Design of a Single Wire Earth Return (Swer) Power Distribution System and Improvement of Its Voltage Profile Using Capacitors

DESIGN OF A SINGLE WIRE EARTH RETURN (SWER) POWER DISTRIBUTION SYSTEM AND IMPROVEMENT OF ITS VOLTAGE PROFILE USING CAPACITORS

NIRERE MARIE SOLANGE

EE300-0014/15

Master of Science in Electrical Engineering (Power Systems Option)

A thesis submitted to Pan African University, Institute for Basic Sciences Technology and Innovation in partial fulﬁllment of the requirement for the degree of Master of Science in Electrical Engineering

2017 DECLARATION

I hereby declare that this thesis is my original work and has not been presented for award of MSc degree in any University.

Signature: ______Date: ______

NIRERE MARIE SOLANGE [EE300-0014/15]

This Thesis has been submitted for examination with our approval as University Su- pervisors.

Signature: ______Date: ______

Dr. C. M. Muriithi

Technical University of Kenya, TUK, (Department of Electrical and Power Engineer- ing)

Signature: ______Date: ______

Dr. C. Wekesa

University of Nairobi, UON, (Department of Electrical Engineering)

i DEDICATION

This thesis is dedicated to my mother Mukakaroli Leonille, my husband Musabyeyezu J Damascens, and my children Senga Davy Axel and Shima Angelo

ii ACKNOWLEDGMENT

I would like to express my gratitude to my supervisors Dr. Christopher Maina Muriithi and Dr. Cyrus Wekesa for their continuous support and guidance, enthusiastic encouragement and useful critiques of this research work and especially for enlightening me from the ﬁrst glance of the research. I also would like to recognize the JICA and AU for funding the research without which it could have been very difﬁcult to achieve the results obtained. I would also like to extend my thanks to the program coordinator of Electrical engineering and all the lecturers of the Department of Electrical Engineer- ing, for the support extended during the study period.

Further I would like to acknowledge the PAUSTI and Jomo Kenyatta University of Agriculture and technology for facilitating the study. It is a great pleasure to remember the kind cooperation extended by colleagues for their encouragement and continuous support during my MSc study and research period. Last but not the least; I would like to thank my family: my mother Mukakaroli Leonille, my husband Musabyeyezu J Damascens, and my children Senga Davy Axel and Shima Angelo for their patience, understanding ,encouragement and support to complete my research. May all glory be to God.

iii ABSTRACT

Access to safe, reliable and cost effective energy is essential for a country to make the economic growth of the population. It is well known that electricity is a service and a key input into economic development and household activities. The cost of extending the main grid to supply remote and rural area is extremely high. Rural areas which are characterized by low load density and scattered load requires a low cost electriﬁcation method; and electriﬁcation using Single wire earth return distribution has shown to be the most economic option . This is more important and more applicable for remote centers in different villages in Rwanda where people are living in settlements. In this research a settlement is considered as one load to be supplied by one distribution transformer.

SWER system is, by deﬁnition a single wire earth return distribution system in which all equipment connected to it, is grounded to earth and the earth is used as a return path where the return current passes through. The purpose of SWER is to power relatively small but relevant loads over a long distance at the least possible cost; SWER is a low capital and maintenance cost distribution system which has ever been chosen and widely built in many parts of the world especially in remote areas where loads are scattered, light and far from the main grid. However,even if this system has shown to be most economical, it has shown challenges especially in terms of voltage regulation and losses that affect the system capacity and the quality of the power to the end users and there is a need to ﬁnd a way of improving this distribution system and providing a reliable and sustainable power to remote rural areas.

This research aims to improve the voltage proﬁle of a single wire earth return distribution system. One of the sector located in rural area of Southern Province of Rwanda was selected, and its demand in electricity was estimated by taking into account several aspects such as type and number of consumers, equipment and expected time of operation for the appliance. After the network was designed a load ﬂow calculation

iv was performed using backward/forward sweep method to determine the system‘s voltage proﬁle, the results from backward and forward sweep method were compared with the results from Direct load ﬂow approach. Both results have shown a poor voltage regulation at all nodes except the starting node which is considered as the slack bus.

To improve the voltage profile capacitors were proposed, These are usually used as reactive power compensators in electrical network. The main benefits of their utilization are to minimize the power losses, improve power factor and maintain best voltage regulations for all load buses. To determine the optimum size and siting of the capacitor, maximum power saving method were used. The findings has shown that the use of capacitor has improved the voltage profile, where voltage at all nodes are within the acceptable limits and both reactive and active losses were reduced.

v CONTENTS

ABBREVIATIONS ...... x

1 Introduction 1

1.1 Background ...... 1

1.2 Statement of the problem ...... 2

1.3 Justiﬁcation ...... 3

1.4 Objectives ...... 3

1.4.1 General Objective ...... 3

1.4.2 Speciﬁc Objective ...... 3

1.5 Scope ...... 4

2 LITERATURE REVIEW 5

2.1 Introduction ...... 5

2.2 Single Wire earth return system ...... 5

2.2.1 Basic conﬁgurations of SWER systems ...... 5

2.2.2 SWER characteristics ...... 9

2.2.3 Mathematical model of SWER distribution line ...... 11

2.2.4 Beneﬁts of SWER ...... 13

2.2.5 SWER limitation ...... 16

2.3 Review on voltage proﬁle improvement in SWER networks ...... 17

2.4 Methods of estimating load demand ...... 19

2.5 Load ﬂow methods used for distribution systems ...... 19

2.5.1 Backward and Forward sweep method ...... 20

2.5.2 Direct Load Flow Approach ...... 25

vi 3 METHODOLOGY 29

3.1 Estimating electricity demand of a rural area ...... 29

3.2 Design of radial distribution network ...... 31

3.2.1 Isolating transformer ...... 32

3.2.2 Distribution transformer ...... 32

3.2.3 Earthing system ...... 33

3.2.4 Conductors ...... 34

3.2.5 Development of the model ...... 37

3.3 Load ﬂow in Radial network and voltage proﬁle determination . . . . 38

3.3.1 Voltage proﬁle improvement ...... 38

3.3.2 4.3.2 Capacitor Sizing and placement determination . . . . . 39

4 RESULT AND DISCUSSION 43

4.1 Introduction ...... 43

4.2 Results from estimation of rural load data ...... 43

4.3 Load ﬂow calculation ...... 49

4.3.1 Results ...... 49

4.3.2 Results with capacitor in the network ...... 50

4.4 SWER cost structure ...... 52

5 CONCLUSIONS AND RECOMMENDATIONS 56

5.1 Conclusion ...... 56

5.2 Recomendations ...... 56

REFERENCES 58

A Detailed daily consumption for selected village 63

vii LIST OF FIGURES

2.1 Single wire earth return system ...... 7

2.2 Isolating transformer with two phase input and single phase output . .8

2.3 Wenner four pin method for measuring soil resistivity ...... 10

2.4 Model of Carson’s line with earth return ...... 11

2.5 Single line diagram of a radial network ...... 21

2.6 Process Flow Chart ...... 23

2.7 A Radial distribution system detailng DLF ...... 25

3.1 Single line diagram of the designed network ...... 37

3.2 Flow chart of the BFSM/MPS ...... 42

4.1 Daily Load proﬁle (Daily Demand variation) for ﬁrst and last centers . 46

4.2 Daily load curve for middle centers ...... 46

4.3 Daily load proﬁle of the whole network ...... 47

4.4 Voltage proﬁle with and without capacitor ...... 51

4.5 Reactive power losses with and without capacitor ...... 52

viii LIST OF TABLES

2.1 Typical Soil Resistivity Values ...... 11

2.2 Relative load capacity of different technologies ...... 15

2.3 Cost of different type of distribution systems ...... 15

2.4 Total investment cost for different distribution systems in USD . . . . 16

3.1 Load classiﬁcation and daily energy demand for each category . . . . 31

3.2 Properties of common SWER conductors [24] ...... 35

3.3 Load location and their maximum demand at peak load ...... 36

4.1 daily hourly demand variation of the whole village ...... 44

4.2 daily demand for 1st and last centers ...... 44

4.3 Daily hourly demand variation of middle centers ...... 45

4.4 Results from BFSMPower ﬂow ...... 49

4.5 Base case results of the test system ...... 50

4.6 Results with capacitor in the system ...... 50

4.7 Cost of 12.7 kV, 1-phase SWER Overhead Line ...... 53

4.8 Cost of 12.7/0.23 kV, SWER, 1 Pole, Distribution Transformer . . . . 54

4.9 Cost of 22/12.7, SWER, 2 Pole, Isolating Transformer ...... 55

A.1 Domestic purposes/ Residential demand ...... 63

A.2 Commercial load ...... 64

A.3 Community load ...... 65

A.4 Secondary school ...... 66

A.5 Small manufacturing units ...... 66

ix ABBREVIATIONS

ACSR Aluminium Conductor Steel Reinforced

BCBV Branch Current and Bus Voltage

BIBC Bus - Injection to Branch Current

BFSM Backward and Forward Sweep Method

DLF Direct Load Flow

GMR Geometric Mean Radius kV Kilovolt kVA Kilovolt-Ampere kVAR Reactive kiloVolt-Ampere

KCL Kirchoff Current Law

KVL Kirchoff Voltage Law kW Kilowatt kWh Kilowatt Hour

LVRs Low Voltage Regulators

MDGs Millennium Development Goals

MV Medium Voltage

MPSM Maximum Power Saving Method

RDN Radial Distribution Network

SWER Single Wire Earth Return

USD United State Dollar

x CHAPTER 1

INTRODUCTION

1.1 Background

Electrification is widely believed to contribute to the achievement of Millennium De- velopment Goals (MDGs), based on the assumption that sustainable access to electricity fosters economic and social development ,and leads to improvement in the quality of life[1]. Yet particularly in rural Sub-Saharan Africa electrification rates are still low, as only 11% of the population use electricity. In rural Rwanda electrification late is considerably lower at 7.7% of the total population [2]. The National target for the country is to increase the overall electrification rate to 70% by 2020. Improving access to electricity has been made a priority in Rwanda, where presently only 24.5% of Rwanda’s households are connected to the grid with a total generation of 193MW. The majority of the population lacks access to electricity, This poses an enormous challenge to the country where most of the rural village consist of a number of isolated populations, with limited financial resources to invest in electrification projects. In addition, with global oil prices rising, the cost of supplying electricity through conventional means has become unsustainably expensive and in some cases unfeasible.

