Fusion grid concept design and techno-economic analysis for developing countries

techno-economic analysis for developing countries

Examiners Professor Olli Pyrh¨onen

Associate professor Tuomo Lindh

Author Iurii Demidov Lappeenranta 2018

Lappeenranta University of Technology LUT School of Energy Systems

Electrical Engineering

Iurii Demidov

Fusion grid concept design and techno-economic analysis for developing countries Master’s thesis

90 pages, 35 pictures and 17 tables

Supervisors Professor Olli Pyrh¨onen D.Sc. Antti Pinomaa

Keywords: Off-grid, energy + connectivity, solar PV plant, battery, scalability, decentralized power stations, design, rural areas, techno-economic analysis, low voltage power distribution.

This Master’s thesis work discuss the issues of concept creation and low voltage (LV) net- work design of off-grid where connectivity is integrated to the very same power system. The work considers implementation of different design approaches of the PV based off-grid distribu- tion system to real grid topologies in rural distant areas in developing markets in Sub-Saharan Africa. Network topology is based on geographical analysis of housing in selected villages of regions of Namibia. Namibia is one of the African country where access to electricity is an issue to be solved.

Only 43% of population has access to electricity in Sub-Saharan Africa countries. This per- centage goes rapidly down if we are talking about rural areas (15%). Most of existing systems are typically either of two types: 1) solar house systems (SHS) - typically its power doesn’t exceed 100 Watts and supply is individual consumer or 2) solar charging systems; (SCS) it can have power rating up to 15 kilowatts and is aimed as a rule to phone charging. However, these systems are not intendent connectivity provision, uninterrupted and power supply with good quality. In order to guarantee such options there is a need to design an off-grid power system, where each consumer is connected to the common grid, which is sourced by PV power plant and backed up by battery storage system and where area of supplied grid is covered by

i

techno-economic point of view. The main emphasis is on uninterrupted power supply, provision of connectivity and high quality electricity to each consumers and cost efficiency of designed system. Fusion off-grid system is one possible solution for this issue. This work considers differ- ent approaches to design of such grid, varying battery types, voltage type, voltage level (within LV limits), power cabling configuration, number of power energy source etc. Perspective of off-grid interconnection and integration with utility grid is also taken into account.

Designed off-grid system consists of PV array, battery storage, inverter or converter, depending on voltage type, transformers in some cases and power cabling part. It includes core initial system configuration, network development and power energy decentralization stages. The first step presupposes grid supplying connectivity source and basic load. The second comprises im- plementation of network part to the off-grid. Three different topologies are considered, which are: dense rural, scattered rural and dense city grid. There are DC bipolar distribution con- figuration, which presupposes usage up to three voltage levels, European 230 VAC [IEC 60038, IEC 50160, CENELEC HD 472 S1, assuming one or three phase and up to two voltage levels, and American 120 VAC [ANSI C84.1-1989], implying two voltage levels. In addition to config- uration comparison, consumers capacity rise is also take into account. The last stage contains analysis of impact of decentralized generation to the total system cost. Topologies from the previous stage is considered. Therefore, in this work from techno-economic point of view the most optimal methodology of Fusion off-grid development and configuration is defined.

ii

This work was carried out at LUT School of Energy Systems, Electrical Engineering, Lappeen- ranta University of Technology (LUT). The work is related to an ongoing collaborative, business Finland funded research project (Fusion Grid) between Lappeenranta University of Technology (LUT), Aalto University , Nokia Oyj, Nordkapp Creative Oy, GreenEnergy Finland (GEF) and University Properties of Finland (SYK) in which the objective is to design off-grid system providing connectivity and electricity, and digital platform to developing countries.

I express my sincere gratitude to Professor Olli Pyrh¨onen and Dr. Tuomo Lindh for the oppor- tunity to work in such an interesting project at LUT and for trusting me to accomplish this master thesis. I would also like to thank my supervisor Dr. Antti Pinomaa and Dr. Andrey Lana for the guidance, encouragement in my master studies and contribution to revise and improve the language of this manuscript.

Finally, I would like to express my heartfelt gratitude to my parents Igor and Olga for your support and care. My girlfriend Anastasiia, thank you for your tolerance and trust.

