Bachelor of Science in Electrical Engine
Bachelor of Science in Electrical Engine
Bachelor of Science in Electrical Engine
Main supervisor
Mr. Kahuma Kenneth
Co Supervisor
Prof. Peter Lating
Submitted in Partial Fulfillment of the Requirement for the award of the Degree of
Bachelor of Science in Electrical Engineering
June 2019
i
DECLARATION
i
APPROVAL
This dissertation entitled "analysis and mitigation of Ferranti effect on long low loaded power
distribution networks" has been done under our supervision and has been submitted to the College
of Engineering, Design, Art and Technology, Makerere University for examination with our
approval as the candidate supervisors.
Main supervisor
Signature: Date:
Mr. Kahuma Kenneth
Co-Supervisor
Signature: Date:
Prof. Peter Lating
ii
DEDICATION
This report is dedicated to my God the Almighty, my dear parents, entire family, all my friends for
the encouragements, love, care and support they have rendered to me through this final year project
and completing the course at large.
To my beloved project supervisors Mr. Kahuma Kenneth and Prof. Peter Lating, project partner
Okeya Jacob, I must say this report is because of your endless help and guidance and therefore
dedicated to you. Your efforts and encouragements have geared me towards achieving my
professional dream.
You have enabled me see the future clearly and may the Almighty God bless you all abundantly.
iii
ACKNOWNLEGEMENT
It’s my pleasure to thank the Almighty GOD for having enabled me to complete my final year
project successfully. All Glory and Honor back to Him.
I wish to thank my dear parents and all well-wishers for their support and facilitation.
My sincere gratitude and special thanks to my supervisors Mr. Kahuma Kenneth and Prof. Peter
Lating for all the support, encouragement and the guidance they gave me since this project process
begun. This has been achievable because of their constant endeavors and guidance.
I would like to extend my appreciation to Dr. Eng. Dorothy Okello & Mrs. Josephine Kakande
Lutalo for their tireless effort in coordinating and guiding the students throughout the project
period.
Special thanks go out to my project partner Okeya Jacob. I am very grateful to him for being so
hardworking and supportive towards the success of this project.
iv
ABSTRACT
Ferranti effect is when the steady voltage at the open end of an uncompensated distribution line is
higher than the voltage at the sending end. This project presented the analysis & mitigation of
Ferranti effect in the Kalangala Infrastructural Services (KIS) power distribution network. The
analysis was based on software simulations with information obtained from the Bukuzindu hybrid
power station operations data log of 1st – 9th of January 2019.
The DIGI SILENT power factory 15.1 software gave the locus of sending end voltage with line
length at lowest loading which showed that receiving end voltage is greater than sending end
voltage.
The MATLAB-Simulink program gave the models of the lumped-up KIS power network & the
devices used in the mitigation of Ferranti effect.
Both SVC (Static Var Compensator) and D-STATCOM (Distribution Static Synchronous
Compensator) are important shunt FACTS equipment for reactive power compensation, which
were compared in response speed, amount of reactive power absorbed, & the amount of voltage
reduced at the same rating. Simulation results were presented as the same controllable rated
capacity for SVC or D-STATCOM was connected on the distribution bay of Bukuzindu power
generation substation.
D-STATCOM is much better than SVC with faster response speed, higher amount of reactive
power absorbed, & higher amount of voltage reduced at the same rating & it was chosen as the
most appropriate shunt FACTS device in the mitigation of Ferranti effect in the KIS power
network.
v
LIST OF TABLES
Table 2-1: Conventional Controllers ............................................................................................ 15
Table 3-1: Network Coverage for KIS.......................................................................................... 22
Table 3-2: Number of Transformers on the Network ................................................................... 22
Table 3-3: Load Flow Data as of 1st January 2019 and 9th January 2019 ................................... 23
Table 4-1: Bugoma Feeder Distance Voltage Values at Different Loading Percentages ............. 30
Table 4-2: Kalangala Feeder Distance Voltage Values at Different Loading Percentages. ......... 31
Table 4-3: Percentage Loading VS Receiving End Voltage(pu) .................................................. 34
Table 4-4: Device Performance Comparison Table...................................................................... 40
vi
LIST OF FIGURES
Figure 1-1: Kalangala and Bugoma feeder single line diagrams .................................................... 3
Figure 2-1: Pi (Π) Model of the Line at No Load ........................................................................... 7
Figure 2-2: Power Factor Triangle ................................................................................................ 12
Figure 2-3: Illustration of Controllability of Power Systems ....................................................... 14
Figure 2-4: basic diagram of a D-STATCOM .............................................................................. 18
Figure 2-5: Typical voltage regulation characteristic of a D-STATCOM .................................... 18
Figure 2-6: Typical Model of an SVC .......................................................................................... 19
Figure 2-7: V-I Characteristics of SVC ........................................................................................ 20
Figure 3-1: Load profile of KIS network ...................................................................................... 24
Figure 3-2: Digisilent power factory software Modeled KIS network ......................................... 25
Figure 3-3: Matlab-Simulink lumped up KIS network & designed mitigation devices model (D-
STATCOM & SVC) ..................................................................................................................... 26
Figure 3-4: D-STATCOM Network Model and controller........................................................... 28
Figure 3-5: SVC Network Model with Controller ........................................................................ 29
Figure 4-1: Kalangala feeder Voltage distance profile for smallest loading (15%) ..................... 32
Figure 4-2: Kalangala Feeder Voltage Distance Profile For Highest Loading (33%) .................. 33
Figure 4-3: Percentage Loading Against the Receiving End Voltage .......................................... 35
Figure 4-4: D-STATCOM Operation Graphs During Compensation .......................................... 36
Figure 4-5: D-STATCOM Mitigation Graph ............................................................................... 37
Figure 4-6: SVC Operation Graphs During Compensation .......................................................... 38
Figure 4-7: SCV Mitigation Graphs ............................................................................................. 39
vii
LIST OF ACRYNOMS, ABBREVIATIONS OR NOMENCLATURE
AAAC All Aluminum Alloy Conductor
AC Alternating Current
DC Direct Current
kV kilo Voltage
kW kilo Watt
MW Mega Watt
viii
VA Voltage Ampere
ix
Table of Contents
DECLARATION ............................................................................................................................. i
APPROVAL ................................................................................................................................... ii
DEDICATION ............................................................................................................................... iii
ACKNOWNLEGEMENT ............................................................................................................. iv
ABSTRACT.................................................................................................................................... v
LIST OF TABLES ......................................................................................................................... vi
LIST OF FIGURES ...................................................................................................................... vii
LIST OF ACRYNOMS, ABBREVIATIONS OR NOMENCLATURE .................................... viii
CHAPTER 1: INTRODUCTION .................................................................................................. 1
