Rural Electrification In Uganda

Master thesis

Rural electrification in Uganda Powering the Masindi district Elektrifisering av landsbyer i Uganda: Masindi District

Stud. techn. Britt-Mari Langåsen Master Program Energy and the Environment Faculty of Information Technology, Mathematics and Electrical Engineering Department of Energy and Process Engineering Spring 2004

I

Rapportnummer EPT-H-2004-34 Gradering

The Norwegian University of Science and Technology POSTADRESSE NTNU INSTITUTT FOR ENERGI OG PROSESSTEKNIKK Kolbjørn Hejes vei 1A N-7491 Trondheim - NTNU

TELEFONER Sentralbord NTNU: Instituttkontor: Vannkraftlaboratoriet:

TELEFAX 73 59 40 00 73 59 27 00 73 59 38 57

Instituttkontor: Vannkraftlaboratoriet:

73 59 83 90 73 59 38 54

Rapportens tittel

Dato

Rural electrification in Uganda: Powering the Masindi District

23.06.04 Antall sider og bilag

Elektrifisering av landsbyer i Uganda: Masindi District Saksbehandler / forfatter

110p incl. 22 p attachemnts Ansv. sign.

Britt-Mari Langåsen Avdeling

Prosjektnummer

Energi og Prosessteknikk ISBN nr.

Prisgruppe

Oppdragsgiver

Oppdragsgivers ref.

Ekstrakt This master thesis is about electrification of a rural district in Uganda. The aim of the thesis is to: ¾ investigate the electricity needs and power characteristics in a rural district in Uganda ¾ to identify and compare different technology options ¾ to recommend how electrification should proceed Three different types of trading centres are investigated, and they differ in the number of inhabitants and load characteristics. Load profiles for the trading centres is made and put into a simulations program called HOMER. This is an optimization program that in this study use solar systems, diesel generation and grid extension as means of supplying a rural area with power. Several situations are simulated; how low the price of solar systems must go to be competitive or how high the diesel price must be to make solar systems able to compete with diesel generation. Different loads are also simulated to show how this affects the technology options. A general discussion about the applicability of grid extension and protection schemes for the network is given. It is found that for a large trading centre, grid extension is the best alternative. For a small and medium community, diesel generation or solar systems are best, depending on diesel price and distance from the main grid.

Stikkord på norsk

Indexing Terms English

Gruppe 1

Elektrifisering av landsbygd

Rural electrification

Gruppe 2

U-land

Developing countries

Egenvalgte stikkord

Teknologi løsninger

Technology options

II

Norges teknisknaturvitenskapelige universitet NTNU

Institutt for energi- og processteknikk

H-2004HOVEDOPPGAVE for Stud.techn. Britt Mari Langåsen Våren 2004 Elektrifisering av landsbyer i Uganda: Masindi District Rural Electrification in Uganda: Powering the Masindi District Background

Like many developing countries especially in Africa, few people in rural areas and towns of Uganda enjoy access to electricity. Since electricity is seen as an important prerequisite for development, Uganda initiated the “Energy for Rural Transformation” program with the support of the World Bank and a number of donor countries. This program aims at increasing the rural electrification rate from 1% today to 10% by 2010. Different options exist to meet this aim. Grid extension, the building of mini-grids based on local generation, and off-grid generation (e.g. PV solar systems) are potential alternative, whose attractiveness depends on various different factors. Aim The aim of this thesis is investigate rural electrification of the Masindi district in Uganda, to identify and compare different options and to recommend how electrification could proceed. The analysis should focus on so-called “trading centers”, small towns, as opposed to the evenly spread farming population. The characteristics of the electric loads should be taken into account, as well as the investment costs and operating expenses of different technical solutions. The analysis should include following elements: ♦ Description of the situation in the trading centers, their electricity needs, current supply, and power characteristics. • Selection and description of a proper analysis framework and modeling tool. • Description of options for rural electrification. • Description/analysis of the local grid and its ability to deal with local power supply. Specification of technical upgrades needed. • Analysis of technical and cost elements of different, “typical” situations and recommendation of specific solutions. Specification of investment needs and power characteristics. • Discussion of the usefulness, applicability and limits of grid extensions and different types of distributed generation.

III

Senest 14 dager etter utlevering av oppgaven skal kandidaten levere/sende instituttet en detaljert fremdrift- og evt. forsøksplan for oppgaven til evaluering og evt. diskusjon med faglig ansvarlig/ veiledere. Detaljer ved evt. utførelse av dataprogrammer skal avtales nærmere i samråd med faglig ansvarlig. Besvarelsen redigeres mest mulig som en forskningsrapport med et sammendrag både på norsk og engelsk, konklusjon, litteraturliste, innholdsfortegnelse etc. Ved utarbeidelsen av teksten skal kandidaten legge vekt på å gjøre teksten oversiktlig og velskrevet. Med henblikk på lesning av besvarelsen er det viktig at de nødvendige henvisninger for korresponderende steder i tekst, tabeller og figurer anføres på begge steder. Ved bedømmelsen legges det stor vekt på at resultatene er grundig bearbeidet, at de oppstilles tabellarisk og/eller grafisk på en oversiktlig måte, og at de er diskutert utførlig. Alle benyttede kilder, også muntlige opplysninger, skal oppgis på fullstendig måte. (For tidsskrifter og bøker oppgis forfatter, tittel, årgang, sidetall og evt. figurnummer.) Kandidaten skal rette seg etter de reglementer og retningslinjer som gjelder ved alle (andre) fagmiljøer som kandidaten har kontakt med gjennom sin utførelse av oppgaven, samt etter eventuelle pålegg fra Institutt for energi- og prosessteknikk. I henhold til Reglement for sivilarkitekt- og sivilingeniøreksamen ved NTNU § 8, forbeholder Instituttet seg retten til å benytte alle resultater i undervisnings- og forskningsformål, samt til publikasjoner. Ett -1 komplett eksemplar av originalbesvarelsen av oppgaven skal innleveres til samme adressat som den ble utlevert. (Det skal medfølge et konsentrert sammendrag på maks. en maskinskrevet side med dobbel linjeavstand med forfatternavn og oppgavetittel for evt. referering i tidsskrifter). Til Instituttet innleveres to - 2 komplette, kopier av besvarelsen. Ytterligere kopier til evt. medveiledere/oppgavegivere skal avtales med, og evt. leveres direkte til, de respektive. Til instituttet innleveres også en komplett kopi (inkl. konsentrerte sammendrag) på CDROM i Word-format eller tilsvarende. Institutt for energi og prosessteknikk, xx.xx.2004

_________________ Ingvald Strømmen Instituttleder ansvarlig/veileder

Kontaktperson(er)/medveileder(e): Olav Bjarte Fosso, Elkraft

______________________ Edgar Hertwich Faglig

IV

Preface This master thesis is written at the Department of Energy and Process Engineering, which is a part of the Faculty of Engineering, Science and Technology at the Norwegian University of Science and Technology, NTNU. It concludes my Master of Science degree in Energy and the Environment and was written in the period January 2004 to July 2004.

I would like to thank Professor Edgar Hertwich and Professor Olav B. Fosso for their guidance during the process of writing this report. I would also like to thank Professor Da Silva at Makerere University and all the people at UEDCL that helped me in finding relevant information. I would also like to thank Rachel Arinda and Richard Okou that helped me during my stay in Uganda and took me into their homes and also thanks to Lars-Petter Bingh for his support and friendly company during our stay in Kampala. Finally I would like to thank Ingvill Horgøien, Idun Skorpa Melvær and Tore Bjølseth for their support and inspiration during this work.

Trondheim, June 23, 2004

Britt-Mari Langåsen

Summary

V

Summary In Uganda, the electrification rate is very low and rural electrification is a part of the government’s energy policy, introduced in the “New Electricity Act” from 1999 and continued in the “Rural Electrification Strategy and Plan” from 2001. The aim of the thesis is to: ¾

investigate the electricity needs and power characteristics in a rural district in Uganda

¾

to identify and compare different technology options

¾

to recommend how electrification should proceed

This has been done by choosing the modelling tool HOMER which is an optimization program for rural electrification with both on-grid and off-grid technologies. Three different “types” of communities have been investigated in this study, three trading centres that have been divided into small, medium and large. These centres represent different sizes (about 2000, 10’000 and 20’000 inhabitants) and thereby different power characteristics. “Typical” load profiles for these centres have been used in the simulations tool with the technology options photovoltaic, diesel generation and grid extension to supply power. Different parameters have been changed during the simulations to see what influences the choice of technology for supplying a rural area with power. Parameters that have been changed are the price of solar systems and diesel prices. Different loads have also been simulated.

The main conclusions that can be drawn from these simulations are: Grid extension is most preferable when there is a relatively large load with a high load factor or if the trading centre is situated close to the existing main grid. For the large trading centre, this is the best option. Diesel generation is the cheapest way of supplying power, but as the diesel prices rise, then photovoltaic power becomes competing. Photovoltaic power systems are best suited for small, dispersed loads in trading centres far away from the main grid. Solar systems are within the energy policy for rural electrification by renewable energy sources promoted by the government. It can also be a supplement in medium and large trading centres. The cost for supplying a rural trading centre is according to this study between US$40’000 and US$7’400’000.

Britt-Mari Langåsen

Spring 2004

Summary

VI

Further studies that are recommended is to investigate a given trading centre and look at the loads on a household level and se what loads should be covered by for example solar home systems and which should be connected to a grid. Another option for further studies is to se what financing mechanisms are best to encourage rural electrification.

Britt-Mari Langåsen

Spring 2004

Sammandrag

VII

Sammandrag I Uganda har en väldigt liten del av befolkningen tillgång på electricitet och elektrifiseing av landsbygden är en del a regeringens energi policy och introducerades i samband med the New Electricity Act 1999 och vidareutvecklades i the Rural Electrification Study and Plan 2001. Målet med den här studien är att: ¾

undersöka elektrisitets behoven och last karakteristiken i ett landbygdsområde i Uganda

¾

att identifiera och jämföra olika teknologi alternativ

¾

att rekommendera hur elektrifiering ska fortskrida

Detta har gjorts genom att välja modelleringprogrammet HOMER som är ett optimiserigs program for elektrifisering av landsbygden med både nät tillknutna och icke nät tillknutna teknologier.

Tre olika typer av samhällen har undersökts i denna studie, tre handels center som har indelats i litet, medium och stort. Dessa center representerar olika storlekar (runt 2000, 10’000 och 20’000 invånare) och därmed olika effekt och energi behov. ”Typiska” last profiler för dessa center har använts i simuleringsverktyget tillsammans med olika tekniska løsningar, solceller, diesel kraftproduktion och nätburen electricitet, för att leverera energi. Olika variabler har ändrats i samband med simuleringarna för att se vad som påverkar valet av teknologi för att förse ett landsbygdsområde med energi. Variabler som har ändrats är pris på solceller och diesel. Det har ockå gjorts simuleingar med olika laster.

De huvudsakliga slutledningarna som kan dras från dessa simuleringar är: Utbyggnad av kraftnätet är att föredra när det finns en relativt stor last med hög last faktor eller om handels centrat ligger nära det existerande kraftnätet. For det stora handels centret är detta den bästa lösningen. Kraftproduktion från diesel generatorer är det billigaste sättet att producera energi, men om diesel priserna stiger så blir solsystem konkurrens kraftiga. Solceller är bäst lämpade för små, utspridda laster i handels centra som ligger långt från kraftnätet. Solsystem är en del av regeringens energi policy for elektrifiering av

Britt-Mari Langåsen

Spring 2004

Sammandrag

VIII

landsbygden genom förnybara energi källor. De kan också vara ett supplement i medium och stora handels center. Kostnaden for att förse ett handels senter på landsbygden med energi kostar enligt den här studien mellan US$40’000 och US$7’400’000.

Vidare studier som är rekommenderade är att undersöka ett givet handels center och se på de olika lasterna och se vilka som borde täckas av till exempel solceller och vilka som borde vara tilknutna kraftnätet. Ett annat fält att undersöka vidare är vilka finansierings mekanismer som verkar bäst för att främja elektrifiering av landsbygden.

Britt-Mari Langåsen

Spring 2004

Abbreviations

IX

Abbreviations DG

Distributed Generation

UETCL

Uganda Electricity Transmission Company Limited

UEDCL

Uganda Electricity Distribution Company Limited

kV

kilo Volt

kW

kilo Watt

kVA

Apparent power

MW

Mega Watt

DC

Direct Current

AC

Alternating Current

PV

Photovoltaic

SHS

Solar Home System

GTZ

Deutsche Gesellschaft für Technische Zusammenarbiet

MUK

Makerere University Kampala

Exchange rate: US$1 = 2000 Ush

Britt-Mari Langåsen

Spring 2004

Table of contents

X

Table of contents 1

Introduction _________________________________________________ 1 1.1

Structure of the report __________________________________________ 2

2

Background _________________________________________________ 3

3

Methodology ________________________________________________ 6 3.1

Preparations and progression of the work _________________________ 6

3.2

Field studies __________________________________________________ 7

3.3

HOMER ______________________________________________________ 9

3.4

Sensitivity analysis ___________________________________________ 13

4

Power generation and technology options ______________________ 14 4.1

Power generation _____________________________________________ 14

4.2

Alternative energy sources _____________________________________ 16

4.3

Rural electrification technology options __________________________ 17

4.4

Conclusion __________________________________________________ 22

5

Analysis of trading centres ___________________________________ 23 5.1

Masindi district _______________________________________________ 23

5.2

Socio-economic characteristics of Masindi________________________ 24

5.3

Load assessment _____________________________________________ 25

6

Basis and optimal solutions __________________________________ 34 6.1

Simulation basis______________________________________________ 34

6.2

Small trading centre ___________________________________________ 37

6.3

Medium trading centre_________________________________________ 48

6.4

Large trading centre___________________________________________ 56

7

Grid extension ______________________________________________ 65 7.1

Applicability _________________________________________________ 65

7.2

Protection schemes ___________________________________________ 67

8

Discussion _________________________________________________ 70

9

Conclusion_________________________________________________ 73

10

Literature references_______________________________________ 76

11

Appendix ________________________________________________ 78

11.1

Interviews ___________________________________________________ 78

11.2

HOMER input variables ________________________________________ 87

11.3

Attached CD with HOMER files and data _________________________ 110

Britt-Mari Langåsen

Spring 2004

Table of contents

XI

List of tables Table 4-1 Suggested places construction of new hydro power generation .................................. 15 Table 4-2 Summary of the technology option................................................................................ 22 Table 5-1 Energy consumption for rural households in Uganda ................................................... 26 Table 5-2 Time usage ( in hours) of electrical appliances in rural households ............................. 26 Table 5-3 Present power need in the Masindi region. Based on data from UEDCL..................... 27 Table 6-1 Optimal solutions for Biizi trading centre, constrained demand .................................... 38 Table 6-2 Optimal solutions for Biizi trading centre with an unconstrained demand .................... 38 Table 6-3 Optimal results for variations in diesel price.................................................................. 43 Table 6-4 Optimal results for variations in diesel price.................................................................. 44 Table 6-5 Solutions when supplying Biizi with PV or diesel .......................................................... 45 Table 6-6 Solutions when powering with PV or diesel .................................................................. 46 Table 6-7 Optimal solutions for Mutunda with a constrained demand .......................................... 48 Table 6-8 Optimal solutions for Mutunda with an unconstrained demand .................................... 49 Table 6-9 Optimal results for variations in diesel price.................................................................. 52 Table 6-10 Optimal results for variations in diesel price ............................................................... 53 Table 6-11 Solutions when powering with PV and diesel ............................................................. 54 Table 6-12 Solutions when powering with PV and diesel ............................................................. 54 Table 6-13 Optimal solutions for Kiryandongo, constrained demand ........................................... 56 Table 6-14 Optimal solutions for Kiryandongo, unconstrained demand ....................................... 57 Table 6-15 Optimal results for variations in diesel price ............................................................... 60 Table 6-16 Optimal results for variations in diesel price ............................................................... 61 Table 6-17 Solutions when powering with PV and diesel ............................................................. 61 Table 6-18 Solutions when powering with PV and diesel ............................................................. 62

Britt-Mari Langåsen

Spring 2004

1 Introduction

1

1 Introduction In Uganda and the Masindi district, like in most developing countries, there is an existing shortage of power. This prohibits economical development. Uganda has in its program Energy for Rural Transformation (ERT) [19] and through the Rural Electrification Strategy and Plan from 2001 [18] stated that by the year 2010, 10% of the population should have access to electricity to enhance their standard of living, as compared to 1% as of now. After the restructuring of the energy sector, independent power producers and distributed generation is being supported and encouraged. In the Rural Electrification Strategy and Plan it is said that “The primary objective of the RF Strategy is to reduce inequalities in national access to electricity and the associated opportunities for increased social welfare, education, health and income generating opportunites.” [18] This is an enormous challenge which Uganda and its government have to face. The solution to this problem is complex where several technology alternatives are available and different technologies may be optimal for different areas. An essential question is what electrification of rural areas will cost and where the financing will come from. There is also a lack of awareness of renewable energy technologies that prohibits a wider use of for example solar systems. In light of this aim, this study might be of help to find out what could be the optimal way of powering a rural community.