Iurii Demidov September 2018

Lappeenranta, Finland

iii

Abstract i

Acknowledgements iii

1 Introduction 4

1.1 Motivation and Objectives . . . 4

1.2 Outline of the work . . . 6

1.3 Contributions . . . 7

1.4 Statement of Originality . . . 8

2 Off-grid technology 9 2.1 Off-grid concept . . . 9

2.2 Photovoltaic array and storage battery sizing . . . 10

2.3 Power electronic and wiring system . . . 21

2.4 Chapter summary . . . 27

3 First steps 28 3.1 Stage 1: Power supply for Kuha base station . . . 28

iv

3.3 Stage 2: Integrated energy source and single consumer . . . 35

3.4 Stage 3: Split energy source and single consumer . . . 42

3.5 Chapter summary . . . 45

4 Power system development and cost analysis 46 4.1 Dense rural grid . . . 47

4.2 Scattered rural grid . . . 53

4.3 Dense city grid . . . 58

4.4 Chapter summary . . . 63

5 Decentralization in off-grid 65 5.1 Dense rural grid . . . 66

5.2 Scattered rural grid . . . 69

5.3 Dense city grid . . . 73

5.4 Chapter summary . . . 75

6 Conclusion 76 6.1 Summary of Thesis Achievements . . . 77

6.2 Future Work . . . 78

Appendix A 78

References 83

v

2.1 Monthly coefficient of Sun peak hours. . . 16

2.2 Load content and operation mode of lead acid batteries. . . 19

3.1 Kuha power supply systems. . . 31

3.2 Load content and operation mode. . . 36

3.3 Stage 2: School case. . . 38

3.4 Stage 2: Mining station case. . . 39

3.5 Stage 2: Mining station case one redundancy day. . . 40

3.6 Stage 3: School case. . . 43

4.1 Total cost of load stages Dense rural case. . . 51

4.2 Price per kWh in Dense rural case. . . 52

4.3 Total cost of load stages Scattered rural case. . . 55

4.4 Price per kWh in Scattered rural case. . . 58

4.5 Total cost of load stages Dense city case. . . 60

4.6 Price per kWh in Dense city case. . . 62 vi

5.2 Total costs optimization of Scattered rural case. . . 72 5.3 Total costs optimization of Dense city case. . . 75

vii

2.1 Common off-grid structure comprising PV, MPPT, battery, power electronic:

step up/down converter for power distribution. a) DC concept, b) AC concept . 11 2.2 Single diode model of a PV module . . . 12 2.3 PV array output current-voltage and power-voltage dependencies at a) fixed

temperature and different insolation b) fixed insolation and different temperature 14 2.4 Sun hours evaluation . . . 15 2.5 Elementary battery scheme . . . 18 2.6 Single phase full-bridge inverter . . . 22 2.7 Converter scheme: a) Buck topology b) Boost topology c) Inverting topology . . 23 2.8 Configurations: (a) DC grid, (b) AC European grid. (c) AC American grid . . . 25

3.1 (a) DC configuration, (b) AC configuration . . . 30 3.2 Total prices of stage 1 system (inverter implementation) . . . 31 3.3 Total prices of stage 1 system (converter implementation) . . . 32 3.4 Implementation of stage 1 system: (a) Totem design, (b) Bus stop design . . . . 32 3.5 System costs comparison assuming equipment lifetime . . . 33 3.6 Wire dimensioning . . . 34

viii

3.7 Total prices of stage 2 (school case) . . . 37

3.8 Prices of one kWh (stage 2 school case) . . . 39

3.9 Total prices of stage 2 (miming station case): (a) one day redundancy, (b) two day redundancy . . . 41

3.10 System implementation with separate Kuha base station PV array and battery storage . . . 42

3.11 Prices of one kWh (stage 3 school case) . . . 44

3.12 Prices of one kWh (stage 3 school case) . . . 45

4.1 Dense rural grid: (a) DC implementation, (b) AC implementation . . . 48

4.2 Dense rural grid total cost: (a) DC 120 VDC implementation, (b) DC 350 VDC implementation, (c) European AC implementation, (d) American AC implemen- tation . . . 50

4.3 Energy production price in Dense rural case . . . 52

4.4 Scattered city: (a) DC implementation, (b) AC implementation . . . 54

4.5 Scattered rural grid total cost: (a) DC 350 VDC implementation, (b) European AC implementation, (C) American AC implementation . . . 56

4.6 Energy production price in Scattered rural case . . . 57

4.7 Dense city grid total cost: (a) DC implementation, (b) AC implementation . . . 59

4.8 Dense city grid total cost: (a) DC 350 VDC implementation, (b) European AC implementation, (C) American AC implementation . . . 61