1.1. Project Background .......................................................................................................... 1
1.2. Problem Statement ........................................................................................................... 3
1.3. Justification ...................................................................................................................... 4
1.4. Objectives ......................................................................................................................... 4
1.4.1. Main Objective.......................................................................................................... 4
1.4.2. Specific Objective ..................................................................................................... 4
1.5. Scope ................................................................................................................................ 4
CHAPTER 2: LITERATURE REVIEW ........................................................................................ 5
2.1. Introduction ...................................................................................................................... 6
2.2. Ferranti Effect in Long Low Loaded Distribution Lines ................................................. 6
2.2.1. Ferranti Effect ........................................................................................................... 6
2.3. Radial Distribution Networks......................................................................................... 10
2.4. Power Concept ............................................................................................................... 10
2.4.1. Power Triangle and Power Factor........................................................................... 12
2.4.1. Sources of reactive power ....................................................................................... 13
2.5. Compensation Techniques in Power Transmission Systems ......................................... 13
2.6. Flexible AC Transmission Systems (FACTS) ............................................................... 14
2.6.1. Benefits Power System Control .............................................................................. 15
2.6.2. Benefits of utilizing FATCS devices ...................................................................... 15
2.6.3. Classification of FACTS Devices ........................................................................... 16
2.7. Matlab/SIMULINK R2016a software ............................................................................ 16
x
2.8. DigSilent Power Factory 15.1 Software ......................................................................... 16
2.9. D-STATCOM................................................................................................................. 17
2.9.1. Operating modes of a D-STATCOM ...................................................................... 18
3.0. Static Variable Compensator (SVC) .............................................................................. 19
CHAPTER 3: METHODOLOGY ................................................................................................ 21
3.1. Introduction .................................................................................................................... 21
3.2. KIS Data ......................................................................................................................... 21
3.2.1. Network Coverage, Transformers, Load Flow Data. .............................................. 21
3.3. Network modeling using DigSilent Power Factory Software ........................................ 24
3.4. Matlab/Simulink KIS Distribution Network Model ...................................................... 26
3.5. Matlab/Simulink Device Modeling ................................................................................ 27
3.5.1. D-STATCOM Network Model ............................................................................... 27
3.5.2. SVC Network Model .............................................................................................. 29
CHAPTER 4: RESULTS AND DISCUSSION ............................................................................ 30
4.1. Introduction .................................................................................................................... 30
4.2. Feeder Distance Voltage Profile Results At 15% And 33% Loading ............................ 30
4.3. Feeder Voltage at Different Load Scales ....................................................................... 34
4.4. Ferranti effect mitigation ................................................................................................ 35
4.4.1. Ferranti effect mitigation using a D-STATCOM.................................................... 36
4.4.2. Ferranti Effect Mitigation Using D-STATCOM Graphs ........................................ 37
4.4.3. Ferranti effect mitigation using SVC ...................................................................... 37
4.4.4. Ferranti effect mitigation using SVC Graphs ......................................................... 39
4.5. Performance Comparison of D-STATCOM and SVC................................................... 39
CHAPTER 5: CONCLUSION, RECOMMENDATIONS CHALLENGES AND FURTHER
RESEARCH.................................................................................................................................. 41
5.1. Conclusion...................................................................................................................... 41
5.2. Recommendation ............................................................................................................ 41
5.3. Challenges ...................................................................................................................... 42
5.4. Future research ............................................................................................................... 42
References ..................................................................................................................................... 42
xi
CHAPTER 1: INTRODUCTION
This chapter presents the project back ground, problem statement, justification, project objectives
and summary of methods.
According to Sebastian Ziani de Ferranti[1] Ferranti effect is a phenomenon where the steady
voltage at the open end of an uncompensated transmission line is always higher than the voltage
at the sending end. Due to capacitive charging current flowing through the inductance of the line
resulting into over voltage increases according to the increase in line length. The factors which
affect the power frequency voltages on the line during normal operation and the increase in
voltages are length of the line, low/light loads and the degree of shunt compensation
This leads to generation of significant capacitive reactive power on the network that needs to be
reduced.
This effect was discovered at the end of 19th century during the installation of an AC based
distribution system in Great Britain. On one installation of an AC transmission system, Ferranti
observed alerted by his installers that by adding additional distribution sections i.e. by increasing
the total length of the transmission line, the voltage on the line increased locally. They observed
first on the Deptford - London line that the luminosity of some carbon fibre lamps increased, when
they attached an additional distribution section. In this case they had a load of only low power
bulbs while having an effective generator power of 935 kW. Ferranti had in fact approximately an
open-ended transmission line. Today, the Ferranti effect is well known in the field of power
transmission over long distances at relatively low frequencies [1].
In relation to KIS distribution network [2]., Kalangala Infrastructure Services Ltd, is a private
limited company operating on Bugala Island, Kalangala District. We provide daily ferry services
between Luuku and Bukakata, generate and distribute electricity, and supply treated water.
Among the four Kalangala Infrastructure Services projects, namely; Ferry, Road, Power and
Water; KIS was licensed to generate and sell Electricity on Bugala Island, Kalangala District; in
fulfilment of this mandate, KIS constructed a 1.6 MW Solar-Thermal hybrid power generating
station with solar and thermal capacity of 0.6MW and 1.0MW respectively, a 2MVA substation
1
with 140km [3]. Medium Voltage/ distribution network to supply over 40 villages on the Island.