This project about rural electrification was initiated owing to collaboration between the Norwegian University of Science and Technology and Makerere University. This collaboration started in 2002. The aim of the thesis is to: ¾

investigate the electricity needs and power characteristics in a rural district in Uganda

¾

to identify and compare different technology options

¾

to recommend how electrification should proceed

The rural area that has been the main object of this study is Masindi district which lies in the north-west of Uganda. The study will try to find out what is the optimal way to supply an area with electricity, through photovoltaic, diesel generation, grid extension or a combination of these.

Britt-Mari Langåsen

Spring 2004

1 Introduction

2

The study included six weeks of research in Kampala, Uganda. There I cooperated mainly with Makerere University, GTZ and UEDCL, but I was also in contact with smaller firms and employees of MEMD. With help from the district population office in Masindi, 44 major communities, trading centres, that are presently without access to grid electricity and that are important for the development of the district was found. Using UEDCL standards for calculating loads and an assessment of economical development in rural areas, the respective expected loads and power characteristics for the trading centres were found. The loads were then, together with data concerning prices etc, used in a simulation program called HOMER to do an optimisation and a sensitivity analysis by changing different variables such as price, distance and load.

1.1 Structure of the report In the second chapter there will be given a background to the problem that has been investigated. A methodology chapter about field studies and the chosen simulation tools will be given in chapter four. A description of the present power situation in the country and the region together with different technology options for rural electrification will be presented in chapter five and also how the government in Uganda plans the future expansion of their power generation. Following in chapter six, the chosen trading centres will be analysed and a load assessment for them will be performed. Different power producing options for the district will be explored using an optimization program called HOMER and a sensitivity analysis will be done to se what impacts changing different variables may have on the final result. These results are presented in chapter seven. A general discussion about grid extension will be given in chapter eight. Finally, in chapter nine and ten, a discussion and conclusion will be drawn from the results presented earlier. In the appendix interviews and meetings have been included and more simulation results are presented. A CD with the program HOMER and simulation files are attached in the end of this report.

Britt-Mari Langåsen

Spring 2004

2 Background

3

2 Background

Picture 2-1 Map over Masindi district

Getting access to electricity is an important part of enhancing the living standard for the rural population in Uganda. Presently only a few percent of the people in rural areas have access to electricity. Since these areas usually have low economic activity as well, they are not high prioritised for extension of the main grid. It has to be taken into account that not all areas are favourable for grid extension because of a scattered population with a low load and people might be mitigating from rural area into towns. Still, several trading centres might benefit from electrification since they are likely to stay inhabited.

Different factors that affect the electrification rate are a low population density, high connection cost, power quality and –security. Shortage of electricity aggravates poverty because it excludes most industrial activities that might develop a rural area by giving more jobs and interest new investors.

During a demographic transition the response in fertility is not as rapid as the response in mortality and the population will rapidly grow due to a longer lifespan and still a high fertility. [20] Because of this, population as an external factor is influencing the energy

Britt-Mari Langåsen

Spring 2004

2 Background

4

consumption. When the population grow, there is an increase in the demand for energy services. But this connection could also be seen the other way, that energy patterns could change the population by supplying better and more accessible energy. This would relieve women and children from the work of gathering fuel wood when they get access to other means of energy than traditional biomass. Giving people access to electricity will also give them energy for lighting purposes. These factors together will give people in developing areas more time to spend on education and income generating activities, which can also be performed during evening time due to better lightning. It is a known fact that women with education and those that has an occupation gives birth later in life and also have fewer children compared to those without any education and profession. When moving up the energy ladder and using for example electricity, then this will relieve the adverse health effects caused by smoke from the burning fuel wood. The fuel wood is burned for cooking purposes and this means that mostly women and children are affected who stay indoor. [20] All in all, getting access to electricity would raise the standard of living and help development in rural areas of the third world.

The main factors that play a role in the transition to more modern energy usage are accessibility, the relative price which decides if people can afford the new services and cultural preferences. [13] An interesting fact is that when people move up the energy ladder they usually do not complete the transition between the energy sources, but they tend to hang on the old ones still and using new ones, hence they use more energy than would have been done if fully transitioned. An example of this is the mother of Rachel Arinda, whom I worked with in my field studies in Uganda, that explained to me that she use an electrical stove, a gas stove and a three stone cooker when preparing food. This is because they al have different positive sides and some foods “have” to be prepared in a certain way.

In this study I have concentrated on a rural district, Masindi, in the north west of Uganda. This area was chosen because of; interest from GTZ who where starting a project on rural electrification of a small trading centre in Masindi together with students from Makerere University Kampala (MUK), interest from my supervisor at MUK, and also because it is an underdeveloped part of the country relying mainly on agriculture and it has few larger industries. Within the area I have looked at 20 trading centres that are

Britt-Mari Langåsen

Spring 2004

2 Background

5

presently not connected to the main grid and that is of economic and social importance for the development for the Masindi district. A deeper study have been dperformed on three different types of trading centres differing in number of inhabitants, distance from the grid and load profile. This has been done to give an understanding on how different variables influence the optimization result for a set of different energy sources.

Britt-Mari Langåsen

Spring 2004

3 Methodology

3

6

Methodology

This chapter is about the methodology used in this study. It includes a description of field studies in Uganda, but first is a presentation of the progression of the work of the study and an explanation of why the used simulation programs have been chosen. It will also give a description of the simulation program used in this study.

3.1 Preparations and progression of the work The work on this study has continued for 22 weeks. Within this time, 6 weeks where spent in Uganda for field studies. The purpose of the trip was to find data and information about distributed generation in Kampala, and how it could be used to improve the grid stability and to help avoid power shortages. This was the original aim of the master thesis. This problem definition included the use of a load flow analysis program called SIMPOW by ABB. The reason for doing a load flow analysis was to se the load structure, how the power flowed and where rehabilitation of the grid and generation input was needed to help improve the power quality. After coming to Uganda and meeting with the supervisors there, it became obvious that information about this “problem” was hard to achieve. Instead the supervisor at Makerere had a project going about rural electrification that he was interested that I take part in. This project was based on finding the optimal way of supplying a rural area with power and was an optimization problem. The project included the use of an optimization program called HOMER by NREL. It was also intended to do a load flow analysis here as well, to see how the grid was affected when new loads from the investigated district was added. Before travelling to Uganda, some preparations where done in the form of collecting literature about rural electrification and to get an idea of the program SIMPOW. During the field studies, information was gathered, some visits to factories where done, but the main activity where to interview people and to collect data and reports that where not available in Trondheim. One important aspect was also for me to actually se the energy situation and to get an understanding of the problems encountered. Back in Trondheim the simulation work began and also some reading of literature about rural electrification. After a while, it became obvious that there weren’t enough time, and to little data to do a load flow analysis over Masindi district.

Britt-Mari Langåsen

Spring 2004

3 Methodology

7

3.2 Field studies This thesis included a six week stay in Uganda and these weeks where spent on field work. The aim of the fieldwork was to receive information by interviewing people and to assess information that is no available on the internet. Because of cultural differences between Scandinavia and Uganda, in some situations there might arise some confusion and misunderstanding in the meeting between the two parts of the world.

There are several things that differ when searching for information in Uganda, compared to Norway. One of the most obvious things is that the time perspective is different. When in Norway one is expected to arrive to a meeting in time, Ugandans seem to have a wider perception of the concept punctual and it’s not strange to arrive half an hour late (according to Norwegian standards). This imprecise understanding of time reflects also in other things, like when sending e-mails of fax. As an example it can be mentioned a fax that I expected. It was meant to come “in a couple of days” but it arrived several weeks later. This may also have its base in the fact that if people don’t feel part or benefit from the result you expect from the data received, then they don’t engage such amount of time in the matter.

Another difference is the technical resources available. Internet connection is mainly slow and this results in that most information is gathered by meeting people face to face, instead of looking it up on the web or contacting people through e-mail. In Masindi district where I was to search for information, they had no access to internet, but where reliant on fax machines. This was not available at the district population office, but was a service that could be bought in the trading centre. Because computer resources are scarce, most documents are in hard copy, and not in pdf forms. This means that if a copy is wanted, some hours in front of a copying machine is to expect. It may also take some days to get access to the reports, since clearances have to be sent to the appropriate persons in a letter form instead of via e-mail. Because of this scarcity of internet access and intranets, it’s likely to believe that information sharing between instances is hard and time consuming as well. Even so, people are very willing to help, even if they do not know about the problem them self. They are most times happy to refer to other persons that may be of interest.

Britt-Mari Langåsen

Spring 2004

3 Methodology

8

Most people have a cell phone so this makes it reasonable easy to get in contact, but there is a problem that the phones are often shut off, the phone numbers received are not working or it’s hard to get a reply if an SMS has been sent or if a voice message is left. This is probably due to difficulties in charging the phones or a shortage in money on the pre paid cards. As an example here, it can be mentioned that people often “beeped” me, so I could call them up instead.

One of the most time consuming activities during the day, is trying to get from one point to another by public transport. Traffic jams are frequent and public transportation operate on the basis that the bus (or taxi as they call them) leave when full. This can take anything from no-time up to half an hour. In rural areas, it can take up to two hours before the bus leave.

All in all, everything usually turns out well in the end, but there might be some frustration before getting there…

Results achieved from the field studies ♦

Solar data from the Meteorological Institute in Kampala. To get the data, we just went to the meteorological institute and met a person which a student I was working with had gone to for data earlier. The data was not up to date because there where no weather station in Masindi presently, but we got what was available. For the data I had to pay a 10 000 Ush which is almost US$ 5.

♦

SingleLineDiagram of the grid from UEDCL. This data where supposed to be used in a load flow analysis. Because of a lack of data and time, this part of the thesis has been changed.

♦

Meeting with local and governmental people and understanding of the way of work in Uganda

♦

Cost of products available locally, for example PV, generators etc. These data where gathered by going to different suppliers and asking for prices. Some where willing to give them to us, others where worried that we would use their retail prices and then give people an incorrect picture of the actual cost for the products. Some price data in this report are from reports from UEDCL.

Britt-Mari Langåsen

Spring 2004

3 Methodology

♦

9

Trading centre data. This was names and population data for the communities that are important for Masindi district development and that are without electricity. The simulations re built upon these data. These figures and names took six weeks to receive. I was promised to have them faxed to Kampala within a week, but after 2 weeks they still hadn’t come. I called and the lady who where looking for the data promised to have the faxed within the same week. After 2 more weeks they still hadn’t arrived and I went up to Masindi town again. I asked for them at the district development office and she said she had given them to the man, Rashid Yawiya, which was going to find the population data for the trading centres. I went up to he district population office and found the Mr Yawiya that where finding the figures and he had received the centre names the day before so he wasn’t finished. This was now four weeks after they had agreed I should have gotten the data. I had to return to Kampala and about a week after I had returned to Trondheim, I got a fax with the data requested.

To think about To get the most out of the field studies, there are some things that are more important than others to think about. Some of these things I did, some I figured out during my stay, and some again, I realized when I got home that I should have done. ♦

Try to get to know some of the cultural differences before going to a foreign country, like what is the norm for meeting in time and dress codes.

♦

Don’t be in a hurry

♦

Be well prepared and ask specific questions

♦

Demand results and interest for what you are doing.

♦

If possible, wait for the data you requested for at the person you where asking. Otherwise my Masindi experience can be experienced. See above.

3.3 HOMER HOMER is a program developed by NREL with the start in 1993 and is an optimization model for distributed power. It was initiated to address the potential electricity opportunities in rural villages and to investigate the technical and financial performance of hybrids given a village load and availability of wind and solar resources etc. [12] It is offered for free by NREL and can be downloaded at www.nrel.gov/homer. HOMER

Britt-Mari Langåsen

Spring 2004

3 Methodology

10

stands for Hybrid Optimization Model for Electric Renewables. It can design power systems for both off-grid and grid-connection, with a variety of applications. A decision that has to be made is what kind of components should be in the system, and what size and quantities they should have. The program has a database with different technology options as a base and does an optimization and sensitivity analysis to make an evaluation of the large number of system configuration options easier. The deciding factor for the optimal system is based on total net present cost. Input parameters that describe the technology options like costs, loads and resource availability are put into the program

In Figure 3-1 the user interface for HOMER can be seen. It’s a windows based program and very user friendly. The energy sources are added and can then be modified by clicking on the buttons and new windows will appear. After all the input parameters are ready, then the simulation can start and finally the result will appear in the result window. The result can be seen either in tabular or graphic form.

Figure 3-1 User interface of HOMER

Britt-Mari Langåsen

Spring 2004

3 Methodology

11

Different objects that where used for simulations in this study are: generators, grid, PV, batteries, converters and loads. To be able to really compare PV and diesel generators with grid extension, all simulations are done with AC loads, meaning that extra components such as inverters are needed for PV systems. There are also applications that run on DC power, but it would be too complicated to compare all these alternatives. To do a simulation that will be viable for the Masindi area different factors had to be taken in to account, among them the solar radiation, diesel costs and grid price, several economic parameters and limitations such as the maximum allowed shortage per year and minimum part of renewable energy.

Three different solutions for powering the trading centres have been used for the simulations, namely solar systems, diesel generators and grid extension. Following in chapter 4.3.1 to 4.3.3 is a description of the systems and advantages and disadvantages of them.

3.3.1 Critique of HOMER HOMER was developed in 1997, and is still under the progress of development. The program has large umber of users, for both school and professional use. It seems to be widely used in university environments, but also among engineers. On HOMERs webpage [24] it can be seen that by now, HOMER have been downloaded by 4405 persons from 157 different countries.

HOMER has a wide variety of input variables to make the simulations as precise as possible, but this also makes the results very sensible for the accuracy of the data that is found and put in to the program. As I have found, there is no possibility to put in a development of the prices for example, for a period of time. It is however a possibility to have several sensitivity variables, so in a way it’s possible to se what system configurations are optimal over a larger time period, given that one has an idea about the progression of prices. This is something that is to be upgraded for grid electricity prices in the next version of HOMER.

The results that come from the simulations in HOMER will not be better than the accuracy on the input variables. This means that if there are some uncertainties in for

Britt-Mari Langåsen

Spring 2004

3 Methodology

12

example price, there will be uncertainties in the simulation result as well. The already existing components can be modified, but when extracting the price curves for the components, it has to be taken into account that it isn’t certain that a larger component may follow the same price curvature as a smaller one. This is especially relevant for solar systems and diesel generators.

A problem with HOMER is weighting environmental concerns. Carbon tax as US$/t and a minimum of renewable energy output can be put in as parameters. Most third world countries have no carbon tax though, but a western standard could be put into simulations or it has to be calculated for qualitative when analysing the results. No carbon tax has been used in the simulations in this study. Neither have any minimum of renewable energy been set. This was to find the absolute minimum cost for systems to supply communities with electricity. Neither is it possible to put in a parameter for an increase in grid extension costs in those cases that there is a rough landscape making grid extension more expensive. This can though be done by putting a sensitivity variable on the cost for grid extension. This has not been done in these simulations.

There have been some problems in finding accurate prices, in particular for the PV panels. The prices that are used in the program are calculated from prices available on the market in Uganda today. These prices are for panels of the size from 50W to 75W. The solar panels that are used in HOMER are in the range from 500W to several hundred kW and the price has been calculated by just multiplying the numbers of panels by the price for a small panel. This calculation may be correct if the panels are used as solar home systems (SHS) with some PV panels at each house. If instead a solution where solar panels are built to a larger unit in one part of a trading centre, then the prices will probably be different, i.e. less. This latter solution have not been investigated and used in this study. The solar data used in these simulations are from the 1960’s, but they where the only one available for Masindi district at the Ministry of Water, Lands and Environment, Department of Meteorology, in Kampala. Assuming that the solar radiation in the district has not changed remarkably the last 20 years, then this should not cause a problem. To see the solar data for Masindi, see the attached CD.

The load profiles used are made by data from UEDCL and assessments done on during what hours different household applications are used. There is some risk that these

Britt-Mari Langåsen

Spring 2004

3 Methodology

13

assumptions may be somewhat wrong, but the same base for the load profiles are used for the different simulations making them as similar as possible.

3.4 Sensitivity analysis Here the concept of sensitivity analysis will be explained and why it is important. There are different questions that need to be answered before anyone is willing to invest in a new system. One of the most critical ones is what the risk for the investor is. Sensitivity analysis is used in the simulation program HOMER. The sensitivity variable used is the PV capital cost multiplier and diesel cost multiplier.