Traditionally in Rwanda, outside of urban areas, where grid extension is not economically feasible, electricity is supplied by decentralized diesel generators. Although, initial capital investment requirements are low, diesel generators are very costly to operate and maintain, which often makes this technology option unsustainable in isolated rural communities where household income is low, and skilled labor rare. Introduc- ing single wire earth return distribution system in remote area will reduce the cost of electrification which is high compared to urban area where loads are heavy and closer. Historically, rural electrification has been a huge challenge and remains so in many parts of the world. The use of standard electrification technologies becomes unviable

1 in rural areas due to the high cost of investment and the low load densities. As a result Single Wire Earth Return (SWER) distribution technology is used to provide a cost-effective way of supplying electricity in rural area where loads are scattered and sparse. SWER systems use light weight high tensile conductors to supply power to rural areas from the main grid network using the earth as return path [1]. SWER allows longer spans, lighter poles and fewer pole top equipment to be used leading to considerable savings on initial investments compared to conventional two wire single phase distribution systems. The technology, initially proposed by Mandeno[3], has proven to be cost-effective in electrifying rural areas with small and scattered loads in countries such as Australia, Namibia, Mozambique, and South Africa. [1]. Although the system is very cost-effective, but it has some disadvantages[4]:

• High losses due to high resistance values of the conductors.

• The system capacity is limited by voltage drops and high voltage regulation;

• Unbalance between the phases of the incoming feeder, due to the fact that SWER is supplied from only two of the three phase lines from the feeder.

1.2 Statement of the problem

The national utility faces a constant problem to supply remote and rural area due to the topography of Rwanda, which is known as a country of thousand hills, it is extremely costly to provide electricity to rural areas through conventional means. Currently, in rural areas, most of the schools, health centers, administration posts, small commercial centers and other communities use solar systems for each home and fuel generators. Instead of providing isolated solar systems for each home or fuel generator, the extension of the grid through the utilization of SWER in distribution network for elec- triﬁcation of the whole community in rural villages would be more economical and reliable. Yet this system has presented challenges especially in terms voltage regulation and losses that are affecting the system capacity and the quality of the power to

2 the end users. There is a need to ﬁnd a way of improving this distribution system and providing a reliable and sustainable power to remote rural areas.

1.3 Justiﬁcation

Electrification is one of the important keys to development as it provides light, heat and power to be used in residential,commercial and manufacturing or industrial sectors. For the country to attain its target, which is to increase the overall electrification rate by 2020 and as part of the government’s effort to provide electricity to all, SWER can play a key role in achieving this vision. The findings of this work will be beneficial to remote communities in rural areas because the cost of electrification will be minimazed and they will be able to afford it. The designed SWER is cost effective due to the use of only one wire/conductor, less pole top equipment, long, hilltop to hilltop spans ,Fewer switching and protection devices and if it is taken from a high voltage three phase transmision line, losses will be minimized and cost of capacitor will be reduced.

1.4 Objectives

1.4.1 General Objective

The main objective of this study is to design a single wire earth return (SWER) power distribution system and analyze the effect of capacitors on the voltage proﬁle of the system.

1.4.2 Speciﬁc Objective

The speciﬁc objectives of this research are:

1. To estimate the load proﬁle of a typical settled rural area

2. To design a radial network to distribute power to the selected area from the grid extension using Single Wire Earth Return.

3. To determine the voltage proﬁle of the network and analyze the effect of capacitor on the voltage proﬁle of SWER.

3 1.5 Scope

The scope of this study ends to the development of the SWER distribution network and determination of the voltage proﬁle as well as improvement of the voltage using switched capacitor. The research scope includes the static model of Single Wire Earth Return distribution networks, whereby load growth is not considered over several time periods and reliability analysis is not included. The analysis of this system will be done by considering the following assumptions/limitations:

• The consumers live conforming to a daily routine coming from the same load cycles every day the peak demand is considered .

• The distribution network is supplied by one SWER isolating transformer substation located at a known point of grid extension. Therefore, substation planning is not considered in the distribution problem.

• The locations of the loads are known beforehand.

• This system is not speciﬁcally located; it will be the optimal conﬁguration for other locations. The same approach can be used for other communities by following the same procedures as it is used throughout this project.

• This study will not discuss the issues related to the stability , protection and control.

4 CHAPTER 2

LITERATURE REVIEW

2.1 Introduction

This chapter presents an overview of the technical aspects and economic aspects of SWER power distribution ,and explain in more details how it works, challenges as well as proposed solutions to solve some of the problems faced by the system. It also highlights other author’s work about the topic.

2.2 Single Wire earth return system

Single Wire Earth Return System (SWER) was ﬁrst invented in New Zealand in 1925, and by the 1940s it was considered as the preferred solution for the economic extension of power distribution networks in the remote rural areas of both New Zealand and Australia. Currently, there are more than 200,000 km of SWER power lines spread throughout the rural areas of both countries .Single Wire Earth Return (SWER) is a single-wire electrical distribution scheme which is used to supply electricity to rural and sparsely populated areas at a low cost. A single wire is used to distribute the electricity at medium voltage (MV) to distribution transformers, and all equipment is grounded to earth. This ground provides a return path for the current through the earth.

2.2.1 Basic conﬁgurations of SWER systems

There are two conﬁgurations of SWER depending on whether the use of an isolating transformer is involved or not. Others depend on these two described below.

• Single phase system without isolation transformer

In this conﬁguration, an only metallic conductor is directly connected to one phase of any three-phase line. The return current ﬂows through earth to the substation earthing.

5 At the consumer entrance, a distribution transformer has its primary winding connected between such single conductor and the ground. This system is usually adopted for feeders that originated from substations, whose transformer has a grounded neutral point at the secondary winding.

The configuration may be considered as simplification of conventional single-phase multi-grounded system, where neutral conductor and the grounding along the line were suppressed. Certainly, this is the cheapest version of the SWER systems. This configuration is not ideal as it does not isolate the SWER network from the main network, and the earth return currents can be of sufficiently high magnitude to operate earth fault protection installed on the main system [5]. Seeing that one of the major costs, complications and technical challenges to implementing SWER systems is the isolating transformer; The South African has adopted a concept of micro SWER consists of allowing SWER extensions without isolating transformers for loads under 5 kVA. This type of supply is principally aimed at supplying remote mobile phone repeater sites.

• Single phase system with isolation transformer

In this typical SWER conﬁguration, two phase conductors from a 33, 22 or 11 kV MV three-phase system are connected to the isolation transformer which then supplies the load at 19.1kV , 12.7 kV or 6.35kV single phase. Consumers are supplied from step-down SWER distribution transformers which have one or two outputs that are center-tapped in a 240-0-240 V arrangement as shown in Figure 2.1 [3].The connected load’s neutral is joined to the earth such that the return current ﬂows into the earth back to the distribution transformer via electrodes embedded deep into the ground. Figure 2-1 shows a schematic diagram of a typical SWER distribution system. [3].

6 Figure 2.1: Single wire earth return system

The use of isolation transformer allows some improvements:

1. To adjust the voltage of SWER system (to standard nominal voltages) : Regard- less on which main MV grid it is tapped, the isolating transformer allows for the selection of voltage for the SWER network independent of that.

2. To raise the voltage level in order to supply longer lines, Although 33 kV and 22 kV three phase lines can be used to supply the equivalent single phase 19.1 kV and 12.7 kV SWER respectively, an isolating transformer can be used to supply 19.1 kV from a 22 or 11 kV MV network.

3. To limit the circulation zone of return current : It restricts the SWER earth currents to the area between the SWER distribution transformers and the supplying isolating transformer which helps to reduce the interference with open wire com- munications

4. To prevent the improper performance of protection device for high impedance phase-ground faults ; Isolating transformer provides sensitive earth fault protection schemes on the primary MV three-phase network. This will help to prevent

7 the sensitive protection schemes from detecting the earth return load currents of SWER as a permanent earth fault.

5. To balance voltage - Supply to the SWER system is taken by two phases of the primary MV network. So the isolating transformer permits better voltage balance on the feeders from the main transformer.

6. To control voltage - Tapping ranges in isolating transformers provide cost effective voltage control on the SWER system. This will allow ﬁxed tap distribution transformers to be used on the SWER network. .

However, it also presents disadvantages, such as, the limitation of branch power to the nominal power of isolation transformer, the need of a very low grounding resistance for the isolation transformer and the additional cost of such transformer. With the high cost of Isolation transformer and higher system losses, SWER is not economically feasible for grid extensions less than 6 km .Figure 2.2 shows a typical SWER isolating transformer in Gerus substation, Namibia with two phase input and single phase out put.[7]

Figure 2.2: Isolating transformer with two phase input and single phase output

8 2.2.2 SWER characteristics

In addition to the conﬁguration detailed above, mentioning Isolating transformer as the main equipment of the SWER network, there are also other parameters that are very important in the distribution using SWER. In these include; the distribution transformer, the type of conductor to be used as single line, the nature of the existing electrical system, the proﬁle of the load to be supplied and the local soil resistivity. Because of the use of ground as return path, the resistivity of the soil has to be closely monitored in order to meet all the earthing requirements and operate the system within safe levels. Soil resistivity may vary widely within short distances in the same area and so several measurements may have to be taken at different locations. The variations in soil type mean that different soils will have different concentrations of electrolytes for conducting electricity. This means, therefore, that soil moisture variations will increase or decrease the soil conductivity. As a general rule, however, the higher the soil moisture content, the higher the soil conductivity [8].