Bugala Island has 89 villages with nucleated settlement, where 80% are connected to electricity
and the 20% awaits network extension and connection. Commercial Operation of this power
system was achieved in 2015 after successful completion of construction and commissioning tests;
and to date, 2,600 customers are on supply 24/7 and also 33 commercial customers, including a
few hotels and small enterprises [4].
The distribution network stretches to about 106kms in its longest part and this part has experienced
a bit of Ferranti effect problem where by the voltage at the receiving end of this stretch is above
the allowable limits of +10% of the supply voltage. This is due to the long network length, low
loads on the network and the degree of shunt compensation.
The power plant is a solar-diesel hybrid system, of 1.6 mw with solar and thermal capacity of
0.6mw and 1.0mw respectively.
The solar component is comprised of solar farm, battery bank and inverters. The solar farm has
2832 solar panels each with 235 Wp rating. The solar modules are connected in series and parallel
configuration to achieve 500vdc and 662.7kwp (desired voltage, and power). The 658kw battery
bank plays a buffer role to the plant operations, when solar component is running the load.
The thermal component of this power plant is comprised of three generators of 500, 320 & 200
KVA capacity, which are designed to switch sequentially to compliment the power from solar.
The main sub-station is composed of two 1 MVA transformers that feed power through a 33kV
distribution system onto two main feeders (Kalangala and Bugoma). The power demand on
Kalangala and Bugoma feeders range at 150-330 kW and 35-70 kW respectively.
2
Bugoma feeder single line diagram
Figures 1.1 show the installed transformers in Kalangala & Bugoma feeders [3]. This medium
length of the Kalangala feeder contributes to the Ferranti effect.
3
Long low/light loaded power networks like for the case of KIS distribution power system network
experiences voltage overshoot of up to 36.44kV at the receiving end which is beyond the
recommended standard error limit of +/-10% of the normal generated supply of 33kV, a
phenomenon referred to us as the Ferranti effect.
1.3. Justification
Kalangala feeder voltage distance profile simulation at lowest (15%) loading showed 36.428kV at
the receiving end. This voltage is above the upper limit of +10% of the supply voltage i.e. (1.1pu)
This is dangerous to the power system in that it affects equipment & load operations, low power
factor, system instability, & poor voltage regulation/power quality, poor efficiency, insulation
failures, risky to Personnel (human life) and losses (both technical & economic)
1.4. Objectives
To analyze the voltage rise and provide possible measures that will lead to standard voltage levels
at the receiving end of a long low loaded power distribution network.
1.5. Scope
4
This project covers KIS power network modeling for Ferranti effect analysis, modeling two
possible shunt FACTS reactive power converting devices for mitigation of this effect and
comparing and selecting the most appropriate device for Ferranti effect mitigation in KIS network
based on performance and effectiveness.
5
2.1. Introduction
This chapter has the overview of the project detailed description and the research made about a
number of items in this project. This gives the basic information and understanding about Ferranti
effect, load profiles, classification of medium transmission lines (regarded as distribution lines),
general line loading, FACTS devices, Dig SILENT Power Factory 15.1 software, R2016a
MATLAB/Simulink software and will be the guide in regards to our methodology.
The effect in which the voltage at the receiving end of the transmission line is more than the
sending voltage is known as the Ferranti effect. Such type of effect mainly occurs because of light
load or open circuit at the receiving end [5].
Ferranti effect is due to the charging current of the line. When an alternating voltage is applied,
the current that flows into the capacitor is called charging current. A charging current is also known
as capacitive current. The charging current increases in the line when the receiving end voltage of
the line is larger than the sending end.
Capacitance and inductance are the main parameters of the lines having a length 240km or above.
On such transmission lines, the capacitance is not concentrated at some definite points. It is
distributed uniformly along the whole length of the line.
When the voltage is applied at the sending end, the current drawnT by the capacitance of the line
is more than current associated with the load. Thus, at no load or light load, the voltage at the
receiving end is quite large as compared to the constant voltage at the sending end.
6
2.2.1.1. Detailed Explanation of the Ferranti Effect by Considering a Nominal Pi (Π)
Model:
Considering the phasor diagram of a transmission line above, OA represents the receiving end
voltage, OD represent the current through the capacitor at the receiving end. The phasor AB
represents the voltage drop across the resistance R and BC represents voltage drop across the
inductance X. The phasor OC represents the sending end voltage under a no-load condition.
It is clear from phasor diagram that OA > OC, that is to say the voltage at the receiving end is
greater than the voltage at the sending end when the line is at no load.
7
𝑍𝑌
𝑉𝑠 = (1 + ) 𝑉 + 𝑍𝐼𝑟 … … … … . … … 1
2 𝑟
At no load, 𝐼𝑟 = 0
𝑍𝑌
𝑉𝑠 = (1 + )𝑉
2 𝑟
𝑍𝑌
𝑉𝑠 − 𝑉𝑟 = (1 + ) 𝑉 − 𝑉𝑟
2 𝑟
𝑍𝑌
𝑉𝑠 − 𝑉𝑟 = ( )𝑉 ………………………2
2 𝑟
𝑍 = (𝑗𝜔𝑙)𝑆 … … … . … … … … … … … … . .4
1
𝑉𝑠 − 𝑉𝑟 = (𝑗𝜔𝑙𝑆)(𝑗𝜔𝑐𝑆)𝑉𝑟
2
1
𝑉𝑠 − 𝑉𝑟 = − (𝜔2 𝑆 2 )𝑙𝑐𝑉𝑟 … … … … … . .5
2
1
For overhead lines, = velocity of propagation of electromagnetic waves on the transmission
√𝑙𝑐
1
√𝑙𝑐 =
3 × 108
1
𝑙𝑐 = … … … … … … … … … … . .6
(3 × 108 )2
Substituting equation 6 in 5,
1 1
𝑉𝑠 − 𝑉𝑟 = − (𝜔2 𝑆 2 ) ( ) 𝑉 … .7
2 (3 × 108 )2 𝑟
8
But 𝜔 = 2𝜋𝑓,
4𝜋 2
𝑉𝑠 − 𝑉𝑟 = − ( × 10−16 ) 𝑓 2 𝑆 2 𝑉𝑟 … . .8
18
Equation 8 shows that (𝑉𝑠 − 𝑉𝑟 ) is negative. That is 𝑉𝑟 > 𝑉𝑠 . From this equation, Ferranti effect
also depends on frequency and the electrical length of the lines [6].