There are several reasons for doing a sensitivity analysis. It is a very good tool when there are doubts about different variables. These might be uncertainties in expected load, diesel price etc. By doing this type of analysis it is possible to determine how important one variable is for the outcome of the simulation and how the result vary with the value of the uncertain variable, i.e. you determine the sensitivity. It is also possible to do a sensitivity analysis to determine what the price on for example PV systems have to be, to be economically compatible with diesel generators and grid extension. This means that by doing a single analysis, it is possible to simulate several different situations with different variables.

Questions that can be answered by doing a sensitivity analysis in HOMER: ¾

How low must the price of solar systems be to be competitive with diesel and grid extension?

¾

How close to the grid must a certain load be for it to be economically viable with grid extension?

¾

How high must the diesel price be to make PV systems competitive?

¾

What happens if the load differs from the assumed one?

Britt-Mari Langåsen

Spring 2004

4 Power generation and technology options

14

4 Power generation and technology options This chapter will try to give an overview of the power generation in Uganda and Masindi district. Hydro power is the major energy source for electricity production, and potential hydro sites that are under planning for future expansion are given in 4.1.3. This chapter will also deal with the power producers that exist in the area and alternative energy resources. Finally the electrification systems used for simulations in HOMER will be discussed.

4.1 Power generation 4.1.1 Present power generation in Uganda Presently the only major power generation is situated around Jinja, consisting of three power stations; Nalubaale 180 MW, Kiira 120 MW and Mubuku 14 MW. The latter one is privately owned and supplies KCCL industries [1]. In the northern regions of Uganda, which are more isolated due to both long distances and because they are unsafe areas, the access to electricity is mainly through diesel generation.

Picture 4-1 The dam at the power station in Jinja

4.1.2 Power producers in Masindi district There are no major power producers within Masindi district except for Kinyara sugar works Ltd. Kinyara Sugar Works are self sufficient regarding electricity due to own power production. This production is based on the burning of bagasse and they also use diesel generators as a supplement. Since there is a substation just outside the factory a connection to the main grid could easily be established. Due to low energy prices it is not economically viable for Kinyara sugar works (according to their own study) to produce more

Britt-Mari Langåsen

Picture 4-2 Kinyara Sugar Works Ltd

Spring 2004

4 Power generation and technology options

15

energy and sell the excess to the utility. It is though an alternative if the energy prices rise. They have recently (March 2004) built a new bagasse furnace that can be transformed to generate power. Kinyara Sugar Works are planning to expand their sugar production and thereby burn more bagasse. [30]

There are some sawmills in the district, but these are very small and consume their power from the grid. They are not of such a size that makes them suitable for electricity generation, and especially not to sell electricity to the grid. [32]

4.1.3 Future power generation There are several potential generation sites within a reasonable distance from Masindi district. These are both hydro and geothermal. The possibilities for wind power are rather scarce as the average wind speed in Uganda is 3m/s [14].

Presently only 380 MW is built at Jinja hydro station but there are plans of building power stations along the Nile. One problem with this is that several of the suggested places for building power stations are situated within national parks. This raises environmental issues and makes financing difficult. Financing are mostly from the World Bank and foreign governments. There have been assessments made for the power that can be recovered by hydropower from the Nile along the stretch between Lake Victoria and Lake Albert and this has come to about 3000 MW. [1] Some 22 places have also been identified for small hydro power generation and some of these far away from the existing grid, making them potential energy sources for minigrids in rural areas. Today the installed capacity from small hydro is 13 MW. [14] Places that are suggested for large scale hydro power generation are: Table 4-1 Suggested places construction of new hydro power generation

Place Kakira

Year of construction 2005

Bujagali

2006-2007

Karuma

2008-2009

Kalagala

2010-2015

Murchison falls

Britt-Mari Langåsen

2015

Spring 2004

4 Power generation and technology options

16

4.2 Alternative energy sources An alternative type of power generation is by geothermal energy. Some research has been done in this area in three places in Uganda, whereby one of he places is not far from the Masindi district. This is Kisoro, close to Lake Albert in Hoima district. The other two are in Katwe and Buranga. [14] Presently the research is still concentrating on localising where the geothermal spots are and doing feasibility studies on these. It is not yet decided what year a geothermal plant may be up and running. No figure for the assumed power produced can be given at this point either. [33] The expected output is somewhere around 450 W for Uganda. [14] This power is meant to supply the surrounding area and the national grid [23]. More detailed investigations are needed to give confidence to the private sector to make them willing to invest in these projects.

Wind power is not a preferable alternative in Uganda, due to relatively low wind speeds. This is valid also in the costal areas. The average wind speed in Uganda is 3 m/s [14]. There are some types of wind generators that work on such low wind speeds, but they are not studied in this thesis. There are some places around Lake Victoria and on hills that has a higher average wind speeds, but the high initial costs involved in wind power systems are also to high for most communities in rural areas.

Britt-Mari Langåsen

Spring 2004

4 Power generation and technology options

17

4.3 Rural electrification technology options Following are the electrification systems that are used as alternatives in HOMER. They will here be presented with their benefits and disadvantages to give a better understanding of the conclusions and recommendations later in this study.

4.3.1 Solar power systems Solar energy is a resource that is readily available in Uganda, but due to high investment cost, it’s not a very widespread technology in rural parts. As with all technologies, solar energy has both advantages and disadvantages and some of then are put up here. Benefits of solar energy: ♦

Solar modules convert freely available sunlight directly into electricity.

♦

Solar systems generate no pollutants or

Picture 4-3 Example of a Solar Home System

exhaust gases when producing energy. ♦

Solar energy systems require little maintenance and have an estimated lifetime of over 20 years.

♦

The price of solar panels have fallen over the past years while the oil price have risen, making the use of solar systems an economic viable solution for powering households.

♦

Solar systems are modular and can easily be expanded when the load increase.

♦

Because off low system voltage the risk of electric chocks are small. The fire risk in homes and schools are also smaller when using solar systems compared to those lit by kerosene lanterns.

Disadvantages of solar energy: ♦

The initial cost for a solar system is high by rural standards which make it hard for people to buy the systems, even though the lifetime cost is lower compared to diesel generators or kerosene.

♦

The performance of a solar system is dependent on the batteries available on the market.

♦

Appliances that run on 12 V (PV system voltage) is not readily available on the market yet.

Britt-Mari Langåsen

Spring 2004

4 Power generation and technology options

18

♦

Solar systems require frequent maintenance.

♦

There are few trained technicians to design, install and maintain solar energy systems.

♦

Solar systems are usually of a very small scale meaning that if there is a relatively high energy demand the investment cost will be very high.

Besides from PV having a very high cost when the systems become very large, they also occupy a large area for the solar panels. This means that if solar systems are to supply a large load, then an option might be to build all the panels as one unit and the customers might come there and get their batteries charged and taken home to the respective homes. A positive thing with this system is that cleaning and maintenance will probably be easier if all the panels are situated at one location. A larger solar system with a mini-grid is also an option. This has not been included in this study. Another option is that only the most important loads get power in a community that is situated far away from the grid. Important loads might be the local health station, the church and the school. Some street light on the main road / along the shops will also make the area more secure and create a natural gathering point for the village’s inhabitants. This option will give smaller systems with less investment cost. Because that the investment costs for PV systems are so high people can’t afford to have large systems meaning that the systems will be too small to generate enough power for any income generating industries, except maybe home based sewing industries.

Britt-Mari Langåsen

Spring 2004

4 Power generation and technology options

19

4.3.2 Diesel generators Diesel generators are commonly used in many parts of Uganda. In several parts of the country diesel generators are the only available electricity source, and in other places they work as backup generation. Benefits of diesel generators: ♦

Modular

♦

Relatively low investment cost

♦

Widely available technology

Disadvantages for diesel generators: ♦

Diesel

generators

give

pollutants

and

exhaust gases when producing electricity. They also produce different levels of noise, depending on type and size. ♦

Diesel is expensive in Uganda, and this is not likely to decrease in the future.

Picture 4-4 Example of a diesel generator

♦

Needs skilled personnel to do frequent maintenance

♦

Dependence on the availability of spare parts

There are many different sizes of generators and the sizes that have been used in the simulations are in the range of 4 kW to several hundred kW. They do not follow a linear cost curve, but become cheaper per kW the higher the rating. Diesel generators work best when they work on a high load ratio. This means that there is no use in over dimensioning because they will then use relatively more fuel because of low load ratio. This is something that has been taken into account in the simulation program. Most probably there will be a load growth in the trading centre, so the choice of generator size is very much an optimization problem.

Diesel generators need proper maintenance to work optimal and have a long lifetime. This requires trained personnel to operate and maintain it. This might be a problem in rural areas. Another problem is the availability of spare parts such as filters etc. But the most obvious problem with diesel generation is that it is not a sustainable source of energy, and will eventually have to be replaced with some other technology. Relying on a system of diesel generators will make a non oil producing country dependent on the

Britt-Mari Langåsen

Spring 2004

4 Power generation and technology options

20

importation of oil and the world price market of oil that is currently rising [10]. For a poor country like Uganda, that has all of its petroleum products imported, this is an undesirable situation. 15% of the country’s export earnings go to import of petroleum, at the expense of development programmes. [14]

Britt-Mari Langåsen

Spring 2004

4 Power generation and technology options

21

4.3.3 Grid extension There are several reasons to expand the transmission and distribution system. [1] Benefits of grid extension: ♦

Rural electrification

♦

To increase the security of the system by introducing a more meshed configured system

♦

To allow the safe input from new power plants

♦

To reduce losses and improve operational economy

Picture 4-5 Rural electrification by grid extension

of the network ♦

To fulfil agreements with neighbouring countries.

There are also negative impacts that may arise from an expanded grid compared to local power production. Disadvantages with grid extension: ♦

Environmental impacts from building and sustaining the grid.

♦

People may illegally tap the grid and thereby cause poorer quality of the power and this can also be hazardous for people.

♦

High investment costs

When planning for a grid extension an anti-theft design must be considered. This may be seen in the choice of conductors. Instead of choosing a cheaper conductor that may be stolen because its ability to be remelted, an expensive but non remeltable transformer may be used. It may also be an alternative to use a higher voltage level on the distribution lines which makes it impossible to use the power directly. (I.e. use 400 V instead of 230 V) [29] There are also several different types of poles that may be chosen, since this is also something that is exposed for theft.

Environmental impacts that a grid extension in the Masindi area may cause are[1]: •

Impacts on flora and fauna

•

Impacts on drainage and water resources

•

Impact on landscape and visual amenity

•

Impact on land use and agriculture

•

Electric and magnetic fields

Britt-Mari Langåsen

Spring 2004

4 Power generation and technology options •

22

Clearances around lines

In general will the environmental impacts of new power lines occur within the line corridor or near to it. There will also be some impacts on the nature if access roads have to be built to get to the line corridor. When travelling around in Masindi district, it can seen that except for around Kinyara sugar works where they farm sugar cane, large areas of the countryside are not cultivated. These are mainly grasslands with low to medium high vegetation except for in Budongo area where there is an old Mahogany forest. In most parts of the district there will probably be few environmental problems with grid extension concerning impacts on landscape, land use and agriculture. If lines are drawn through sugar cropping or otherwise where there is high vegetation, there must be a high limit of 1.8 m for the crops. This is a safety limit that should be followed. There are also some impacts from the environment on the transmission lines, for example growth of vegetation that can touch the conductors and during rainfalls, trees can fall on the lines. Wind I s also a problem that can cause trees to fall on the lines in line clearances are not followed. In some areas with salty soils, this can harm the steyers that support the electric poles that are mostly made out of wood.

4.4 Conclusion A summary of the technology options are shown in Table 4-2. Table 4-2 Summary of the technology option

Technology

PV

Application area Usual size Average cost

SHS, small scale applications 50WUS$280-

Diesel generation Stand-alone systems, back-up power 7kW-200kW US$ 7000-34500

+

Sustainable technology

Widely available technology

-

High investment costs, frequent maintenance

Pollution, noise, frequent maintenance

Britt-Mari Langåsen

Grid extension Areas with a high load, domestic and industrial US$20000 / km Available technology, no maintenance High investment cost, maintenance, deforestation and use of agricultural landscape

Spring 2004

5 Analysis of trading centres

23

5 Analysis of trading centres This chapter will give an overview of the socio-economic situation present in Masindi district and the expected development. It will also give the criteria that where used to find the different trading centres that where used in this study. Finally the background for the assumed load profiles will be given and the profiles used in the simulation program will be presented.

5.1 Masindi district Masindi district is one of 56 districts in Uganda. It is found in the north-west, close to Lake Albert, and this district covers an area of approximately 9442.9 sq km. The district has a total population of 469,865 where the urban population is 0.8% and the rural population is 99.2%. The main economic activities carried out in the district are agriculture, trading and others and the household main sources of livelihood are: subsistence farming, commercial farming, trading, employment income, family support and others. [2]

Considering that electrification work best as a mean for economic development when the overall conditions are right for rural income growth and when it is complemented by social and economic infrastructure development [28], the trading centres where chosen because of their importance in developing Masindi district. These trading centres, which are small communities, where chosen with the help from the district population office in Masindi.

The trading centres studied in this report where chosen because of: ♦

Their importance for the development of Masindi district

♦

They are presently not connected to the main electrical grid and are not likely to be in a near future

♦

They are likely to continue growing as trading centres

The last point is important as it is no idea to invest in powering areas that are likely to be mitigated from in a not to distant future, by grid extension. Three different types of trading centres will be studied further. These have different number of inhabitants and

Britt-Mari Langåsen

Spring 2004

5 Analysis of trading centres

24

thereby different loads. The different centres have about 20’000, 10’000 and 2000 inhabitants.

5.2 Socio-economic characteristics of Masindi A large portion of the population within Masindi district lacks skills or is semiskilled, so most of the labour force is engaged in farming. Cottage industries engage very few people and young men therefore turn to petty trade or motorcycle business. Young women don’t engage in the visible economic activities to a high degree because of a lack of education.

The most important source of livelihood in Masindi district is agriculture and about 76% of the population is engaged in farming activities. Most of the farmers are family based with an area of 1-2 hectares under cultivation and only 1 % of the population is employed in commercial farming. [1] The only major mono cropping plantation within the district is Kinyara Sugar Works who crop sugarcane.

There is only one large industry in the area (Kinyara

Picture 5-1 Small trading centre

sugar works) but there are several small-scale industries coming up. The existing and new ones are agro-based processing industries which include maize and oil milling, rice hulling and furniture making. At the moment some of these are out of production due to a lack of spare parts. [1]

When it comes to trading petty trading engages about 6.9% and formal trading employs about 3% of the economically active population. [2] The commodity traded is crop within the district and salted and sun-dried fish across the border to the Democratic Republic of Congo.

The need for electricity in rural areas are mostly for lighting, TV and radio and in the richer areas also for fans, refrigeration and hot plates. Today candles and kerosene are used for lighting and the food that is used is bought fresh or needs no refrigeration.

Britt-Mari Langåsen

Spring 2004

5 Analysis of trading centres

25

When cooking, charcoal and fuel wood are used. Battery driven radios are used, but replacement batteries can be hard to get and are relatively expensive. When I visited Biizi trading centre, a community with a bit more than 2000 inhabitants, the water pump was driven by hand power and there seemed to be no lights outside any shops.

5.3 Load assessment Load assessments have been done for the three types of trading centres investigated. These assessment have been done with data about time usage and type of appliances from UEDCL, and by assuming the usage of appliances during the day. To see a more thorough description of the load profiles, see the CD in appendix 11.3. Using data from UEDCL (see Table 5-2) the energy consumption for different types of rural households has been calculated. These trading centres have been investigated further as to represent a small, medium and large trading centre. Estimates given in Table 5-1 are used for energy consumption for rural households in the simulations in HOMER. Two different load profiles have been found for the three trading centres, one for constrained demand and one for an unconstrained demand. Constrained demand can signify different demands depending on the context. Generally it is a system demand limited by generation supply for those that are already connected to the grid. This demand does not really reflect the actual load growth over a given period of time. Unconstrained demand is what they would use if there where enough generation capacity and no power shortages. [31] Constrained demand can also signify the power use according to a certain income level, and that is the meaning I have chosen to use here. This has been done to represent different loads that can be expected. Unconstrained use will in this case mean the power use if the customers where at a higher income level. These two load profiles have been used in the simulations to find the difference in optimal systems to supply the trading centre with energy. The constrained demand is not preferred situation, but one that the customers could “manage” if that is the only alternative.

Britt-Mari Langåsen

Spring 2004

5 Analysis of trading centres

26

Table 5-1 Energy consumption for rural households in Uganda

Type of household

Energy consumption

(kWh)

Constrained demand

Unconstrained demand

Rural high

3.453

6.077

Rural medium

0.612

1.556

Rural low

0,044

0,108

Following are the data from UEDCL for the time usage for different appliances in rural households. It can be seen that there is a large difference between a rural high and a rural low household and thereby will the load profiles that are built on these figures be quite different. There is also quite a large difference between the constrained and unconstrained demand.

