and
power injected by DVR (PD, QD) can be written as under:
PU = VU I L Cos( + )
(5.8)
QU = VU I L Sin( + )
(5.9)
PL = VL I L Cos( )
(5.10)
QL = VL I L Sin( )
(5.11)
PD = VD I L Cos( + )
(5.12)
QD = VD I L Sin( + )
(5.13)
The above equations are written using the active sign convention for sources and the
passive sign convention for load. Therefore positive values of active/reactive powers for
sources mean power delivered and negative sign implies absorption of active/reactive
power.
139
Equation (5.12) implies that active power contribution from DVR shall remain
positive as long as the cosine term remains positive. Mathematically, this condition can
be expressed as:
(5.14)
90 + 90
The equations (5.3~5.14) set out guidelines for sag mitigation and ultimately the dynamic
voltage restoration using leading series voltage injection to attain regulated load bus
voltage. The choice of is driven by the active power available at the DVR input and its
value can be increased or decreased accordingly.
It is evident that for symmetrical sag, an increase in will result in a higher injection
voltage. However, this will reduce the requirement of active power from DVR and
increase the reactive power contribution from DVR. This result is significant as
distributed energy resources (DER) like solar and wind is dependant on environment and
available active power may vary over a large range, provided enough storage capacity is
not provided. The injection voltage phase-angle may be varied to achieve the saving on
available active power and prevent the collapse of DVR voltage.
5.3
Power Export
If there is sufficient distributed energy capacity present and the load power factor is
kept near unity with the help of power factor correction devices, phase-angle of the load
current will become the same as that of the load voltage. Then the angle difference
between injected voltage and load current will be (). To have () at 90 or greater,
VU has to be greater than VL in magnitude which is not possible. Hence, for a unity power
increased, the magnitude of VD increases and also augments. A point will be reached
when attains the value of 90, active power from the utility side becomes zero, and
DVR will be providing all the load active power as well as reactive power to the utility.
If is further increased such that becomes greater than 90, the utility active power will
become negative. It indicates that power is being supplied by the DVR to the utility
which is also supplying active power to the total load.
Maximum active power supplied to the utility by the DVR occurs when both and
attains the value of 180. At this point, some of the maximum power is picked up by the
load and rest is exported to the utility. This is a new concept for distribution system
application, as power export to utility is generally characterized by a shunt connection to
the bus. Another point to emphasize is that current magnitude during the variation of
power export by DVR remains the same; variation in power is obtained by changing
voltage magnitude and phase-angle. Maximum voltage required during this variation is
twice the rated load voltage minus the sag voltage on the utility side. It is also evident
from the Fig. 5.5.
The above described arrangement for power export will also work even if there is no
voltage sag; maximum voltage required to inject maximum power in the system will be
twice the rated load voltage, and maximum power will be twice the rated load power,
half of which will be taken up by the load and the other half will be exported to utility.
Active power contribution from DVR is only possible if energy storage reservoir is
considered on the DVR side. The DC-bus of DVR may be supplied with energy from the
main supply through a rectifier-supported DVR or from engine-driven generators.
Alternate energy resources like solar, wind or fuel cells can also be utilized as energy
141
source at the DC-bus in DG-supported DVR. In the absence of an energy reservoir, the
DC-bus includes capacitors in DVR which will have limited ability to mitigate PQ
problems. However, it is assumed in this research work that a PV source with battery
storage is connected to the DC-bus of DVR to have regulated DC voltage. PV source
consists of solar panels with a charge control module which works in a voltage regulation
mode of charging.
Fig. 5.5 Phasor diagram showing maximum possible voltage injection capability of
DVR during sag.
gate signals to the switching devices in the voltage-sourced converter (VSC) according to
the modulated signal [23, 77]. In this work, the voltage sag mitigation strategy and power
export concept is illustrated by the simulation examples shown in the next section.
Table 5.2 gives a summary of design specifications and component values for the
simulated examples.
Table 5.2
Summary of design specifications for a single-phase simulated system
Value/Range
Value/Range
Sr. No.
Component/Parameter description
(Case I)
(Case II)
1
Rated load voltage
240 V
240 V
2
Maximum load apparent power
707 kVA
500 kVA
3
Maximum load current
2945.8 A
2083.3 A
4
Load power factor
0.707 lag.
unity
5
VSC DC-bus voltage
700 V DC
700 V DC
6
System frequency
50 Hz
50 Hz
7
Maximum sag depth
15%
15%
8
Sag detection and mitigation time
< 20 ms
< 20 ms
9
10
11
5.4
1:1
PWM
Low pass LC
Simulation Results
Fig. 5.6 shows the block diagram of simulation set up used for investigation of
proposed LSVI by the DVR for sag mitigation and power export. A simulation of one
phase of a system compensated by DVR is carried out for 15% sag in utility voltage by
varying
different
injection
voltage
angles
in
the
MATLAB/SIMULINK
The calculated values are obtained using mathematical model of section 5.2. Then
various plots for above mentioned variables against injected voltage phase-angle () are
generated using MS Excel to have comprehensive analysis of LSVI scheme regarding
voltage restoration and power export. Time simulation of voltage dip and its mitigation
for different types of load is also included to show dynamic behavior of DVR during sag
condition. The DVR consists of a PWM-IGBT based VSC with a filter supplied from a
DC-bus. The filter output is connected to the primary winding of the transformer whose
secondary is connected in series to the utility line.
Fig. 5.6 Block diagram showing DC source as DG injecting voltage in series with utility supply.
Following
two
cases
are
simulated
in
MATLAB/SIMULINK
SimPowerSystems environment.
Case I:
In this case an inductive load with 0.707 lagging power factor is used. Active power
rating of the load is 500 kW and reactive power is also 500 kVAR. The results are
presented in Figs. 5.7, 5.19, 5.22 and 5.28, where relevant utility, DVR and load
variables are plotted against .
144
Case II:
Voltage (V)
500
400
DVR Voltage
300
Utility Voltage
200
Load Voltage
100
0
0
30
60
90
120
150
180
Fig. 5.7 Voltage variation with injected voltage phase-angle (Case I).
Figs. 5.7 and 5.8 show the variation in magnitude of injected DVR voltage, and its
effect on load voltage magnitude which is held almost constant at the rated value i.e. 240
V for a 15% sag in utility voltage. Both figures show that DVR injected voltage
magnitude (VD) increases significantly for the values between 75~165 and VD
reaches to its maximum value of 444 V at =180. This shows that maximum value of
VD is twice the rated load voltage minus the sag voltage on the utility side.
145
Voltage (V)
400
DVR Voltage
300
Utility Voltage
200
Load Voltage
100
0
0
30
60
90
120
150
180
Fig. 5.8 Voltage variation with injected voltage phase-angle (Case II).
Voltages (Nonlinear Load)
500
Voltage (V)
400
DVR Voltage
300
Utility Voltage
200
Load Voltage
100
0
0
30
60
90
120
150
DVR injected voltage phase angle (degree)
180
Fig. 5.9 Voltage variation with injected voltage phase-angle (nonlinear load).
To investigate the effect of nonlinear load on the performance of the proposed LSVI
scheme, a diode rectifier bridge with an RL load is used. Fig. 5.9 shows plot for DVR,
utility and load voltage for varying DVR injected phase-angle. Results in Fig. 5.9 show
that there is insignificant effect of nonlinear load on the performance of DVR for voltage
restoration.
146
In addition to the measurements taken for the steady-state value of voltage and power for
different injection voltage phase-angles, the dynamic response of DVR, utility and load
with respect to their voltage and power is also examined by considering 15% sag.
However, time simulation of voltage dip and its mitigation with 45 injection voltage
phase-angle is shown only in Figs. 5.10-5.12. Other plots for different injection voltage
phase-angle can be obtained on a similar analogy.
Approximately 66% voltage sags are that of the single-phase voltage sags [217]. Figs.
5.10-5.12 show 15% single-phase symmetrical sag initiated at 0.04 s in the utility voltage
which sustain until 0.1 s for 0.06 s. The peak value of utility voltage under normal
circumstances is 339.4 V, whereas, it is 288.4 V during sag. In case of leading voltage
injection during sag, DVR responds by injecting 51 V to compensate the sag in the utility
voltage.
Fig. 5.10 Dynamic Voltages with 45 injected voltage phase-angle (Case I).
147
Fig. 5.11 Dynamic Voltages with 45 injected voltage phase-angle (Case II).
Fig. 5.12 Dynamic Voltages with 45 injected voltage phase-angle (nonlinear load).
Voltage injection after sag termination is zero as evident from Figs. 5.10-5.12. DVR
is connected between the utility and load to inject compensation voltage quickly to
148
restore load voltage to its rated value. The first two cycles of injected voltage are prior to
the DVR series connection with the line. They are used to energise primary winding of
the injection transformer of DVR circuit. According to the simulations, the load voltage
is evidently restored in 10 ms against a target of 20 ms (one-cycle) as suggested by ITIC
curve and SEMI-F47 standard.
In a DVR, there are two main considerations in its working performance: the
compensation capability and the output voltage quality. In other words, there are two
main factors relating to the capability and performance of DVR working against voltage
sag in a certain power system: the sag severity level and the total harmonic distortion
(THD). Both of these are determined by the DC source and the firing control scheme of
DVR VSC. DC source in this work is assumed ripple free. According to IEEE Std. 5191992 THD limit for the voltage is 5% [128], whereas, harmonic content in the
compensated load voltage of Figs. 5.13-5.15 are measured through FFT analysis of load
voltage. They are found to be 0.06 % (THD=0.06%) in Case I and 0.13% for Case II.
THD of load voltage after series compensation for nonlinear load is measured 0.29%.
These THD values are within the permissible limits. Hence, DVR with proposed LSVI
scheme can tackle the voltage harmonics due to PWM VSC and nonlinear load.
Figs. 5.13-5.15 also show time simulation of sag and its mitigation. In this case, 15%
sag is initiated at 0 s which persists for 0.25 s (15 cycles). DVR is connected between
utility and load at 0.04 s during sag to inject compensation voltage equal to 51 V to
restore load voltage to its rated value. Again, the first two cycles of DVR injected voltage
shown in Figs. 5.13-5.15 are prior to the DVR series connection with the utility line.
149
Fig. 5.13 Dynamic Voltages with 45 injected voltage phase-angle (Case I).
Fig. 5.14 Dynamic Voltages with 45 injected voltage phase-angle (Case II).
150
Fig. 5.15 Dynamic Voltages with 45 injected voltage phase-angle (nonlinear load).
Fig. 5.16 Load current without sag and without compensation (nonlinear load).