𝑉𝑠 = 𝐴𝑉𝑟 + 𝐵𝐼𝑟
At no load, 𝐼𝑟 = 0, 𝑉𝑟 = 𝑉𝑟𝑛𝑙
𝑉𝑠 = 𝐴𝑉𝑟 = 𝐴𝑉𝑟𝑛𝑙
|𝑉𝑠 |
|𝑉𝑟𝑛𝑙 | =
|𝐴|
For any long line, A is less than unity, and decreases with the increase in the length of the line
which makes the receiving end voltage at no load greater than the voltage at sending end (𝑉𝑟𝑛𝑙 >
𝑉𝑠 ). Also as the line length increases further, the rise in the no load voltage at the receiving end
becomes more predominant [7] [8].
By placing the shunt reactive power compensating devices optimally. These devices include
Dynamic Voltage Restorers (DVR), Thyristor-controlled reactors (TCR), Static Voltage Restorers
(SVR) among others.
A shunt reactor is an inductive current element connected between line and neutral to compensate
the capacitive current from transmission lines. When this effect occurs in long transmission lines,
shunt reactors compensate the capacitive Var of the lines and therefore the voltage is regulated
within the prescribed limits.
By running the distribution line with higher load i.e if you there a number of lines carrying loads
on each sides, we can switch off one transmission line and the remaining load can be diverted to
the line with Ferranti effect [9].
9
Ferranti effect has adverse effects on the system which include; equipment damage, and their
windings can burn because of high voltage, reduces the power system power factor, leads to
instability of the power system, the power system becomes inefficient and it reduces the load
current.
Note: Voltage rise is directly proportional to the square of the length of a line and is more in short
transmission cables because their capacitance is high.
Radial distribution system is a system whereby power is received at the utility supply voltage level
by a single, incoming substation. Through a series of step downs and splits, the power is converted
for individual end-use equipment [10].
It is the cheapest to build, and is widely used in sparsely populated areas. A radial system has only
one power source for a group of customers. A power failure, short-circuit, or a downed power line
would interrupt power in the entire line which must be fixed before power can be restored [11].
Power is the time rate of expending or absorbing energy where energy is the capacity to do work.
P=vi where v is the voltage or potential defined as the force needed to move electrons and I is the
current defined as the rate of flow of charges per second through a material when the voltage is
applied.
In AC circuits, power is instantaneous and is measured in watts. It is the product of the
instantaneous voltage drop across the load and the instantaneous current into the load.
𝑣(𝑡) = 𝑉𝑚 cos 𝑤𝑡 … … … … … … … … … … … … … . .1
𝑖(𝑡) = 𝐼𝑚 cos ( 𝑤𝑡 + 𝜃) … … … … … … … … … . … … 2
𝑝(𝑡) = 𝐼𝑚 𝑉𝑚 cos 𝑤𝑡 cos ( 𝑤𝑡 − 𝜃) … … … … … … . .3
P(t) defines the instantaneous power absorbed by the circuit. Using trigonometry identities in
equation 3
10
𝑝(𝑡) = 1⁄2 𝑉𝑚 𝐼𝑚 cos 𝜃(1 + cos 2𝑤𝑡) + 1⁄2 𝑉𝑚 𝐼𝑚 𝑠𝑖𝑛𝜃 sin 2𝑤𝑡 … … . .4
The first part of equation 4 is time independent and represents the active, real or average power
(P). The second part of the equation is sinusoidal function with frequency 2w which is twice the
angular frequency of the voltage or current. Its known as the instantaneous reactive power (Q).
𝑃 = 1⁄2 𝑉𝑚 𝐼𝑚 cos 𝜃 … … … … … … … … … . .5
𝑄 = 1⁄2 𝑉𝑚 𝐼𝑚 sin 𝜃 … … … … … … … … … . .6
Q has a unit of Volt-Ampere Reactive (VAR) and P has a unit of a Watt. Reactive power is present
when the voltage and current are not in phase. Its developed when the current waveform leads the
voltage waveform (leading power factor) and consumed when the current waveform lags the
voltage waveform (lagging power factor). Given the phasor expressions for both current and
voltage, real and reactive power can be calculated and expressed in complex form i.e 𝑉 = |𝑉| <
∝ 𝑎𝑛𝑑 𝐼 = |𝐼| < 𝛽, then the comples apparent power absorbed by the circuit is the product of the
voltage and the complex conjugate of the current.
𝑆 = 𝑉𝐼 ∗ = |𝑉||𝐼| <∝- 𝛽 = |𝑉||𝐼|cos (∝- 𝛽) + 𝑗|𝑉||𝐼|sin (∝- 𝛽)……….7
𝑆 = 𝑃 + 𝑗𝑄 … … … … … … … … … … … … … . . . … … .8
11
2.4.1. Power Triangle and Power Factor
Active power P(kW) is responsible for the useful work and is associate with the portion of the
current that is in phase with the voltage, reactive power Q (Kvar) sustains the electromagnetic field
used to make devices like motors operate as an energy exchange between reactive components of
the electrical systems while apparent power S (Kva), gives a geometrical combination of the active
and reactive powers i.e the total power drawn from the network.
The ratio between the active power and reactive power is the power factor or cos𝜃 and gives the
measure of how efficient the utilization of electrical energy is. It is always expressed as a decimal
value or as a percentage. It defines the phase andle between the magnitudes of the rms values of
voltage and current [12].