4.9 Energy production price in Dense city case . . . 63

5.1 Dense rural case one bus configuration . . . 67

LtoL implementation, (c) Tier 5 LtoL implementation, (d) Tier 1 One bus im- plementation, (e) Tier 3 One bus implementation, (f) Tier 5 One bus implemen- tation, (g) Tier 1 Two bus implementation, (h) Tier 3 Two bus implementation, (i) Tier 5 Two bus implementation . . . 68 5.3 Scattered rural case one bus configuration . . . 70 5.4 Scattered rural grid cost distribution: (a) Tier 1 LtoL implementation, (b) Tier

3 LtoL implementation, (c) Tier 5 LtoL implementation, (d) Tier 1 One bus im- plementation, (e) Tier 3 One bus implementation, (f) Tier 5 One bus implemen- tation, (g) Tier 1 Two bus implementation, (h) Tier 3 Two bus implementation, (i) Tier 5 Two bus implementation . . . 71 5.5 Dense city case one bus configuration . . . 73 5.6 Dense city grid cost distribution: (a) Tier 1 LtoL implementation, (b) Tier 3 LtoL

implementation, (c) Tier 5 LtoL implementation, (d) Tier 1 One bus implemen- tation, (e) Tier 3 One bus implementation, (f) Tier 1 Two bus implementation, (g) Tier 3 Two bus implementation, (h) Tier 5 Two bus implementation . . . 74

x

Symbols

A_h actual battery pack capacity A_h0 initial battery capacity CEnergy-tariff estimated energy tariff

C_Battery price of a battery

CBattery-replacement battery replacement costs CBattery-material material cost of energy storage CPE-replacement power electronic replacement costs

C_PE-losses costs of losses in power electronic equipment for 20 years CPE-material-installation power electronic material and installation costs

C_per-kWh price per kWh produced energy

C_PV price of a PV module

CPV-material material cost of PV plant

C_system PV and energy storage installation price

CSystem-total sum costs of all system component CSystem-installation system installation costs

CTL-installation transmission line installation costs

C_TL-losses costs of losses in transmission line for 20 years CTL-material transmission line material costs

dV voltage drop

dV_cable-fact factual transmission line voltage drop EBattery day energy per day through a battery E_ES capacity of storage energy

E_PV^max maximum PV daily generated energy E_PV^min minimum PV daily generated energy E_PV energy per day produced by PV plant EPV-20-years energy produced for 20 years by PV plant

Es battery bank energy

I current

I_cable transmission line current

Id diode current

I_sh shunt current

I_s saturation current

K Boltzmann constant

Kconfiguration coefficient of configuration 1

k_discharge coefficient of possible depth of discharge Ksunpeakhours coefficient of sun peak hours

Kpeaksunhours^max maximum monthly coefficient of sun peak hours Kpeaksunhours^min minimum monthly coefficient of sun peak hours kprice-decrement coefficient assuming price reduction for 20 years

K_sim simultaneously factor

L_cable line length

N series connected cells (modules) number

N_Battery total number of batteries

N_Cycles number of battery cycles

N_parallel number of parallel PV panels

N_PE number of power electronic equipments

N_PV number of PV modules

N_series series number of PV panel

P_Converter converter power

p_Cu specific resistance of copper

P_PE nominal power of power electronic unit

P_Inverter inverter power

P_load total load capacity

P_PE power electronic nominal power

P_PV PV plant power

P_module PV module nominal power

P EPrice-material-installation price including material and installation of power electronic equipment

q the electron charge

R_cable resistance of a wire

R_in internal battery resistance

R_load load resistance

R_s series resistance

R_sh shunt resistance

Sload.capacity line transmission capacity

S_line wire cross section

STL-losses-W capacity losses in transmission line

STL-losses-Wh energy losses in transmission line per day

T cell temperature

tPE-operation power electronic operation period tBattery-years battery lifetime

t_load load duration

T LPrice-installation TL installation price per unit T LPrice-material TL material price per unit

V voltage

V_AC nominal AC voltage

V_cable nominal transmission line voltage

V_DC nominal DC voltage

V_ES voltage rating of battery string

V_panel nominal voltage of a PV module

Voutput req. required output PV voltage

V₀ no-load battery voltage

Greek alphabet

η_B battery efficiency

η_Converter converter efficiency

η_ES round-trip efficiency of batteries η_Inverter inverter efficiency

η_PE efficiency of power electronic Abbreviations

AC alternate current

b.h. big house

DC direct current

ES energy storage

FZM flexi-zone micro

LFP lithium iron phosphate

LTE long-term evolution

LtoL line to load

LV low voltage

MPPT maximum power point tracker

PE power electronic

PV photovoltaic

PWM Pulse-width modulation

sh.g. shack group

SCS solar charging systems

s.h. small house

SHS solar house systems

SoC state of charge

SWA steel wire armoured

TL transmission lines

V voltage

Introduction

It is impossible to imagine life of modern people without electricity and Internet. Electrification of developed countries converge to 100% and it is hard to find place where there is no Internet access. These services are ordinary for us but it is really rare for majority of citizens in developing countries. Therefore, there is a need for the system which combine both electricity and connectivity issues and based on developing countries specific conditions such as weather and grid topologies. Design of such kind of system is considered in this Master’s thesis.