Soil resistivity is also affected by temperature. The resistivity of soil, given constant moisture content, reduces with increasing temperature and increases with decreasing temperatures. At lower temperatures, for example, when soil moisture freezes, the soil resistivity increases almost threefold compared to its unfrozen values [8]. This is especially detrimental to conductivity in clay or cement-based soils which rely solely on moisture to conduct current. The relationship be-tween soil resistivity and temperature is given by the Steinhart-Hart equation in (2.1) [4].

1 = A + Bln(ρ) +C [ln(ρ)]3 (2.1) T

Where, T is temperature (K), ρ is the soil resistivity (Ω − m)and A, B, C are Steinhart- Hart coefﬁcients which vary depending on the temperature range in question. From the relation between resistivity, moisture content and temperature, it can be deduced that resistance of any grounding system will vary greatly at different times of the year depending on the season. Therefore, good measurement of soil resistivity prior to

9 SWER system installation is essential. The soil resistivity measurements are typically done using resistivity meters that is based on the Wenner four-pin method [4]. In this method, four rods are driven into the earth spaced at equal distances from each other in a straight line as shown in Figure 2.3. A current is then driven through the two outer rods (C1 and C2) and the voltage across the two inner rods (P1 and P2) is measured.

Figure 2.3: Wenner four pin method for measuring soil resistivity

From this, the resistance of the soil can be calculated using Ohm’s law and the resistivity,ρ, is calculated using (2.2) [4].

4πaR ρ = (2.2) 1 + √ 2a − √ 2a a2+4B2 4a2−4B2

Where a the distance between the rods (m), B the depth of the rods (m), R the earth resistance (Ω) and ρ the soil resistivity (Ω −m). For cases where a > 20B, the formula is simpliﬁed to (2.3). Table 2.1 shows typical values of resistivity for different soil types [3].

ρ = 2πAR (2.3)

10 Table 2.1: Typical Soil Resistivity Values

Type of Soil Resistivity (Ω.m) Clay 5-150 Chalk, Limestone 90-400 Sand and Gravel 100-3000

Hence, soils with low earth resistivity allow for higher loads to be supplied. The I2R losses in form of heat through the earth may cause the earth to dry. This will increase its resistivity further which in turn will increase the I2R losses leading to further drying and so on .Therefore, loads in these systems are generally restricted to 450 kVA with current limited to 25 A at 19.1 kV.

2.2.3 Mathematical model of SWER distribution line

SWER distribution line is based on Carson’s model which considers a single overhead conductor a of unit length parallel to the earth and carrying a current Ia with return path g − g’ underneath the earth surface. Earth is assumed to be a single solid conductor of inﬁnity length, uniform resistivity and a geometric mean radius (GMR) of 1m [12]. Considering the ﬁgure below

Figure 2.4: Model of Carson’s line with earth return

The voltage of both overhead conductor and ground return are formulated as follows:

11 [10]

    V 0 V −V 0  aa   a a    =   (2.4) Vgg0 Vg −Vg0

    z¯ z¯ I  aa ag   a  =   ∗   (2.5) z¯ag z¯gg −Ia

Where all voltage are at the same reference and Vg = 0, subtracting above equations we get

Va = (z¯aa + z¯gg − 2¯zag)Ia = zaaIaa (2.6)

From which the resultant SWER overhead conductor impedance zaa is obtained as

zaa = z¯aa + z¯gg − 2¯zag Ω/km (2.7)

Withz ¯aais the line self-impedance,z ¯gg the ground self-impedance, andz ¯ag the mutual impedance between the line and earth. The factor (z¯gg − 2¯zag) represents the impedance correction due to the earth presence. The self-impedance of the overhead conductorz ¯aa, the self-impedance of ground conductorz ¯gg and the mutual-impedance between overhead and ground conductors are then formulated using the modiﬁed Car- son line model [4] as follows:

−4 2ha z¯aa = Ra + j4π ∗ 10 f ln (2.8) GMRa

With Ra the resistance of the conductor in (Ω/km), f the frequency (Hz),ha the height of the conductor above the earth in (m), and GMRa the Geometric Mean Radius of overhead conductor(m).

12 2 z¯ = π2 ∗ 10−4 f − j0.0386 ∗ 8π ∗ 10−4 f + j4π ∗ 10−4 ∗ f ln (2.9) gg 5.6198 ∗ 10−3

  −4 ha z¯ag = j2π ∗ 10 ln (2.10) q ρ  f

Whereρ is the earth resistivity in Ω.m

Line shunt admittance is derived using shunt capacitance C which is used to express the capacitance reactance Xc=1/( jωC) the shunt admittance Y = 1/Xc in S/km.

1 h C−1 = ln a (2.11) 2πε0 ra

It can be seen that increase in separation between earth and overhead conductor results in high capacitance reactance, this causes higher charging current in SWER than in conventional system. Regarding the use of steel conductor which are characterized by a high resistance, SWER network faces a big challenge of increase in losses and voltage drop and that leads to a very poor voltage proﬁle. Thus a need to improve the voltage regulation.

2.2.4 Beneﬁts of SWER

The main advantage of SWER systems over conventional distribution is the considerably reduced cost both in initial investment and maintenance. The single conductor allows for much longer span lengths enabling use of fewer and lighter poles. This results into a marked reduction in material and labor costs. It also requires a lot fewer insulators as well as switching and protection devices compared to three-phase lines of similar length. The design simplicity and easy of construction of SWER networks ensures a shorter implementation time as well as low maintenance costs after installation. Given the simple protection schemes required, the major complexity is ensuring the low resistance earthing at the isolating and distribution transformers. Utilities op-

13 erating SWER systems estimate that the maintenance costs for SWER are about 50% those of conventional systems [1].

The reliability of SWER networks is higher than that for conventional three-wire three- phase and two-wire single phase networks. This is mainly because the single conductor eliminates instances of conductor clashing, both physically and magnetically, and requires fewer components. The former feature reduces bush ﬁres prevalent in rural areas supplied by multi-wire distribution systems. Incidences of lightning strikes are reduced by using simple surge arrestors. Detailed descriptions of some of the above advantages are given below.

1. Reduced capital cost – SWER system use only one conductor. As a result, there are less pole top equipment (one insulator and no cross arms). Very long spans can be achieved thus requiring fewer poles, insulators and other materials result- ing in lower labor and material costs.

2. Design simplicity - It is a simple single-wire system supported on basic poles and with basic electrical protection. The only major concern in the design is ensuring that low-resistance earths are achieved both at the isolating transformer and the distribution transformers.

3. Ease of construction - With only one wire and simple basic pole supports, construction is much easier. Sagging and separation of conductors is not an issue. Many of the SWER lines in New Zealand and Australia have been erected by farmers with no previous experience in electing power lines.

4. Reduced maintenance costs - SWER has fewer components than traditional systems, so clearly there are less things to go wrong. With a single wire, there are no problems with line clashing. Tree and vegetation management problems are also minimal. The only signiﬁcant maintenance issue is the testing of earths.

The study done on the cost aspect of the SWER in [15] has shown that SWER is sometimes a reasonable alternative, if the loads are low. The utilization of only one

14 conductor decreases the investment costs of the system, but on the other hand, losses are high and any isolating transformers needed increase the costs. SWER also requires proper earthing and earth surface with adequately low resistivity year round.

Table 2.2: Relative load capacity of different technologies No Type of Technology Relative Difference in load Capacity(%) 1 Three Phase Three Wire 100 2 Two phase Two Wire 50 3 Single phase Two Wire 26 4 SWER 29

The table indicates that three-phase system has the greatest load capacity. The two phase system has twice the load capacity of the single phase two-wire system, which again has very similar ﬁgures to those of SWER depending on the earth resistivity , earthing and the power factor. The investment costs for three-phase lines are higher than for two or single phase lines because they have three conductors, even though they can have a smaller cross sectional area for the same load carrying capacity. SWER has the lowest investment costs, but high losses and some disadvantages dealt with later. Further comparison between two phase and single phase lines indicates that savings in the insulation level of single phase lines will be partly offset by the larger cross sectional areas of the phase conductors required to deliver the same load with equal voltage quality. There are no great differences in the substation structure, except sometimes in the earthing of the neutral point of the HV/MV transformer.

Table 2.3: Cost of different type of distribution systems No Type of Technology Cost (USD)/km MV line 1 Three Phase three wire 13000 2 Two Phase two wire 9500 3 Single Phase Two wire 8600 4 SWER 5200

This table shows the cost of constructing a one km of line by selected four technologies. From the Table it can be clearly seen that the lowest cost technology is SWER and Single-phase two-wire system, two-phase two-wire system are the next low cost solutions respectively.

15 Depending on the location of consumers and loads to be supplied, different solutions could be used. Below, in Table 24 the investment costs for different type of villages are compared. This can be considered as a rough estimation for all the countries in the African region. Local conditions may have an impact on these costs.

Table 2.4: Total investment cost for different distribution systems in USD

The most economic and technically feasible distribution solution depends on the number of customers to be connected, the distance between the connection points and the grid and the type of connection as discussed above. For electriﬁcation of lightly populated rural areas located at least 30 km from the grid, SWER could be used, as long as the total load is below 450 kVA. SWER could be a temporary or permanent solution depending on the development of the load. The main advantage with this solution is that it is easy to erect with low investment and if needed can be upgraded in steps, to two-phase and/or three-phase by adding cross-arms and more poles without squander- ing the initial investment [15].

2.2.5 SWER limitation

Single wire earth return is not used as a standard distribution strategy especially in urban area due to the fact that urban loads are high compared to rural load. In urban areas three phase balancing is required whereas SWER is causing unbalancing from the substation is taken from. Load densities in SWER are typically less than 0.5 kVA per kilometer of line with a maximum demand per customer of 3.5 kVA. A large system may supply up to 80 distribution transformers with unit ratings of 5 kVA, 10 kVA and 25 kVA. SWER lines tend to be long, with high impedance; typically, conductors

16 have a small diameter and high strength, and are made of aluminum/steel or steel cable. Sections at the sending end of the line often are 3/4/2.5 ACSR (34mm2) or are similar. Tee-offs and lightly loaded sections often are 3/2.75 SC/GZ (17.8mm2) [1]. These high impedance conductors in long SWER lines can cause the total impedance at the end of the feeder to be 1000 ohms. This causes the voltage drop along the line to be often a problem, causing poor regulation. Also disparities in demand cause variation in the delivered voltage.