Figs. 5.16-5.18 show RMS value of load current without sag, with sag and after series
compensation, respectively for the nonlinear load. These plots show that DVR response
151
to stabilize load current after sag is visibly slower as compared to load voltage restoration
for a nonlinear load. Although it is observed in Case I and II that, load current is
stabilized quickly in similar fashion as load voltage stabilization. Active power plots for
Case I and II are shown in Figs. 5.24 and 5.25, respectively. Where, it is obvious that
DVR response to sag mitigation is very fast.
Fig. 5.17 Load current with sag and without series compensation (nonlinear load).
Fig. 5.18 Load current with sag and with series compensation for =45 (nonlinear load).
152
200
150
100
Load Voltage
50
Load Current
0
-50
-100
0
30
60
90
120
150
180
Fig. 5.19 Phase-angle variation of load voltage and current with injected voltage phase-angle (Case I).
200
150
Load Voltage
100
Load Current
50
18
0
15
0
12
0
90
60
30
Figs. 5.19-5.21 indicate the phase-angle of load voltage and current against the
variation of injection voltage phase-angle for Case I, II and nonlinear load, respectively.
Plots of load voltage and current phase-angles coincide in Fig. 5.20 as these are in-phase
153
for a compensated load, whereas, load current phase-angle behaviour shown in Fig. 5.21
is of oscillatory nature for most of the time due to the nonlinear current flowing in the
load.
Phase angles ( Non-linear Load)
250
150
50
Load Voltage
Load Current
-50
-150
0
30
60
90
120
150
180
Figs. 5.22 and 5.23 present the active power plots for different DVR voltage phaseangles for Case I and II, respectively. It is obvious from Fig. 5.22, DVR active power
contribution (PD) to load is decreasing with the increasing value of injection voltage
phase-angle () and it becomes negative for the values of between 60~130; this is the
same interval during which utility active power (PU) becomes greater than load active
power (PL). It is the point where reverse power flow occurs through the VSC. However,
PD contribution to utility and load increases significantly and PU contribution to load
deceases rapidly beyond 120. At =150, PU and PD contribution is equal to meet rated
load power whereas PD = PL and PU =0 for =165. Hence, overall analysis of Fig. 5.22
reveals that DVR either supplies or absorbs active power to compensate sag in the utility
154
voltage for value of between 0~165. For value beyond 165, DG supported DVR
not only compensates sag in the utility voltage but also supplies total load active power
and exports excess power to utility.
Active Powers (Inductive Load)
1,000,000
500,000
DVR
Utility
Load
0
-500,000
0
30
60
90
120
150
180
Fig. 5.22 Active power variation with injected voltage phase-angle (Case I).
DVR
500,000
Utility
Load
18
0
15
0
12
0
90
60
30
-500,000
0
1,000,000
Fig.5.23 Active power variation with injected voltage phase-angle (Case II).
155
Usually in DC capacitor supported DVR and Rectifier supported DVR; the reverse
power flow may increase the DC-link voltage large enough to damage the DC storage
capacitors and/or switching devices. Usually, to tackle reverse power, a resistor or battery
is connected in parallel to the DC capacitor. In case of assumed PV source with battery
storage system, a built-in protection for the DVR components against reverse power flow
is available.
The analysis in Fig. 5.23 shows that PD contribution to load is increasing with the
increasing value of and it never becomes negative as is the case in Fig. 5.22. However,
PD contribution to load and utility increases significantly and PU contribution to load
deceases rapidly beyond 135. At =115, PU and PD contribution is equal to meet rated
load power whereas PD = PL and PU =0 for =135. Hence, overall analysis of Fig. 5.23
reveals that DVR only supplies active power to compensate sag in the utility voltage for
value of between 0~135. For value beyond 135, DG supported DVR not only
compensates sag in the utility voltage but also supplies total load active power as well as
exports excess power to utility.
DVR in its compensation mode also helps in smoothing active and reactive power
drawn by the load as shown in the Figs. 5.24-5.27. More plots for different injection
voltage phase-angles can be obtained on a similar analogy. Figs. 5.24 and 5.25 show that
active power drawn by the load during sag is reduced. But when DVR is brought in at
0.04 s then it not only mitigates voltage sag but also brings the load power to its rated
value. Prior to DVR entry, load power demand is met by utility only.
156
Fig. 5.24 Active power variation with 45 injected voltage phase-angle (Case I).
Fig. 5.25 Active power variation with 45 injected voltage phase-angle (Case II).
Fig. 5.26 indicates that both the DVR and utility are contributing reactive power to
meet load reactive power demand during sag. Whereas the entire load reactive power
demand is served by utility prior to the start of sag mitigation operation by the DVR. By
contrast, Fig. 5.27 shows that the load does not absorb reactive power because it is
157
compensated by power factor correction measures and all the reactive power supplied by
the DVR flows towards the utility to support the utility grid to meet reactive power
requirement.
Fig. 5.26 Reactive power variation with 45 injected voltage phase-angle (Case I).
Fig. 5.27 Reactive power variation with 45 injected voltage phase-angle (Case II).
158
Figs. 5.28 and 5.29 describe the reactive power contribution from all elements for
Case I and II, respectively. Fig. 5.28 indicates that any growth in injection voltage due to
increasing value of will increase the DVR reactive power (QD) contribution until
attains the value of 160. Further increase in will again bring about a reduction in QD.
In contrast, utility reactive power (QU) contribution to load decreases with the increasing
value of and this situation continues till attains the value of 160, but further increase
will again bring about rise in value of QU. At =80, QU and QD contribution is equal to
meet rated load power whereas QD = QL and QU =0 for =105. However, QD
contribution to load and utility increases significantly and QU contribution decreases
rapidly beyond 105. Hence overall analysis of Fig. 5.28 reveals that DVR only supplies
reactive power to compensate sag in the utility voltage for values between 0~105. For
value beyond 105, DG supported DVR not only compensates sag in the utility voltage
but also meets all load reactive power demand and exports excess power to utility.
500,000
Utility
Load
-500,000
-1,000,000
0
30
60
90
120
150
180
Fig. 5.28 Reactive power variation with injected voltage phase-angle (Case I).
159
Fig. 5.29 depicts that all the reactive power supplied by the DVR for values of
0~135 is exported to utility with the assumption that load is being compensated by a
power factor correction device. DVR will always supply reactive power in the angle
interval of interest; however, this will start to decrease beyond =135 and becomes zero
at =180.
Reactive Powers (Resistive Load)
600,000
400,000
200,000
DVR
Utility
Load
-200,000
-400,000
-600,000
0
30
60
90
120
150
180
Fig. 5.30 shows the variation in magnitude of injected DVR voltage, and its effect on
load voltage magnitude which is held almost constant at the rated value i.e. 240 V for a
zero value sag in utility voltage. This figure show that DVR injected voltage magnitude
(VD) is zero for values between 0~90. However, its value increases significantly for
the values between 90~180 and VD reaches to its maximum value of 480 V at
=180. This shows that maximum value of VD can be twice the rated load voltage to
serve total load power and export available surplus power to utility.
160
600
Voltage (V)
500
400
300
200
DVR
Utility
Load
100
0
0
30
60
90
120
150
180
Fig. 5.30 Voltage variation with injected voltage phase-angle (Case I-without sag).
800
600
Utility
400
DVR
200
Load
0
-200
-400
-600
0
30
60
90
120
150
180
Fig. 5.31 Active power variation with injected voltage phase-angle (Case I).
The analysis in Fig. 5.31 shows that PD contribution to load is zero for value of
between 0~90. Hence DVR goes to standby mode. However, PD contribution to load
and utility increases significantly and PU contribution to load deceases rapidly beyond
90. At =120, PU and PD contribution is equal to meet rated load power. Whereas PD =
PL and PU =0 for o =135 where o is DVR injection voltage phase-angle when PU =0.
161
Therefore, DVR is in power balancing mode at o =135. Hence, overall analysis of Fig.
5.23 reveals that DVR is in standby mode for value of between 0~90. For value
beyond 90 and equal to 135, DVR is in power sharing mode meaning that both utility
and DVR share load to meet load power demand. Whereas, for value beyond 135 DG
supported DVR enters into power export mode. In it, DVR not only supplies total load
active power but also exports excess power to utility.
600
400
200
DVR
Utility
-200
Load
-400
-600
0
30
60
90
120
150
180
Fig. 5.32 Reactive power variation with injected voltage phase-angle (Case II).
Fig. 5.32 depicts that all the reactive power supplied by the DVR for values of
90~135 is exported to utility with the assumption that load is being compensated by a
power factor correction device. DVR will always supply reactive power in the angle
interval of interest; however, this will start to decrease beyond =135 and becomes zero
at =180. Whereas, QD contribution to utility decreases after =135 and becomes zero
for the value of =180.
162
Fig. 5.33 Dynamic voltage variation without sag with =150 (Case II).
Figs. 5.33-5.35 show power export concept using DG supported DVR for the
proposed LSVI scheme when no sag is involved in the utility voltage. These figures show
active and reactive power of load, utility and DVR for value of 150. The plots of PL,
PU, PDVR and QL, QU, QDVR for different values can be obtained on a similar analogy.
Fig. 5.33 shows utility voltage without sag. Load voltage sees a slight disturbance for less
than 15 ms when DVR is brought in at 0.04 s to serve total load power demand and
export excess power to utility. However, Fig. 5.34 show that the disturbance seen by load
does not affect power supply to the load significantly because, power export by the DVR
to load and utility starts in less than 0.02 s. Prior to DVR involvement, all load power
demand is met by utility as evident by Fig. 5.34. As mentioned in section 5.3, power
export to utility through series-connected generator (DVR) is a new concept, because
power export to utility is generally characterized by a shunt connection to the bus. Also
163
load current magnitude during the variation of power export by DVR remains same;
variation in power is obtained by changing voltage magnitude and phase-angle.
Fig. 5.34 Dynamic active power variation without sag with =150 (Case II).
Fig. 5.35 Dynamic reactive power variation without sag with =150 (Case II).
Fig. 5.35 shows that all reactive power supplied by DVR is exported to utility as it is
assumed that load is compensated by a power factor correction device and the power
factor is unity. Hence, in power export mode, DVR supported by DG acts as secondary
utility which supplies to both the load and the primary utility. Results in Figs. 5.33-5.35
164
show dynamic voltage variations and power export concept for Case II without sag.
However, plots for voltage, active and reactive power of utility, DVR and load can be
obtained for Case I without sag on the similar analogy for any desired injected voltage
phase-angle. These plots for value of 165 are shown in Figs. 5.36-5.38. More plots for
other values can be obtained on similar analogy.
Fig. 5.36 Dynamic voltage variation without sag with =165 (Case I).