1.1 Motivation and Objectives

Only 43% of population has access to electricity in Sub-Saharan Africa countries.This percent- age goes rapidly down if we are talking about rural areas (15%) [1]. Moreover, 89% [2] of electricity is generated by coal-fired power stations, while Africa is one of the most favorable places for the implementation of solar photo-voltaics (PV) energy production. According to the ”World Sunshine Map” it gets the more sun days per year than any other continent in the world [3, 4].

In the last years many of countries have started to develop in solar power energy sector. By 2020, 30% of Senegal rural population will supplied with electricity by solar power stations [5].

Mozambique invested more than $500 millions in solar photovoltaics-based energy production 4

[6]. In rural areas of Rwanda there are many off-grids projects which supply local schools and farmers [7]. Africa Mini-grid Developers Association was created in order to boost growth of mini-grids projects. It is expected that by year 2020 the number of renewable mini-grids rise from 12 000 to more than 145 000 [8].

Namibia is the country in Sub-Saharan Africa with high potential for off-grid projects devel- opment, where only 51.8% of total population has electricity access and 28.7% in rural area conditions [1]. The energy system of Namibia covers only industrial and city areas of the coun- try, while most of the country territory remains out of electricity access [9]. The maximal power demand of the country is 657 MW (2015). The installed power generation capacity is 511 MW and only 1.85% of it is Solar PV, in addition to lack of generation capacity, 100% of liquid fuels are imported [10], while Namibia has the second highest solar annual irradiation level in the world at 3000 kWh/m2 over most part of the country. In some regions, solar irradiation level reaches more than 5.8 kWh/m2 [11].

As for communication side of Namibia, the situation is even worse than with electrification.

Only few big city areas are covered with 3G and 4G mobile networks [12]. Therefore, most rural citizens have no access to the Internet and digital world and services. They are lacking behind in business side, education and rise of social development as well.

In order to solve both these issues there is a need for a system which delivers both electricity and Internet access designed for African rural conditions. The system should be constructed to be scalable in order to fit different environments with different characteristics as range and cov- erage and distances etc. It also should serve different consumers’ needs, electric safety should be guaranteed and in addition it should be adaptive. Adaptive system presuppose that the system with its topology should reconfigure and recover itself automatically after faults. In addition sun hours provision could be implemented, which forecasts future weather (thanks to Internet/Satellite access or using machine learning technology), and regulates amount of distributed energy. In condition of sun hour shortage, system should have possibility to auto- matically drop less important consumer loads or limit its power consumption.

1.2 Outline of the work

This Master’s thesis based on Electrical Network Design Methodology, which presuppose fol- lowing [13]:

• Issue data collection- problem, required equipment and design approach identification

• Preparation of the preliminary single-line diagram- setting obligatory requirement for the system and compilation of single-line diagram

• Technical studies and single-line diagram validation- techno-economical optimiza- tion, calculation comparison of different network typologies

• Selection of equipment and components

• Choice and setting of power grid protection devices

• Choice and installation of a control and monitoring system

Chapter 1 gives brief description of designed system components. It is observed such equip- ment as PV array, batteries, converter, inverter, maximum power point tracker (MPPT) con- troller and different approaches to grid topology configuration.

Chapter 2 illustrates starting stage of electricity-connectivity off-grid and possible ways of system development. Two case studies are proposed in this part: 1) School - is a reference point for the development of the project from the social side, involving government as one of investors; 2) Mine - is a reference point for the development of the project from the business side, making the system focused on business purposes. In addition this chapter compares dif- ferent components types and suggest probable solution of cost reduction.

Chapter 3 is a continuation of previous part and presents further power system development, including energy distribution and load rise estimate. Different voltage types and approaches are compared and optimal voltage configuration is identified, mapping three cases: Dense rural, Scattered rural and Dense city grid. Each case has specific numbers of consumers and location.

Chapter 4 depicts implementation of distributed generation and its economical efficiency.

Two solar PV power station emulate power system decentralization. Three cases from previous chapter are presented for showing reasonableness of distributed energy source compared with conventional one. This part can be used either for optimal system development applying two energy source or implementation of the second power station as a next step of system improve- ment.