Table 5-2 Time usage ( in hours) of electrical appliances in rural households

Fan Flat iron Fridge Hot plate Lighting Radio TV

Rural high

Rural medium

Rural low

Constrained Unconstrained demand (h) demand (h)

Constrained demand (h)

Unconstrained demand (h)

6 0,5 10 2 4 4,9 1,6

15 0,8 14 3,5 8 7 5

5

15

6

14

2 3 2

6 6 4

42 1300 45 1000 8 10 35

W W W W W W W

Constrained demand (h)

Unconstrained demand (h)

3 2

6 6

Power rating for appliances: Fan Flat iron Fridge Hot plate Lighting Radio TV

Typical loads in a rural household are lights, small televisions and radio sets and irons. Other rural loads are electric fencing, water pumping, small industries and institutions, telecommunication, lighting at schools and churches, small shops and health centre vaccine refrigeration. In the simulations only the loads from households have been used. Usually the shops in a trading centre function as homes as well, with the same Britt-Mari Langåsen

Spring 2004

5 Analysis of trading centres

27

standards, so these are included. Excluded are churches, schools and health centres. These where considered as comparably small loads that will not affect the results to a high degree.

Based on data from the census of 2002 in Uganda, there are on average 4.8 persons per household [25]. This information together with Table 5-1 and Table 5-2 will give the following table for the expected power consumption for the chosen trading centres. The expected electric energy growth according to a study made by SWECO, is 7.9% per year during the period 2001-2016 [1] and the consumption for 2010 is also given. Below is given an explanation of the colours and from what load profiles the demands calculated.

Table 5-3 Present power need in the Masindi region. Based on data from UEDCL

Kiryandongo Bweyale Kigumba Katulikire Kabango Butiaba Masidi Town Mutunda Apodorwa Biiso Kyatiri Nyabyeya Pakanyi Karuma Nyakabale Wanseko Kigezi Bugoigo Buliisa Kijura Biizi

Energy consumption 2004 Constrained demand (kWh) 4339 3773 2097 2060 645 604 577 548 531 518 516 494 469 325 315 306 202 201 180 176 20

Britt-Mari Langåsen

Energy consumption 2010 Constrained demand kWh 6846,55 5953,60 3309,96 3250,19 1017,60 953,14 909,98 865,22 838,03 816,90 813,78 779,81 739,77 513,01 496,87 483,26 318,76 316,48 284,15 277,62 32,20

Energy consumption 2004 Unconstrained demand kWh 8563 7446 4400 4321 1634 1530 1461 1389 1346 1312 1307 1252 1188 822 797 775 510 506 454 444 50

Energy consumption 2010 Unconstrained demand kWh 13513,49 11751,02 6943,83 6818,45 2578,52 2415,17 2305,82 2192,41 2123,50 2069,95 2062,05 1975,97 1874,53 1297,83 1257,00 1222,58 804,17 798,42 716,85 700,38 79,04

Spring 2004

5 Analysis of trading centres Pink

0.2*rural high+0.4*rural medium+0.4*rural low

Purple

0.1*rural high+0.4*rural medium+0.5*rural low

Blue

0.4*rural medium+0.6*rural low

Turquoise

0.3*rural medium+0.7*rural low

Green

0.2*rural medium+0.8*rural low

Yellow

1.0*rural low

Britt-Mari Langåsen

28

Spring 2004

5 Analysis of trading centres

29

5.3.1 Small trading centre A typical load profile for a small trading centre can be seen in Figure 5-1. Two demand profiles have been done, a constrained demand and an unconstrained demand. As can be seen from the profile below, the maximum demand will occur during evening time and there will be a smaller peak in the morning. The minimum load is after breakfast and during night time. The peak is almost three times higher than the base load. The reason for this is that there are very few large loads on during the day time in a domestic area that a small trading centre mainly is. Most of the energy in rural areas is for lighting purposes and entertainment meaning the energy will be consumed after sunset and before sunrise i.e. between 6 p.m. and 7 a.m. The wattage behind the peak is mainly for lighting and radio and a maybe a few television sets. There is almost no load during night time and that is because that most lights are off besides from security lights and larger loads like fridges are turned off during the night due to high energy prices. Expensive loads like security lights and fridges may not even exist in a small rural area that most often is a poor area as well. This information is found from talking to Ugandans and after visiting some trading centres and seeing where they may live and the standard under which they live in.

Load profile 10,00

kWh

8,00 6,00 4,00 2,00 0,00 1

3

5

7

9

11

13

15

17

19

21

23

Time (hours) Constained

Unconstrained

Figure 5-1 Load profiles for Biizi trading centre

Britt-Mari Langåsen

Spring 2004

5 Analysis of trading centres

30

The peak effect in the constrained situation is 12.8kW and the energy is 20kWh/day, calculated from the simulation program HOMER. The same data for the unconstrained situation is 12.8kW and 55kWh/day, respectively. The peak effect is the same in both situations, but the energy consumed is largest in the unconstrained situation. This is natural, because the consumers will not have more appliances in the unconstrained situation, but they will use the ones they have for more hours per day.

Britt-Mari Langåsen

Spring 2004

5 Analysis of trading centres

31

5.3.2 Medium trading centre The load profile for a medium trading centre can be seen in the figure below. The medium trading centre is represented by Mutunda, a community with about 10 000 inhabitants. The medium centre is assumed to consist of households that consume energy according to rural medium and rural low energy pattern. The rural medium pattern has more appliances than rural low, which can be seen in Table 5-2. The load profile is created on the basis that the community consist of 60% rural low and 40% rural medium households.

Load profile 150,00

kWh

100,00 50,00 0,00 1

3

5

7

9

11

13

15

17

19

21

23

Time (hours) Constrained

Unconstrained

Figure 5-2 Load profiles for Mutunda trading centre

In the load profile for Mutunda trading centre it can be seen that the energy consumption in the unconstrained situation is more than twice the amount in the constrained condition. (1.393MWh/day compared to 545MWh/day). This difference is mostly due to a wish to have a fan on at night and also a refrigerator on at day in the unconstrained situation compared to the constrained. The relatively high base load during night time that weren’t there in the small example is due to the wish for having a fan on a night. The unconstrained situation shows that much more power is consumed during the day compared to the other situation. This is due to a wish for having radios and fans on during day time.

Britt-Mari Langåsen

Spring 2004

5 Analysis of trading centres

32

5.3.3 Large trading centre A load profile that is developed for a large trading centre can be seen in the figure below. To represent a large trading centre, Kiryandongo with about 20 000 inhabitants have been chosen. In this profile, all the alternatives for rural energy demands have been included. The compositions of demands are 10% rural high, 30% rural medium and 60% rural low households. Explanation of these alternatives can be found in Table 5-2.

Load profile Kiryandongo 1000,00

kWh

800,00 600,00 400,00 200,00 0,00 1

3

5

7

9

11

13

15

17

19

21

23

Time (hours) Constrained

Unconstrianed

Figure 5-3 Load profiles for Kiryandongo trading centre

In this final profile the load profiles are more complex due to a more complex compound of households. It can be seen here that the scale on the y-axis is very different from the two previous examples and even though the trading centre have only twice as many inhabitants as the medium example, the power drawn is 3 times as much. This trading centre represents the one with most inhabitants, about 20 000, and it is assumed to be the richest. This means that they have most appliances and also use them the longest hours. It’s assumed that not all people in the trading centre have a job and a good income, but the standard is though higher that in smaller trading centres in more rural areas. The use of hotplates during lunch and dinner time is spread on four hours since they will not all be used at the same time, even though Table 5-2 states that hotplates are only used two hour per day. In this way it will seem that half the population use it two hours each. The use of hotplates gives part of the explanation of the high peaks as they are very power consuming. The base load is not much different from the load for the

Britt-Mari Langåsen

Spring 2004

5 Analysis of trading centres

33

medium trading centre. The consumers have a fan on during day time or night time, depending on if it is a constrained or unconstrained demand. The fridge is on during day time only.

5.3.4 Summary for the trading centres When seeing the load profiles for these trading centres, the problems with dimensioning a system to supply these kinds of trading centres can be seen. Because of the low base load and high peaks that arise under a relatively short period of time, grid extension is not automatically the best way of supplying power even though there are quite a lot of inhabitants. It will be seen in the simulation results from HOMER that the price of solar systems and diesel will have a great impact on the optimal solution and the break-even grid extension distance.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

34

6 Basis and optimal solutions The results from the simulations will be presented in this chapter. There will be a more thorough presentation for the three different types of trading centres investigated and finally a general discussion about the results found. First the equations that the result from the simulation program is based on are presented. To get a better understanding of the results, and to see the detailed simulation input variables, see the HOMER files on the attached CD in appendix 11.3.

6.1 Simulation basis From HOMER several results can be found; optimal system configuration, total net present cost (NPC) and break-even grid extension distance among others. As described in chapter 3.4 sensitivity analyses is a tool in HOMER that can be used to find results in risk analysis and system configuration based on the development of prices. The demand in the simulations is divided in constrained and unconstrained demand.

The main economic output from HOMER is the total NPC. The NPC decides the ranking of all system options and is calculated according to Equation 1. Net present cost can be understood as the amount one would have to deposit in a bank today, for the amount to match a given value some given time from now. If the interest rate is high, then the required time for the amount to grow to the given value becomes shorter. This means that for a high interest rate the more short-term projects become preferred. This is usually bad for alternative energy sources that have a high investment cost and long pay back time. [21] Equation 1

C NPC =

C ann ,tot CRF (i, R proj )

where: Cann,tot = total annualized cost [$/yr] CRF() = capital recovery factor = interest rate [%] i Rproj = project lifetime [yr]

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

35

As can be seen from the equation above, the NPC is calculated from the total annualized costs. The total annualized cost is in turn calculated from the sum of the annualized costs of each component in the system plus other annualized costs. The annualized cost consist of the operating-, capital- and replacement cost over the project lifetime.

The capital recovery factor (CRF) that is used in the calculations is found according to Equation 2. The CRF is used to calculate the present value of an annuity. Equation 2

CRF (i, N ) =

i (1 + i ) N (1 + i ) N − 1

where: i = interest rate N = number of years The break even grid extension distance is the distance from the existing grid that gives the same net present cost of extending the grid as the net present cost for a stand-alone system. [8] This equation is important as is gives an idea of when to expand the grid and when to build a stand-alone system. The breakeven grid extension distance is calculated according to Equation 3 [8]: Equation 3

D grid =

C NPC * CRF (i, R proj ) − C power * Ltot C cap * CRF (i, R proj ) + C om

where CNPC CRF() i Rproj Ltot cpower ccap com

= total net present cost of the stand-alone power system [$] = capital recovery factor = interest rate [%] = project lifetime [yr] = total primary and deferrable load [kWh/yr] = cost of power from the grid [$/kWh] = capital cost of grid extension [$/km] = O&M cost of grid extension [$/yr/km]

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

36

As can be seen from the simulations and Equation 3, the higher the cost for grid extension, the shorter the break even grid extension distance gets and a stand-alone system may be optimal for a trading centre that is situated far away. This of course is dependent on the load of the trading centre. If the load is rising, then grid extension may be optimal after all. Unfortunately, the simulation program does not have the feature of a rising load demand. This makes it difficult to make realistic future prognoses.

The levelized cost of energy (COE) is the average cost of producing electricity and is another output from HOMER. It is calculated using the following formula: Equation 4

COE =

C ann ,tot E prim + E def + E grid , sales

where: Cann,tot = total annualized cost of the system [$/yr] = primary load served [kWh/yr] Eprim = deferrable load served [kWh/yr] Edef Egrid,sales = total grid sales [kWh/yr] In the simulations that will be done, Egrid,sales will be equal to zero since there is no connection to the grid.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

37

6.2 Small trading centre Biizi trading centre has been chosen to represent a small, rural community. Biizi has 2206 inhabitants according to the census of 2002 [25]. The characteristics for its load profile is a very peaky load and very little or no load during large parts of the day. The load profile can be seen in Figure 5-1. A problem with a small trading centre is that if some consumers decide not to get connected to the new electricity supply system, then the relative change in load and thereby the risk for the investor will become larger compared to in a large trading centre where each customer represent a smaller part.

Following are the simulation results for Biizi trading centre. The results will be presented with the optimal solution given different capital multipliers for the photovoltaic systems. (PV Cap. Mult.) The first column, “PV Cap. Mult.”, can be understood as the part of the actual price for PV systems. This is done to represent a change in price for PV systems and to see how low the price must fall on solar systems for them to be comparable with diesel systems. The next columns, “PV”, “Gen” and “Converter” columns, give the effect of the optimal energy sources. “Total capital” is the investment cost for the solution and “Total NPC” is the total net present cost which is explained in chapter 6.1. “COE” is the levelized cost of energy is that is the average cost of producing electricity and the column “Renewable fraction” gives how much of the total installed energy is renewable. In the simulations a cost of US$0.09 / kWh is used. [26] For other input variables, see appendix 11.2. “Capacity shortage” is the difference between the actual operating capacity that the system can provide and the required operating capacity. Finally, the “Diesel” column tells how much diesel is used and the “Gen (hrs)” tells how many hours the respective diesel generators are running.

As can be seen in Table 6-1 and Table 6-2, the amount of PV depend strongly on the PV price. Later it will be seen that they also depend on the diesel price and the load.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

38

6.2.1 General solutions The optimal solutions for Biizi when there is a constrained demand are as given in Table 6-1. In the simulations, energy demand of 20kWh/d and 12.8kWp has been used. Input parameters can be seen in appendix. Table 6-1 Optimal solutions for Biizi trading centre, constrained demand

PV Cap. Mult. 1.00 0.50 0.10 0.01 PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW)

18.0

Gen1 (kW) 4 4 4 4

Gen2 (kW) 4 4 4 4

COE ($/kWh) 0.453 0.453 0.453 0.446

Renewable fraction 0.00 0.00 0.00 0.84

Capacity shortage 0.13 0.13 0.13 0.13

Converter (kW)

4

Total capital $ 8,000 $ 8,000 $ 8,000 $ 9,811

Total NPC $ 40,015 $ 40,015 $ 40,015 $ 39,414

Diesel (L) 2,379 2,379 2,379 2,216

Gen1 (hrs) 1,095 1,095 1,095 1,095

Gen2 (hrs) 944 944 944 766

The different results give almost the same total NPC, independent of the PV capital multiplier. The reason that the different solutions have almost the same total NPC is that the optimal systems are the same, except for the case when the PV capital multiplier is 0.01. The price for 20 kW PV is in that option very small and will affect the total cost little. The optimal power generation in these options will therefore be diesel generation.

The optimal solutions for Biizi trading centre when there is an unconstrained demand are as in Table 6-2. The simulation program has calculated with an energy use of 55 kWh/d and 12 kWp. Other input parameters can be seen in appendix.

Table 6-2 Optimal solutions for Biizi trading centre with an unconstrained demand

PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW)

15 35

Gen1(kW) 7 7 7 7

Britt-Mari Langåsen

Converter (kW)

7 7

Total capital $ 7,000 $ 7,000 $ 20,245 $ 12,285

Total NPC $ 94,075 $ 94,075 $ 91,818 $ 75,130

Spring 2004

6 Basis and optimal solutions PV Cap. Mult. 1.00 0.50 0.10 0.01

COE ($/kWh) 0.414 0.414 0.398 0.323

Renewable fraction 0.00 0.00 0.67 0.84

39

Capacity shortage 0.16 0.16 0.15 0.14

Diesel (L) 6,282 6,282 5,032 4,549

Gen1 (hrs) 3,285 3,285 2,747 2,398

For the unconstrained demand, the solutions are a bit different. It can be seen that the investment cost increase when the PV capital multiplier decrease, but the total NPC decrease at the same time. The reason for the increased capital cost is the increase in installed solar power. It can be seen from these results that when PV is simulated with its actual, present cost, no PV power will be used and only a 7 kW diesel generator will produce energy. The capacity shortage is 0.16 and the total Net Present Cost (NPC) is $ 94,075. To have no capacity shortage, the total NPC will arise to $ 108,677 and two diesel generators, 4kW and 7kW respectively, will be used. (See Biizunconstrained.hmr on the CD) The cost for PV systems has to come down to a tenth of the original cost for it to be optimal to use PV according to the simulation. In this solution 15 kW of PV will be used together with two 7 kW diesel generators. The total capital cost is more than twice the first option, but the total NPC is lower and this decides the optimum in HOMER. The reason that the net present cost goes down even though the capital cost rise, is that as more PV comes into the system, the operating costs will go down, compared to a system with diesel generation. A diesel system has a relatively low investment cost, but high operating cost. Therefore the annualized costs that are the base for the net present cost will be high for a diesel system, but lower for the solar system.