Fig. 5.37 Dynamic active power variation without sag with =165 (Case I).
165
Fig. 5.38 Dynamic reactive power variation without sag with =165 (Case I).
Figs. 5.39-5.44 show power export concept using DG supported DVR for the
proposed LSVI scheme considering sag in the utility voltage for Case II. These figures
show active and reactive power of load, utility and DVR for values of 130 and 150.
The plots of PL, PU, PDVR and QL, QU, QDVR for other values can be obtained on a
similar analogy. Figs. 5.39 and 5.42 show utility voltage with sag. Load voltage sees a
slight disturbance for less than 15 ms when DVR is brought in at 0.04 s to serve total
load power demand and export excess power to utility. However, Figs. 5.40 and 5.43
show that the disturbance seen by load does not affect power supply to the load
significantly because, power export by the DVR to load and utility starts in less than
0.02 s. Prior to DVR involvement, all load power demand is met by utility as evident by
Figs. 5.40 and 5.43.
166
Fig. 5.39 Dynamic voltage variation with sag with =1300 (Case II).
Fig. 5.40 Dynamic active power variation with sag with =1300 (Case II).
167
Fig. 5.41 Dynamic reactive power variation with sag with =1300 (Case II).
Fig. 5.42 Dynamic voltage variation with sag with =150 (Case II).
168
Fig. 5.43 Dynamic active Power variation with sag with =150 (Case II).
Fig. 5.44 Dynamic reactive power variation with sag with =150 (Case II).
169
Figs. 5.45-5.47 show power export concept using DG supported DVR for the
proposed LSVI scheme considering sag in the utility voltage for Case I. These figures
show active and reactive power of load, utility and DVR for value of 165. The plots of
PL, PU, PDVR and QL, QU, QDVR for other values can be obtained on a similar analogy.
Fig. 5.45 shows utility voltage with sag. Load voltage sees a slight disturbance for less
than 15 ms when DVR is brought in at 0.04 s to serve total load power demand and
export excess power to utility. However, Fig. 5.46 show that the disturbance seen by load
does not affect power supply to the load significantly because, power export by the DVR
to load and utility starts in less than 0.02 s. Prior to DVR involvement, all load power
demand is met by utility as evident by Fig. 5.46.
Fig. 5.45 Dynamic voltage variation with sag with =165 (Case I).
170
Fig. 5.46 Dynamic active power variation with sag with =165 (Case I).
Fig. 5.47 Dynamic reactive power variation with sag with =165 (Case I).
Power flow from the DG supported DVR is twice to that of rated load power rating at
==1800 according to developed LVSI model equation 5.12 and V
DVR
will than be
equal to twice the rated load voltage according to developed LVSI model equation 5.5.
171
Active power plots for the rated load of 500 kW and 1000 kW (twice to rated load)
considering no sag in the utility voltage for Case II are now shown in Fig. 5.48 and 5.49
for ==1800 value.
Fig. 5.48 Dynamic active power variation without sag with =180 when load is 500 kW (Case II).
Fig. 5.49 Dynamic active power variation without sag with =180 when load is 1000 kW (Case II).
172
The presented power export concept (when VL=VU) is valid with the constraint that
line impedance be considered and further change in resulting angles shown in Fig. 5.4 be
also considered. Due to these line impedance angles (for a practical system), power
export concept is valid for only specific values.
5.5
detection by researchers. Some of the commonly used sag detection methods are:
Fourier-transforms, phase-locked-loop, software phase-locked-loop, peak value detection
of the supply waveform and wavelet transformation. Each method has its associated
advantages and disadvantages. However, a time based ramp at a clock rate of 50 Hz is
used to obtain three phase reference signal that is compared with actual three phase utility
voltage to obtain an error signal. If the error in each phase of the utility voltage is greater
than the zero then appropriate control signals are generated. The switching devices in
VSC are switched accordingly to compensate voltage sags, swells and phase jumps in the
utility voltage that propagates to load. For the mitigation of voltage sag, swell and phase
jumps, the unipolar SPWM control is used. The proposed control system succeeds in sag,
swell and phase jump detection and mitigation in less than 10 ms against a target of 20
ms as suggested by ITIC curve and SEMI-F-47 standard.
Conventional voltage compensation techniques utilized by the DVR are Pre-sag, Inphase and Phase advance [19], [187]. Depending upon DVR application environment,
each technique has its own merits and demerits. This work, however, evaluates in-phase
compensation technique for the proposed closed loop control strategy to mitigate voltage
sag, swell and phase jumps in the utility voltage with mathematical equations and the
same have been validated by results obtained from MATLAB/SIMULINK
SimPowerSystems simulations. In-phase voltage compensation strategy injects
minimum voltage vector in the sagged voltage to have regulated load voltage. So size of
the energy storage device and injection transformer is reduced as compared to pre-sag
compensation technique. Phase advance compensation technique, however, uses only
reactive power to mitigate voltage sag. But, all types of sags cannot be mitigated without
174
real power injection especially when power factor of the load is high. Usually, it is stated
in the literature that in-phase compensation scheme compensates only voltage magnitude
and is not able to mitigate phase jump in the utility voltage hence load voltage observes
phase-shifts. However, the proposed closed loop control scheme is also investigated to
tackle phase jumps in the utility voltage and it responded excellently by compensating the
phase jumps.
DVR follows same principle as for sag compensation to counter voltage swell so that
load voltage can be restored to its nominal value [23]. Hence DVR has the ability to
generate active and reactive power during sag as well as absorb active and reactive power
during swell. When active power contribution from DVR is required to compensate
larger voltage sags, active power available from any DG source can be utilized to inject
the voltage difference that compensates the load voltage to required level [215]. A DG as
an energy source connected to DC-link of DVR is shown in Fig. 5.50.
175
Then in phasor notation, the load voltage for the system of Fig. 5.50 can be written as
[187], [218]:
VU + VDVR = VL
(5.15)
(5.16)
(5.17)
Where
V * DVR = Reference voltage phasor for DVR
176
As in-phase injection technique is used for this work hence, it is required from the
DVR to inject missing voltage that is in-phase with utility voltage. This concept is shown
in Fig. 5.51.
Although, DVR injects minimum voltage vector for in-phase injection scheme [90],
however, active power supply from some kind of energy storage device is required from
DVR to mitigate sag using the scheme. The DC-bus of DVR may be supplied with
energy from the main supply through a rectifier or may be fed from engine driven
generating sets fueled by gas, gasoline or diesel. Alternate energy resources or DG in the
form of solar, wind or fuel cells can also be utilized as energy source of DC-bus.
However, the assumption made earlier, that PV source with battery storage is connected
to DC-link of DVR as DG to maintain the constant DC-link voltage, is also valid for this
strategy. This configuration is shown in Fig. 5.50. In the absence of an energy reservoir,
the DC-bus consists of capacitors and the ability to mitigate PQ problems is limited.
Active and reactive power flow equations (5.8~5.13) of LSVI mathematical model at
utility bus (PU, QU), load bus (PL, QL) and power injected by DVR (PDVR, QDVR) can be
modified for in-phase injection scheme as under:
177
PU = V U I L Cos
(5.18)
QU = V U I L Sin
(5.19)
(5.20)
(5.21)
PL = V L I L Cos
(5.22)
QL = V L I L Sin
(5.23)
As utility and DVR supplies active and reactive power during sag to serve load power
demand, hence equations (5.22) and (5.23) can also be written as
PL = PU + P DVR
(5.24)
QL = QU + Q DVR
(5.25)
Substituting equation (5.18) and (5.22) in equation (5.24) to obtain equation for PDVR as
PDVR = (V L - V U ) I L Cos
(5.26)
Now substituting equation (5.19) and (5.23) in (5.25) and solving for QDVR gives
QDVR = (V L - V U ) I L Sin
(5.27)
The equations (5.18~5.27) set out guidelines for dynamic voltage restoration using inphase voltage injection to attain regulated load bus voltage. This will inturn result in
smooth and clean power delivery to load.
179
In order to control the switching of the VSC, a three-phase sinusoidal reference signal
of similar amplitude and phase as to that of utility voltage has to be generated for
comparison with actual utility voltage in the control system. Block diagram for the
reference signal generation is shown in Fig. 5.52 where first of all a time based ramp at a
clock rate of 50 Hz is generated. For an in-phase injection scheme, DVR injects voltage
in phase to the utility voltage. Hence three reference signals are generated so as to follow
the phase sequence of three-phases of utility voltage i.e. reference signals are phase
locked with utility voltage.
180
Generation of error signal is a key step in the control system design. Subtraction of
three-phase actual utility voltage from the generated three-phase reference provides a
three phase error signal. This error signal is then separated for each phase through
demultiplexer block. Each of the error signal drives the corresponding phase pulse
generator which in turn provides control signals (pulses) to the corresponding PWMIGBT VSC switch to turn on and off according to requirement.
An equation for error signal in phase-a of utility voltage can be written as
v Ra (t) - vUa (t) = ea(t)
(5.30)
Similarly for other two-phases equations for error signal can be written as:
v Rb (t) - vUb (t) = eb(t)
(5.30)
(5.30)
181
The generated three phase sinusoidal reference signals have constant amplitude and
each separated by 120 phase apart whereas the utility voltage follows the reference
signals so that there is minimum error in each phase. The bottom line, therefore, for the
control is to try and make the error signal zero. There are various methods that can be
used to implement such a control. However, this research work utilizes unipolar SPWM
Control to obtain the control targets [1].
In the unipolar control law, the states of four switches (IGBTs) of H-bridge
configuration of Fig. 4.2 will be as shown in the Table 5.3. In it, the two sinusoidal
modulated signals (here error signals) Vmod are compared with triangular (carrier) wave
(Vcarr) so that switching signals can be generated for the H-bridge VSC valves. These
switching signals are generated according to following unipolar SPWM control law.
If modulating signal (+Vmod) is greater than the carrier signal (Vcarr) then switch
Ta+ is ON otherwise switch Ta+ is OFF. However, if modulating signal (+Vmod) is
less than the carrier signal (Vcarr) then, switch Ta- is ON otherwise switch Ta- is
OFF.
If modulating signal (-Vmod) is greater than the carrier signal (Vcarr) then switch
Tb+ is ON otherwise switch Tb+ is OFF. However, if modulating signal (-Vmod) is
less than the carrier signal (Vcarr) then, switch Tb- is ON otherwise switch Tb- is
OFF.
182
Sr.
No.