Therefore, in this thesis requirement for the system is identify:

• voltage level - due to remoteness from engineering services and small load size, low voltage (LV) levels are applied;

• voltage type - today efficiency of direct current (DC) or alternating current (AC) imple- mentation is discussed issue;

• centralized or decentralized generation - usage of several power source in off-grid is poorly studied, but it could significantly influence on system price and reliability;

• power equipment - there is no innovative equipment application in the system, all price and settings is taken from real sellers and Internet equipment.

As an Internet access, a standard, but specially for rural areas designed 4G LTE Kuha base station by Nokia is used [14]. It provides several alternatives for back-haul connection, such as satellite connection, microwave link, or 3G/4G connection to existing mobile networks. There is different range of coverage options and different number of simultaneously connected users depending on the Kuha base station size and features, including RF characteristics, antennas etc.

1.3 Contributions

Master’s thesis is done within a research project - Fusion Grid - funded by Business Finland BEAM program (2018-2020). The main contribution of the work is an identification of optimal electricity-connectivity off-grid system construction from the techno-economical point of view.

In addition, actual design issues such as battery type choosing, grid configuration application, energy source decentralization are studied and could be used for further implementation of electricity-connectivity system. Thesis depicts step by step power system development and allows to predict costs for different system topology, load size and grid configuration cases. It also has a practical importance. All condition as close as possible to the real that allows to construct pilot, using this Master’s thesis as technical specification for the Fusion Grid research project.

1.4 Statement of Originality

There are many already existing off-grid concepts and projects, as well as researched for Sub- Saharan African conditions. However most of these projects are small scale PV and power bank solutions, which are supplying separated house and utilized only for some basic need such as phone charging, typically its power doesn’t exceed 100 Watts. These are called solar house systems (SHS) [15]. Another example of existing projects is more powerful solar station where the locals can charge their battery banks, it has power rating up to 15 kilowatts and are called Solar charging systems (SCS) [16]. As a rule, above mentioned off-grids doesn’t presuppose uninterruptedly providing consumers with a good quality electricity. Moreover, none of these system provide connectivity side. In this thesis, the system not only provide with electricity to each customer but also with connection to the Internet, providing great opportunities for development, giving chance to broaden one’s mind and business opportunity is designed. Therefore, the integration of Internet connection leads to the system that may rise consumers revenue and may solve system profitability issues.

Off-grid technology

This chapter overviews basic design methodology and equipment of off-grids. Therefore, it is introduced what off-grid is and what it consists of. Design principles, challenges and issues connected to off-grids is described. This part presents background knowledge connected to technical side of the Master’s thesis.

2.1 Off-grid concept

Off-grid system is power system isolated from the conventional utility grid. It includes stan- dalone, self-powered and sustained energy source, and in this design integrated with energy storage, for power generation and accumulation, power electronics for power conversions, elec- tricity grid for power delivery, protection and control devices and end consumers and loads.

Commonly, renewable energy sources only or combined with diesel generator are used for en- ergy generation in off-grids. There are such system as wind generator combined with solar PV plant/diesel generator; one source wind generator/solar array or all three together solar PV + wind generator + diesel generator [17]. In this thesis, only solar energy source (as starting point) will be considered because, the work is done for Fusion grid project, it focuses on sys- tem location in Sub-Saharan Africa where implementation of solar energy is especially efficient.

However, system supplied from solar energy source, does not consider single generation that 9

means system can include several PV plant.

In initial design phase the solar hours per day, peak power of PV plant, battery size considering days of autonomy and efficiency of the equipment should be taken into account. Talking about a system isolated from the utility grid we also should estimates the future development of en- ergy distribution system of a country where the system is intended to implement. Therefore, the possibility of future integration and moving from Off to On grid or interconnection together of several small off-grid cells into bigger one, forming infrastructure, also should be taken into account.

The interest to the solar-power-based off grids has risen for the last few decades. Such system have been successfully implemented in many Asian, South American and African countries, especially for rural areas. Such interest is motivated by high renewable potential of countries, poor electrification and high energy tariffs of utility energy system. High PV panel demand leads to increase of supply toward this product, as a result development of its production and item price reduction. International Renewable Energy Agency broadcast the 57% PV panel cost decrease [18].

2.2 Photovoltaic array and storage battery sizing

Micro off-grid is an off-grid system that involves small-scale electricity generation and which serves a limited number of consumers via a distribution grid [19]. Usually it is installed right close to consumer (on the roofs or in the garden), and consequently power lines aren’t included into system. However, in this thesis uninterrupted energy supply is a base demand set which is ensured by connection all consumers to common network, supplied by centralized or decen- tralized solar PV power station(s). Figure 2.1 illustrates common off-grid structures with DC and AC network part implementations.