6.2.2 Load profiles and unmet load In the following figures, unmet load, capacity shortage and excess electricity is shown compared to the load. Unmet load is the electrical load that cannot be met because of to little generation. The capacity shortage is, as explained above, the difference between the actual operating capacity that the system can provide and the required operating capacity. A system has often a certain amount of operating reserve to cover unexpected loads. The excess electricity is surplus energy that occurs when there is a surplus of power being produced and no load is to be served. The excess electricity has to be dumped in a load called a dump load, if there isn’t enough battery to absorb it all. A dump load may be a set of light bulbs or a resistive heater.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

40

The loads in the constrained situation are peaky and there are only two peaks per day, a smaller in the morning and a large in the evening. When looking at these figures from the simulations with the constrained demand in Figure 6-1 and Figure 6-2, it can be seen that there is almost no unmet load and capacity shortage, in both cases. In the case of PV multiplier 0.01, there will be an over dimensioning of the system and a lot of excess energy. This will

Figure 6-1 Load profile from simulations with constrained demand. PV multiplier 1.0

actually mean that a lot of the produced energy has to be used in a capacitor bank and without any real use. This means that the price of PV doesn’t need to be that low for PV to be an economically viable alternative.

Figure 6-2 Load profile from simulations with constrained demand. PV multiplier 0.01

The following figures are from a simulation with an unconstrained load profile. For such a small trading centre, it can be seen that the load is very “peaky”. It can also be seen that in the system that has a PV multiplier of 0.01, there will be a lot of excess electricity, compared to none in the case with a PV multiplier of 1.0. At the same time the unmet load and the capacity shortage will be less in the case with the low PV price. This is because the system has more installed capacity.

Britt-Mari Langåsen

Figure 6-3 Load profile from simulations with unconstrained demand. PV multiplier 1.0

Spring 2004

6 Basis and optimal solutions

41

In Figure 6-3 it can be seen that during evening peaks, there is both capacity shortage and unmet load. This result is for a simulation with the actual cost of PV. When the cost for PV is down to a hundred, then the situation is different as can be seen in Figure 6-4. When the PV price is down to a hundred, four times as much solar power is installed. A 4kW diesel generator is also installed in this situation. All together there will be excess electricity that has to be dumped during the low loads during the

Figure 6-4 Load profile from simulations with unconstrained demand. PV multiplier 0.01

day.

6.2.3 Break even grid extension distance It can be seen from the break even grid extension distance that the distance changes with a changing PV price. Logically, the break even distance becomes smaller when the PV price decreases. Since it is relatively expensive with grid extension, it will become more economically viable to build PV and diesel generator systems closer to the grid if the price for these stand-alone systems gets lower.

The break even grid extension distances are in the constrained condition 1.25 km and 1.23 km for PV multiplier 1.0 and 0.01 respectively. The total NPC is around $ 40 000 for both situations.

In the results from the constrained demand, the total NPC is almost the same weather the PV multiplier is 1.0 or 0.01. This is because the load is Figure 6-5 Break even grid extension distance.

small and the systems are the same, except for the Constrained demand, PV multiplier 1.0 PV. (See Table 6-1) The PV prices

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

42

are so low though, that they barely have an impact on the result.

When there is an unconstrained demand, the break-even grid extension distances become longer. The break even grid extension distance for this small trading centre is between 2 and 3 kilometres with the given scenario. Figure 6-6 Break-even grid extension distance.

If the results from the simulations with unconstrained Constrained demand, PV multiplier 0.01 demand are compared, it can be seen that the total NPC is lower when the PV capital multiplier is 0.01 even though more diesel generation capacity is installed. This is because in this situation the system has a lot of PV, which has a very low cost (1/100 compared to the actual situation) but also diesel generators to cover the load as much as possible. With a low cost on PV, the net present cost will be lower

Figure 6-7 Break-even grid extension distance.

than in the situation with only diesel generators who Unconstrained demand, PV multiplier 1.0 consume (expensive) diesel.

If the constrained demand is compared to the unconstrained demand, it can be seen that the break-even grid extension distance is shorter in the constrained situation. This is reasonable because in the constrained demand, the load is smaller and the total NPC is lower than in the unconstrained demand but the investment cost for grid extension is still the same and relatively high for such a low load. The pay back time for grid

Figure 6-8 Break-even grid extension distance. Unconstrained demand, PV multiplier 0.01

extension will be very long with few customers.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

43

6.2.4 Variations in diesel price When the diesel price rise, then PV systems become more and more economically viable for small trading centres. The systems become more expensive if it’s calculated with the present price of solar systems, so accordingly the break-even grid extension distance become longer. It is shown by simulations that as diesel prices rice, standalone systems based on solar systems are a good option. As the price for solar systems a probable to decrease as well, this makes the solution of supplying a trading centre with PV even better. Unfortunately, complicated simulations with so many variables can’t be done in HOMER yet. Therefore, the following simulations are done with an increase in diesel price but otherwise the same input parameters as in the previous simulations. The diesel prices have been simulated up to three times the present price. From today’s price of US$ 0.8, through US$ 1.6 up to US$ 2.4. (1DP = 1 times diesel price = US$0.8 etc.)

US$

Cost as a function of diesel price Constrained

50000 45000 40000 35000 30000 25000 20000 15000 10000 5000 0 Total capital

DG1

Total NPC

DG2

DG3

Figure 6-9 Cost as a function of diesel price when there is a constrained demand Table 6-3 Optimal results for variations in diesel price

PV (kW) 1DG 2DG 3DG

5.0 5.0

Gen1 (kW) 4

Gen2 (kW) 4

Gen3 (kW)

Battery

Converter (kW)

20 20

8.0 8.0

When there is a constrained demand, the optimal solutions will change when the diesel price become twice the price today (US$0.8). As can be seen in Figure 6- the investment costs will be much higher (more than three times) when the diesel price rise, but the net present cost will be almost the same even though the diesel price is higher. This is because there will now be a much less usage of diesel which constitute the major part of

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

44

the operating costs. The running costs for solar systems are almost non-existing. The reason for the high investment costs is that solar power is presently expensive to purchase.

Cost as a function of diesel price Unconstrained

250000

US$

200000 150000 100000 50000 0 Total capital DG1

Total NPC DG2

DG3

Figure 6-10 Cost as a function of diesel price when there is an unconstrained demand Table 6-4 Optimal results for variations in diesel price

PV (kW) 1DG 2DG 3DG

Gen1 (kW) 7 7 7

Gen2 (kW)

Gen3 (kW)

Battery

Converter (kW)

In the unconstrained condition, the solutions are not affected by a change in diesel price up to three times the present price. This means that according to these simulations energy from PV systems will not be an optimal way of supplying energy to Biizi trading centre when there is an unconstrained demand. An objection to this is that when looking at the simulation results in HOMER, it can be seen that the NPC is not so much higher when 5kW PV is installed and the break-even grid extension distance is almost the same, about 8 km. (See BiiziunconstrainedDG3.hmr on the CD) Adding environmental concerns into the calculations, PV systems may be an option for powering Biizi at a unconstrained demand after all.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

45

6.2.5 Only PV or diesel There have also been simulations done on what the cost would be to supply Biizi trading centre if only diesel generation or solar systems are available. This have been done to see the differences and to be able to better compare the two options.

For Biizi trading centre with a constrained demand, the following results have been found. Table 6-5 Solutions when supplying Biizi with PV or diesel

PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW) 5.0 5.0 5.0 30.0

Gen1 (kW) 4

Gen2 (kW) 4

Battery 20 20 20 10

Total capital $ 8,000

Converter (kW) 10 10 10 10

Total capital $ 25,040 $ 15,040 $ 7,040 $ 5,940

Total NPC $ 42,666 $ 31,570 $ 22,693 $ 20,417

Total NPC $ 40,015

COE ($/kWh) 0.453

Renewable fraction 0.00

COE ($/kWh) 0.492 0.364 0.262 0.244

Renewable fraction 1.00 1.00 1.00 1.00

Capacity shortage 0.13

Diesel (L) 2,379

Capacity shortage 0.12 0.12 0.12 0.20

Gen1 (hrs) 1,095

Gen2 (hrs) 944

From this table it can be seen that supplying the community with solar power almost give the same net present cost as for diesel generation. The capital cost is though three times higher which for a poor community may be a high obstacle to overcome. Since the net present cost is almost the same for the two options the break even grid extension distance is almost the same for diesel and PV 1.0. When looking at the simulation results, it can be seen that for PV multiplier 1.0 there is little capacity shortage and very little unmet load. There is a little excess electricity in the middle of the day. For the diesel option there is little unmet load and capacity shortage as well, and no excess electricity.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

46

Similar results have been found in simulations with an unconstrained demand. Table 6-6 Solutions when powering with PV or diesel

PV Cap. Mult. 1.00 0.50 0.10 0.01

Gen1 (kW) 7

PV (kW) 15.0 15.0 15.0 25.0

Battery 35 35 35 25

Total capital $ 7,000

Converter (kW) 7 7 7 7

Total NPC $ 94,075

COE ($/kWh) 0.414

Total capital $ 115,695 $ 61,695 $ 18,495 $ 8,155

Renewable fraction 0.00

Total NPC $ 157,701 $ 97,396 $ 49,152 $ 36,822

Capacity shortage 0.16

COE ($/kWh) 0.714 0.441 0.222 0.167

Diesel (L) 6,282

Renewable fraction 1.00 1.00 1.00 1.00

Gen1 (hrs) 3,285

From this table it can be seen that supplying a larger load as with the unconstrained demand, the NPC for the diesel system is much lower than the PV system with multiplier 1.0. The price of PV systems must be half the present price to be able to compete with diesel systems. Because of the much higher net present cost for PV with multiplier 1.0, the break even grid extension distance will be longer for this option compared to the diesel system. From the simulation results it can be seen that the unmet load and capacity shortage is almost the same for the two options, but that the excess electricity is much higher during the day for the PV option.

As a small conclusion of the solutions for a small trading centre, the following can be said. Diesel generation is the cheapest alternative, but if the price of diesel rise (which is very likely with today’s oil situation) then solar systems may be a cost effective way of producing electricity. Especially for a small trading centre, with a load profiles with large peaks and little ground load, PV systems are a viable option if the relative high investment cost can be covered. When adding environmental concerns, diesel generation comes out negatively. If the trading centre is out of a certain distance from the main grid, grid extension is not a good alternative as a start. It can be seen by the simulation result that the break even grid extension distance is very short meaning that for such a small community and scattered load, stand alone systems is a very good alternative.

If an investment is done in this trading centre according to the unconstrained demand, but the customers only can afford according to the constrained demand, alternatively,

Britt-Mari Langåsen

Spring 2004

Capacity shortage 0.19 0.19 0.19 0.19

6 Basis and optimal solutions

47

that fewer customers connect to the system, then this will represent a risk for the investor. From the simulation results it will be seen that the total capital cost will increase from $7000 to $8000, but the total NPC will decrease from $94000 to 40000. This is quite a difference and constitutes a large risk. In this case a positive risk, but if the investment costs where based on the constrained demand, price difference could constitute a large problem. In this case the system dimensioned for a constrained demand would probably be able to serve an unconstrained demand with the difference that the generators would have longer operating hors and consume more diesel. Some more unmet load could be calculated. The cost of energy for the constrained demand is $0.454 compared to $0.414 for the unconstrained demand.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

48

6.3 Medium trading centre To represent a medium trading centre, Mutunda trading centre has been chosen. This trading centre has according to the census of 2002 9704 inhabitants [25]. To see the load profile for Mutunda, se Figure 5-2. Mutunda trading centre has in the unconstrained demand a load during most parts of the day, but for the constrained demand there is a small peak during the morning, a larger one during the middle of the day and a high peak during evening time. In the medium trading centre, each consumer does not represent a large part of the total number of consumers, as it did in the small trading centre.

Following are the simulation results for the medium community. Here it will be shown that the optimal solution is much more dependent on the diesel price than the small trading centre. But still the fraction of solar energy is very much dependent on the price of photovoltaic as well.

6.3.1 General solutions The optimal solutions when there is a constrained demand are as given in Table 6-7. The energy demand used in the simulations is 545kWh/d and a 194kWp. The input parameters can be seen in appendix 11.2. Table 6-7 Optimal solutions for Mutunda with a constrained demand

PV Cap. Mult. 1.00 0.50 0.10 0.01 PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW)

100 350

Gen1 (kW) 66 66 66 66

COE ($/kWh) 0.372 0.372 0.250 0.192

Renewable fraction 0.00 0.00 0.76 0.93

Britt-Mari Langåsen

Gen2 (kW) 66 66

Capacity shortage 0.06 0.06 0.12 0.11

Battery

Converter (kW)

250 250

50 50

Diesel (L) 78,32 78,32 21,671 17,598

Gen1 (hrs) 4,38 4,38 1,147 932

Total capital $ 33,000 $ 33,000 $ 157,095 $ 102,055

Total NPC $ 917,137 $ 917,137 $ 588,957 $ 453,560

Gen2 (hrs) 733 733

Spring 2004

6 Basis and optimal solutions

49

As can be seen from the result, there will be no PV in the two first systems, resulting in the exact same solutions. When the price is down to a hundred there will be a lot of excess electricity as will be seen later and the system will be over dimensioned. The optimal solution for Mutunda trading centre when there is a constrained demand is therefore to supply it by diesel generation.

The optimal solutions when there is an unconstrained demand are as in Table 6-8. The demand in the calculations has been set to 1393kWh/d and 190kWp. The energy demand in the unconstrained demand is then more than twice the energy demand than in the previous example. This can foremost be seen in the number of operating hours for the diesel generators. Table 6-8 Optimal solutions for Mutunda with an unconstrained demand

PV Cap. Mult. 1.00 0.50 0.10 0.01

PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW)

Gen1 (kW) 40

50 300 450

COE ($/k Wh) 0.323 0.328 0.253 0.197

Renew able fractio n 0.00 0.21 0.70 0.79

Gen2 (kW) 40 66 66 66

Capacity shortage 0.17 0.20 0.17 0.17

Battery 1 10

Diesel (L) 157,785 134,881 92,469 86,723

Converter (kW) 0.6 50.0 150.0 150.0

Total capital $ 27,835 $ 243,495 $ 330,595 $ 108,355

Gen1 (hrs) 8,760

Gen2 (hrs) 5,344 8,376 5,961 5,569

Total NPC $ 1,856,216 $ 1,824,035 $ 1,421,868 $ 1,106,953

In this solution it can be seen that PV power is used in all solutions except the first option. Owing to the high energy demand, the generators have high levels of operating hours. It can be seen, that like in the previous examples, the price of solar systems must come down quite a bit to compete with the diesel systems.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

50

6.3.2 Load profiles and unmet load Following are figures to illustrate the load profile, unmet load, capacity shortage and excess electricity. For an explanation of these terms, see chapter 6.2.2. From Figure 6and Figure 6- it can be seen that for the PV multiplier 1.0 the load is almost fully met and with capacity shortages only at the highest peaks during the evening. There is some excess electricity, but not nearly as much as when the PV multiplier is 001.

Figure 6-11 Load profile from simulations with constrained demand. PV multiplier 1.0

Figure 6-12 Load profile from simulations with constrained demand. PV multiplier 0.01

For the unconstrained demand there is a much higher fraction of unmet load and capacity shortage when the PV multiplier is 1.0 compared to the same multiplier in the constrained demand. The load profile also has a higher load factor and consumes much energy. According to this simulation there will be many customers that have to be load shed during the high peaks. In the example where the PV multiplier is 0.01 there is a lot of excess electricity that has to be dumped.

Figure 6-13 Load profile from simulations with unconstrained demand. PV multiplier 1.0

Britt-Mari Langåsen

Figure 6-14 Load profile from simulations with unconstrained demand. PV multiplier 0.01

Spring 2004

6 Basis and optimal solutions

51

6.3.3 Break even grid extension distance The break even grid extension distances are naturally longer for this medium trading centre compared to the small community. There are also large differences in distances between the actual cost of PV and the option with PV multiplier of 0.01. (28 km compared to 9 km) It can though be seen that for a trading centre of this size, the grid can e extended quite far and it is still more economic than to build a stand alone system.

Figure 6-15 Break even grid extension distance. Constrained demand, PV multiplier 1.0

For the unconstrained demand the break even distances are even longer (50 km and 22 km) because the systems are here even larger and therefore the systems more expensive. Because of its

Figure 6-16 Break even grid extension distance. Constrained demand, PV multiplier 0.01

size and high energy demand, the community use a lot of diesel with gives the high net present cost. This means that grid extension is really a viable option, given that the load profile will be more or less as expected during the simulations. The fact that the trading centre has a relatively high load factor during the day when if there is an unconstrained demand, also makes it a good candidate for grid extension. To have a high load factor means that the average load is close to the maximum load. The load factor is not very high here, but definitely

Figure 6-17 Break even grid extension distance. Unconstrained demand, PV multiplier 1.0

higher that in the other options.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

52

If the break even distances for the constrained demand

is

compared

to

those

for

the

unconstrained demand it can be seen that they are

half

the

distance

compared

to

the

unconstrained. This seems logical since the load is less than half to.