1
2
3
4
Table 5.3
States of four switches of H-bridge VSC configuration
VSC output voltage
Switch Ta+ Switch Tb+ Switch Ta- Switch Tb(VAB)
Status
Status
Status
Status
ON
ON
OFF
OFF
OFF
ON
OFF
ON
OFF
OFF
ON
ON
ON
OFF
ON
OFF
+Vdc
0
0
-Vdc
Control actions generate the firing signals for each VSC switch with controllable
amplitude, phase, and frequency, whenever sag, swell and phase jump is detected. These
firing signals are high frequency pulses with polarity as controlled by the control action
at the output of VSC. However, DVR has to inject a voltage having sinusoidal waveform
at 50 Hz frequency. This requires that high frequency pulses at the output of VSC have to
undergo filtering process. Hence they are passed through a series LC filter. The per
phase inductance value as well as capacitance value chosen for the LC filter after
extensive simulations are 0.8 mH and 300 F, respectively.
The Table 5.4 provides a summary of design specifications and different component
183
184
185
Table 5.4
Summary of design specifications for a three-phase simulated system
(per phase values)
Sr. No. Component/Parameter description
Value/Range
1
415 V
Rated load voltage (3)
3
Load power factor
0.85 lagging
4
Rated load voltage (per phase)
240 V
Maximum load apparent power (per phase)
750 VA
Maximum load current (per phase)
3.125 A
5
VSC DC-bus voltage (per phase)
120 V DC
6
System frequency
50 Hz
7
Maximum sag depth
50%
- 3 balanced sag
- 3 unbalanced sag
15%
Phase-A
25%
Phase-B
35%
Phase-C
- 3 unbalanced sag
45%
Phase-A
45%
Phase-B
No sag
Phase-C
8
9
10
11
12
13
5.6
< 20 ms
1:2
SPWM
Low pass LC
0.8 mH
300 F
Simulation Results
The simulation setup for the block diagram of Fig. 5.55 is developed according to the
concept of the DVR and the control strategy employed. A MATLAB/SIMULINK
simulation for three-phases of a system compensated by DVR in each individual phase is
carried out for different balanced as well as unbalanced sag/swell and 25 phase jumps
in utility voltage using SimPowerSystems environment. The system frequency is 50
Hz. Set up of DVR includes DC-bus, PWM-IGBT based voltage-sourced converter
(VSC), harmonic (LC) filter and injection transformer of 1:2 ratio. The rated voltage per
phase of the system is 240 V. This means that the desired load voltage per phase is 240
V. Though, DVRs are usually used for voltage sag mitigation where consumer supplied
186
voltage is in general 11 kV, the selected voltage level is suitable for a prototype design in
a laboratory environment. The per phase load power consumption is chosen as 750 VA
which is chosen for simplicity of calculation. This corresponds to a rated current of 3.125
A at 240 V. The volt-ampere rating of injection transformer per phase has been taken as
750 VA which is the same as the maximum load power. Voltage rating of the transformer
connected to the main circuit is 240 V whereas that of the winding connected to the VSC
side as 120 V. DC-bus has been taken as an ideal DC source of 120 V. It is assumed that
significant power from DG is available for conversion to AC and contribution to the main
system.
Three different cases have been simulated using proposed feed-forward VSC control
strategy.
Case I: Compensation of voltage sags in utility voltage.
Case II: Compensation of voltage swells in utility voltage.
Case III: Compensation of phase jumps in utility voltage.
generated after comparing utility voltage with generated three-phase reference is shown
in Fig. 5.57, which also shows the corresponding sag magnitude in p.u. Now according to
the feed-forward VSC control strategy, the DVR quickly injected desired three-phase
voltage components in the utility voltage by following the error signal and compensated
the sag while load voltage is maintained constant as shown in Fig. 5.58. Simulation result
shows that the load voltage is restored in less than 10 ms against a target of 20 ms as
suggested by ITIC curve and SEMI-F47 standard.
Large spikes in load voltage are visible in Fig. 5.58 during transient period. These
spikes occur when DVR is suddenly brought in for sag mitigation as well as brought out
when sag disappears. These transient disturbances are caused by sudden injection and
removal of energy by switching. Time duration of the spikes in load voltage at the start of
sag is 0.81 ms. However, the duration is 0.85 ms when sag terminates. When the DVR is
suddenly involved for sag mitigation then spikes in load voltage reaches -618V at the
188
beginning of sag and -844 V at the end of sag recovery period. Although these spikes in
load voltage are not encouraging but their effect on load voltage waveform is negligible.
Harmonic content in the compensated load voltage is 0.25% (THD=0.25%) that lies
within the acceptable range as suggested by IEEE Std. 519-1992.
Fig. 5.57 Error signal generated due to 50% balanced three-phase sag.
Fig. 5.58 Compensated load voltage for 50% balance three-phase sag.
189
error signal to mitigate the sag in each phase of utility voltage. In this way, load voltage
is kept unchanged and balanced throughout the simulation. This shows that, an
unbalanced voltage disturbance can be treated in a similar manner as for a balanced
voltage disturbance. However, in case of unbalanced sag, the unbalanced source voltage
is decomposed into their positive, negative, and zero-phase sequence components.
.
Fig. 5.60 Error signal generated due to unbalanced sag in all three-phases.
Fig. 5.61 Compensated load voltage with unbalance sag on all three-phases.
191
Negative and zero-phase sequence components are eliminated by the negative and
zero- phase sequence injection voltages, so the magnitudes of the negative- and zerophase sequence injection voltages is same as their source side counterparts but with a
phase displacement of 1800. Voltage restoration is implemented with the positive-phase
sequence component of injected voltage. Energy dissipation in this case is related only to
positive-phase sequence voltage injection. Harmonic content in the compensated load
voltage is measured 0.35% (THD=0.35%) and is found within satisfactory range.
Frequency response curve of the LC filter used for the DVR system shows that, the
filter has desirable characteristics near fundamental frequency region, but unsatisfactory
response at harmonic frequency range such as the amplification of the magnitude and the
phase-angle delay [219]. This response curve also reveals the resonant peak that may
cause a large overshoot during the transient period. Although literature survey revealed
that by employing repetitive control algorithm, these problems can be compensated.
However, for the proposed strategy, THD calculated above and in all other cases of this
work is within permissible limit. Hence, LC filter characteristics at harmonic frequency
range do not significantly hamper the feed-forward control scheme performance and no
additional measures are required at this stage to rectify above mentioned LC filter
problems.
a three-phase utility voltage is simulated. The utility voltage with 45% sag in its twophases is shown in Fig. 5.62. The error signal and compensated load voltage are shown in
Figs. 5.63 and 5.64, respectively. Fig. 5.64 shows that the DVR responded quickly by
generating required voltage components for the two-phases to mitigate 45% sag in the
utility voltage. In this way, it maintains a balanced and constant load voltage. THD for
the compensated load voltage is 0.25% and is within permissible limits.
193
Fig. 5.64 Compensated load voltage (line-to-ground) for 45% sag on two-phases.
Fig. 5.65 shows 45% voltage sag initiated in phase-A at 0.06 s which is kept until
0.12 s, with sag duration of 0.06 s. Both utility and load voltage plots overlap other than
this sag duration. Dashed and solid line in the plot show utility voltage with sag and with
compensated load voltage, respectively. However, voltage injected by the DVR is
represented by dotted line in the same figure. As a result of control strategy employed for
the DVR injected voltage, DVR voltage injection is almost zero during normal operation
but it responds quickly by injecting required voltage magnitude to compensate the sag in
utility voltage after its detection and regulates the load voltage. It is evident from the
simulation that the voltage is restored in less than 10 ms against a target of 20 ms as
suggested by ITIC curve and SEMI-F47 standard. DVR in its compensation mode also
smoothen the active and reactive power drawn by load as shown in Fig. 5.66 and 5.67.
194
Fig. 5.65 Dynamic voltages for 45% sag in Phase-A of utility voltage.
Fig. 5.66 Dynamic active powers for 45% sag in Phase-A of utility voltage.
195
Fig. 5.67 Dynamic reactive powers for 45% sag in Phase-A of utility voltage.
The dynamic voltages, active powers and reactive powers for load, utility and DVR
are plotted for only phase-A of three-phase system to show sag mitigation by the DVR
for the feed-forward control strategy. However, plots of dynamic voltages, active powers
and reactive powers for the other two-phases can be obtained on the similar analogy.
with reference is shown in Fig. 5.69, which also shows the corresponding swell
magnitude in p.u. Now according to the feed-forward VSC control strategy, the DVR
quickly injected desired three-phase voltage components in the utility voltage by
following the error signal and compensated the swell while load voltage is maintained
constant as shown in Fig. 5.70.
Fig. 5.69 Error signal generated due to 50% balanced three-phase swell.
197
Large spikes in load voltage are visible in Fig. 5.70 during transient period. These
spikes (or transient disturbances) occur when DVR is suddenly activated for swell
mitigation and sudden injection of energy takes place by switching. Time duration of
spikes in load voltage at the start of swell is 0.81 ms. When the DVR is suddenly
involved for swell mitigation then spikes in load voltage reaches -844 V at the beginning
of swell. Although spikes in load voltage are not encouraging but, they have not effected
load voltage waveform considerably. Harmonic content in the compensated load voltage
is found 0.28% (THD=0.28%) that lies within the acceptable range.
Fig. 5.70 Compensated load voltage for 50% balance three-phase swell.
Fig. 5.71. This figure shows the utility voltage with 15 %, 35% and 25% sag in phase-A,
B and C, respectively. Voltage swell in the three phase utility voltage is initiated at 0.06 s
and kept until 0.12 s.
Fig. 5.72 shows the generated error signal after comparing three-phase utility voltage
with the reference signal whereas the compensated load voltage is shown in Fig. 5.73.
The peak value of utility voltage under normal conditions is 586 V, whereas it is 498 V
for phase-A, 439 V for phase-B and 381 V for phase-C during sag. As seen from Fig.
5.73, DVR responds quickly by injecting required each phase compensating voltage
according to the error signal to mitigate the swell in each phase of utility voltage. In this
way, load voltage is kept unchanged and balanced throughout the simulation. Harmonic
content in the compensated load voltage is measured 0.37% (THD=0.37%) and is found
within satisfactory range.
199
.
Fig. 5.72 Error signal generated due to unbalanced swell in all three-phases.
Fig. 5.73 Compensated load voltage with unbalance swell on all three-phases.
200
201
Fig. 5.76 Compensated load voltage (line-to-ground) for 45% swell on two-phases.
202
Fig. 5.77 Dynamic voltages for 45% swell in phase-A of utility voltage
In addition to the measurements taken for the dynamic behavior of different voltages
and powers for 45% sag in Phase-A of utility voltage, dynamic voltages and powers for
45% swell in Phase-A of utility voltage are also shown in Figs. 5.77-5.79, respectively.