(a)

(b)

Figure 2.1: Common off-grid structure comprising PV, MPPT, battery, power electronic: step up/down converter for power distribution. a) DC concept, b) AC concept

The first step of solar off-grid design is sizing of the PV array, which is the power source that transforms solar light into electricity. It typically consists of solar panels connected in series and/or parallel. Each solar panel consist of basic PV cells. Panels connected in series increase the voltage of the PV array, and parallel connection increase the output current of the PV array. The same rules suits for the PV cells in one panel. Figure 2.2 illustrates single diode model of a PV module.

I ^ph

Diode

Rsh Rs

V I ^d I ^sh I

Figure 2.2: Single diode model of a PV module

According to the Kirchhoff’s current law, the load current is :

I =I_ph−I_s·(e^{q·V+N·K·Tq·Rs·I} −1)− (V +I·Rs)

R_sh , (2.1)

whereI_s is saturation current;I_ph - current in the PV array,V - voltage across PV module; R_s - series resistance; Rsh - shunt resistance; K = 1.380·10⁻²³J ·K⁻¹ - the Boltzmann constant;

N - series connected cells (modules) number; T - cell temperature and q - the electron charge [20].

Diode current is

I_d =I_s·(e^{q·V+N·K·Tq·Rs·I})−1. (2.2) Shunt resistor current is

I_sh = (V +I·R_s)

R_s . (2.3)

Series resistor current is

I_s =I_ph−I_d−I_sh. (2.4)

Therefore, from above mentioned equations, resulting load current is

I = (R_sh·I_s)

R_sh ·(1−( V

R_sh·I_s)−e^N·K·Tq·V ) +Iph. (2.5) The output power of the PV array largely depends from the solar irradiation and ambient temperature. These dependencies are shown in Figure 2.3. From the plots it is seen that there is a Catch 22 (mutually conflicting or dependent conditions), when the higher solar radiation the higher output power and the higher ambient temperature. However the higher ambient temperature the lower output power. In order to maximize PV plant output power, special PV modules supported structure can be applied [21]. It tracks the sun, rotating and changing angle of solar panel to the horizon, but such technical solution is optional. Maximum Power Point Tracing (MPPT) is a device [22], which is used in all modern solar power systems for maximization of power produced by the PV array. There are several MPPT algorithms:

• Perturbation and observation (P&O): This algorithm disturbs the operating voltage to ensure maximum power.

• Incremental conductance: This algorithm, compares the incremental conductance to the instantaneous conductance in a PV system. Depending on the result, it increases or decreases the voltage until the maximum power point is reached.

• Fractional open-circuit voltage: This algorithm is based on the principle that the maxi- mum power point voltage is always a constant fraction of the open circuit voltage. The open circuit voltage of the cells in the PV array is measured and used as in input to the controller.

0 50 100 150 200 250 300 350 Voltage (V)

0 100 200 300 400

Current (A)

Array type: SunPower SPR-305E-WHT-D;

5 series modules; 66 parallel strings

1 kW/m²

0.5 kW/m²

0.25 kW/m²

0 50 100 150 200 250 300 350

Voltage (V) 0

5 10 15

Power (W)

10⁴

1 kW/m²

0.5 kW/m²

0.25 kW/m²

(a)

0 50 100 150 200 250 300 350

Voltage (V) 0

100 200 300 400 500

Current (A)

Array type: SunPower SPR-305E-WHT-D;

5 series modules; 66 parallel strings

0 ^oC 25 ^oC 50 ^oC

0 50 100 150 200 250 300 350

Voltage (V) 0

2 4 6 8 10 12

Power (W)

10⁴

0 ^oC 25 ^oC 50 ^oC

(b)

Figure 2.3: PV array output current-voltage and power-voltage dependencies at a) fixed tem- perature and different insolation b) fixed insolation and different temperature

From Figure 2.3 we can see that changing solar insolation leads to reduction of output power.

To maintain the maximum possible power, there is a need to decrease output voltage.

Combination of MPPT and charge controllers is the most effective in the following cases:

• In cloud and cold weather - PV module works better in cold temperature conditions and MPPT produces the maximum available power;

• With significantly discharged battery - the output solar module voltage must be greater than the battery voltage for charging battery. When battery voltage is low, it is possible to decrease PV panel voltage in order to extract more current and charge the battery faster.