2.3 Review on voltage proﬁle improvement in SWER networks

In a practical power system network especially in distribution system the system oper- ators are always obligated with voltage levels of each customer bus within the satisﬁed limits. To ensure good voltage proﬁle in distribution systems, several standards have been established to provide recommendations and stipulations. In general many electrical power supply companies try to maintain/control the distribution voltage variations within the range of±5%. Distribution systems normally consist of a main feeder and lateral distributors. They acts as a link between high voltage transmission line and the low voltage consumers. The low voltage and high current characteristics of distribution system leads to high power losses compared to that of transmission system. About 13% of total power generated is consumed as power losses at the distribution system [19]. In SWER several issues are associated with in the system that need to be addressed. One of the key issues pertaining to SWER feeders is that of voltage regulation. Another is high losses due to the long distances and high resistance of the overheard conductor.

In [12], the autor have proposed Distributed generation along SWER lines to improve voltage regulation and reduce feeder losses. This strategy is effective at controlling voltages and reducing feeder losses.These solutions are more technically complex but are certainly achievable. In [29]Central Queensland University has been examining method of applying controlled reactors as an intermediate approach to improve voltage regulation of SWER systems at a lower capital cost. Also low voltage regulators

17 (LVRs) were proposed; these are power conditioning units designed to supply power at a settable voltage and at unity power factor within a broad range of input voltages and power factors. They are likely to be used to:

1. Support low voltage levels at a customer supply point.

2. Reduce voltage variation at a customer supply point to fall within±7%;

3. Improve power quality by removing peaks and sags.

During very light load periods the SWER line capacitance creates voltage rise toward the end of the feeder. To combat this issue, ﬁxed shunt reactors are used to control voltages. Unfortunately, these reactors add to the load during peak load periods. As a result, excessive voltage drop limits the load capacity of the feeder. Recently, devices incorporating power electronics have been designed to give greater control over the operation of reactive devices. A switched reactor was proposed in [10], which consists of the conversion of the ﬁxed reactors into dynamic elements that switch in and out depending on the voltage level. These reactors are used to reduce over voltage by consuming the surplus of reactive power.

In[13] , a thyristor controlled reactor is proposed, which has been in used on the Stan- age Bay SWER in Central Queensland as a trial. The thyristors act as switches which can switch the inductive coils in or out according to the input voltage. This has the effect of improving voltages during light load conditions by switching the reactor in. In heavy load conditions when voltage rise is not an issue, the reactive coils are typically switched out.

This research proposes the use of capacitor as one of the most effective and useful methods in reducing the power losses in distribution networks and improve the voltage profile. By using shunt capacitors, the reactive power needed for loads is provided so that besides the reduction of losses, the voltage profile of nodes is also improved. The capacitor will reduce the line current which is drawn by the load in the system which results in voltage profile improvement, line loss reduction, better reliability and stability of the distribution system.

18 2.4 Methods of estimating load demand

At the rural community level, electricity is used for lightening and household purposes. In addition, electricity is also used for the purpose of farming operations, agriculture, and for other small and medium scale industry operations . In order to effectively plan and operate the rural electric power utility system, the load demand of the rural community must be accurately estimated. Estimation of load constitutes the major step for the planning engineer of distribution systems. Indeed, load estimation in- ﬂuences transformer sizing, capacitor banks, conductor size, and peak load demand period. This means that accurate load estimation is essential to determine the number, locations, and capacities of future substations[33]. In order to estimate the load in rural areas different methods have been used. Electric utilities have utilized several methods of analyzing KWHR consumption which have yielded various levels of accuracy.

One approach was the estimation of the diversiﬁed demand as a function of the average KWHR per customer and the diversity as a function of the number of customers. In another approach, multiplying factors which are arbitrarily called “K factors” are applied to the number of KWHR consumed by each consumer to estimate the maximum demand of each consumer. In[36]an approach was proposed . It is based on fuzzy set theory to estimate the loads in a distribution system using operator experience and ex- pert knowledge. The load patterns are characterized by linguistic variables using fuzzy set notation. In this method the load points within the same category are assumed to have the same hourly load pattern in a day. The hourly load pattern of a particular load category is employed to approximate the load pattern of any branching point within that load category. However, this is only an approximation because the exact historical record of the branching point does not exist.

2.5 Load ﬂow methods used for distribution systems

Load ﬂow studies are performed on Power Systems to understand the nature of the installed network. Load ﬂow is used to determine the static performance of the system.

19 The distribution power flow involves, first of all,finding all the node voltages. From these voltages, it is possible to compute current directly, power flows, system losses and other steady state quantities . Some applications, especially in the fields of optimization of distribution system, and distribution automation (i.e., VAR planning, network optimization, state estimation, etc.), need repeated fast load flow solutions.[19].

A radial network leaves the station and passes through the network area with no nor- mal connection to any other supply. In [32] A. Augugliaro and L. Dusonchet proposed an improved Backward /Forward sweep load ﬂow algorithm for radial distribution systems which includes the backward sweep and the decomposed forward sweep. Back- ward sweep uses KVL and KCL to obtain the calculated voltage at each upstream bus . The properties of the backward/forward sweep method with different line X/R ratios and the convergence conditions are explained in [22]. The Forward-Backward Sweep Method (FBSM) in [19]. is easy to program and runs quickly. Apart from BFSM there is also Direct Load Flow(DLF) and each is explained in the next subsection..

2.5.1 Backward and Forward sweep method

The backward and forward sweep method is used to solve the power ﬂow analysis of the radial distribution systems with recursive equations. The method known as the modiﬁed Ladder iterative technique was proposed by the Autors in [20, 21] and the convergence of the method was explained in [23]. The backward and forward sweep method is based on the Kirchoff’s voltage and current law and in each iteration, two computation stages occur, the forward path and the backward walk as described below.

Backward path : The purpose is to find the current flowing in each branch in the tree by considering the constant value of voltages found in the previous iteration while a flat voltage value is assumed in the initial iteration. The backward path starts from the last node to the source node.

Forward path : It starts from the source node to the far end node aims to calculate the voltages at each bus while keeping the current obtained from the previous walk meaning that the current obtained in the backward walk will be held constant during

20 the forward propagation. The calculated voltages are compared with the speciﬁed voltage and if the error is within tolerance limits, the process is stopped and the power line losses are computed otherwise, the process is repeated until criteria conditions are met.

Figure 2.5: Single line diagram of a radial network

Considering the ﬁgure above a branch is connected between the nodes ‘y’ and ‘y +

1’.The effective active (Py) and reactive(Qy) are sending-end powers that are transfered 0 from node ‘y to node ‘y+1’ and PLy+1 and QLy+1 are load connected to node y+1 ,Iy is the current magnitude ﬂowing from node y to node y+1. Ry and Xy are the resistance and reactance of the line respectively with , Vy and Vy+1 are the voltage magnitude at sending and receiving node respectively. The value of the current is calculated

Py + jQy Iy = con j (2.12) Vy

The voltage at the receiving end is calculated as follow:

Vy+1 = Vy − Iy(Ry + jXy) (2.13)

Nodal voltages are updated in a forward sweep starting from branches in the ﬁrst layer toward those in the last. For each branch, the voltage at node y + 1 is calculated using the updated voltage at node y , and the branch current calculated in the preceding backward sweep.

21 PTLand QTL are the total active and reactive power losses in the system respectively and are calculated :

N 2 PTL = ∑ Iy ∗ Ry (2.14) y=1

N 2 QTL = ∑ Iy ∗ Xy (2.15) y=1

Algorithm for Backward and Forward method

Step 1: Read Bus data(P,Q) and line resistance and reactance data

Step 2: Read base MVA and base KV and calculate the per unit values of the data loaded.

Step 3: Backward walk from end node to source node to ﬁnd all branch currents by using equation 12 while keeping constant ﬂat initial voltages

Step 4: Forward walk from source node to the far end node, to ﬁnd all voltages using equation 13 while updating the constant current values obtained in the previous iteration and check for convergence criterion.

Step 5: Check if the mismatch of the speciﬁed and calculated voltages at the substation is less than the convergence tolerance. If yes, go to next step. Otherwise, repeat step 3 and step4.

Step 6: Calculate the total active and reactive line losses using equations 14 and 15 with the currents and voltages obtained from the backward and forward sweep method.

Step 7: Print the result of all bus voltage and total loss in the system

Step 8: Stop

22 Figure 2.6: Process Flow Chart

23 Backward /forward sweep method in SWER

The SWER load ﬂow formulation based on the forward/backward sweep method for radial distribution networks with earth return is derived in [12]. It is described as follows/:

In the ﬁrst step, all nodal current injections due to loads, capacitor banks, if any, and shunt elements are calculated based on initial voltages. In subsequent iterations, updated voltages are used to calculate the nodal currents. The calculation of nodal currents is given by:

 k     (k−1) sia Iia (k−1) Yia Via    via       =   −    (2.16) Iig −Iia 0 Vig

Where, Iia and Iig are the current injections at node i for the overhead line and earth return path respectively, Sia is the speciﬁed complex power load at node i, Via and

Vig are the complex voltages at node i for the overhead conductor and earth return respectively, Yia is the shunt admittance of the overhead line at node i, and k is the iteration index.

The second step is the backward sweep which calculates branch currents starting from the end nodes of the radial distribution network (RDN) backwards to the source node following Kirchoff’s Current Law. The current J from l is calculated as:

 k  k  k J I J  ia   ja   ma    = −  + ∑   (2.17) Jig Ijg meM Jmg

With j is the end node of branch l and M is the set of all branches connected down- stream from node j.