Power factor of the load has been taken 0.85 lagging as mentioned in Table 5.4. Dotted
line in Fig. 5.77 shows the injected voltage that is generated by DVR according to the
developed control strategy to correct load voltage. The load voltage is kept near to rated
value as evident by solid line in the plot.
Because DVR follows same principle as for sag compensation to regulate load
voltages during swell, hence simulation results in Figs. 5.78 and 5.79 show that DVR
absorbs excess power during swell to regulate load voltage to smoothen the power flow
to load. The comparison of above two cases (5.6.1 and 5.6.2) shows that the DVR reacts
203
instantly during sag/swell to correct load voltage to its rated value according to developed
VSC control scheme.
Fig. 5.78 Dynamic active powers for 45% swell in utility voltage.
Fig. 5.79 Dynamic reactive powers for 45% swell in utility voltage.
The dynamic voltages, active powers and reactive powers for load, utility and DVR
are plotted for only phase-A of three-phase system to show voltage swell mitigation by
204
the DVR for the feed-forward control strategy. However, plots of dynamic voltages,
active powers and reactive powers for the other two-phases can be obtained on the
similar analogy.
Fig. 5.80 Dynamic voltages for +25 phase jump in Phase-A of utility voltage.
only a phase jump. Fig. 5.80 shows dynamic voltages for a +25 phase jump. It is evident
from this figure that there is injection of voltage from DVR until phase jump persists so
that an effect of utility voltage phase-shift on load voltage is nullified. After the
termination of the phase jump, there is zero injection of voltage from DVR and normal
operation is restored.
Fig. 5.81 Dynamic active powers for +25 phase jump in Phase-A of utility voltage.
Figs. 5.81 and 5.82 show active and reactive power plots of load, utility, and DVR
during +25 phase jump in utility voltage. As power factor is inversely proportional to
the phase-angle hence an increase in value of angle between utility voltage and load
current due to +25 phase jumps will decrease active power supply and increase reactive
power supply from utility source. Hence DVR will try to inject active power as well as
absorb excess utility reactive power during the +25 phase jump to smoothen the power
flow to load. However, Fig. 5.81 shows 100% rated load power is not maintained during
phase jump. Moreover, active power from utility decreases from 637.5 W to 400 W
206
during +25 phase jump. To compensate deficient active power flow to load, DVR
reacted quickly to supply deficient active power i.e. 237.5 W, but it could supply 176.5
W. Hence active power plot of load seems somewhat below than its rated value. Fig. 5.82
shows that 100% rated load reactive power is not maintained during +25 phase jump.
Moreover, it shows reactive power from utility increases from 395 VAR to 588.5 VAR.
To compensate excess reactive power supplied by utility, DVR reacted quickly to absorb
surplus reactive power i.e. 193.5 VAR, but it absorbed 232.5 VAR. Hence reactive power
plot of load seems slightly below than its rated value.
Fig. 5.82 Dynamic reactive powers for +25 phase jump in Phase-A of utility voltage.
Fig. 5.83 Dynamic voltages for -25 phase jump in phase-A of utility voltage.
Figs. 5.84 and 5.85 show active and reactive power plots of load, utility, and DVR
during -25 phase jump in utility voltage. As power factor is inversely proportional to the
phase-angle hence decrease in the value of angle between utility voltage and load current
due to -25 phase jump will increase active power supply and decrease reactive power
supply from utility source. Hence DVR will try to absorb excess utility active power as
well as supply reactive power during the -25 phase jump to smoothen the power flow to
load. However, Fig. 5.84 shows that 100% rated load power is not maintained during 25 phase jump. Moreover, it shows active power from utility increases from 637.5 W to
750 W. To compensate excess active power flow to load, DVR reacted quickly to absorb
excess utility supplied power i.e. 112.5 W, but it absorbed 91 W. Hence active power plot
of load seems slightly raised than its rated value. However, it is just 3.5 % above than the
rated value. Fig. 5.85 shows that 100% rated load reactive power is not maintained
during -25 phase jump. Reactive power from utility decreases from 395 VAR to 125.5
VAR during -25 phase jump. DVR responded quickly to supply deficient reactive power
208
i.e. 269.5 VAR, but it supplied 283.2 VAR. Hence reactive power plot of load seems
slightly above than its rated value. However, it is just 3.2 % above than the rated value.
Fig. 5.84 Dynamic active powers for -25 phase jump in Phase-A of utility voltage.
Fig. 5.85 Dynamic reactive powers for -25 phase jump in phase-A of utility voltage.
209
5.7
Hence due to above mentioned power reliability and quality events, there are two
types of costs associated with each event.
Direct cost
Indirect cost
210
Table 5.6
Examples of PQ problems, costs, and solutions in various industries*[222]
Sr.
No
.
1
Customer
segment
Industrial
cement plant
Industrial
chemical plant
Picture tube
manufacturerindustrial
Electronics
part for
Automotive Industrial
Aircraft
Engine
manufacturerindustrial
Paper and
pulp-Industrial
Paper and
pulp-Industrial
Type of
equipment
problem
Computer and
programmable
logic
controller
failures
Equipment
shutdowns
Type of
power line
disturbance
Voltage
sags
Solution
Cost of PQ
problem
Source of
information
DVR
Loss =
$1,000,000/plant
shutdown
Utility and
industrial
customer
Power line
disturbance
Loss=
$1,000,000/minute
Utility
Downtime,
loss of
production due
to process
interruption
Downtime,
loss of
production due
to process
interruption
Downtime,
loss of
production due
to process
interruption
Loss of
productivity+
Equipment
damage +
Start-up time
Voltage
sags and
momentary
interruption
s
Voltage
sags
Information
not
available
DVR
$100,000/momentary
interruption,
$40,000/sag
CustomerEPRI/PEA
C
DVR
$25,000 per
process
interruption
CustomerEPRI/PEA
C
Voltage
sags
DVR
CustomerEPRI/PEA
C
Voltage
sags
DVR
CustomerEPRI/PEA
C
Turnover
losses
Voltage
sags and
momentary
interruptions
DVR,
Large scale
reactive
power
$492,750 for 7
process
interruptions in
1996 includes
$100,000
equipment damage
Total $2,740,000
turnover losses at
8 paper and pulp
mill in South
Africa in one year
compensatio
n
Electronics
manufacturing
-Industrial
Component
damage + Loss
of production
Voltage
sags and
momentary
interruptions
Semiconductor
or fabrication
processing industrial
Wafer losses
Voltage
sags and
momentary
interruption
s
*Compiled by EPRI/PEAC
211
DVR
$50,000/year for
board replacement
and $50,000/year
in loss of
productivity
Annual wafer
losses $3.1 million
CustomerSouth
African
Power
quality
conference
EPRI PQ
database
From
customerSilicon
Systems
Inc.
Direct cost is borne by utility and this cost is simply calculated by multiplying the
rate of units to the number of units of electricity that are not being utilized by customer
during interruption of supply. However, an indirect cost varies industry to industry but
some costs variables are common. That is,
Production stoppage
Cost of repair
In some industries indirect costs due to sag/swell or outage is enormous and are million
folds to that of cost of electricity (direct cost). Table 5.6 shows this aspect.
According to IEEE Std. 493-1997, average cost of single power interruption for
industrial plants is as shown in Table 5.7 [223]:
Table 5.7
Average cost of single power interruption for industrial plants
Type of plant with respect to demand of energy
For all plants
For plants having max demand > 1MW
For plants having max demand < 1MW
Also IEEE Std. 493-1997 reveals average cost of single power interruption for
commercial and office buildings as shown in Table 5.8[224]:
Table 5.8
Average cost of single power interruption for commercial and office buildings
Type of load
For all commercial buildings
For all office buildings
212
Table 5.9 shows how varying length interruptions, costs the most common industrial
sectors, that is: continuous process manufacturing (CPM), Fabrication and essential
services (FES), and digital economy (DE) [223]. This table also shows that CPM
industrial sector interruption overall average costs are higher than other two sector
whereas DE sector is less affected with interruption.
Table 5.9
Cost per outage for CPM, FES and DE sectors for different outage duration [223]
Sr.
No.
1
Type of consumer
Continuous process
manufacturing
Fabrication
and
industrial services
Digital economy
2
3
Cost
of
1
second
outage
($)
12,500
Cost
of
Recloser
operation ($)
2,500
Cost of
minute
outage ($)
17,500
Cost of 1 hour
outage ($)
500
7,500
1000
13,000
1,000
500
1,500
5,000
14,600
Table 5.10 shows an EPRIs Consortium for Electric Infrastructure for a Digital
Society (CEIDS) survey results that compares the cost of a one hour outage to industrial
sector with and without specific equipment and processes [222].
Table 5.10
Selected industries equipment and processes average cost/hour outage [222]
Sr.
No.
1
2
Electronic circuits
Computer assisted production
systems
Variables speed drives
Refrigeration
Electric process heating
3
4
5
25,000
26,000
20,500
2,500
11,000
11,000
Table 5.11 shows an industry specific annual cost of power quality within the CPM
sector that helps in macro level planning purpose.
213
Table 5.11
Selected industries equipment and processes average cost/hour outage*
Sr. No.
1
2
3
4
5
6
7
8
9
10
11
Cost in $/kW
8.0
11.7
1.7
20.6
3.0
8.0
15.5
58.3
10.0
28.6
3.7
* Compiled by EPRI/PEAC
To observe the annual cost of outage events per establishment, Table 5.12 is
presented below that shows the larger establishments have largest outage costs.
Table 5.12
Cost of outage events per establishment by annual kWh*
Sr. No. Per establishment energy utilization
1
Less then 50 MWh
2
Between 50-499 MWh
3
Between 500 MWh-4.9GWh
4
Above 5 GWh
* CEIDS survey
CEIDS survey reported that for 2 million U.S. industrial and DE establishments, total
annual outage losses are $45.7 billion. However, the establishments that consume energy
greater than or near to 5GWh are almost 58,000 and they have share of $9.6 billion
outage losses that accounts 21% of all the losses that occurs due to power outage.
So far different aspects of cost of power outage have been presented. However, Table
5.13 shows costs for QRA events such as sags, swells, and harmonics, when there is no
complete interruption of power. Again, similar to outage cost analysis, the establishments
that consume energy greater then or near to 5GWh have $60,000 average annual cost of
PQ events whereas EPRIs CEIDS survey reported that for 2 million U.S. industrial and
214
DE establishments, annual PQ losses are $3.5 billion out of total $6.7 billion and that
accounts for 52% of all the PQ losses.