Figure 6-18 Break even grid extension distance. Unconstrained demand, PV multiplier 0 01

6.3.4 Variations in diesel price For the medium trading centre there have also been done simulations for what happens when the diesel prices rise. The same increases in diesel price as in chapter 5.3.1 have been used. Cost as a function of diesel price Constrained 2500000 2000000

US$

1500000 1000000 500000 0 Total capital

Total NPC DG1

DG2

DG3

Figure 6-19 Cost as a function of diesel price when there is a constrained demand Table 6-9 Optimal results for variations in diesel price

PV (kW) 1DG 2DG 3DG

5 5

Gen1 (kW) 40 66 66

Gen2 (kW) 66 40 40

Gen3 (kW)

Battery

Converter (kW) 5 5

With the constrained demand in the medium trading centre, the simulations show that when the diesel price rises, there will be possibilities for solar systems to be economically viable. In the constrained demand the installed PV will not cover much of

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

53

the load and it can be seen from Figure 6- that the total NPC will be high for the energy systems as the diesel price rise. This is as mentioned before mostly due to the operating cost for the diesel generators.

Cost as a function of diesel price 5000000 4500000 4000000

US$

3500000 3000000 2500000 2000000 1500000 1000000 500000 0 Total capital

Total NPC DG1

DG2

DG3

Figure 6-20 Cost as a function of diesel price when there is a constrained demand Table 6-10 Optimal results for variations in diesel price

1DG 2DG 3DG

PV (kW) 5 5 90

Gen1 (kW) 40 4

Gen2 (kW) 40 7 40

Gen3 (kW) 66 24

Battery 5 2 10

Converter (kW) 10 5 75

For the unconstrained demand there is a dramatic change in the installed capacity of PV when the price of diesel rise to three times the present. As can be seen from Figure 6the total NPC has not increased as much between change in diesel price from US$1.6 to US$2.4 as between the price lift from US$0.8 to US$1.6. This is because now half of the peak load can be covered by the installed effect of PV, which has very low running costs that in turn affect the net present cost.

6.3.5 Only PV and diesel The following results are from simulations done if only diesel generation or solar systems where available. The main conclusion that can be drawn from this is that in general, diesel generation is the cheapest alternative.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

54

If there is a constrained demand, the following results was achieved: Table 6-11 Solutions when powering with PV and diesel

PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW) 125 125 125 200

Gen1 (kW) 40

Battery 400 400 400 250

Gen2 (kW) 66

Converter (kW) 100 100 100 100

Total capital $ 30,240

Total capital $ 1,170,995 $ 632,995 $ 202,595 $ 89,855

Total NPC $ 777,266

COE ($/kWh) 0.340

Total NPC $ 1,558,089 $ 956,495 $ 475,220 $ 352,859

Renewable fraction 0.00

COE ($/kWh) 0.717 0.440 0.219 0.162

Capacity shortage 0.15

Renewable fraction 1.00 1.00 1.00 1.00

Diesel (L) 65,917

Gen1 (hrs) 3,222

Capacity shortage 0.19 0.19 0.19 0.19

Gen2 (hrs) 1,891

The total NPC for supplying the Mutunda community with only solar power is much higher than with diesel, about twice the cost, and the upfront investment cost is much higher.

If there is an unconstrained demand the following results is achieved: Table 6-12 Solutions when powering with PV and diesel

PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW) 325 325 325 675

Gen1 (kW) 88

Battery 600 600 600 400

Total capital $ 19,412

Converter (kW) 150.0 150.0 150.0 100.0

Total NPC $ 2,048,261

Total capital $ 2,978,495 $ 1,560,495 $ 426,095 $ 154,155

COE ($/kWh) 0.344

Total NPC $ 3,814,217 $ 2,228,461 $ 959,856 $ 642,310

Renewable fraction 0.00

COE ($/kWh) 0.703 0.411 0.177 0.117

Capacity shortage 0.13

Renewable fraction 1.00 1.00 1.00 1.00

Diesel (L) 178,878

Gen1 (hrs) 8,760

For the unconstrained demand the solution is more or less the same as in the above example. The total NPC is much higher for the PV system and the investment cost is almost infinitely higher. A thought is though that in these simulations the costs for the solar systems is the cost for solar home systems. In reality, if such a large trading centre would be supplied with power from solar power, it would probably be done in another way.

Britt-Mari Langåsen

Capacity shortage 0.19 0.19 0.19 0.20

Spring 2004

6 Basis and optimal solutions

55

If investments where made and another demand than the predicted would happen for the medium trading centre, then a similar result as he one in the small trading centre would occur, but with larger mounts. If the demand is calculated to be the unconstrained, but the constrained occur, then the investment costs would be more ($33000 to $28000), but the net present cost would be calculated to be much higher. As in the previous example, the net present cost would be only half, $917000 compared to 1 850 000. If the opposite occurs, then the expenditures will be higher, but the revenue from more customers will also be there. As there is two 66 kW generators in the constrained demand, but only two 40 kW generators in the unconstrained demand, it can be assumed the load from an unconstrained demand could be covered with the constrained solutions and that like previous, the generators would operate for longer hours. The cost of energy is more expensive for the constrained demand, $0.372 compared to $0.323 based on respective demands.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

56

6.4 Large trading centre To represent a large trading centre, Kiryandongo has been chosen. This trading centre had according to the census of 2002 21852 inhabitants. [25] The load profile for Kiryandongo can be seen in Figure 5-3. Kiryandongo has two large peaks during morning and evening and high energy consumption.

6.4.1 General results The optimal solutions for Kiryandongo with a constrained demand are given in Table 6-13 and in the simulations a energy demand of 3.8MWh/d and a peak of 1241kWp have been used. The input parameters can be seen in the appendix.

Table 6-13 Optimal solutions for Kiryandongo, constrained demand

PV Cap. Mult. 1.00 0.50 0.10 0.01 PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW)

Gen2 (kW) 120 120 200 200

Gen3 (kW) 66 66 200 200

Converter (kW)

1000 2500

Gen1 (kW) 500 500 160 160

COE ($/kWh) 0.293 0.293 0.258 0.183

Renewable fraction 0.00 0.00 0.74 0.89

Capacity shortage 0.13 0.13 0.19 0.19

Diesel (L) 440,163 440,163 245,409 198,662

500 800

Total capital $ 86,160 $ 86,160 $ 1,154,325 $ 601,485

Total NPC $ 4,738,761 $ 4,738,761 $ 3,888,303 $ 2,756,877

Gen1 (hrs) 1,748 1,748 2,875 2,037

Gen2 (hrs) 3,170 3,170 1,489 1,112

Gen3 (hrs) 4,430 4,430 845 766

The optimal solutions for Kiryandongo with a constrained demand are given in Table 6-13 and in the simulations a energy demand of 6.3MWh/d and a peak of 1196kWp have been used. The input parameters can be seen in the appendix.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

57

Table 6-14 Optimal solutions for Kiryandongo, unconstrained demand

PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW) 5 5 1000 1200

Gen1 (kW) 160 160 40 40

Gen2 (kW) 120 120 160 160

Gen3 (kW) 200 200 120 120

Battery 75 75 75 50

Converter (kW) 100 100 1000 1000

Total capital $ 158,135 $ 148,135 $ 1,309,975 $ 533,985

PV Cap. Mult. 1.00 0.50 0.10 0.01

COE ($/kWh) 0.298 0.298 0.259 0.227

Renewable fraction 0.00 0.00 0.65 0.70

Capacity shortage 0.20 0.20 0.20 0.20

Diesel (L) 661,66 661,66 355,625 339,778

Gen1 (hrs) 4,943 4,943 3,541 3,333

Gen2 (hrs) 6,377 6,377 3,564 3,476

Gen3 (hrs) 2,767 2,767 4,191 3,952

Total NPC $ 7,398,032 $ 7,388,032 $ 6,379,532 $ 5,588,750

The most apparent with the results from the simulations of Kiryandongo trading centre, is that the break even grid extension distance is very far. This indicates that supplying a rural trading centre of this size with a stand alone system as designed in this simulation is very expensive. This becomes more obvious when comparing the energy demand to that of Biizi trading centre, which represents a small community. The energy demand here is under the constrained condition over 4 MWh/day, compared to xxx. This gives a much higher demand for energy which is a corner stone for making grid extension economically viable.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

58

6.4.2 Load profiles and unmet load During the high peaks in the evening there seem to be some unmet load and capacity shortage for a system with constrained demand and a PV multiplier of 1.0. When the PV price goes down to a hundred there is a lot of excess energy but also some capacity shortage and unmet load even here. This means though that there is a lot of solar power and diesel generation installed, it is not enough to fully cover the accumulated effect. In the simulations a maximum annual capacity shortage is set to 20%.

Figure 6-21 Load profiles for simulation with constrained demand. PV multiplier 1.0

Figure 6-22 Load profiles for simulation with constrained demand. PV multiplier 0.01

When there is an unconstrained demand, it seems to be a higher fraction of capacity shortage and unmet load. When the PV price is low, there is not so much excess energy as with the constrained demand. It can be seen from the tables above that there is less capacity intalled in the unconstrained situation, which can be seen from these figures.

Figure 6-23 Load profiles for simulation with unconstrained demand. PV multiplier 1.0

Britt-Mari Langåsen

Figure 6-24 Load profiles for simulation with constrained demand. PV multiplier 0.01

Spring 2004

6 Basis and optimal solutions

59

6.4.3 Break even grid extension distance In the constrained option, the break-even grid extension distance is shorter, and this is reasonable because the energy consumed is less, even though the peak effect is almost the same.

Figure 6-25 Break even grid extension distance. Constrained demand, PV multiplier 1.0

Figure 6-26 Break even grid extension distance. Constrained demand, PV multiplier 0.01

As can be expected the break even grid extension distances from the simulations with an unconstrained demand are very long. This means that the best way of supplying power to a trading centre with a load demand as this will be by grid extension. At least if the community is within 200 km from the existing grid. This will be the situation for most trading centres in Uganda. If Figure 6-27 Break even grid extension distance. Unconstrained demand, PV multiplier 1.0

the price on solar systems fall there is still a good chance that the best way for supplying electricity is through grid extension since the break even distance is 120 km when the PV multiplier is down to 0.01.

Figure 6-28 Break even grid extension distance. Unconstrained demand, PV multiplier 0.01

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

6.4.4

60

Variations in diesel prices

As in the previous chapters, simulations with a changing diesel price have been done for Kiryandongo trading centre as well. The prices are still the same, US$0.8 to US$2.4.

Cost as a function of diesel price 16000000 14000000 12000000 US$

10000000 8000000 6000000 4000000 2000000 0 Total capital

Total NPC DP1

DP2

DP3

Figure 6-29 Cost as a function of diesel price when there is a constrained demand Table 6-15 Optimal results for variations in diesel price

PV (kW) DP1 DP2 DP3

100

Gen1 (kW) 500 500 500

Gen2 (kW) 120 66 66

Gen3 (kW) 66 120 120

Battery

Converter (kW)

100

For the profile with a constrained demand the price have to be three times the present price for it to be economically viable with photovoltaic. The load will even though be covered mostly by diesel generators. By looking at the simulation results in HOMER it can be seen that there will be a little excess electricity in the morning and some capacity shortage and unmet load in the high peaks in the evening when the diesel price is three times the present price. The break even distance for the same diesel price is 477 km.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

61

Cost as a function od diesel price Unconstrained 25000000

US$

20000000 15000000 10000000 5000000 0 Total capital

Total NPC DP1

DP2

DP3

Figure 6-30 Cost as a function of diesel price when there is an unconstrained demand Table 6-16 Optimal results for variations in diesel price

DP1 DP2 DP3

PV (kW) 5 75 350

Gen1 (kW) 160 120 200

Gen2 (kW) 120 160 24

Gen3 (kW) 200 200 160

Battery 75 75 100

Converter (kW) 100 50 300

As in the previous examples, solar systems are more viable when there is a large load that has to be covered by diesel generators that demand a lot of fuel. For a diesel price three times todays price, there will be quite a lot of unmet load and capacity shortage during the evening peaks. This can be seen also from Table 6-16 where the installed effect is under 450kW. This can be compared to the peak effect of 1241kWp for Kiryandongo trading centre. The break even distance for grid extension is 675 km.

6.4.5 Only PV or diesel These are the results from the simulations with a constrained demand where only PV or diesel generation supply power. Table 6-17 Solutions when powering with PV and diesel

PV Cap. Mult. 1.00 0.50 0.10 0.01

PV (kW) 900 900 900 2200

Batteries 1900 1900 1900 1000

Britt-Mari Langåsen

Converter (kW) 700 700 700 750

Total capital $ 8,425,995 $ 4,477,995 $ 1,319,595 $ 605,855

Total NPC $ 10,820,137 $ 6,404,914 $ 2,872,735 $ 1,963,513

COE ($/kWh) 0.719 0.426 0.191 0.131

Renewable fraction 1.00 1.00 1.00 1.00

Spring 2004

Capacity shortage 0.20 0.20 0.20 0.20

6 Basis and optimal solutions

Gen1 (kW) 500

Gen2 (kW) 66

Gen3 (kW) 120

Total capital $ 86,160

Total NPC $ 4,738,761

62

COE ($/kWh) 0.293

Renewable fraction 0.00

Capacity shortage 0.13

Diesel (L) 440,163

Gen1 (hrs) 1,748

Gen2 (hrs) 4,430

Gen3 (hrs) 3,170

From the results of the simulations done to see the optimal solutions when supplying power with only PV or diesel generation, it can be seen that diesel generation is the most economical alternative. Diesel generation has less than half the net present cost compared to PV with its present cost. This is a realistic result since the investment costs for diesel is much lower than solar systems and it is a large load to be covered.

The results from the simulations for Kiryandongo with an unconstrained demand and only PV systems or diesel generation can be seen below. Table 6-18 Solutions when powering with PV and diesel

PV Cap. Mult. 1.00 0.50 0.10 0.01

Gen1 (kW) 200

Gen2 (kW) 160

PV (kW) 1500 1500 1500 1800

Gen3 (kW) 160

Converter (kW) 900 900 900 900

Batteries 2400 2400 2400 2000

Total capital $ 99,960

Total NPC $ 7,406,441

Total capital $ 13,850,995 $ 7,262,995 $ 1,992,595 $ 773,155

COE ($/kWh) 0.296

Total NPC $ 17,539,196 $ 10,171,485 $ 4,277,317 $ 2,895,017

Renewable fraction 0.00

COE ($/kWh) 0.718 0.417 0.175 0.118

Capacity shortage 0.19

Renewable fraction 1.00 1.00 1.00 1.00

Diesel (L) 679,925

Gen1 (hrs) 2,628

The same conclusion as above can be drawn for the results from the simulation with an unconstrained demand. Diesel is by far the cheapest option if only one energy source is to be used. It though seem like there is less unmet load for the PV system.

Like in the previous chapters for a small and medium trading centre it is considered what would happen if the load would differ from the assumed one. Here the results are different and if a constrained demand would occur instead of the unconstrained, then both the investment cost and the NPC would be lover, more or less half the cost for he unconstrained solution. ($86000 and $4 700000 compared to $158000 n $7 400 000 for the solution for the unconstrained demand.) The reason for the high investment cost is the use of photovoltaic.

Britt-Mari Langåsen

Capacity shortage 0.20 0.20 0.20 0.20

Spring 2004

Gen2 (hrs) 8,379

Gen3 (hrs) 3,218

6 Basis and optimal solutions

63

The solution or the unconstrained demand has less installed effect, but since the peak effect it is derived from is almost the same as for the constrained demand, then it is assumed that it could cover the load for the constrained demand and vice versa with the difference in operating hours or the diesel generators. The cost of energy is almost the same for the two options.

Britt-Mari Langåsen

Spring 2004

6 Basis and optimal solutions

64

General results After seeing the results from all these simulations some major conclusions can be drawn. These are: For a small trading centre, the best way to supply it with power is through diesel generation. For a medium and large trading centre, the best way if grid extension is no option, is through diesel generation with a bit of solar power. If the diesel price rise, then the larger trading centres that has a large load and use a lot of diesel, should install some solar power. This will give quite high investment costs, but will lower the net present cost. Some diesel generation will still be installed. If only solar power or diesel generation is available, then diesel generation is by far the cheapest alternative with the current prices on photovoltaic. If the prices on PV systems go down to 50%, then these systems can compete with diesel generation. Still the investment cost will be much higher than for the diesel option, but the net present cost will then be about the same and there are also positive effects because of no pollution and noise. The risk for an investor is quite large if the load will differ like between the constrained and unconstrained demand.