𝑤𝑎𝑡𝑡𝑠 𝑃 𝑉𝐼𝑐𝑜𝑠𝜃
𝑃𝑜𝑤𝑒𝑟 𝐹𝑎𝑐𝑡𝑜𝑟 = = = = 𝑐𝑜𝑠𝜃
𝑣𝑜𝑙𝑡 − 𝑎𝑚𝑝𝑒𝑟𝑒𝑠 𝑆 𝑉𝐼
12
2.4.1. Sources of reactive power
This defines all devices with fixed value ratings, which can be connected or disconnected as system
conditions change for example a capacitor or an inductor. It can be permanently connected or
switched throughout the day. The switching action can be made in a few seconds using automatic
controls or manual operation (slow response speed).
These are active devices that can provide variable amounts of reactive power in a few milliseconds.
Common examples are static Var compensators (SVCs), static synchronous compensators
(STATCOMs), synchronous condensers, and conventional generators. They are capable of rapid
dynamic response.
Both the passive reactive power sources and dynamic reactive power sources differ in
characteristic’s, that is dynamics and speed of response, ability of voltage changes, capital costs,
operating costs and opportunity costs [13] [14].
13
2.6. Flexible AC Transmission Systems (FACTS)
Flexible AC Transmission Systems or FACTS are devices which allow the flexible and dynamic
control of power systems. These controllers are used for solving various power system steady state
control problems [15].
FACTS are static equipment’s and are used for electrical energy transmission in form of an
Alternating Current by enhancing controllability and increase power transfer capability.
Power systems involve generation, transmission distribution and electrical utilization in which its
stability can be improved by using these FACTS devices. In a power system, there exists some
constraints that limit the amount of power that can be transferred of a power system. These include
steady state power transfer limit, voltage stability limit, dynamic voltage limit, transient stability
limit, power system oscillation damping limit, thermal limit, short circuit current limit and others.
In power systems, there are variables that can be impacted by control and these are voltage, angle
and impedance and this can be depicted in figure 2-3 below.
𝐸𝑠 𝐸𝑟
𝑃= 𝑠𝑖𝑛𝛿
𝑋
where P is the power, Es is the sending end voltage, Er is the receiving end voltage, 𝛿 is the control
angle. This controllability can be achieved by using conventional equipment and FACTS
controllers. Conventional controllers are outlined in the table 2-1.
14
Table 2-1: Conventional Controllers
Facts controllers include static synchronous compensator (STATCOM), Static VAR Compensator
(SVC), Unified Power Flow Controller (UPFC), Convertible Series Compensator (CSC),
Interphase Power Flow Controller (IPFC), Static Synchronous Series Controller (SSSC) which
have an impact on voltage, impedance and/or angle.
Thyristor Controlled Series Compensator (TCSC), Thyristor Controlled Phase Shifting
Transformer (TCPST) and Super Conducting Magnetic Energy Storage (SMES) which control
impedance, angle and voltage and power respectively.
Increased loading and more effective use of transmission corridors, power flow control, improved
power system stability, increased system reliability, additional flexibility in starting new
generation and elimination for the need of new transmission lines.
15
2.6.3. Classification of FACTS Devices
Depending on the type of connection to the network, we have got serial controllers, derivation
controllers, serial to serial controllers and serial-derivation controllers.
Depending on technical features, there are first generation controllers that use thyristors with
ignition controlled by gate (SCR) for example SVC, TCSC, TCPS and second generation
controllers that use semiconductors with ignition and extinction controlled by gate (GTO’s,
IGBTS, IGCTS etc) for example STATCOM, SSSC and UPFC [15] [17].
The name MATLAB stands for the Matrix Laboratory. MATLAB is a very fast performing
language in technical fields. It integrates computation, visualization and programming in an easy-
to-use environment where problems and solutions are demonstrated in intimate mathematical
notation. In this project, Ferranti effect mitigation in distribution lines is analyzed by using
Matlab/SIMULINK software.
The name DigSilent stands for "Digital Simulation and Electrical Network calculation program''.
Power Factory, is a powerful calculation program (as written by DigSilent GmbH) for power
system analysis. DigSilent Power Factory is the leading high-end tool for applications in
generation, distribution and industrial systems. It provides optimal organization of data required
for performing any type of calculation, memorization of settings or software operation options
[18].
The desire to improve modeling algorithms and use of modern software technologies prompted its
development in1993. In 1997, Power Factory new generation software for analysis of electrical
systems was developed. This software has numerous advantages that ensure efficiency in the use
of simulation software to solve system problems, such as enhanced system stability calculations,
improved graphical user interface etc.
16
This software can be integrated in into other open standard systems like the GIS, RMS and DMS.
In this project, Ferranti effect in distribution lines is analyzed with help of DigSilent Power Factory
(2015) Version 15.1 software.
2.9. D-STATCOM
Distribution Static Compensator (D-STATCOM) is a shunt FACTS device that can regulate
network voltage at its connection point, mitigate the impact of disturbing loads such as voltage
variations, unbalance and harmonics and compensate load reactive power [19] [20]. It is termed as
a D-STATCOM when used in distribution systems.
It achieves dynamic compensation by adjusting the AC-side voltage of the connected bus and
absorbing and/or emitting reactive power depend on the voltage change, it connects to the
distributed grid through a transformer and VSC that converts the DC voltage to AC voltage
according to the voltage at its connected bus such as when the grid voltage is greater than the D-
ST A TCOM voltage, D-ST A TCOM will absorb reactive power and when the grid voltage is less
than the D-STATCOM voltage, it will inject reactive power to the distributed [20].
D-STATCOM therefore is a fast-compensating reactive power source that is applied on the
transmission or distribution system to reduce voltage variations such as sags and surges along with
instability caused by rapidly varying reactive power demand. It can generate a sinusoidal voltage
with any required frequency, and phase angle and the converter is made of some kind of energy
storage which supplies the converter with a DC voltage.
It consists of mainly the dc capacitor, voltage source converter, coupling transformer and a
controller as can be seen in figure 2-4.
17
Figure 2-4: basic diagram of a D-STATCOM
A static VAR compensator is a static Var generator whose output is varied so as to maintain or
control specific parameters say voltage or reactive power of bus of the electric power system. They
are used to control the AC voltage in transmission and distribution networks [22]. It consists of a
TCR, TSC, controller and a coupling transformer as in figure 2-6 bellow.