There are several methodology how to scale the size of the PV array such as Intuitive, Numer- ical, Analytical, Computer tools and Hybrid [23]. In this study Intuitive methods for optimum sizing PV system is applied. It assumes using coefficient of sun peak hours and assuming re- quired energy per sun and ’dark’ hours. In addition, the efficiency of power electronics and battery is considered. The solar modules efficiency isn’t applied because the usage of MPPT is estimated, consequently it tend to be one. The battery efficiency isn’t taken into account while sun hours, since we guess that load is supplied directly from the PV and excess of energy charges the battery.

Time of Day 1kW/m²

Time of Day 1kW/m²

Equal area under the curves

a) b)

W/m² W/m²

Figure 2.4: Sun hours evaluation

Solar radiance is the power per unit area received from the Sun in the form of electromagnetic radiation in the wavelength range of the measuring instrument [24]. The coefficient of sun peak hours (Ksunpeakhours) is used as approximation that allows evaluating the amount of energy per day, in order to simplify its calculation. It presents sun radiation curve, shown in Figure 2.4a, as a strict number of hours per day when sun has maximum insolation, the rest of the day it is assumed that sun doesn’t shine at all. Therefore, if the solar radiation per day is 3.9 kWh/m2, it is possible to present that there is a peak solar radiance is 1 kW/m2 for 3.9 hours per day. Figure 2.4b shows the day solar insolation using Ksunpeakhours. The square under the curves represent the total amount of solar energy. The area under the curves in Figure 2.4a and Figure 2.4b are equal [25]. For Namibian location where the first system is aimed and to be designed for in Fusion Grid project, the coefficients of solar radiation on horizontal plane per day are presented in Table 2.1 [26].

Table 2.1: Monthly coefficient of Sun peak hours.

Month Daily solar radiation, kWh/m²/day

January 7.12

February 6.69

March 6.06

April 5.83

May 5.31

June 4.87

July 5.14

August 5.87

September 6.83

October 7.31

November 7.52

December 7.69

According to modified equations (result efficiency is varied, depending on time of a day) of Intuitive methods for optimum sizing PV system [27], required output power of PV panels is defined as:

P_PV = P_load·Kpeaksunhours

η_PE·Kpeaksunhours

+P_load·(24−Kpeaksunhours) η_PE·η_ES·Kpeaksunhours

, (2.6)

where PPV is the required PV power, ηPE is efficiency of power electronic converter, ηES is round-trip efficiency of batteries, Kpeaksunhours in kWh/m²/day is solar insolation for PV power plant location for the worst month in a year.

Minimum PV daily generated energy is

E_PV^min =PPV·ηPE·Kpeaksunhours^min , (2.7)

whereKpeaksunhours^min = 4.68 [kWh/m²/day] is the minimum monthly coefficient of sun peak hours.

Maximum PV daily generated energy is

E_PV^max =PPV·ηPE·Kpeaksunhours^max , (2.8)

where Kpeaksunhours^max = 7.41 [kWh/m²/day] is the maximum monthly coefficient of sun peak hours. By the reason that there is only one energy source in the system in this case, having no additional redundancy suppliers, consequently we have to consider the worst possible condition, which considers lack of the sun for two days.

After defining generation capacity there is a need to identify the number of parallel and series- connected panels according to above-mentioned principle. Firstly, we calculate the total quan- tity of modules:

N_PV = P_PV

P_module, (2.9)

where P_module is a nominal power of one PV module.

The number of parallel panels depends on required output voltage of the PV array.

Nparallel = Voutput req.

V_panel , (2.10)

whereVoutput req. is a required output PV voltage; V_panelis a nominal voltage of one PV module.

The series module quantity is

N_series = NPV

N_parallel. (2.11)

All numbers are rounded up to an integer.

Next step in off-grid designing is batteries scaling. Batteries are used as energy storage that provides system with power in the absence of the sun. It consist of voltaic cells and converts chemical energy to electrical one and vice versa. Lithium-ion and lead-acid types of batteries are considered in this thesis. Since there is no need to assume dynamic effect of the battery, it is possible to present battery in a simplest way constituted by a function of ideal voltage source from the State of Charge (SoC) -V₀(SoC) and internal resistanceR_in(Figure 2.5). This scheme represents lithium-ion and lead-acid batteries and considers suits only for steady-state behavior, excluding thermal one [28].