To sum up the findings in this result chapter, a table over the results for the general solutions for the three trading centres will be given. These are all results with a PV multiplier of 1.0 and represent the cost for supplying a rural trading centre with the given assumptions in this report. Table 6-19 Summary of the solutions for rural electrification, PV multiplier 1.0 PV (kW) Biizi constrained Biizi unconstrained Mutunda constrained Mutunda unconstrained Kiryandongo constrained Kiryandongo unconstrained

Gen1 (kW)

Gen2 (kW)

4

4

Gen3 (kW)

Battery

Converter (kW)

7

5

66

66

40

40

500

120

66

160

120

200

Britt-Mari Langåsen

1

75

0.6

100

Total capital

Total NPC

COE ($/kWh)

$ 8,000

$ 40,015

0.453

$ 7,000

$ 94,075

0.414

$ 33,000

$ 917,137

0.372

$ 27,835

$ 1,856,216

0.323

$ 86,160

$ 4,738,761

0.293

$ 158,135

$ 7,398,032

0.298

Spring 2004

7 Grid extension

65

7 Grid extension In this section there will be a general discussion about grid extension. The applicability of grid extension is discussed and protection systems for the network parts are drafted. Some questions for further studies if a grid extension is relevant are stated finally.

7.1 Applicability Grid extension is probably the most viable alternative in the

Picture 7-1 Warning sign at a grid pole in Kampala

long term in a larger electrification program for Uganda. Stand alone systems may be the optimal technology option for small and remote villages, but as the trading centres become larger it can be seen that the break-even grid extension distance is quite long, specially if the load is relatively high and has a high load factor. It is said in Uganda’s Strategy for Accelerating Grid-based Renewable Power Generation for a Clean Environment that “The major challenge for Uganda is to ensure that the grid network is strengthened to have the necessary capacity to handle the additional power.” [19] Additional power will probably come from construction of new hydro power sites. Considering that Uganda in several parts of the country have a problem with sustaining voltage levels and high transmission losses, there is lot of work to be done on the rehabilitation of the grid. Some of this work has been initiated by UETCL with the help of SWECO that have done a study on the transmission and sub-transmission network in Uganda. According to this study, the potential load growth in rural areas is high and is determined by the changes in electrification policy more than changes in consumer behaviour. [1] As mentioned in chapter 4.3.3, the influx from people moving from rural areas to urban areas will change the need for sub transmission lines and grid extension to certain communities may be unnecessary.

The national grid in Uganda is already now under high stress because of little generation capacity and transmission lines that have been without proper maintenance. Masindi district is far from the main power generation sites that are situate not far from Kampala and therefore have problems with keeping the bus bar voltage at an acceptable level. The long and radial lines are able to carry very little load before they cause severe

Britt-Mari Langåsen

Spring 2004

7 Grid extension

66

voltage problems. For this reason autotransformers of 5 MVA are installed at Hoima and Busunjo to boost up the voltage [31] but these have only short-term positive effects. Since UEDCL is to be contracted out to the private sector, no formal expansion plans are available because these will be the responsibility of the new owners.

Many rural areas face high transmission and distribution costs for several reasons: the capacity of power lines is inefficiently used because of low population, densities and demands are low, villages may have very peaky demand profiles and line losses tend to be high. [28] The existing 33kV and 11kV subtransmission network is basically a distribution network with long radial feedings and characterized with low, rural loads.

There are several reasons to expand the transmission system as mentioned in chapter 4.3.3. Some of these were rural electrification, to increase security and to reduce losses and improve operational economy. The existing network may need to be upgraded and reinforced with new lines, substations and voltage equipment due to the fact that the lines are going to be overloaded or have an unfeasible capacity because of an increased transfer demand. Some other reasons is that voltage level is too low for the transfer demand or the reactive balance in not optimal which in turn increase losses in the system. [1]

As mentioned a reason for the transmission losses in the Ugandan grid is due theft of electricity directly from the grid. About 30% of the losses in the Ugandan grid are caused by non-technical losses like theft. There is also a problem with theft and vandalism of conductors and poles in some areas. It is therefore important with an anti-theft design when planning for new lines. This can be to use a higher voltage level than most electrical appliances work on so that it will be impossible to use the electricity without transformers. [29] Other ways to protect from theft are:[1] ♦

To use anti-vandalism bolts in the lattice steel structures combined with a friction bond. These bolts cannot be removed after being put into place and the tower cost is affected with less than 1%.

♦

To use tubular steel towers that cannot be disassembled. These towers use little land and are good for populated areas, but due to their slim design they cost up to 30-50% more than standard lattice steel structures

♦

To use wooden poles with metal crossbars. They are not attractive for thefts and they also have the lowest investment cost. These poles require high maintenance and can

Britt-Mari Langåsen

Spring 2004

7 Grid extension

67

become rotten or eaten by termites, but this can be minimised by chemical treatment. They are not recommended for transmission system level. ♦

To use concrete poles. These poles have a high cost, 50% more than lattice steel towers, and must be manufactured in one piece, but they are not attractive for theft and are slim and use little land.

♦

To use welded lattice steel structures that cannot be disassembled and hence not so easily stolen. They are expensive because they have to be welded at the manufacturer making transport difficult. They are not common and favoured by utilities.

♦

To avoid ACSR conductors where the pure aluminium is attractive. It is possible to use AAAC conductors instead which cannot be used in the process for pure aluminium. These conductors are more costly and for a transmission line the use of these will increase the total line cost by 3%.

The most realistic anti-vandalism solutions in Masindi is to use anti-vandalism bolts in steel lattice towers or to use wooden poles. Also the use of AAAC conductors is recommended. [1]

7.2 Protection schemes To have a more reliable system, a meshed system should be considered. This will demand protection schemes so that only the part of the grid with the fault will be out, and the meshed system will ensure electricity to such a large portion of the grid as possible. When extending the grid, several protection schemes are needed for the various network elements. The selection of protection schemes has to be made with the consideration of availability of protection systems in Uganda and it should also be cost effective. For the 132kV and 220kV system two independent protection systems based on different measuring principles should be used. The 132kV lines are presently equipped with the following protection [1]: ♦

A main protection that is a distance protection with three forwarding zones, reverse zone and autoreclosing relay operation on three-phase autoreclosing scheme. This scheme has a teleprotection scheme for operation of the remote feeder breaker in case of zone one fault.

♦

A back-up protection that consist of static three-phase over current protection and a directional earth fault protection scheme. 1

♦

Supplementary protection equipment like a fault locator for indication of the distance to a fault. This is part of the overhead line protection scheme.

Britt-Mari Langåsen

Spring 2004

7 Grid extension

68

These protection schemes are acceptable and comply with safety requirements normal for the voltage level. As a modernisation multi functional microprocessor controlled line protection relays or additional line protection systems could be introduced. An increased telecommunication capacity could allow upgrading of existing protection systems or implementation of additional protection schemes that could increase protection reliability and selectivity and to decrease tripping time.

For the 33kV lines, the protection scheme could be chosen depending on the importance of the line. If the line is working as a transmission and supply line for a larger area with substations, a protection system as one for a 132kV could be considered. If the line is less important, a downgrading of the protection system to only a distance or over-current protection may be an option. It is recommended to supplement the protection scheme with an auto-reclosing function to avoid total shut down in case of temporary faults.

There is also a need for protection of the transformers for transmission and subtransmission level. The existing transformers are equipped with the following main protection functions and this is considered sufficient [1]. ♦

Current differential relay

♦

Restricted earth fault relay for star windings

♦

Three phase over current and earth fault protection

♦

Bucholz relay

♦

Suddenoil pressure relay

♦

Winding temperature

♦

Oil temperature

The reliability in the Ugandan grid is low compared to European figures, but it could be improved by relatively simple measures like bush cuttings, keeping the way of lines clear of trees that may fall over the lines and replacement of wooden poles. According to the district manager for UEDCL in Masindi, the most common fault that makes lines fall out is rain, thunder, wind etc. There are most external faults that make Masindi district loose power. When the new line from Apac to Masindi will be conducted, there will be some more security concerning power reliability and hopefully less loss of mains. There are several issues that need to be addressed when extending the transmission and sub-transmission grid. Some of these may be:

Britt-Mari Langåsen

Spring 2004

7 Grid extension

69

•

How does new load and new lines affect the existing grid?

•

How do added generating units affect the existing grid?

•

Will there be a need for more power or is rehabilitation of the grid enough with booster transformers etc when new loads in Masindi is added?

These were questions that were thought to be answered by doing a load flow analysis with the program SIMPOW by ABB. The data for the grid would be put in to a single line model and then a load flow analysis could be run. By adding new branches it is possible to se how new loads affect the grid and from that it would also possible to se in what order the grid should be expanded. Instead, this could be recommended to do as a further study if a rural electrification by grid extension is to proceed.

Britt-Mari Langåsen

Spring 2004

8 Discussion

70

8 Discussion Rural electrification is included in the energy policy of Uganda and it is stated that they wish to enhance the living conditions of the rural population and to reduce the inequalities in national access to electricity and social welfare. Rural electrification by means of renewable energy and the use of the natural resources in Uganda is also promoted. This means principally the use of water, solar and biomass. Rural electrification programmes are all over the world and it is a big challenge for the countries concerned to meet the future energy demands in a sustainable way.

A change in technology is linked to economical change which in turn is linked to social change. [21] Given that most people may at first be reluctant to social change, it is not hard to understand that a change in technology options is hard to drive through. This can be seen in for example the use of charcoal stoves that are used even though some people have access to both electrical stoves and gas stoves. In a larger scale, people may be worried that introducing new forms of energy will affect places of work. For example it can be mentioned that if the technology of making coal from bagasse would be subsidised and promoted, then several people that are occupied with the work of making charcoal would be out of work. Of course it would be more environmentally sound to use coal from bagasse, but as mentioned people are reluctant to change, especially when jobs are in danger.[30]

There are several reasons why rural electrification programs have a hard time becoming viable. Some of these are [21]: •

Several projects compete for the same capital

•

The projects are evaluated according to net present cost

•

The cost of capital depend on associated time and risk

All together this means that projects that are known and considered safe, may get capital. In this context, this would mean that diesel generation and grid extension would be favoured compared to photovoltaic. Connected with a high interest rate, solar systems that have a high initial cost get a non-favourable position as well. This means that the Government, as an incentive to promote sustainable energy generation, could, if

Britt-Mari Langåsen

Spring 2004

8 Discussion

71

possible, put a lower interest rate on renewable energy projects as compared to diesel system projects.

Uganda is in a special position as power generation is concerned because almost all of the power produced in Uganda is from hydro power which is considered as a renewable energy source. One problem is though that several of the new hydro sites that are under consideration are situated in national parks. The energy that will be produced there will be renewable, but will it be considered sustainable? Several environmentalist groups have protested against this project. There is also a problem of getting financing from the World Bank and foreign governments for the building of the hydro station at Bujagali falls for example. The reason for this is the political risk connected to the governmental stability in Uganda. Of course some of the planned hydro power sites are outside national parks and not as large as the Bujagali falls project. Considering all this, solar systems is put into a more favourable position compared to grid extension where the power is produced from hydro power. However, more electricity could be produced from other energy sources like biomass or geothermal energy, making grid based electricity both renewable AND sustainable.

A problem one has to face when investing in new energy systems is concerning who should be responsible for operation and maintenance, who will pay for the system and will the customers come? This problem is relevant for all the technologies. For grid extension, it is relatively easy to say that the grid utility will pay for the cost of extension, and the customers will pay an interconnection fee and the price of energy used. But should the interconnection fee be subsidised to promote rural electrification? And when it comes to the electrical installations in the houses, who should pay for that? Should it be included in the interconnection fee or separate? To ensure as many customers as possible to connect to a new grid, all cost should probably be included as one. This usually makes fees look less, and it becomes a bit like a plug-and-play package. Solar systems are another question. If small solar home systems are considered, then each customer could be responsible for their own system. Because of the high investment cost, an official payment plan from an investing bank could be an option. This could be a solution if a rural electrification program by solar power is promoted by the government .A solution with a large PV system with a mini grid or where people can come and charge their batteries, will still give a cost for the customer since the electric

Britt-Mari Langåsen

Spring 2004

8 Discussion

72

installations in the house has to be done, and a battery has to be bought, but the large investment cost for the solar panels will be in the hands of a large investor. Power from diesel generation should probably be phased out since it is a nonsustainable solution. On the other hand it is an available technology in Uganda and has also the lowest investment costs. If the diesel prices do not rise much in the near future, it can be seen from the simulations that diesel is the most viable alternative. It could function as a start for raising the living conditions for rural people and be phased out with sustainable technologies further ahead.

A problem is for an investor to assess how much he is willing to pay to avoid a power shortage. Since the area to be powered is mainly a residential area, with few or no industries, then this sum is probably quite small. If there had been large industries dependent on electricity for their production, they might have been willing to pay more for power security and then there can be larger investments. Now, the customers will be regular households that still have not got the same standard and power requirements as the western world. There is for example no cost for undelivered energy in Uganda.

Some discussion could be made of the applicability of HOMER as a simulation tool and the input parameters that have been used. There is always an uncertainty of prices for the technology options and the resources. What is also introducing a large insecurity of the results in this study are the load profiles that the simulations are based upon. These have been done partly by collected data from Uganda, but also by own experience from daily life in Scandinavia and Uganda. As a justification, it can be said that the same grounds are used for all the load profiles so at least the simulations will be based on the same assumptions. The fact that HOMER can not have a development of prices in its simulations, have made the simulations time consuming and not always compatible with real life situations.

Britt-Mari Langåsen

Spring 2004

9 Conclusion

73

9 Conclusion When looking at the energy situation in the world today, and the third world in particular, it can easily bee seen that supplying people with reliable, clean and inexpensive energy is a major challenge. Even so, giving people in poor areas access to energy is a big step closer towards giving them a better future. Access to energy will lead to better health care, better education possibilities and maybe even lead to small scale industries that will give an economic development. Together, these things will lead to a better quality of life for the people concerned.

It can also be seen from contracts being created between both companies and governments that rural electrification is an important question in today’s world. As an example can be mentioned a contract between the Tanzanian energy company Tanseco, and the consultant company SWECO that are going to investigate and try to find solutions for rural electrification in Tanzania. [27] SWECO has also as a task to develop the transmission network in Uganda. [1] Other organizations that deal with this type of questions are the World Bank, NORAD and Danida, to mention some.

Power from photovoltaic generation is most applicable in smaller communities that are situated far from the main grid and have a dispersed load. Another application area for photovoltaic systems is when there are a number of small but important loads that have to be covered, for example vaccine refrigeration at a local health station, lighting at a school or security lights along the main street in a trading centre. A negative side with PV systems is the high investment cost for a relatively low effect. Many who install solar systems become disappointed because they do not know the limitations of the system and feel that the usage area of the electricity is smaller than wanted. For example does a kettle for heating water usually use more effect (about 1000 W) than a solar home system can produce. Other, larger, solar systems with mini-grids may be an option for rural electrification and could be subject to further studies.

Diesel generation is a widely available technology that is known in most areas. Diesel generators are used as back up system in many trading centres that are already connected to the grid. They are available in many different sizes and it is relatively easy

Britt-Mari Langåsen

Spring 2004

9 Conclusion

74

to find a generator that can fit the actual load. For a trading centre far away from the grid, there can be difficulties with the transport of diesel, making diesel generation less preferable in such a situation. If transport problems are an issue for a trading centre, then solar systems may be preferable. Diesel generation also have a negative environmental impact, with pollution and noise. Frequent maintenance is required for a diesel generator. It should though be easier to find a technician that can repair and do the maintenance of a diesel generator compared to a solar system because of the frequent use of generators.

Grid extension is most applicable when there is a large load to be covered or if the trading centre is very close to the main grid. As it could be seen by the simulations, the break even grid extension distance was quite long for both the medium and large trading centre. (From 9 km and up.) Grid extension may have environmental effects like deforestation and use of agricultural landscape but in Masindi this should not be a large problem. As mentioned before the Ugandan grid is already under stress because of too low generation capacity and transmission overloads. A new line is though being built from Apac in the east to Masindi, giving Masindi a higher level of security in case of congestions or loss of mains on the way up to Masindi, since it will now be fed with power from two different directions. The rural loads that might be added to the main grid through a grid extension are quite small. They will in any case not be built earlier than new generation capacity is planned by the government to be built.

As a further study, rural electrification of a certain trading centre could be an option. Then each household load etc. could be modelled and put in to another simulation tool also developed by NREL. This tool is called VIPOR and is an optimization tool for village power electrification system. This program, given a map over the village and data about the loads, can design what loads should be supplied by for example SHS and what loads should be connected to a grid. It is preferably used together with HOMER.

Another option is to further investigate the possibilities of using solar systems in a larger scale for rural electrification. Not only as solar home systems, but also as power generation for mini grids. Maybe solar thermal generation could be an alternative.

Britt-Mari Langåsen

Spring 2004

9 Conclusion

75

There are a lot of questions about financing rural electrification, but these have been partly out of this scope in this thesis. Means of financing and investment plans and investment risks could be the aim of further research.