The location of an SVC is so much important for best performance and effectiveness and usually
its located at the electrical center or the midpoint of the distribution line. It offers voltage regulation
at the terminals by controlling the amount of reactive power injected into or absorbed from the
power system i.e when the system voltage is low, it generates reactive power (capacitive nature)
and when the system voltage is high, it absorbs reactive power (inductive nature).
Reactive power variation is achieved by switching three phase capacitor banks and inductor banks
connected on the secondary side of the coupling transformer. Each capacitor bank is switched on
and off by three thyristor switches (TSC) while reactors are switched on or off by TSR or Phase
Controlled Reactor (PCR). An SVC has go voltage current characteristic as in figure 2-7.
19
Figure 2-7: V-I Characteristics of SVC
The term static means that the SVC has no moving parts. This device is an automated by impedance
matching and is designed to bring the system closer to unity power factor. When the system is
capacitive, the SVC uses reactors (usually in form of TCR) to consume VARs from the system
hence lowering the system voltage. Under inductive conditions (lagging), capacitor banks
automatically switch on thus providing a higher system voltage [15].
20
CHAPTER 3: METHODOLOGY
3.1. Introduction
This chapter describes the steps taken in order to achieve the project specific objectives. This
included obtaining research about the problem, consultations, data collection, loading calculations
and other processes. Most of the KIS network data was obtained through KIS operations
management.
The data of KIS was obtained through research on internet and site visits to Kalangala Island.
By visiting the site it was observed that the network is lightly loaded, the available loads are un
evenly distributed most especially on existing three phase transformers, most single phase
transformers connected to the red and blue phases of the network and also small clearance between
phase and earth on the old network and lv network for the combine bare systems which could also
be the source of the Ferranti effect on their network.
However it is clear that the major factors leading to Ferranti effect are low loads, network length
and degree of shunt compensation.
Tables 3-1, 3-2 and 3-3 below show the network coverage for KIS, transformers on the network
and the load flow data that were given to us.
21
Table 3-1: Network Coverage for KIS
Lv network 114.6
22
Table 3-3: Load Flow Data as of 1st January 2019 and 9th January 2019
23
Based on the load flow data given to us in table 3-3, the following load profile (figure 3-1) was
plotted.
Load(kVA)
Figure 3-1: Load profile of KIS network
KIS distribution network was modeled using DigSilent Power Factory Software for Ferranti effect
analysis. This enabled us to investigate the voltage distance profile of the entire network from the
generating station to the user end point. This network model can be seen in figure 3-2 below.
24
1.01pu 33.467kV(
Bugoma feeder 1.014pu)
1.0pu
36.428kV
substation (1.104pu)
1.013pu
1.03pu Kalangala feeder 1.073pu
1.059pu
1.07pu
1.011pu
Bugoma feeder
substation
(1.075pu)
Kalangala feeder
25
It’s made of the generating station, two feeders (Bugoma and Kalangala feeders) and terminal
points for voltage measurement. At each node, the voltages verses the distances were noted and
results recorded as seen in the results under table 4-1
KIS distribution network was lumped and modeled as can be seen in figure 3-3 below. This model
is composed of the generating source, two distribution feeders, connecting and disconnecting
breakers and the compensating devices.
Figure 3-3: Matlab-Simulink lumped up KIS network & designed mitigation devices model (D-
STATCOM & SVC)
26
The D-STATCOM and SVC are used to regulate voltage on the 33Kv distribution network. These
devices regulate the voltage at the receiving end of the network by absorbing reactive power from
the network when there is a voltage rise beyond regulatory limits. Two feeders Kalangala
(105.91km) and Bugoma (20.05km) are used to supply the loads. Breakers 1 and 2 are open to test
the network at 15% loading. When they are closed, the network is at 33% loading and there is no
Ferranti effect.
If the system is at 15% loading and is to be compensated using D-STATCOM, breakers 1, 2 and
4 are open while breakers 3 and 5 are closed. When the voltage is within limits, the D-STATCOM
floats and does not absorb any reactive power. The D-STATCOM absorbs reactive power only
when there is a voltage rise at the receiving end of the network.
If the system is at 15% loading and is to be compensated using SVC, breakers 1, 2 and 5 are open
while breakers 3 and 4 are closed. The SVC floats under normal operating conditions (not
absorbing any reactive power) and absorbs reactive power when there is a voltage rise at the
receiving end of the network.
Figure 3-4 shows a D-STATCOM model and its controller modeled using Matlab/Simulink. As
earlier discussed, it achieves dynamic compensation by adjusting the AC-side voltage of the
connected bus and absorbing and/or emitting reactive power depend on the voltage change, it
connects to the distributed grid through a transformer and VSC that converts the DC voltage to
AC voltage according to the voltage at its connected bus such as when the grid voltage is greater
than the D-STATCOM voltage, D-STATCOM will absorb reactive power and when the grid
voltage is less than the D-STATCOM voltage, it will inject reactive power to the distributed.
27
D-STATCOM Network Model
D-STATCOM controller
SVC is used to control the AC voltage in transmission and distribution networks [22]. As seen in
figure 3-5, it consists of a TCR, TSC, controller and a coupling transformer. It offers voltage
regulation at the terminals by controlling the amount of reactive power injected into or absorbed
from the power system i.e when the system voltage is low, it generates reactive power (capacitive
nature) and when the system voltage is high, it absorbs reactive power (inductive nature).
Reactive power variation is achieved by switching three phase capacitor banks and inductor banks
connected on the secondary side of the coupling transformer. Each capacitor bank is switched on
and off by three thyristor switches TSC while reactors are switched on or off by TSR or TCR. An
SVC has go voltage current characteristic as in figure 2-7.