In the third step, i.e. the forward sweep, bus voltages are updated using the branch currents obtained from the backward sweep, starting at the root node towards the end nodes. Nodal voltage calculations in the forward sweep are calculated using

24  k  k   k V V z z J  ja   ia   aa ag  la    =   −    (2.18) V jg Vig zag zgg Jlg

This methodology can then be formulated as an optimization algorithm to obtain a solution to the network load ﬂow following convergence. With an objective of minimize the difference between the speciﬁed and calculated load power injections at each bus.

2.5.2 Direct Load Flow Approach

This method performs the load ﬂow analysis for radial distribution system under bal- anced operating condition employing constant power load model. Three important steps are considered in this approach, namely

1. Equivalent current injection.

2. Formulation of BIBC matrix .

3. Formulation of BCBV matrix

Figure 2.7: A Radial distribution system detailng DLF

The method is explained in [25] and using the ﬁgure above, BIBC matrix and BCBV ∗ Py+ jQy matrix are formed . using equation(12), Iy = Vy power injections can be con- verted to equivalent current injections and a set of equations can be written by applying Kirchhoff’s Current Law (KCL) to the distribution network. Formulation of branch

25 currents can be done as a function of the equivalent current injections. For example the branch currents B5, B4,B3 ,B2 and B1can be expressed as :

B5 = I6 (2.19)

B4 = I5 + I6 (2.20)

B3 = I4 + I5 + I6 (2.21)

B2 = I3 + I4 + I5 + I6 (2.22)

B1 = I2 + I3 + I4 + I5 + I6 (2.23)

Moreover, the Bus-Injection to Branch-Current (BIBC) can be obtained as,

      B1 1 1 1 1 1 I2              B2   0 1 1 1 1   I3               B  =  0 0 1 1 1  ∗  I  (2.24)  3     4               B4   0 0 0 1 1   I5        B5 0 0 0 0 1 I6

[B] = [BIBC] ∗ [I] (2.25)

Then the BCBV matrix is accountable for the relationship between the branch currents and the bus voltages. By use of BCBV matrix, the respective variation of the bus voltages which is generated by the variation of the branch currents can be established directly as :

26 V2 = V1 − B1Z12 (2.26)

V3 = V1 − B1Z12 − B2Z23 (2.27)

V4 = V1 − B1Z12 − B2Z23 − B3Z34 (2.28)

V5 = V1 − B1Z12 − B2Z23 − B3Z34 − B4Z45 (2.29)

V6 = V1 − B1Z12 − B2Z23 − B3Z34 − B4Z34 − B5Z56 (2.30)

It can be observed that the bus voltage can be expressed as a function of the branch currents, line parameters and substation voltage. By utilizing same procedure for rest of the buses, BCBV matrix can be derived as,

        V1 V2 Z12 0 0 0 0 B1                  V1   V3   Z12 Z23 0 0 0   B2                   V  −  V  =  Z Z Z 0 0  ∗  B  (2.31)  1   4   12 23 34   3                   V1   V5   Z12 Z23 Z34 Z45 0   B4          V1 V6 Z12 Z23 Z34 Z45 Z56 B5

[∆V ] = [BCBV ] ∗ [B] (2.32)

Algorithm for Direct Load Flow Approach

A concise idea of how bus voltages can be obtained for a radial distribution system is summarised below:

Step 1: Input data.

Step 2: Form the BIBC matrix.

27 Step 3: Form the BCBV matrix.

Step 5: Set iteration count k = 0.

Step 6: Set iteration count k = k + 1.

Step 7: Solve the equations iteratively and update voltages

∗ k Pi + jQi Ii = (2.33) Vi

k h ki [∆V ] = [DLF ] Ii (2.34)

k+1 k If Ii − Ii >tolerance, go to Step-6 else print result.

28 CHAPTER 3

METHODOLOGY

3.1 Estimating electricity demand of a rural area

First of all, it is necessary to determine the daily load profile of the village. There is no variations of the load profile due to season changes because due to the equatorial location there are no distinct summer or winter seasons in Rwanda. Here, the calculation of the load profile of the village is done via self-performed survey that I could perform due to my familiarity with this region. In addition, I used the results from survey forms for households grid connected which have been conducted on other rural villages connected to national grid one year ago. I used parameters such as, the number of households and public utilities, family income, predisposition and readiness to purchase electrical appliances and potential small businesses that can emerge with the availability of electricity. However, a reasonable assumption were used in case where to get the data from site survey is not possible in order to estimate the load curve.

Rongi sector is one of Muhanga district located in remote areas of southern province in Rwanda. From the main grid to the main center of this sector you pass by other four developed centers which are considered as major loads. These ones are localized at different distance about 10 to 15km one another. The distance from the three phase transmission line/main grid to the ﬁrst center is 10km and 60km to the last. Electricity demand in Rongi sector village has been estimated and classiﬁed into residential load, community load, commercial load, and small-scale industrial load.

• Residential load

Load analysis in this particular category was accomplished by using common home appliances in rural communities that are used essentially for lighting, entertainment, and communication purposes. Appliances such as ﬂuorescent tubes, energy saver lamps,

29 TV sets, radio sets, mobile phone charger and iron boxes and refrigerators were considered in determining electricity demand. In order to establish the electricity demand for residential, an assumptions were made: individual residential consists of 10 high income families,40 medium income families and 100 low income families. The detailed daily consumption for selected village and the daily power hourly distribution can be seen in the appendix. The total energy consumption was obtained by summing up the energy consumption of the various income groups.

• Community load

Community load is comprising of dispensary, village executive office, primary school, secondary school, churches, and mosque. The variation in electric consumption between community load is due to the type of electrical equipment available in the build- ings as well as the hours of operation. Load profile for individual consumer was established according to the size of the building and energy requirement by the prospective consumer. Not all the five centers have the same community load. Because some does not have secondary school and church. An assumption made is that middle centers are similar ; first and the last are similar too.

• Commercial Load

Commercial load included grocery, bars, shops, saloon (female), and Barber shop (male).

• Small-scale manufacturing units

Small-scale industrial load included workshop such as carpentry, welding. This category of load comprises also of small processing plants for cassava ﬂour, sorghum ﬂour and rice).

Energy demand evaluation for each category has been conducted, acconding to customer’s appliances consists of power rating of the appliances and their time of operation. Load analysis is obtained from:

30 Load = power ratingo f appliance(W) ∗ Time(h). (3.1)

Table 3.1: Load classiﬁcation and daily energy demand for each category No Consumer Type Number Daily cunsumption in KWh high income families 10 33 domestic load medium income families 40 46 low income families 100 23.3 shops 5 17.7 commercial load bars 2 14.6 administartion posts 2 4.6 primary school 1 4.5 community load churches 1 2,9 Secondary school 1 11 Health center 1 34.2 small-scale industrial load small manufacturing units 3 32.4 total 217.2

For each consumer type a table with types of appliances, their ratings and time by which they are connected was made and they are found on appendix. All the tables were compiled in one table for the two categories of centers and daily electricity demand proﬁle are shown in chapter 4. Based on these, a typical daily load curve with hourly resolution has been derived for this village and it is given in chapter 4.

3.2 Design of radial distribution network

To design a typical rural electriﬁcation system the necessary information to be considered is the load average variation during one day. A basic short line transmission model is developed in Etap. The load demand determined the size of transformer and cable to be used. The SWER system which is used to supply single phase power to the estimated rural load from the main grid MV network (11kV), is considered to be supplied by one isolating transformer which is the most important component in the system. The substation is located at the point of grid extension from main three phase system.SWER system component characteristic and size are described as follow:

31 3.2.1 Isolating transformer

The isolating transformer is providing earth fault protection from the MV network. Without it the return current would ﬂow back to the main three phase system.

This transformer is sized using load data of the village to be electriﬁed. The maximum demand of the whole network is taken into consideration as the size of the isolating transformer. The primary side of this transformer has two phases from the MV (11kV).and grounded on its secondary side with one phase, which supplies power to loads, through distribution transformers.

Modelling Isolating transformer

• Voltage = 6.35 kV line to ground this is at the secondary side as from the primary side is 11kV

• Rating = 160kVA

• Earthing Resistance RE = 1 ohm

• ThereforeIbase = 160/6.35 =25.19A

• Voltage drop over RE = 25.19A x 1 ohm = 25.19

3.2.2 Distribution transformer

This last is sized based the peak/maximum demand of each individual center from the estemated load . The location of the distribution transformer in the network is considered as location of the load also considered as a node or bus. Each distribution transformer has its primary connected between the SWER line and ground, and its secondary provides two low voltage phases of 220 volts and a neutral, which is grounded. The ground terminal on the primary winding can be speciﬁed as fully insulated at an extra cost, which allows the distribution transformer to be upgraded to a three-phase system when required [9]

Modelling of Distribution transformer

32 • HV Line Voltage = 6.35 kV

• LV Line Voltage = 220 V

• Rating = 30 kVA

• Earthing Resistance RE = 1Ω

30 • Therefore Ibase = 6.35 = 4.72A

3.2.3 Earthing system

Since earth is used as the return path, appropriate earthing is one of the most important requirements in SWER. Thus earthing must be carried out carefully in order to provide safe and efﬁcient system operation. At least two load current carrying earths are needed for proper functioning; one on the supply side and the other on the load side [26]. This earthing system must be able to conduct both continuous load current and the occasional network fault currents. Before a model can be created, line pa- rameter calculation theory applicable to SWER feeders needs to be calculated . The resistance of the earth return path and inductive reactance of the SWER feeder circuit are calculated using simpliﬁcations of Carsonís equations detailed in [15].