Table 5.13
Cost of PQ events per establishment by annual kWh*
Sr. No.
1
Less then 50 MWh
2
Between 50-499 MWh
3
Between 500 MWh-4.9GWh
4
Above 5 GWh
* CEIDS survey
$ 174
$2,698
$2,768
$59,601
P. T. Cheng and his colleagues reported that each sag event in Taiwanese
semiconductor foundries that are mostly located in Hsinchu Science-based Industrial
Park (HISP) results in US $100,000 to US $10, 00,000 financial losses because of clean
up work, down time, and faulty products [225]. Authors presented a Table 5.14 that
depicts the contract capacity, backup generators capacity and UPS capacity of different
foundries in HISP. But, in 1999 and 2000, 62.7% of voltage sags still affected the
production of foundry A according to Table 5.15 although, backup generator and UPS
have been provided as backup power sources. The cost effectiveness of these back-up
power sources due to these unclear sag events remained unknown.
Fig. 5.88 shows that DVR provides most cost effective solution although low speed
flywheel has same cost/kVA as that of DVR but this solution requires again voltagesourced converter to convert its DC output to AC, as generally loads are operated at AC
voltage. Hence requirement of additional VSC will increase the solution cost of low
speed flywheel having DC output as compared to DVR.
215
Table 5.14
Backup power sources capacity of semiconductor foundries in Taiwan [225]
Sr. No.
Foundries
Contract capacity
1
A*
15
2
B
55
3
C
12
4
D
12
5
E
16.5
6
F
31
7
G
23
8
H
16
* Table 5.18 below is recorded by foundry A
7.5
22.5
6.0
7.5
6.0
21
32
8.0
UPS capacity
(MVA)
9.4
36.1
3.72
4.8
19.0
14
23.5
18.6
Table 5.15
Power disturbances recorded by foundry A in Table 5.17 [225]
Sr. No.
Type of disturbance
1997
1998
1999
2000
1
2
Outages
Sags
2
12
1
11
3
27
2
16
Fig. 5.86 shows mitigation devices and price in $/kVA of each solution provided by these
devices [222].
700
600
500
400
300
200
100
Mitigation Devices
216
5.8
Summary
Different DVR control strategies that eliminate propagation of voltage sags, swells
and phase jumps in the utility voltage to load are presented and evaluated through
simulations. The calculations and simulation results for LSVI scheme show that DVR
can contribute towards the mitigation of sag as well as share power in a manner which
suits a specific customer. In case of abundant availability of generated power from a
distributed energy resource, maximum power can be shared while in case of scarcity,
minimum power sag mitigation can be managed.
In addition, a new concept of power export through a series generator is presented,
whose power export capability is managed by injected voltage magnitude and phaseangle instead of the current. The current always remain constant in magnitude and is
equal to the load current. This may act as a useful tool when there is power generated at
consumer premises and the consumer may get the benefit of price obtained therein. In
developing countries, there is always shortage of power on the grid, electric utilities may
benefit from injection of power from the consumer side. However, the approach of
mitigating voltage sag using LSVI scheme is suitable for those locations where the
phase-shift in the voltage will not cause any problem.
Simulation results are shown for a single-phase DVR but these results can be
extended for a three-phase DVR by considering sag in three-phase utility voltage with
three-phase load using proposed LSVI VSC control strategy. In this case, three-phase
VSC is made by three single-phase converters (VSC) connected to star connected
primary of interface transformer. The proposed strategy should also compensate sag and
export power to utility for three-phase case as well.
217
Next, a feed-forward control strategy for the VSC of DVR is presented that is based
on unipolar SPWM switching strategy to detect and mitigate sag, swell and phase jumps
in the three phase utility voltage. The proposed strategy is although simple but works
efficiently so that the VSC not only converts DC to AC but also mitigates sag, swell and
phase jumps in the utility voltage, acting as power flow controller. The unipolar SPWM
control strategy continually monitors the supply and makes correction if sag, swell or
phase jump is detected. This method regulates load voltage approximately within 10 ms
such that DVR either injects or absorbs required voltage that is in-phase with the utility
voltage. The proposed control strategy succeeds in achieving load voltage THD limit
within 5%. The power shared by the utility and DVR to meet load demand is as per
requirement to compensate sag or swell considering the capacity of available DG. With
the help of proposed control scheme of DVR, the dynamic performance capability of
DVR increases the sensitive load penetration in the distribution system. Feed-forward
method of compensation shows promise in keeping the load voltage at the correct level
and phase-shift free. The advantage of the in-phase compensation is that the magnitude of
injected DVR voltage is minimial for constant load voltage magnitude.
Different voltage sag detection techniques have been utilized in DVRs such as
Fourier- transform, phase-locked-loop, software phase-locked-loop/vector control, peak
value detection of the supply waveform and wavelet transformation. Each technique has
its own merit and demerit. However, in this research work, instantaneous voltage
comparison of actual utility voltage and generated three-phase reference voltage has been
carried out to detect sag, swell and phase jump in utility voltage using in-phase injection
218
scheme for feed-forward VSC control strategy. The sag detection technique responded
excellently as evident from the presented simulation results.
Cost benefit analysis study reveals that there is enormous amount of indirect cost that
the each industrial and commercial establishment has to face due to disturbances such as
sags (swells) or outages. The DVR is one of the CPD that have the ability to mitigate
voltage sags, swells, phase jumps and reduce harmonics. It provides a cost effective
solution as compared to other mitigation solutions. Its application helps in regulating load
bus voltage to avoid interruption of critical industrial processes and maintains smooth
and clean power delivery to the utility load.
219
CONCLUSIONS
The calculations and simulation results for LSVI scheme show that DVR can
contribute towards mitigation of sag as well as share power in a manner suitable to a
specific customer. In case of abundant availability of generated power from a distributed
energy resource, maximum power can be shared while in case of scarcity, sag mitigation
can be achieved with minimum power. The power export through a series generator is
managed by injected voltage magnitude and phase-angle instead of current. The current
always remains constant in magnitude and is equal to load current. This may act as a
useful tool when there is power generated at consumer premises and consumer may get
the benefit of extra revenue from power export to utility. The approach of mitigating
voltage sag using LSVI is suitable for those applications where the phase-shift in the
voltage do not disturb load operation.
Feed-forward VSC control strategy for the DVR utilizing a Unipolar SPWM control
is used to detect and mitigate PQ issues in the utility voltage. The feed-forward control
system succeeds in sag, swell and phase jump detection and mitigation in less than 10 ms
against a target of 20 ms. THD in load voltage is within the limits of 5%. These results
show that control strategy employed works satisfactorily to provide excellent voltage
regulation with acceptable harmonic distortion.
The control strategies successfully restored and regulated load bus voltage to its
nominal value against disturbances in the utility voltage. The results of simulation reveal
that each strategy is beneficial to utility and consumers, especially, commercial and
industrial consumers. However, each strategy depends upon its application requirement.
By using these control strategies of DVR, interruption of critical industrial processes
220
during disturbance in utility voltage can be avoided. Hence, by opting any of the above
VSC control strategies for the DVR, QRA of the utility voltage to consumers can be
enhanced. This inturn, reduces direct cost for utility as well as indirect cost of consumers.
221
FUTURE RECOMMENDATIONS
Synchronization of the reference signal with the utility voltage for the feed-forward
control strategy is done by generation of time based ramp at a clock rate of 50 Hz.
Although, THD in load voltage is less than 3%, however, alternate synchronization
methods may be explored for enhancing the performance of the strategy. Simulation
results for active and reactive power plots while mitigating phase jumps in three-phase
utility voltage for feed-forward control strategy of DVR show that rated load power do
not flow to load. This problem can be solved by adding a load voltage feedback in the
strategy. This will further improve the voltage regulation at the load bus and rated load
power may flow to the load.
It is assumed that PV source with battery storage system is connected to DC-link of
the DVR to maintain constant DC-link voltage. However, in the simulations, an ideal DC
voltage source is used to simplify the analysis. Therefore, a mathematical model of PV
source in MATLAB/SIMULINK SimPowerSystems tool box needs to be
developed. Moreover, a control of DC-DC converter would be required to control the
charging of battery banks.
In LSVI strategy, DVR goes to standby mode when IL=0. This limits DVR usage.
Hence in order to increase application of DG type DVR, closed loop control of can be
designed so that DG type DVR automatically sets its values corresponding to IL. In this
way, if IL=0 then will be so adjusted that it supplies power to grid without going into
standby mode. Moreover, simulation results for the LSVI control strategy may be
extended to a three-phase DVR by considering sag in the three-phase utility supply with
three-phase load.
222
223
REFERENCES
[1] A. Ghosh and G. Ledwich, Power Quality Enhancement Using Custom Power
Devices, Kluwer Academic Publishers, The Netherlands, 2002.
224
[11] E. Lakervi and E. J. Holmes, Electric Distribution Network Design, IEE Power
Engineering Series, London: Peter Peregrines Ltd., 1989.
[12] H. L. Willis, Power Distribution Planning Reference Book, 2nd Edition, CRC
Press, USA, 2004.
[13] S. O. Uwaifo, Electric Power Distribution Planning and Development, The Nigerian
Experience, Lagos State, Nigeria: Hanon Publishers Limited, 1998.
[14] T. A. Short, Electric Power Distribution Equipment and Systems, CRC Press, Taylor
and Francis Group, USA, 2006.
[15] T. A. Short, Electric Power Distribution Handbook, CRC Press, Taylor and Francis
Group, USA, 2004.
[16] C. W. Brice, Comparison of approximate and exact voltage drop calculations for
distribution lines, IEEE Transactions on Power Apparatus and Systems, vol. PAS101, no. 11, pp. 4428-4431, Nov. 1982.
[17] A. O. Al-Mathnani, A. Mohammad and M. A. M. Ali, Photovoltaic based Dynamic
voltage restorer for voltage sag mitigation, The 5th Student conference on Research
and Development, Malaysia, Dec. 2007.
[18] M. H. J. Bollen, Understanding Power Quality Problems: Voltage Sags and
Interruptions, IEEE Press, New York, 2000.
[19] A. Rai and A. K. Nadir, Modeling and simulation of Dynamic Voltage Restorer
(DVR) for enhancing voltage sag, Sensors and Transducers Journal, vol. 87, no. 1,
pp. 85-93, 2008.
225
[20] J. Wang, S. Chen and T. T. Lie, Estimating economic impact of voltage sags,
Proceedings of IEEE International Conference on Power System Technology,
[22] A. Baggini, Handbook of Power Quality, John Wiley and Sons, Ltd. West Sussex,
England, 2008.