It can be concluded that rural electrification is an important task to enhance the living standard for millions of people and that there is a lot of further research to be done to find optimal solutions for rural electrification in both Masindi district, Uganda and in general in the third world. In this study simulations to give an idea of the costs have been done, but to be really applicable, closer studies of the area to be electrified has to be done, and more requirements have to be given for the proceeding of an electrification of a rural area.

Britt-Mari Langåsen

Spring 2004

10 Literature references

76

10 Literature references [1]

“Consultancy services for UEB. Transmission and subtransmission study”, NDF 103-3 Power III, SWECO, Dec 2003

[2]

Arinda R., Okou R., “Technology options for powering Biizi trading centre – Masindi district”, Kampala, Uganda, March 2004

[3]

Bahati Godfrey, “Geothermal energy in Uganda, country update”, Entebbe, Uganda, International Geothermal Conference, Reykjavik, Sept 2003

[4]

Redwood-Sawyerr Jonas A S, “Widening access in the context of Power sector reform – an overview of the institutional challenges in Africa”, UNEP-IEA/AIE Meeting 21-22 May 2002, Paris, France

[5]

Solar radiation data for Masindi district, Ministry of Water, Lands and Environment, Department of Meteorology, Kampala, March 2004

[6]

Electricité de France, “Uganda load forecast review (Update 2001)”, January 2001

[7]

“HOMER, Getting started guide, Version 2.0”, May 2003, NREL, USA

[8]

Help in HOMER

[9]

Oparaku O. U., “Rural area power supply in Nigeria: A cost comparison of the photovoltaic, diesel/gasoline generator and grid utility options”, Renewable Energy 28 (2003) 2089-2098, Elsevier Science Ltd

[10] Christofides Constantinos, “Autonomous photovoltaic power system or connection with electrical grid? A preliminary feasibility study for small and isolated communities”, 1989, Solar Cells 26 (1989) 165-175, Elsevier Sequoia [11] Martzoukos S. H., Teplitz-Sembitzky W., “Optimal timing of transmission line investments in the face of uncertain demand, an option valuation approach”, 1992, Butterworth-Heinmann Ltd [12] Flowers Larry, “Renewables for sustainable village power”, Presentad at International Conference of Village Elctrification through Renewable Energy, New Dehli India, NREL, March 1997 [13] “Energy and Poverty”, IEA: world energy outlook 2002 [14] Kamese Geoffrey, “Renewable energy technologies in Uganda: The potential for Geothermal Energy Development”, March 2004 [15] Ijumba N. M., “Application of distributed generation in optimised design and operation of rural power supply networks” [16] Daley James M., Siciliano Robert L., “Application of emergency and standby generation for distributed generation: Part 1 – Concepts and hypotheses”, IEEE Transactions in industry applications, vol. 39 No. 4, July/August 2003

Britt-Mari Langåsen

Spring 2004

10 Literature references

77

[17] The Government of Uganda, “New strategy plan & implementation plan”, June 1999, Uganda [18] MEMD, “Rural Electrification Strategy and Plan Covering the Period 2001 to 2010”, The Government of Uganda, February 2001 [19] Bbumba S. N. M., “Uganda’s strategy for accelerating grid-based renewable power generation for a cleaner environment”, Uganda [20] “Energy after Rio: Prospects and Challenges”, Chapter 2 – Energy and Major Global Issues, UNDP, www.undo.org/seed/energy [21] Gether Kaare, “Transition to Large Scale Use of Hydrogen and Sustainable Energy Services, Choices of technology and infrastructure under path dependence, feedback and nonlinearity”, Doctoral thesis at NTNU 2004:71 [22] Langåsen, Britt-Mari, “Distributed generation in developing countries”, Fall 2003, NTNU [23] Bahati Godfrey, “Geothermal energy in Uganda, country update”, presented at the International Geothermal Conference, Reykjavik, September 2003, Entebbe Uganda. [24] HOMER webpage, www.nrel.gov/homer [25] Mean household size, OBOS, http://www.ubos.org/appendix3prov.pdf [26] Electricity Regulatory Authority (ERA) webpage, http://www.era.or.ug/prices.asp [27] SWECO webpage, http://www.sweco.se/templates/Page____11616.asp [28] Goldemberg José, “Rural energy in developing countries”, chapter 10, WEA: Energy and the challenge of sustainability [29] Andrew Mubonga, District Manager, Masindi district, UEDCL, personal contact [30] Terry Jobling, Engineering Manader, Kinyara Sugar Works Ltd, personal contact [31] Elisabeth Kabagante, Planning Engineer, UEDCL, personal contact [32] Nyanzi Joseph Kubo, Principal Planning Engineer, UEDCL, Uganda, personal contact [33] Fred Tugume, A.G. Coordinator for Geothermal Activities, Entebbe, Uganda, personal contact

Britt-Mari Langåsen

Spring 2004

11 Appendix

78

11 Appendix 11.1 Interviews This chapter will include typed interviews and meetings with persons met in Uganda. Addresses and other contact information where accurate in March 2004, but cannot be guaranteed in the future. Monday 16 feb: Meeting with Dr. Okure and Dr. da Silva. Dr. Izael Pereira da Silva Lecturer Dept. of Electrical Engineering Makerere University Tel: +256-41-250415 Mob: +256-77-5057902 [email protected] This was the first meeting at the university and we met Dr Okure and Dr Da Silva. They had expected someone studying solar power (Torill) and had made a program for that purpose during the day. It turned out that the communication between Makerere and NTNU had been somewhat insufficient and that they had no idea who where coming and what we where going to do research on and they had other expectations of our projects than we had. This made things a bit insecure and a not so promising start.

Monday 16 february: Meeting with Philippe Simonis Energy Advisor Amber Hose, Room A205 c/o GTZ Office Kampala P.O Box 10346 Tel/Fax: +256-41-234165 Mob: +256-75-791268 [email protected] Mr Simonis had expected someone studying solar energy. He was a bit disappointed when it turned out that no one studying that came to Makerere. He was not interested in my original project because he didn’t think it was feasible. Mr Simonis instead suggested that I should work with two of Dr. Da Silva’s students. They where going to look at a cost analysis for electrifying a trading centre in the Masindi district. I was a bit disappointed

Britt-Mari Langåsen

Spring 2004

11 Appendix

79

that they wanted to change my project and said that I had to talk to my advisor first. They had an understanding for this, but emphasised that they probably wouldn’t be able to help me with my original project, but if I wanted to go through with it, they would try to help me get in contact with the right people. Tuesday 17 february: Meeting with Rachel Arinda and Richard Okou. Rachel Arinda Mob: +256-71Richard Okou Mob: +256-71This was the first meeting with Rachel and Richard. They are two final year bachelor students at Institute for Electric Engineering at Makerere University. They where going to do a project for GTZ on the optimal way of electrifying a trading centre in Masindi. I received some papers about the simulation program HOMER that they where going to use and they told me what they knew about the project. They were going to travel up to Masindi already the following weekend, and we decided to meet again the next day to go to Mr Simonis as GTZ and to travel together up to Masindi on Saturday. I thought it was a bit strange to go during a weekend, but being new in the country I thought they knew best.

Thursday 19 february: Meeting 2 with : Philippe Simonis Energy Advisor Amber Hose, Room A205 c/o GTZ Office Kampala P.O Box 10346 Tel/Fax: +256-41-234165 Mob: +256-75-791268 [email protected] Meeting with Rachel and Richard at Mr. Simonis office about the project. During this meeting I got some more information about the project in Masindi, and we also talked a bit on how my project could be incorporated with this. We talked about that I maybe could look at a larger part of Masindi and several trading centres. I.e doing the same as Rachel and Richard on a larger scale. We decided that Mr Simonis should write an email to Edgar Hertwich and Olav Fosso and tell them about the project.

Britt-Mari Langåsen

Spring 2004

11 Appendix

80

Friday 20 february: Meeting 3 with Mr. Simonis Talked about my project and decided that I would look at the whole Masindi district, but otherwise do the same as the two students. I was also going to do a load flow analysis over the district.

Friday 20 february: Meeting at UEDCL Franklin Kizito Oidu Manager Technical Services Amber House Plot 29/23 Kampala Road P.O.Box 7059 Kampala Tel: +256-31-360704 Fax: +256-41-349565 Mob: +256-41-259716 [email protected] [email protected] Met Mr. Oidu with R&R and spoke about the information we wanted for our project. He seemed very helpful and became even more interested when he learnt that I studied at NTNU where he had been for a year in 1986.

Nyanzi Joseph Kubo Principal Planning Engineer Amber House Plot 29/23 Kampala Road P.O. Box 7059 Kampala Tel: +256-31-360600 Fax: +256-41-349565 Mob: +256-77-404878 Mr. Kubo was the one helping us to get the information about the grid and also about costs for grid extension. While in the office a man came down the roof, twice. He was up there fixing wires. We got single line diagrams over the Masindi district and over the total Uganda grid. He promised he would help me later with the load flow analysis. We talked about IPPs but he didn’t know about any other in the area than the Kinyara sugar factory. They had plans about expanding. As of now the produce 15 MW which they use for their own consumption. He didn’t know about any plans for hydro power and the area wasn’t suited for geothermal activities.

Britt-Mari Langåsen

Spring 2004

11 Appendix

81

Monday 23 february: Meeting with the LC5 in Masindi. The LC5 is the chairman for the district of Masindi. We went to the LC5 to get information about the population in Biizi and also about the economic activities. He seemed “maatligt” interested in what we where doing and probably thought that we should have been more prepared. He sent us to another person in the same building, Mr Rashid Yawiya. He had the information we wanted about the trading centres but the problem was that he didn’t know which where electrified and not. To get that information we had to go to the planning office and then get back in touch.

The planning office promised to help finding important trading centres. The lay helping us didn’t have time to do the work when we where there, but promised to find the information, give it to the population office and then fax the information to GTZ. (After 3 weeks the info has still not come. Will go up there again.)

Week 9 Searched for information about prices for PV and generators. Went to Mantrac to find prices about generators (why not the internet?) and to UltraTec to get prices for PV. At Mantrac we spoke to a lady that where going to help us and find the data we wanted about different generators. She couldn’t do it momentarily but she said she would have it done within a couple of days. The person at Ultratec didn’t want to give us the pricelist because he was afraid that we would publish his figures that are wholesale. I asked if it is ok to get the info if the report has a confidential part but he didn’t seem to understand or didn’t want to give out the price anyway. We didn’t get any copies but had to write down the figures he said about different costs and life spans of products.

Wednesday 10 march: Elisabeth Kabagante Planning Engineer UEDCL Amber House, Room B209 Tel: +256-31-360600 [email protected] Ms. Kabagante works as a planner at UEDCL and will help me get information about the load data for the district of interest. Decided that I would write an e-mail about the

Britt-Mari Langåsen

Spring 2004

11 Appendix

82

information I wanted and that we would make an appointment after that. Seemed interested in what I am doing. She uses the programs ETAP and EPSS for load flow analysis. She also gave us a copy of a part of an report which gave us figures about operating hours that UEDCL use when planning and doing load profiles.

Wednesday 10 March: Meeting with Luzze Fred Senior Technician Solar Energy for Africa Ltd. Plot 40, Bombo Road P.O. box 4155 Kampala Tel: +256-41-250125 Tel/Fax: +256-41-250131 Mob: +256-77-564405 [email protected] During this meeting we got access to old orders that had been made by SPA. This gave us an idea about the cost for solar and also what wattage one should put on different devices. The man we talked to did not know how to design a system by him self, but he used software that seemed excel based. We didn’t get any copies here either of systems, but had to write down the data that were presented to us.

Friday 12 March: Meeting with Dr. Izael Pereira da Silva Lecturer Dept. of Electrical Engineering Makerere University Tel: +256-41-250415 Mob: +256-77-5057902 [email protected] Talked about the collaboration between Makerere University and NTNU, what have been good and what needs to be improved. We said that communication prior to our arrival was a key ingredient that where missing a bit this time. A better communication would be in interest of both MUK and NTNU, the student and the professors if the collaboration is to continue.

Tuesday 16 March: Meeting with Elisabeth Kabagante Planning Engineer

Britt-Mari Langåsen

Spring 2004

11 Appendix

83

UEDCL Amber House, Room B209 Tel: +256-31-360600 [email protected] This meeting was a result of an e-mail I had sent with questions to Elisabeth the 11 of March. The questions where about data for a load flow analysis.

First I got some

information at her office in Amber House and then we went to Lugugo and the UEDCL office there. In Lugogo she showed me the load flow analysis program ETAP which she used for analyses.

Tuesday 16 March: Meeting with Nyanzi Joseph Kubo Principal Planning Engineer Amber House Plot 29/23 Kampala Road P.O. Box 7059 Kampala Tel: +256-31-360600 Fax: +256-41-349565 Mob: +256-77-404878 This was a “spontaneous” meeting and we talked a bit about different problems with grid extension in rural areas. Some of the problems when constructing new lines are: -Theft of expensive parts (transformers are stolen and remelted. This means that the cheaper aluminium transformers can’t be used, but instead they have to use the triple SC (?) transformer) -Vandalism

There are no costs for the distribution company when the power is not delivers, as in a load shedding situation or a loss of means. Sometimes they pay if damage on devices or loss of production can be proven. The larger industries wish for compensation for loss of power and because of that production losses.

Different areas in Kampala have different load shedding priority, with the Mulago Hospital coming first. This is not a priority you can pay for, but a natural selection.

Britt-Mari Langåsen

Spring 2004

11 Appendix

84

Wednesday 17 March: Meeting with: Terry Jobling Engineering manager P.O Box 179 Masindi Tel.: +256 036 600 410 [email protected] Mr Jorling works at Kinyara Sugar Works Ltd and is the engineer there. Kinyara Sugar Works is almost self sufficient with electricity and it is produced from both the burning of bagasse which is a residue from the sugar production and from diesel generation. They are at this moment building a new furnace for bagasse, since they are expanding their production of sugar, but they will not utilise more of the bagasse for electricity generation. The furnace that is being built can though be rebuilt with a generator to recover steam for electricity generation. According to studies done by Kinyara Sugar Works, it is not economically viable for them to sell excess energy to the utility. There is a sub station just outside the factory, so connecting to the grid is not a problem, but Mr Jorling said that if they where to sell electricity to the grid, they would have to make contracts with certain amounts of power to sell. If they for some reason should produce less energy and have contracts on energy to sell, they would have to produce it by diesel generation. This would be to expensive. If it where to be economically sound to sell excess electricity to the grid, the grid price would have to be about twice the amount it is today.

Monday 22 March: Meeting with Andrew Mubonga, DM for UEDCL, Masindi Andrew Mubonga District Manager, Masindi district, UEDCL Mob: +256-77-490596 This meeting was to find out some things about the reliability of the lines supporting Masindi. Got some statistics from September to February, with some months missing. During the rainy season there are often more outages because of falling trees, rain, thunder etc. Also got some general costs for grid extension. If there is an extension with no pole or one pole, e.g. very close to the existing grid, then the customer can contact the district

Britt-Mari Langåsen

Spring 2004

11 Appendix

85

office and the will handle the extension. If there is a larger extension, then the main office have to do a research of the investment.

The cost for a no-pole extension: 150 000 Ush The cost for a one-pole extension: 350 000 Ush These prices include a 50 000 Ush security deposit which the customer will get back if there is a problem (?)

23 March: Meeting with Herman Ssenyondwa Principal planning engineer UETCL P.O. Box 6088 Kampala [email protected] I met with Mr Ssenyondwa and told hi about my project which where going to include a load flow analysis over Masindi district, part from doing a cost analysis of electrifying a trading centre. Got access to a report made by the Swedish company SWECO about the transmission and subtransmission system in Uganda. But to get permission to have a copy of this report I had to get a clearance from the head at UETCL. He was away on business and I got in touch with … instead. The day after a got a clearance and could copy parts of the report. Unfortunately it wasn’t in soft copy.

Wednesday 24 March: Meeting with Fred Tugume Fred Tugume A.G Coordinator for geothermal activities Geological surveys & Minerals Department Entebbe, Uganda [email protected]

There are three different areas in Uganda around Kibiro that are being explored for geothermal activities. They are still doing feasibility studies on these sights but there is problems getting funding for the research. That is why there have been little development of the studies since 1994. Mr Tugume said that the project was prioritised by the World Bank that where funding the studies. The steam from he geothermal

Britt-Mari Langåsen

Spring 2004

11 Appendix

86

activities can be used for drying fish or tobacco in the area and that it is compatible with small hydro power generation. He didn’t know how much power that where expected to be used from the geothermal spots.

Britt-Mari Langåsen

Spring 2004

11 Appendix

87

11.2 HOMER input variables

Britt-Mari Langåsen

Spring 2004

11 Appendix

110

11.3 Attached CD with HOMER files and data

Britt-Mari Langåsen

Spring 2004