Resistance and inductance of the ground (Earth) return path

According to Rudenberg in [27], the effective resistance of the ground return current path is given by:

R = (π2 f l) × 10−7ohms

Where

f = supply frequency in Hz

1= length of ground path in meters

It should be noted that the resistance of the earth path is dependent upon the frequency of current and length of path but is independent of earth resistivity. and

The inductance of the ground return path is:

33 s 562.8 ρ L = 2ln ∗ 10−4H/km (3.2) h f

With h= height of conductor above ground in meters

ρ= resistivity of earth in ohm-meter

At power frequency of (50Hz), when h is 10 meters , l is 5meters and earth resistivity (ρ) is 100 ohm-m, the value of inductance (L) is 0.88mH/km and R = 2.46 ∗ 10−7ohms/km

3.2.4 Conductors

Due to the fact that the load current on SWER feeders is relatively low compared to the three-phase system; smaller, cheaper conductors are used which have a reduced load carrying capacity. In installed SWER system commonly used conductor are Steel Cored Galvanized Zinc (SC/GZ) or Steel Cored Aluminium Clad (SC/AC) [26]. Other conductor types such as Aluminium Conductor Steel Reinforced (ACSR) cables are also used in limited applications. Cost of the steel conductor is low and it has signif- icant tensile strength for its weight. Therefore it permits higher spans which results reduced number of poles per kilometer. Disadvantages of steel include higher resistance and corrosion. Corrosion can be minimized by using galvanized conductor. Conductors incorporating aluminium clad steel for reinforcement have lower electrical resistance and provide better protection against corrosion than those using galvanized steel. Aluminium-clad steel has an aluminium cladding with a radial thickness not less than 5% of the overall wire diameter. It has also been noted that angular cross sec- tioned conductors such as are available in the below table e.g. 3/2.75 are less prone to vibration damage than round ones[18]. Properties of some common SWER conductors are listed in Table below.

34 Table 3.2: Properties of common SWER conductors [24]

Calculation of line impedance of SWER

The impedance of a SWER line Using Carson‘s approach as explained previously in chap.2 is calculated as:

Zaa = Z¯aa + Z¯gg − 2Z¯ag (3.3) with

2ha z¯ = R + j4π ∗ 10−4 f ln (3.4) aa a GMRa

2 z¯ = π2 ∗ 10−4 f − j0.0386 ∗ 8π ∗ 10−4 f + j4π ∗ 10−4 ∗ f ln (3.5) gg 5.6198 ∗ 10−3

  −4 ha z¯ag = j2π ∗ 10 ln (3.6) q ρ  f

GMR is a function of physical and magnetic property of the conductor and is constant for any given conductor provided that its property remains unchanged due to current ﬂow. It is considered as the radius the conductor assumed to have no internal ﬂux but with same inductance as actual conductor of radius r.

− 1 GMR = re 4 = 0.7788r (3.7)

35 Having chosen SC/AC conductor

r 10.26 GMR = 0.7788 ∗ = 1.408 (3.8) 3.14

Assuming the height of the conductor above the earth in (m), ha = 10m and ρ the soil resistivity inΩm at average soil which is about 250 Wm

−4 2∗10 z¯aa = 5.75 + j4π ∗ 10 50ln 1.856 = 5.75 + j0.149 z = 2 ∗ −4 ∗ − j . ∗ ∗ −4 ∗ + j ∗ −4 ∗ ln 2 ¯gg π 10 50 0 0386 8π 10 50 4π 10 50 5.6198∗10−3 = 0.0493 + j0.364 ! −4 10 z¯ag = j2π ∗ 10 ln q 250 50 = j0.000941

The conductor impedance is now Zaa = 5.7992 + j0.512Ω/km

Table 3.3: Load location and their maximum demand at peak load

36 3.2.5 Development of the model

With a known distance from one load to another and after calculating the line pareme- ter , esch section or branch of the line is speciﬁed in terms of resistance and reactance respectively. This part documents the construction and assumption made on the model along with an analysis on the performance of the model. While assessing the performance of the model some considerations are used.

• Main system voltages and power factor of the whole network

• Loads and Distances of each Section

• Technical speciﬁcations of each section of the line

1. Resistance per km

2. Reactance per km

• Power loss and Voltage drop for each section

Figure 3.1: Single line diagram of the designed network

37 The SWER network was modelled using ETAP software package by isolating the system from the rest of the network at the isolating transformers and adding an inﬁnite bus to supply the required power at 1 per unit. Loads were assumed to be proportional to the size of the distribution transformers and uniformly at 0.9 power factor. It ignores any voltage drop between the distribution transformer and the customer connection point. In the assumptions within the system designed along with an analysis of the performance of the network; line impedance and length as well as supplying voltage and the load to be supplied at each node were the major consideration for the next step of the load ﬂow calculation.

3.3 Load ﬂow in Radial network and voltage proﬁle determination

Radial distribution is the type of power distribution where the power is delivered from the main branch to the sub branches then it splits out from the sub-branches again. The radial structure implies that there are no loops in the network and each bus is connected to the source via exactly one path. The distribution power flow involved, first of all, finding all the node voltages. From these voltages, it was possible to compute current directly, power flows, system losses and other steady state quantities. the load flow for this study was based on Backward forward sweep method, and was compared by the Direct load flow approach. Details on these two methods are described in chapiter 2 and respective results are presented in chapter 4.

3.3.1 Voltage proﬁle improvement

SWER line is with high impedance, so the voltage drop along the line is a problem, causing poor voltage regulation. To combat that, the proposed capacitor has to be applied to improve the voltage profile of the network.The main benefits of using capacitor are to minimize the power losses and maintain best voltage regulations for all load buses. Normally, for radial distribution systems, all active power demands and losses must be supplied by the source at the main node. Addition of shunt capacitors can compensate some of the reactive power and thus, the losses associated with the reactive power flowing through the branches can be minimized.

38 However, capacitors are normally added according to the local situation of buses to increase the voltage level at that point. Studies have indicated that as much as 13% of total power generated is consumed I2R as losses at the distribution level. Reactive currents account for a portion of these losses. However, the losses produced by reactive currents can be reduced by the installation of shunt capacitors. Effective capacitor installation can also release additional KVAR capacity from distribution apparatus and improve the system voltage proﬁle [27]. With regard to the power losses in the feeders, capacitor installations have demonstrated their effectiveness in reducing the overall current by canceling part of the reactive current supplied by the substation as in [31].

3.3.2 4.3.2 Capacitor Sizing and placement determination

Many techniques have been used in sizing and placement of the capacitor in radial distribution. In this study a maximum power saving method is used, which is based on branch current formula, it consisted of reducing the line losses considering the current ﬂowing in that line. a) Mathematical expression

• Power loss

N 2 QTL = ∑ Ij ∗ Xj (3.9) j=1

WhereQTL is the total reactive power loss in the system,Ij the branch current and

Xj the reactance of the branch. The current Ij ﬂowing has a real and imaginal part component, thus above equation can be rewritten with

Ij = Ia j + Ir j (3.10) as:

N N 2 2 QTL = ∑ Ia j ∗ Xj + ∑ Ir j ∗ Xj (3.11) j=1 j=1

• Bus voltage limit

39 The bus voltage should be within the tolerable range of ±5%.

|Vj min| ≤ |Vj| ≤ |Vj max| (3.12)

• Power ﬂow

The Power ﬂow in each branch must be equal or less than maximum capacity rating in order to respect the thermal capacity limit of the line.

|Ij| ≤ Ij max, j = 1,2,3...... N (3.13)

The total reactive power injected (Q jCap) must be equal to the sum of total reactive power loss (QLoss)and the total load (Q jLoad) .

∑Q jCap = Qloss + ∑Q jLoads (3.14)

If the capacitor current is injected at bus k, the current ﬂow from the source to the bus k is affected by the current injected but beyond the node k, the ﬂow remains the same, and the mathematical expression is given below:

N N N 2 2 2 QTLcapk = ∑ Ia j ∗ Xj + ∑ (Ir j + akICapk) ∗ Xj + ∑ Ir j ∗ Xj (3.15) j=1 j=1 j=k+1

WhereQLTcapk is the total reactive power loss with the capacitor cuurent injected at node k, Ia j and Ir j is the real and imaginary components of the current ﬂowing from the base load ﬂow, Icapk is the capacitor current injected at node k and X j is the line reactance. ak= (sign) tan (cos-1(PF cap)), sign =+1 if capacitor is injecting the reactive power and sign= -1 if it is consuming the reactive power from the network.

Now the saving at each node is calculated by subtracting the total power loss with capacitor from the total based loss without the capacitor as shown below:

Saving(SS) = QTL–QTLcapk (3.16)

40 The maximum value of power saving is found by equating to zero the derivative of the power saving with respect to its equivalent capacitor current injected at node k and considering only the imaginary component of the current ﬂowing for reactive power injection.

k k 2 SS = −2Icapk ∑ Ir j ∗ Xj − ICapk ∑ Xj (3.17) j=1 k=1

δSS k k = 0≈ − 2 ∑ Ir j ∗ Xj − 2ICapk ∑ Xj (3.18) δICapk j=1 k=1

From the derivative equation, the value of the current injected at each node can be evaluated respectively and these computed current values are replaced in the maximum saving equation for all nodes, then the node with higher power saving is identiﬁed and selected as candidate for capacitor placement. The expression for the capacitor current at each node is given below:

k ∑ j=1 Ir j ∗ Xj ICapk = − k (3.19) ∑ j=1 Xj

The optimal size of capacitor at selected node k is calculated using the capacitor current injected at optimal branch and its corresponding voltage magnitude.

k ∑ j−1 Ir j ∗ Xj QCapk = ICapk ∗ |Vk| = |Vk| ∗ k (3.20) ∑ j=1 Xj

This method is combined with the backward and forward sweep method and a load ﬂow is performed after capacitor is placed [17]. below is the ﬂowchart of the algorithm used.

41 Figure 3.2: Flow chart of the BFSM/MPS

42 CHAPTER 4

RESULT AND DISCUSSION

4.1 Introduction

This chapter presents the results obtained during the estimation of the rural load, design and sizing of the SWER system for the estimated load. The results also include system simulation results using power flow algorithm. Using the load data estimated, the researcher was able to size the system that would adequately supply electricity to a that load. The designed system required additional power supply from the capacitor to improve the voltage profile of the network and an algorithm was used to find the optimal size and placement for that capacitor .