SVC Controller
29
CHAPTER 4: RESULTS AND DISCUSSION
4.1. Introduction
This chapter presents the simulation results of Ferranti effect analysis in the KIS distribution
network. The results were obtained with help of DigSilent simulation software and
Matlab/SIMULINK software for both different load conditions. We are able to see under what
conditions that Ferranti effect occurs and also how it can be mitigated by using the modeled
FACTS devices.
4.2. Feeder Distance Voltage Profile Results At 15% And 33% Loading
With help of DigSilent software, the first objective of this project was achieved by first modeling
the KIS power distribution network after which the voltages were determined at different parts of
the network and the results were as shown in table 4-1 and 4-2 below for Bugoma and Kalangala
feeder respectively at 15% and 33% loading.
Table 4-1: Bugoma Feeder Distance Voltage Values at Different Loading Percentages
Distance (km) Voltage (pu) @ 15% loading Voltage (pu) @ 33% loading
0.000 1.010 1.008
1.311 1.011 1.009
1.498 1.011 1.009
1.909 1.011 1.009
2.835 1.011 1.009
3.695 1.012 1.009
5.843 1.013 1.010
8.841 1.013 1.010
11.265 1.014 1.011
13.765 1.014 1.011
13.965 1.014 1.011
16.025 1.014 1.011
18.025 1.014 1.011
20.025 1.014 1.011
20.050 1.014 1.011
30
Table 4-2: Kalangala Feeder Distance Voltage Values at Different Loading Percentages.
Distance (km) Voltage (pu) @ 15% loading Voltage (pu) @ 33% loading
0 1.03 1.024
0.147 1.031 1.024
1.647 1.033 1.025
3.847 1.036 1.028
6.847 1.039 1.029
8.437 1.04 1.029
10.487 1.043 1.031
11.236 1.044 1.032
14.501 1.047 1.033
15.107 1.048 1.033
15.377 1.048 1.034
15.841 1.048 1.034
16.413 1.049 1.035
16.985 1.05 1.035
18.185 1.052 1.036
20.685 1.055 1.038
21.782 1.056 1.039
23.882 1.059 1.041
33.882 1.07 1.049
35.442 1.071 1.05
36.538 1.072 1.051
37.915 1.073 1.052
40.515 1.076 1.053
50.515 1.085 1.06
53.015 1.086 1.061
56.515 1.089 1.063
66.515 1.093 1.065
67.015 1.093 1.065
67.865 1.093 1.065
71.249 1.095 1.067
73.449 1.096 1.068
77.949 1.098 1.068
89.949 1.102 1.07
97.949 1.104 1.073
105.949 1.104 1.074
105.981 1.104 1.075
31
Using the data obtained from Digisilent software simulated simulations in Table(s) 4-1 and 4-2
above, Kalangala feeder voltage distance profile for lowest loading (15%) was plotted as seen
below.
Voltage(pu)
Figure 4-1: Kalangala feeder Voltage distance profile for smallest loading (15%)
From figure 4-1 above, it is seen that the voltage values increase with increasing line distance. It
can be seen that the voltage is above the recommended error limits of +10% of the system voltage.
The voltage at the receiving end of Kalangala feeder increased to 1.104pu (above upper limit of
1.1pu) as illustrated in figure 4-1, while for Bugoma rose to 1.014pu (within limit) as can be seen
in the table 4-1.
32
At 33% loading, a plot of the voltage distance values shows no Ferranti effect as can be seen in
figure 4-2.
Voltage(pu)
Figure 4-2: Kalangala Feeder Voltage Distance Profile For Highest Loading (33%)
The above figure (s) 4-1 and 4-2 show that Ferranti effect arises as the line length increases.
Bugoma feeder has no Ferranti effect hence no plots made as our concern is now Ferranti effect
on the Kalangala feeder.
33
4.3. Feeder Voltage at Different Load Scales
With help of DigSilent power factory software also, the load was scaled using different percentage
loading and corresponding voltage values recorded as can be seen in table 4.3.
The plot in figure 4-3 below shows that at minimum loading (15%), there is Ferranti effect that is,
the voltage at the end of Kalangala feeder is beyond the allowable voltage range. Bugoma feeder
has no Ferranti effect and the voltage is within limits.
34
Voltage(pu)
Figure 4-3: Percentage Loading Against the Receiving End Voltage
This gives us another factor leading to Ferranti effect in Kalangala distribution network and hence
the need for mitigation measures.
To mitigate Ferranti effect as the third objective of this project, two devices were modeled and
there performance compared to determine the most effective one. From figure 3-3, the source
voltage was programmed to three voltage values as 1.0pu, 1.104pu and 1.0pu. Ferranti effect
voltage value obtained during simulations was 1.104 and is the voltage value between 0.1s and
0.2s in the upper graph of figure 4-5 and between 0.1s and 0.3s in the upper graph of figure 4-7.
The lower graphs in both figures show the performance of each device in reducing this voltage
value to a value within limits (see in the performance comparison table 4-4). It can be seen that
the high voltage value was reduced by some value for each device.
35
4.4.1. Ferranti effect mitigation using a D-STATCOM
Figure 4-4 shows the D-STATCOM operation graphs during compensation. The first graph shows
the controller current of the D-STATCOM, second graph shows when the reactive power
absorption, graph three shows the VSC voltage and graph four shows the voltage and current of
D-STATCOM.
From 0 to 0.1s, the D-STATCOM is initially floating, at 0.1s – 0.2s, there is a voltage rise, the
capacitive current flows into the device and gets in phase with the voltage during which there is
reactive power absorption. +0.25Mvar is absorbed (blue trace) to bring line voltage close to 1pu.
After 0.2s, the device again gets into the floating mode and there is no current into the device there
for no reactive power absorption. Reactive power absorption is happening when there is a voltage
rise beyond the limits.
From figure 4-5 below, graph two shows the amount by which the voltage has been reduced during
the mitigation time (between 0.1 and 0.2s).
36
4.4.2. Ferranti Effect Mitigation Using D-STATCOM Graphs
Figure 4-6 shows the SVC operation graphs during compensation. Graph one shows the
Vmeasured and the Vreference of SVC, graph two shows the firing pulse of the SVC-TCR, graph
three shows the amount of reactive power absorbed and graph four shows the voltage and current
characteristic of the SVC.