[23] S. S. Choi, J. D. Li and D. M. Vilathgamuwa, A Generalized voltage compensation
strategy for mitigating the impacts of voltage sags/swells, IEEE Transactions on
Power Delivery, vol. 20, no. 3, pp. 2289-2297, July 2005.
[24] C. S. Lam, M. C. Wong and Y. D. Han, Voltage swell and over voltage
compensation with unidirectional power flow controlled Dynamic Voltage Restorer,
IEEE Transactions on Power Delivery, vol. 23, no. 4, pp. 2513-2521, 2005.
[29] A. Tuladhar, H. Jin, T. Unger and K. Mauch, Parallel operation of single phase
inverters with no control interconnections, Proceedings of IEEE Applied Power
Electronic Conference, vol. 1, 1997, pp. 94-100.
pp. 526-532.
[32] E. J. Davison and B. Scherzinger, Perfect control of the robust servomechanism
problem, IEEE Transactions on Automatic Control, vol. 32, pp. 689-702, Aug. 1987.
[33] S. J. Chiang and J. M. Chang, Parallel control of the UPS inverters with frequencydependent droop scheme, Proceedings of IEEE Power Electronics Specialist
Conference, vol. 2, 2001, pp. 957-96l.
227
[44] T. V. Van and R. Belmans Distributed generation overview: current status and
challenges, International Review of Electrical Engineering, vol. 1, no. 1, pp. 178189, 2006.
[45] P. A. Daly, and J. Morrison, Understanding the potential benefits of distributed
generation on power delivery systems, Proceedings of IEEE Rural Electric Power
Conference, USA, pp. 1-13, 2001.
[46] DR012012-4 Draft Australian Standard for Grid Connection of Energy Systems via
Inverters.
[47] IEEE Standard P1547/D07, Draft Standard for Interconnecting Distributed
Resources with Electric Power Systems, The Institute of Electrical and Electronics
multi-type
FACTS
devices,
Electric
Power
Systems
Research,
229
[55] H. W. Ngan and W. Fang, Coordinated power control strategy for Flexible AC
Transmission System, Proceedings of IEEE International Conference on Power
Electronics and Drives systems, vol. 2, pp.677-682, 1999.
230
[61] L. Gyugyi, K. K. Sen and C. D. Schauder, The interline power flow controller
concept: a new approach to power flow management in transmission systems, IEEE
Transactions on Power Delivery., vol. 14, no. 3, pp. 1115-1123, July 1999.
231
[70] R. Cao, J. Zhao, W. Shi, P. Jiang and G. Tang , Series power quality compensator
for voltage sags, swells, harmonics and unbalance, Proceedings of IEEE
Transmission and Distribution Conference and Exposition, vol. 1, pp. 543-547, 2001.
[73] A. Sannino and M. H. J. Bollen, Mitigation of voltage sags and short interruptions
through distribution system design, International Symposium EPETM, 2000, Capri,
Italy.
[74] F. Abrahamsen and A. David, Adjustable speed drives with active filtering
capability for harmonic current compensation, Proceedings of IEEE-PESC, USA,
vol. 2, pp. 1137-1143, 1995.
[75] J. G. Nielsen, M. Newman, H. Nielsen and F. Blaabjerg, Control and testing of a
dynamic voltage restorer (DVR) at medium voltage level, IEEE Transactions on
Power Electronics, vol. 19, no. 3, pp. 806-813, 2004.
232
[76] A. Elnady and M.M.A Salama, Unified approach for mitigating voltage sag and
voltage flicker using the DSTATCOM, IEEE Transactions on Power Delivery., vol.
20, no. 2, pp. 992-1000, 2005.
[77] M. Vilathgamuwa, A. A. D. Ranjith Perera and S. S. Choi, Performance
improvement of the dynamic voltage restorer with closed-loop load voltage and
current-mode control, IEEE Transactions on Power Electronics, vol. 17, no. 5, pp.
824-834, Sept. 2002.
[78] D. M. Vilathgamuwa, H. M. Wijekoon and S. S. Choi, A novel technique to
Compensate Voltage Sags in Multiline Distribution System The Interline Dynamic
Voltage Restorer, IEEE Transactions on Industrial Electronics., vol. 53, no. 5, pp.
1603-1611, 2006.
[79] M. R. Banaei, S. H. Hosseini and G. B. Gharehpetian, Inter-line dynamic voltage
restorer control using a novel optimum energy consumption strategy, Simulation
Modeling Practice and Theory, vol. 14, pp. 989-999, 2006.
[80] H. Kim and S. K. Sul, Compensation voltage control in dynamic Voltage restorers
by use of feed-forward and state feedback scheme, IEEE Transactions on Power
Electronics, vol. 20, no. 5, pp. 1169-1177, Sept. 2005.
233
Improvement, IEEE Transactions on Power Delivery., vol. 24, no. 1, pp. 277-284,
2009.
[83] P. R. Snchez and E. Acha, Dynamic voltage restorer based on flying capacitor
multilevel converters operated by repetitive control, IEEE Transactions on Power
Delivery, vol. 24, no. 2, pp. 951-960, 2009.
[84] B. J. Quirl, B. K. Johnson, and H. L. Hess, Mitigation of voltage sags with phase
jump using a dynamic voltage restorer, IEEE Proceedings of 38th North American
Power Symposium, 2006, pp. 647-654.
dynamic voltage restorer, IEEE Transactions on Power Delivery, vol. 19, no. 3, pp.
1511-1518, 2004.
[90] A. K. Jindal, A. Ghosh, and A. Joshi, Critical load bus voltage control using DVR
under system frequency variation, Electric Power Systems Research, vol. 78, pp.
255-263, 2008.
[91] A. Ghosh and A. Joshi, A new algorithm for the generation of reference voltages of
a DVR using the method of instantaneous symmetrical components, IEEE Power
Engineering Review, vol. 22, no. 1, pp. 63-65, 2002.
[92] Il-Yop Chung, D. J. Won, S. Y. Park, S.-Il Moon and J. K. Park, The DC-link
energy control method in dynamic voltage restorer system, Electrical Power and
Energy Systems, vol. 25, pp. 525-531, 2003.
235
[101] B. King and K.J. Olejniczak, An evaluation of static VAR compensation strategies
for voltage flicker mitigation in electric distribution systems, Seventh International
Conference on Harmonics and Power Quality, Las Vegas, pp. 6-11, 1996.
[102] A. Sannino, Static transfer switch: analysis of switching conditions and actual
transfer time, IEEE Power Engineering Society Winter Meeting, USA, vol. 1, pp.
120-125, 2001.
[103] P. Samuelsson, G. Gensini, and G. Mavridis, Danarc technology for electric arc
furnaces, Iron Steel Engineering, vol. 72, no 6, pp. 21-25, 1995.
[104] D. L. Dickmander, B. H. Thorvaldsson, G. A. Stromberg, D. L. Osborn, A.E.
Poitras and D.A. Fisher, Control system design and performance verification for the
chester, main static VAR compensator, IEEE Transactions on Power Delivery, vol.
7, no. 3, pp.1492-1503, 1992.
236
[105] M. Goto, T. Nakamura, Y. Mochinaga and Y. Ishii, Static negative phasesequence current compensator for railway power supply system, International
Conference on Electric Railways in a United Europe, pp. 78-82, 1995.
[106] A. Werme, P. Hyvonen and R. Grunbaum, Improved power quality and economy
by rolling mill by means of an SVC, Steel Times International, pp. 44-46, 1998.
[107] H. Frank and K. Pettersson, Raising the production of arc furnaces by stabilizing
the voltage with thyristor switched capacitors, ASEA Journal, vol. 50, no. 1, pp.9-16,
1977.
[108] R. Grunbaum, T. Gustafsson, T. Johansson, and U. Olsson, Energy and
environmental savings in steel making by means of SVC light, Electric Furnace
Conference Proceedings, 2000.
[110] C. Schauder et al., AEP UPFC project: Installation, commissioning and operation
of the -160MVA STATCOM (Phase I), IEEE Transactions on Power Delivery, vol.
13, pp. 15301535, Oct. 1998.
[111] I. Papic, Power Quality improvement using distribution static compensator with
energy storage system, Proceedings of 9th International Conference on Harmonics
and. Quality Power, pp. 916920, 2000.
[112] O. A. Lara and E. Acha, Modeling and analysis of custom power systems by
PSCAD/EMTDC, IEEE Transactions on Power Delivery, vol. 17, no. 1, pp. 266272, 2002.
237
[116] A. Ghosh and G. Ledwich, Structures and control of a dynamic voltage regulator
(DVR), Proceedings of IEEE-PES Winter Meeting, USA, 2001.
[117] A. Ghosh and G. Ledwich, Load compensating DSTATCOM in weak AC
systems, IEEE Transactions on Power Delivery, vol. 18, no. 4, 2003.
[118] A. Ghosh and G. Ledwich, The concept and operating principles of a mini custom
power park, IEEE Transactions on Power Delivery, vol. 19, no. 4, 2004.
[119] W. Freitas, A. Morelato, W. Xu and F. Sato, Impacts of AC generators and DSTATCOM devices on the dynamic performance of distribution systems, IEEE
Transactions on Power Delivery, vol. 20, no. 2, 2005.
[120] A. Elnady and M. A. Salama, Unified approach for mitigating voltage sag and
voltage flicker using the D-STATCOM, IEEE Transactions on Power Delivery, vol.
20, no. 2, 2005.
[121] N. Mariun, H. Masdi1, S. M. Bashi, A. Mohamed and S. Yusuf, Design of a
prototype D-SATCOM using DSP controller for voltage sag mitigation, Proceedings
of IEEE power India conference, India, 2006.
238
[122] B. Blazic and I. Papic, Improved D-STATCOM control for operation with
unbalanced currents and voltages, IEEE Transactions on Power Delivery, vol. 21,
no. 1, 2006.
[123] A. Shukla, A. Ghosh and A. Joshi, Hysteresis current control operation of flying
capacitor multilevel inverter and its application in shunt compensation of distribution
systems, IEEE Transactions on Power Delivery, vol. 22, no. 1, 2007.
[124] A. Shukla, A. Ghosh and A. Joshi, State feedback control of multilevel inverter
for D-STATCOM applications, IEEE Transactions on Power Delivery, vol. 22, no.
4, 2007.
[125] A. Shukla, A. Ghosh, and A. Joshi, Control schemes for DC capacitor voltages
equalization in diode-clamped multilevel inverter-based D-STATCOM, IEEE
Transactions on Power Delivery, vol. 23, no. 2, 2008.
[128] IEEE Std. 519-1992, IEEE recommended practices and requirements for harmonic
control in electrical power systems, 1992.