4.2 Results from estimation of rural load data

In this section load data was analyzed in readiness for the design. The data from all the estimated types of the load were combined in one table showing the load connected at each hour of the day. hourly electricity usage was summed up to obtain the total daily consumtion. A key aspect was to know accurately for which load to plan for. The highest instantaneous peak power observed was 31.47kW and 29.82kW respectively for the ﬁrst-last centers and middle centers which occurred at 7 PM for an hour. For the purpose of design, a power factor of 0.9 was assumed and kVA power was used for the sizing of the ransformers.

43 Table 4.1: daily hourly demand variation of the whole village

Table 4.2: daily demand for 1st and last centers

44 Table 4.3: Daily hourly demand variation of middle centers

The three middle centers are assumed to be similar in terms of demand and the last and the first center are similar. Based on these table, typical daily load curves with hourly resolution are derived for this village and there are shown in Figure 4.1 and Figure 4.2; after a detailed estimation of all the assumptions made on each category of the mentioned load type (customer type) in the table above. figure 4.3 shows the load profile of the whole network.

45 Figure 4.1: Daily Load proﬁle (Daily Demand variation) for ﬁrst and last centers

Figure 4.2: Daily load curve for middle centers

46 Figure 4.3: Daily load proﬁle of the whole network

The method has taken into account several aspects such as type and number of consumers, equipment and expected time of operation for the appliance. Daily electricity demand for Rongi sector was found to be 1478.35 kW. Peak electrical demand of 152.9 kW occurred at 7 pm while lowest load demand of 23.43 kW occurred between 12 pm to 5am. Lack of recorded data on energy requirement especially in rural areas is among of the challenges of rural electriﬁcation . Thus, availability of this information is important for rural electriﬁcation programs at Rongi sector. Residential premises were found to have the highest electricity consumption.

Average demand = total daily demand of all appliance divided by the time period of 24hours; give:

1478.35 P = (4.1) Av 24

= 61.59kW

Load factor equals the Average demand divided by the total Demand

47 61.599 load f actor = ∗ 100 (4.2) 152.9

= 40.28%

Taking into consideration of a power factor (cosθ) of 0.9, according to the fact that ,the inductive load are not many in the network. The size of transformers and rated current I in amperes of each distribution transformer to serve different loads are estimated by the following equations:

P S = (4.3) Cosθ

S I = (4.4) V

With:

S total load demand in kVA ratings, P total load demand in kW andcosθ power factor, S15is the total load demand for each of the ﬁrst and last centers and adversity factor(DF) of 1.2 was applied to each S power for proper sizing of the transformer.

31.7 S = = 29.38KVA (4.5) 15 (0.9 × 1.2)

S234= the total load demanded for each of the middle centers

29.3 S = = 27.21KVA (4.6) 234 (0.9 × 1.2)

St= the total maximum demand of the whole system

152.9 S = = 145.85KVA (4.7) t (0.9 × 1.2)

48 4.3 Load ﬂow calculation

The SWER load ﬂow formulation is based on the forward/backward sweep method for radial distribution networks with earth return as derived in [12].It is described again in chap.3 In the ﬁrst step, all nodal current injections due to loads, are calculated based on initial voltages. In subsequent iterations, updated voltages are used to calculate the nodal currents.

4.3.1 Results

Table 4.4: Results from BFSMPower ﬂow

By using backward propagation the power of each branch has been calculated and the voltage magnitudes at each node are calculated in forward propagation. The voltage magnitude at different nodes of this system is given by Table 4.4, the minimum voltage of proposed system is 0.8281 p.u at node 3. Real and reactive power losses of 6bus system is obtained . The power loss for each branch is also shown in Table 4.4

The result from the power ﬂow shows a very high power losses especial active power and voltage deviation which needs to be improved. The next section explains more on the capacitor sizing and placement which needs to improve the voltage proﬁle.

49 4.3.2 Results with capacitor in the network

With the result presented in the above subsection a capacitor is placed to improve the voltage and minimize the losses by Maximum power saving method. The proposed method validity and performance was tested with the 6bus SWER RDN developed previously, the study was done under Matlab R2013a and the based load ﬂow losses and the voltage using the BFSM load ﬂow analysis method are shown in table below

Table 4.5: Base case results of the test system

Buses 6 Total load 141.57KVA Toatl reactive power loss 8.4456KVar Total active power loss 90KW Maximum Voltage 1.0pu Minimum Voltage 0.8281pu

Applying the maximum power saving formula in the base load flow results above, the node 3 was identified for the optimal placement of the capacitor with the saving of 6.24kVar. It was then selected for the maximum reactive power loss reduction and the optimal size of the capacitor is calculated. Another load flow is carried out with the capacitor injected to get the new system line loss and the new voltage profile. The results are presented in the table below

Table 4.6: Results with capacitor in the system

Buses 6 Optimum sitting node 3 optimum size 55.218KVar Total reactive power loss with capacitor 4.KVar Total active power loss with capacitor 35KW Minimum Voltage 0.952pu

50 Figure 4.4: Voltage proﬁle with and without capacitor

The above ﬁgure shows that system is much healthier with the approach used. As before and after placement of capacitor, there is a lot of change in system values(loss and voltage). ﬁgure 4.5 depicts the detailed change in the system.

51 Figure 4.5: Reactive power losses with and without capacitor

With regard to the power losses in the feeders, capacitor installations have demonstrated their effectiveness in reducing the overall current by canceling part of the reactive current supplied by the substation.In addition to powerloss reduction, the voltage proﬁle is improved.

4.4 SWER cost structure

With lower consumer density in rural areas, the cost of delivering power to a new customer has increased. Also these new consumers have low income and purchase less electricity. With high construction costs per customer, low revenues and other associated costs while managing rural systems, electric utilities have found it even more difﬁcult to meet demand for electricity in rural areas. SWER systems are used where load density is extremely low and its main beneﬁt is reduced construction cost. Various sources [31] report that savings of between 25% and 40% can be made over standard three-phase construction costs. This makes SWER a valuable tool in providing electricity supply to remote areas for the lowest possible cost. Developing nations are placing a strong reliance on single-phase distribution systems such as SWER to provide economic access to electricity for remote communities. The following table updates the cost structure for SWER so that indicative costs may be derived [31].

52 Cost of 12.7 kV, 1-phase SWER Overhead Line, Cost per kilometre

Table 4.7: Cost of 12.7 kV, 1-phase SWER Overhead Line Material Quantity Unit price Total price USD USD Pole,12metre 6 119 716 Concret m3,foundation 3 35 105 Line post insulators 6 10.89 65 Tension insulators 4 17.05 68 Tension insulator hardware 4 2.77 11 Braces and Bolts 6 4.40 26 Stays 2 38 75 Misc 1 200 200 1267 Installation Line route clearing 1 100 100 Pole top assembly 6 10 60 Polesetting 6 10 60 Foundation, Anchors 1 50 50 Guy assembly 1 50 50 Grounding Installation 1 44 44 Conductor stringing 1 60 60 Transportation and Tools 1 215 215 639 Other Design, survey,Stacking 1 376 376 Conductor SC/AC 1150 0.40 460 Total cost 2742

It is based on 6 poles per kilometre, including 2 strain poles and one angle pole. The labour costs, overhead costs, taxes and duties . The construction is based on 24 kV material so that there is a future upgrade path to two-wire 22 kV single phase distribution, with the addition of intermediary poles cross arms, insulators and a second conductor.

Cost of 12.7/0.23 kV, SWER, 1 Pole, Distribution Transformer

It is based on the cost of 25 kVA SWER distribution transformer. The labour costs, overhead costs, taxes and duties may be greater depending on the country situation.

53 Table 4.8: Cost of 12.7/0.23 kV, SWER, 1 Pole, Distribution Transformer Material Quantity Unit price Total price Pole12 metre 1 119 119 Concrete m3 0.5 35 18 Crossarms 2 9 18 Line post insulator 1 10.89 11 Fuse cut out 1 48.4 48 HV surge arresters 1 55.11 55 LV insulator 4 1 4 LV surge arresters 2 6 12 Earthing 1 10.23 10 HV and LV connectors 1 0.75 1 Braces and Bolts 4 3.124 12 Stay 1 37.51 38 HV conductor m 10 3.41 34 LV conductor 12 1.2 14 Earthing conductor m 20 0.78 16 LV Fuse switch 2 31 62 Misc 1 1 50 522 Installation Pole top assembly 1 10 10 Pole settings 1 10 10 Foundation,Anchors 1 50 50 Ground installation 1 44 44 Transportation and tools 1 538 538 652 Other Design ,Survey,Staking 0.5 376 188 188 Transformer Single phase 25kVA 1 822 822 Total cost 2184

Cost of 22/12.7, SWER, 2 Pole, Isolating Transformer

It is based on the cost of a two pole isolating transformer. The labour costs, overhead costs, taxes and duties may differ depending on the country situation.

54 Table 4.9: Cost of 22/12.7, SWER, 2 Pole, Isolating Transformer Material Quantity Unit price Total price Pole12 metre 2 119 238 Concrete m3 1 35 35 Crossarms 2 9 18 Line post insulators 1 10.89 11 Tension insulators 6 59 354 Fuse cut out 3 48.4 145 HV surge arresters 3 55.11 165 LV insulator 6 1 6 LV surge arresters 4 6 24 Earthing 3 10.23 31 HV and LV connectors 1 0.75 1 Braces and Bolts 4 3.124 12 Steel work 1 134 134 Stay 1 37.51 38 HV conductor m 10 3.41 34 LV conductor 12 2.15 26 Earthing conductor m 80 0.78 62 LV Fuse switch 2 31 62 Misc 1 1 50 1446 Installation Pole top assembly 2 10 20 Pole settings 2 10 20 Foundation,Anchors 1 50 50 Ground installation 1 44 44 Transportation and tools 1 538 538 672 Other Design ,Survey,Staking 0.5 376 188 188 Transformer Transformer 160kVA 1 2009 2009 Total cost 4315

55 CHAPTER 5

CONCLUSIONS AND RECOMMENDATIONS

5.1 Conclusion

The alternative grid based rural electriﬁcation technology dealt with in this study, re- ferred to as SWER, allows electricity to be distributed at MV to community located far away from the HV transmission line or main grid. In rural areas houses are scattered in a very large area. Using a single large distribution transformer may cause higher loss due to very long low voltage lines from transformer to rural houses. Rather than going for a single large transformer one can use several single-phase transformers for several houses which are in closer proximity. but to avoid this, this study implemented SWER in a densely settled rural area. This will be much easier in Rwanda where people live in settlement to facilitate distribution of ﬁrst need service. By this way huge losses in lengthy low voltage lines are minimized.