37
SVC operation during compensation
Initially from 0 to 0.1s, the SVC is floating, at 0.1s there is a voltage rise that is sensed and makes
the TCR to trigger on thereby starting to absorb reactive power. Reactive power absorption goes
on up to 0.3s when the voltage returns almost to 1pu value and the device stops absorbing reactive
power when the measured voltage is the same as the referenced voltage. Reactive power absorbed
is -0.22Mvar for the voltage to be brought close to 1pu.
For both devices, the amount of reactive power absorbed differs based on the device performance
and effectiveness and this can be as seen in the measurement graphs.
38
4.4.4. Ferranti effect mitigation using SVC Graphs
As discussed, D-STATCOM and SVC were used to mitigate this effect. According to Ding Lijie,
Lui Yang and Miao Yiqu [10], a comparison in their performance was done to determine the most
appropriate for adoption.
Table 4-4 shows the parameters that were chosen for the general performance comparison from
which the best device was chosen to be the most effective for Ferranti effect mitigation.
39
Table 4-4: Device Performance Comparison Table
Response time Faster (0.103 - 0.1 = 0.003s) Slow (0.118 – 0.1 = 0.018s)
(from figure 4.4, graph four- (from figure 4.6, graph four-
blue trace, scale 1) blue trace, scale 1)
(from figure 4.5, scale 1&2) (from figure 4.7, scale 1&2)
40
CHAPTER 5: CONCLUSION, RECOMMENDATIONS CHALLENGES AND FURTHER
RESEARCH
This chapter covers the project conclusions, recommendations, challenges and future research. The
conclusion presented has been obtained based on the results and how the devises have performed
in mitigating Ferranti effect in KIS distribution network.
5.1. Conclusion
This project presented the Ferranti effect on long low loaded power distribution lines. From the
discussion above, Ferranti effect on Kalangala infrastructure services power distribution network
(33kV) was analyzed with help of the DigSilent Power Factory 15.1 software and using the data
provided by KIS. It was discovered that there was a voltage overshoot of up to 1.104pu (36.44Kv)
at the receiving end of Kalangala feeder and this value is beyond the recommended standard error
limit of +/-10% of the normal system voltage.
Two devices SVC and D-STATCOM were modeled also with help of Matlab/Simulink. Both SVC
and D-STATCOM are important shunt FACTS equipment for reactive power compensation,
which were compared in response speed, amount of reactive power absorbed, & the amount of
voltage reduced at the same rating. Simulation results were presented as the same controllable
rated capacity for SVC or D-STATCOM was connected on the distribution bay of Bukuzindu
power generation substation and their performance and effectiveness analyzed.
From the performance comparison of the two devices, installation of 0.3Mvar D-STATCOM at
the distribution bay of Bukuzindu hybrid power station will mitigate Ferranti effect in the KIS
network and lead to the standard voltage limits of between 0.9pu & 1.1pu i.e. within the +/- 10%
of the supply voltage.
5.2. Recommendation
41
Have a D-STATCOM installed on the network to reduce on Ferranti effect.
Reconfigure all the single-phase transformers on both feeder to ensure that all the phases are well
balanced that in this case you will not that all the transformers are picking from red and blue phase
leaving out yellow thus current imbalance and its side effects.
Ensure that the loads on these 3-phase transformers are all balanced because the considered losses
are due to current imbalances.
5.3. Challenges
There were challenges during data collection from KIS. KIS is far and the transport means
involving water and road, at times we had to spend night at the island because ferry services seize
to exist beyond 6pm.
There was also a challenge of accessing and learning to use the different software such as
MATLAB/SIMULINK and DigSilent which were never taught in class. It was really a challenge
as one could not use the software’s before studying and learning them which consumed a lot more
time.
Cost benefit analysis of the two devices has not been done. This is because many parameters to do
the cost benefit analysis most especially for Kalangala geographical area, maintenance schedules,
manufacture procedures were not known. Therefore, for a secure investment, further research
should be done to determine the most cost-effective device in Ferranti effect mitigation if installed.
References
42
[1] G. Deb, "Ferranti Effect in Transmission Line," International Journal of Electrical and
Computer Engineering (IJECE), pp. Vol.2, No.4, pp. 447~451, ISSN: 2088-8708, August 2012.
[6] Sreejith.S, Upama Bose, K. Muni Divya Sree Vachana, Vallathur Jyothi,, “Application of
[7] Nguyen Van Minh, Bach Quoc Khanh, Pham Viet Phuong, “Comparative Simulation Results of
DVR and D-STATCOM to,” 2017 International Conference on System Science and Engineering
[8] International Journal of Innovative Research in Electrical, Electronics, Instrumentation and Control
[10] Venkata Padmavathi S, SaratKumar Sahu, A Jayalaxmi, “Modeling and Simulation of Static
43
Var Compensator to Enhance the Power System Security,” Hyderabad, A.P., India, 2013.
[11] Ding Lijie, Liu Yang, Miao Yiqun, “Comparison of high capacity SVC and STATCOM in real
power grid,” 2010 International Conference on Intelligent Computation Technology and Automation,
for compensating voltage fluctuations, in IEEE Transactions on Power Delivery, vol. 16, no. 2,
[16] S. M. B. N. M. F. Huweg, Application of inverter based shunt device for voltage sag mitigation due
https://books.google.co.ug/books?id=Gw9pCwAAQBAJ&pg=PA173&lpg=
PA173&dq=According+to+IEEE-CIGRE+co-definition+AN+SVC&source=bl&ots=DV2sk
QMTyQ&sig=ACfU3U3oApXXdS0gibSKEzIotWv39aV6Yw&hl=en&sa=X&ved=2ahUKE
wjG4MS8xsPiAhVCXRUIHe1IAV4Q6AEwAnoECAkQAQ#v=onepage&q=.
44
[Accessed 1 June 2019].
[18] KIS, "Bukuzindu hybrid power station operation data log of 1st-9th January 2019," Kalangala, 2019.
45