[129] M. B. Brennen and B. Banerjee, Low cost, high performance active power line
conditioners, Proceedings of 3rd International Conference on Power Quality: EndUse Applications and Perspectives, EPRI, Amsterdam, pp. 24-27, Oct. 1994.
239
[130] E. J. Stacey and M. B. Brennen, Active power conditioner system, U.S. Patent
4651265, 1987.
[131] H. Fujita and H. Akagi, Voltage regulation performance of a shunt active filter
intended for installations on power distribution system, IEEE Industry Applications
Society 40th annual meeting, 2005.
[135] L. Gyugyi, Unified power flow control concept for Flexible AC Transmission
Systems, Proceedings of IEE, vol. 139(C), pp. 323-331, 1992.
[136] V. Khadkikar and A. Chandra, A new control philosophy for a unified power
quality conditioner (UPQC) to coordinate load-reactive power demand between shunt
and series inverters, IEEE Transactions on Power Delivery, vol. 23, no. 4, 2008.
[137] A. Ghosh and G. Ledwich, A unified power quality conditioner (UPQC) for
simultaneous voltage and current compensation, Electric Power Systems Research,
vol. 59, pp. 55-63, 2001.
240
[141] H. Nasiraghdam and A. Jalilian, Balanced and unbalanced voltage sag mitigation
using DSTATCOM with linear and nonlinear loads, International Journal of
Electrical, Computer, and System Engineering, pp. 86-91, 2007.
[144] R.S. Thallam and S. Eckroad, Multimode battery energy storage system for
custom power applications, Proceedings of IEEE Power Engineering Society Winter
Meeting, vol. 2, pp. 1147-1150, 1999.
241
[149] S.M. Schoenung, W.R. Meier, and R.L. Bieri, Small SMES technology and cost
reduction estimates, IEEE Transactions on Energy Conversation, vol. 9, no. 2,
pp.231-237, 1994.
[150] C. Perry and D. Dorr, Custom power choices abound, [Online]. Available
(August, 22, 2009): http://tdworld.com/mag/power_custom_power_choices/
[151] M. M. Granaghan, T. Grebe, G. Hensley, T. Singh and M. Samotyj, Impact of
utility switched capacitors on customer systems, part II- adjustable speed drive
concerns, IEEE Transactions on Power Delivery, vol. 6, no. 4, pp. 1623-1628, 1991.
[152] M. M. Granaghan, T. Grebe, G. Hensley, T. Singh and M. Samotyj, Impact of
utility switched capacitors on customer systems, part I- adjustable speed drive
concerns, IEEE Transactions on Power Delivery, vol. 7, no. 2, pp. 862-868, 1992.
[153] C. S. Chang and P. C. Loh, Integration of fault current limiters on power systems
for voltage quality improvement, Electric Power Systems Research, vol. 57, pp. 8392, 2001.
242
[154] P.C. Loh, C. Hong, S. Mok, C. S. Chang and S. Elangovan, Voltage quality
enhancement with Wavelet- Artificial Intelligence Current Limiting Devices,
Electric Machines and Power Systems, vol. 28, pp. 811-822, 2000.
[156] H. Mokhtari, S. B. Dewan, M.R. Iravani, Analysis of a static transfer switch with
respect to transfer time, IEEE Transactions on Power Delivery, vol. 17, no.1,
pp.190-199, 2002.
[157] A. Sannino, Power quality improvement in an industrial plant with motor load by
installing a static transfer switch, Proceedings of IEEE Industry Applications Society
Annual Meeting, USA, vol. 2, pp. 782-788, 2001.
[158] H. Mokhtari, S.B. Dewan, M.R. Iravani, Effect of Regenerative Load on a Static
Transfer Switch Performance, IEEE Transactions on Power Delivery, vol. 16, no. 4,
2001.
[159] T. Larsson, R. Innanen, and G. Norstrom, Static electronic tap-changer for fast
phase voltage control, IEEE Proceedings of International Conference on Electric
Machines and Drives, pp. TC3/4.1-TC3/4.3, 1997.
243
[161] R. Buxton, Protection from voltage dips with the dynamic voltage restorer,
Proceedings of IEE Half Day Colloquium on Dynamic Voltage Restorers: Replacing
Those Missing Cycles, pp: 311-316, 1998.
using
DVR,
[Online].
Available
(August,
23,
2009):
http://www.scribd.com/doc/2876632
[165] C. Zhan, A. Arulampalam and N. Jenkins, Four wire dynamic voltage restorer
based on a three dimensional voltage space vector PWM algorithm, IEEE
Transactions on Power Electronics, vol.18, no.4, pp.1093-1102, 2003.
[166] H. Kim and S. K. Sul, Compensation voltage control in Dynamic voltage restorers
by use of feed-forward and state feedback scheme, IEEE Transactions on Power
Electronics, vol. 20, no.5, pp. 1169-1177, 2005.
[173] M. H. Rashid, Power Electronics Handbook. San Diego: Academic Press, 2001.
[174] N. Kularatna, Power Electronics Design Handbook: Low-Power Components and
Applications. Boston: Newnes, 1998.
[178] S. Bernet, State of the art and developments of medium voltage converters an
overview, Przeglad Elektrotechniczny (Electrical Review), vol. 82, no. 5, pp.1-10,
245
2006.
[179] D. M. Vilathgamuwa and A. A. D. Ranjith Perera, Voltage sag compensation with
energy optimized Dynamic voltage restorer, IEEE Transactions on Power Delivery,
vol.18, no. 3, pp.928-936, 2000.
[180] P. T. Nguyen and T. K. Saha, Dynamic voltage restorer against balanced and
unbalanced voltage sags: Modeling and simulation, Proceedings of IEEE Power
Engineering Society General Meeting, vol. 1, pp. 639-644, 2004.
246
[186] S. Kumar Dynamic voltage restorers (DVR) and their control [Online]. Available
(August 23, 2009):http://sureshks.netfirm.com/article/dvr/dvr.
[187] M. Tumay, A. Teke, K. C. Bayindir and M. U. Cuma, Simulation and modeling of
a dynamic voltage restorer, [Online]. Available (August 23, 2009):
http://www.emo.org.tr/ekler/ee5605917626676_ek.pdf.
[188] V. K. Ramachandaramurthy, A. Arulampalam, C. Fitzer, C. Zhan, M. Barnes and
N. Jenkins Supervisory control of dynamic voltage restorers, IEE Proceedings of
Generation, Transmission and Distribution, vol. 151, no. 4, pp. 509-516, July 2004.
247
[200] Nielsen, G. J., Blaabjerg, F. and Mohan, N., Control strategies for dynamic
voltage restorer compensating voltage sags with phase jump, Proceedings of 16th
IEEE Applied Power Electronics Conference and Exposition, Volume 2, pp. 1267-
[201] M. B. Hughes and J. S. Chan, Canadian national power quality survey results,
Proceedings of IEEE Transmission and Distribution Conference, Canada, 1996, pp.
45-51.
[202] M. F. Akram and S. M. Bajwa, A sample power quality survey for emerging
competitive electricity market in Pakistan, Proceedings of IEEE International Multi
Topic Conference, Pakistan, pp. 38-44, 2001.
[204] Recommended practice for monitoring electric power Quality, IEEE Std. 11591995 USA, 1995.
[205] M. S. Illindala, P. Piagi, H. Zhang, G. Venkataramanan and R. H. Lasseter
Hardware development of a laboratory scale microgrid phase 2: operation and
control of a twoInverter microgrid National Renewable Energy Laboratory report,
Chapter 2, pp. 10, USA, March 2004.
[206] J. Wang, S. Chen and T. T. Lie, Estimating economic impact of voltage sags,
Proceedings of IEEE international conference on Power System Technology,
Singapore, vol.1, pp. 350-355, 2004.
[207] J. Wang, S. Chen and T. T. Lie, System voltage sag performance estimation,
IEEE Transactions on Power Delivery, vol. 20, no. 2, pp. 1738-1747, 2005.
[208] Basics of power quality and surge protection, Florida Power Corp., 1998.
[209] P. T. Nguyen and T. K. Saha, Dynamic voltage restorer against balanced and
unbalanced faults: modeling and simulation, IEEE-School of Information
249
2004.
[210] K. Piatek, A new approach of DVR control with minimised energy injection,
Proceedings of 12th International Conference on Power Electronics and Motion
Control, Slovenia, pp.1490-1495, 2006.
[211] G. T. Heydt, W. Tan, T. LaRose and M. Negley, Simulation and analysis of series
voltage boost technology for power quality enhancement, IEEE Transactions on
Power Delivery, vol. 13, no. 4, pp. 1335-1341, 1998.
[217] J. Stones and A. Collinson Power Quality, Power Engineering Journal, vol. 15,
no. 2, pp. 58-64, April 2001.
[218] S. V. R. Kumar and S. S. Nagaraju, Simulation of D-STATCOM and DVR in
power systems APRN Journal of Engineering and Applied Sciences, vol. 2, no. 3,
pp. 7-13, 2007.
[219] P. K. Lim and D. S. Dorr, Understanding and resolving voltage sag related
problems for sensitive industrial customers, IEEE power Engineering Society Winter
Meeting, vol. 4, pp. 2886-2890, 2000.
251
252
Pertinent Publications
1. A. Hanif and M. A. Choudhry, Dynamic voltage regulation and power export in a
distribution system using distributed generation, Journal of Zhejiang University
Science-A (Springer), vol. 10, no.10, 2009, pp.1523-1531. (HEC recognized impact
factor journal)
2. A. Hanif and M. A. Choudhry, PWM-IGBT inverter based power flow control
strategy at the load bus in the presence of utility supply, European Journal of
Scientific Research, Austria, vol. 18, no. 3, 2007, pp. 514-531.
3. A. Hanif and M. A. Choudhry, Power flow control strategy at the load bus in the
presence of dispersed generation, World Scientific and Engineering Academy and
Society Transactions on Power Systems, Greece, vol. 1, no. 8, 2006, pp. 1475-1483.
253
VITA
Aamir Hanif was born in Rawalpindi, Pakistan on December 03, 1973. He received
Engineering and Technology Taxila, Pakistan in 1997 and 2001 respectively. Since 1997,
he has been with the Department of Electrical Engineering, University of Engineering
and Technology Taxila, Pakistan. His employment experience includes the teaching at
University of Engineering and Technology Taxila. His special fields of interest include
custom power devices applications in electric power distribution system, distributed
generation, renewable energy systems, power flow control, and power quality.
At present he is pursuing the Ph.D. degree in electrical engineering at University of
Engineering and Technology Taxila, Pakistan. He is also member of IEEE power
engineering society since 2006.
Aamir Hanif
254