An optimal placement of capacitor in the distribution system, using backward and forward sweep method, based on the analytical maximum power saving technique for the maximum power loss reduction and voltage proﬁle improvement were used in this study. The validity and performance the proposed method was tested on the 6bus system developed for this study.The maximum power saving method had a goal of minimizing the power losses of the network and improve the voltage proﬁle by determining the optimal placement and size of the capacitor to be used in the network. The place and size of the capacitor were determined. Both active and reactive power losses were reduced and the voltage of the network were improved.

5.2 Recomendations

Shunt capacitors are used in raising the line voltage. Unfortunately, a SWER line has a highresistance relative to the line inductance; therefore,the resistive line loss will

56 remain high in any case. This reseach recommand further study on loss reduction by applying a an active power source.

Further reseach are also recomended on cost beneﬁt analysis of the reactive power compensation in SWER.

In the context of rural electrification for developing countries, SWER is still an ideal technology to employ in the initial stages of electrification. The important issue is to plan the network with an upgrade path from SWER to two-wire and then three- wire three-phase distribution. The upgrade path includes the use of common materials and equipment that remain in service as demand develops and the electrification ratio increases.

57 REFERENCES

[1] R. Karhammar, A. Sanghvi, E. Fernstrom, M. Aissa, J. Arthur, J. Tul-loch, I. Davies, S. Bergman, and S. Mathur, “Sub-saharan africa: Introducing low cost methods in electricity distribution networks,” Energy Sector Management Assis- tance Program (ESMAP), Tech. Rep. 104/06,October 2006.

[2] https://www.usaid.gov/sites/default/ﬁles/documents/1860/Rwanda Country Fact Sheet-2016. FINAL.pdf

[3] ] P. J. Wolfs. Capacity improvements for rural single wire earth return systems. In Proc. Power Engineering Conference, 2005.IPEC 2005. The 7th International, pages 1–8, 29 2005-Dec. 2 2005.

[4] G. Bakkabulindi, I. P. Da Silva, L. Soder, and M. Amelin. Rural electriﬁcation practicalities of using single wire earth return as a low cost method for grid extension: the case of ntenjeru, Uganda. In Proc. Int. Conf. Energy and Sustainability, Newark, DE, August 9–12, 2009.

[5] N. Hosseinzadeh and J. Rattray. Economics of upgrading SWER distribution systems. In Proc. Australian Universities Power Engineering Conf, 2008.

[6] B.S. Samra, “An analysis of single wire earth return (SWER) system for rural electriﬁcation” (1972). Masters Theses. Paper 5062

[7] Power to the People Describes use of SWER in the rural electriﬁcation of Mozambique. Transmission & Distribution World, 2009. Accessed 2011-8-10

[8] C. W. Holland. Single wire earth return for remote rural distribution, reducing costs and improving reliability.Technical report.

[9] T. R. Brooking, N. Janse van Rensburg, and R. Fourie. The improved utilization of existing rural networks with the use of intermediate voltage and single wire

58 earth return systems. In AFRICON ’92 Proceedings. 3rd AFRICON Conference, pages 228 –234, September 1992.

[10] Hesamzadeh, MR, Hosseinzadeh, N & Wolfs, PJ 2008, ’Design and Study of a Switched Reactor for Central Queensland SWER System’, 43rd International Universities Power Engineering Conference.

[11] Kashem M, Ledwich G. Distributed generation as voltage support for single wire earth return systems. IEEE Trans Power Deliv 2004; 19(3):1002–11.

[12] P. J. Wolfs, N. Hosseinzadeh, S. T. Senini, “Capacity Enhancement for Aging Single Wire Earth Return Distribution Systems”, IEEE Power Engineering Soci- ety Annual General Meeting, Tampa Florida, 24-28 June 2007.

[13] Mayer, J and Hosseinzadeh, N and Wolfs, Peter. Modelling of Voltage regulation in SWER Systems using PSCAD/EMTDC, in Kalam, A. Australasian Uni- versities Power Engineering Conference AUPEC. Melbourne Victoria: Victoria University. Dec 10 2006.

[14] N. Madjissembaye and et al. Optimal siting and sizing of single SPV system in radial distribution network for loss reduction based on maximum power saving technique. International Journal of Engineering Modern Research, IJMER. 6: 7, 30-37, July 2016.

[15] Mahanthege Sampath Dayarathne. Technological and cost effective selection procedure for rural electriﬁcation systems. Masters thesis. Department of Elec- trical Engineering, University of Moratuwa, Sri Lanka, June 2015.

[16] Odax Ugirimbabazi. Analysis of Power System Options for Rural Electriﬁcation in Rwanda. Masters Thesis, Faculty of Engineering and Science, University of Agder Grimstad, 25 May 2015.

[17] Ng. HN and Salama MMA and Chikhani AY. Classiﬁcation of capacitor allocation techniques Analysis. IEEE Trans Power Del. 15:1, 2000.

59 [18] W.H.Kersting and W.H. Distribution System Modeling and Analysis. CRC Press. 2002.

[19] J. A. M. Rupa and S. Ganesh. Power Flow Analysis for Radial Distribution Sys- tem Using Backward, Forward Sweep Method. International Journal of Elec- trical, Computer, Energetic, Electronic and Communication Engineering.Vol:8, No:10, 2014

[20] Berg and R.Hawkins and E.S and Plelines and W.W. Mechanized Calculation of Unbalanced Load Flow on Radial Distribution Circuits. IEEE Trans. on Power Apparatus and Systems. 4: 86. 415-421. 1967.

[21] Michael McAsey and Libi Mou. Convergence of the Forward Backward Sweep Method in Optimal Control. IL 61625.

[22] Bompard and E. et al. Convergence of the backward, forward sweep method for the load-ﬂow analysis of radial distribution systems. International journal of electrical power and energy systems. 22:7. 521-530. 2000.

[23] Vijay Laxmi Mishra and Manish Kumar Madhav and Dr. R. S. Bajpai. A compar- ative Analysis of Distribution system load ﬂow for 33bus system. International conference of Innovative trends in Science, Engineering and Managment. 27 may 2016.

[24] R. Nobbs. Development of advanced SWER models for the Ergon Energy network. dissertation for bachelor of power engineering.University of Southern Queensland Faculty of Engineering and Surveying, 2012.

[25] P. Simonds. Designing and testing low-resistance grounding systemsk. Power Engineering Review, IEEE. 20:10. 19 - 21, Oct 2000.

[26] A. A. E. Shammah and A. M. Azmy and P. Simonds and A. A. El-Ela. Optimal Sitting and Sizing of Capacitor Banks in Distribution Networks using Heuristic Algorithms. J. Electrical Systems. 9:1. 1-12. 2013.

60 [27] Mohamed M. Hamada and Mohamed A.A.Wahab and Abou-ashema M.El-Sayed and Husam A. Ramadan. A New Approach for Capacitor Allocation in Radial Distribution Feeders. The Online Journal on Electronics and Electrical Engineer- ing (OJEEE). 1:1. 24-28.

[28] National Rural Electric Cooperative Association (NRECA) International Ltd. Reducing the Cost of Grid Extension for Rural Electriﬁcation. Report on Joint UNDP/World Bank Energy Sector Management Assistance Programme (ESMAP)). February 2000.

[29] N. Hosseinzade. J.E. Mayer, P.J. Wolfs Rural Single Wire Earth Return distribution networks – Associated problems and cost-effective solutions. 13 August 2010 Elsevier Electrical Power and Energy Systems 33 (2011) 159–170

[30] Peter Wolfs, Steven Senini, Nasser Hossein-Zadeh, Dawit Seyoum, Antho- nyLoveday,Jon Turner, “Reactor based Voltage Regulator for Single Wire Earth ReturnSystems”, Australian Universities Power Engineering Conference, 11 th -13 th December, Melbourne, Australia, 2006.

[31] A.Hemasekhar , Chevireddy Harika ,Voltage Proﬁle Improvement by Capacitor Placement and Control in Unbalanced Distribution Systems Using Differential Evolution Algorithm, International journal of advanced in Electrical, Electronics and instrumentation engineering. Vol. 3 , Issue 8 , August 2014

[32] A. Augugliaro, L. Dusonchet,” A backward sweep method for power ﬂow solution in distribution networks” Electrical Power and Energy Systems 32 (2010) 271–280.

[33] Nabeel I. Tawalbeh , A.M. Al-Fahed Nuseirat, “ Peak and daily proﬁle estimation for substations in rural networks”Electrical Power and Energy Systems 30 (2008) 202–206

[34] Nazarko J and Tawalbeh N. Statistical properties of the diversity factor and its use

61 in load estimation. I Ogolnopolskie Sympozjum Naukowe, Bialystok – Suprasl, 18–19 Pazdziernika; 1995.

[35] Isaka J. Mwakitalima, Cecil K. King’ondu, Electricity Demand Evaluation for Rural Electriﬁcation. International Journal of Engineering Research & Technol- ogy (IJERT) ISSN: 2278-0181 IJERTV4IS060726 Vol. 4 Issue 06, June-2015

[36] Falcao DM, Henriques HO. Load estimation in radial distribution systems using neural networks and fuzzy set techniques, panel summary. In: Proceedings of the 2001 IEEE power engineering society summer meeting, Vancouver, Canada; 2001

62 APPENDIX A

DETAILED DAILY CONSUMPTION FOR SELECTED VILLAGE

Table A.1: Domestic purposes/ Residential demand

63 Table A.2: Commercial load

64 Table A.3: Community load

65 Table A.4: Secondary school

Table A.5: Small manufacturing units