V

(SOC)

R

V I

Figure 2.5: Elementary battery scheme

The internal resistance is calculated as follows:

R_in= E_s·m·(1−η_B)

A_h0·I(A_h0)·(1 +η_B), (2.12) whereE_s is the battery bank energy;η_B is the battery efficiency;A_h0 is the initial capacity and I(A_h0) is the current characteristic to initial capacity. The electric circuit is described by the equations:

V =V₀(SoC) +R_in·I, (2.13)

SoC = Z

Idt+SoC_i, 0≤SoC ≤A_h, (2.14) where V is battery voltage; V₀ is no-load battery voltage; I is battery current; SoC_i is initial

Table 2.2: Load content and operation mode of lead acid batteries.

Depth of discharge Deep-cycle battery

100% 150-200 cycles

50% 400-500 cycles

30% 1000 cycles

state of charge; A_h is actual battery pack capacity.

Battery capacity, lifetime and efficiency depends on many factors and gradually decrease during operation. The battery end life is reached when it has worked stated number of cycles and its capacity has decreased to 80% from initial. Lead-acid battery nominal operation temperatures is from 10 to 50^◦C (battery has full nominal capacity), however 25^◦C is the most optimal one because every 8^◦C rise leads to twice reduction of battery lifetime. Deep discharge also cut the battery capacity. After single discharge below specified voltage value, the battery lose 5%

of nominal capacity. For the further calculation, it will be accepted 50% depth of discharge for lead-acid battery. The number of cycles of lead-acid battery varies from the manufactors and features of production, but the average quantity is 500. Table 2.2 presents dependence of number of cycles from depth of discharge. Efficiency of a battery depends on the time of discharge. Nominal capacity (100%) is represented for 20 hours rate, but in case of high discharging the efficiency decreases: 80% - 4 hours rate, 60% - 2 hours rate [29]-[30].

Lithium-ion battery chemistry can subdivided into lithium iron phosphate (LFP, LiFePO4) and metal oxides (NCM, NCA, Cobalt, Manganese). In current studies the second material type is applied.

Li-ion battery nominal operation temperature is from 5 to 45^◦C. It is possible to charge the battery in 0–5^◦C temperature range, but in condition of low current rate. Temperature higher than 45^◦C of battery operation can deteriorate its performance and lead to capacity and lifetime reduction. Depth of discharge for li-ion battery is accepted at 75%. Assuming the normal temperature conditions, the number of cycles is 2000. Therefore, considering depth of discharge and life cycles, in order to get comparable lifetime of both battery types, lead-acid pack must have three times more capacity. Efficiency of li-ion battery also depends of the time of discharging. However, the efficiency reduction is less than for lead-acid type: 99% 4 hours

rate, 92% - 2 hours rate. [31, 32].

In order to provide system of enough amount of energy in case of sunless period, there is a need to scale the battery storage. Methods used for the PV array scaling are also applicable to battery storage sizing methodology. However, first of all, the size and duration of system load should be determined. After that, it is possible to calculate the required battery storage energy. However, the temperature, sun hours, battery type, on or off-grid system type and number of energy source also should be considered. One more issue is the number of dark days redundancy for the system. Most of the companies engaged in off-grid system installing, dimension battery storage capacity assuming two days of reserve [33, 34]. One of the possible batteries dimensioning equations assume exceed of generated PV energy per day:

E_ES=E_PV− P_load·Kpeaksunhours

η_PE , (2.15)

Another way of sizing the worst case considered when there is no sun for several days (two in our case).

E_ES = 2·t_load·P_load

η_PE·η_ES·k_discharge. (2.16)

Therefore the required capacity in Ah is:

E_ES^Ah =EES/VES, (2.17)

where EES is a required battery storage energy [Wh], EPV is energy per day produced by PV plant [Wh], k_discharge is a coefficient of possible depth of discharge, P_load is a total load capacity, t_load is a load duration, V_ES is a voltage rating of battery string.

Next step is to determine the number of parallel and series-connected batteries in energy storage.

Such kind of sizing is based on Kirchhoff laws and is the same as for the PV array. Final stage of battery sizing is calculation of duration of battery life. These equations describe amount of energy spent by the battery storage per day and number of years of operation.

EBattery day= P_load·(24−Kpeaksunhours) ηPE·ηES

, (2.18)

tBattery-years= 0.9·E_Battery·N_Cycles

(EBattery day/(V_ES·N_Battery))·365.2, (2.19) where E_Battery [Wh] is capacity of a battery,N_Battery is the total number of batteries, N_Cycles is the number of cycles of a battery.

More detailed comparison of lead-acid and Li-ion battery type’s especially with economic ben- efit side will be presented in further chapter.

After defining the PV array and battery storage sizes, there is a need to choose required power electronic equipment which ensures required voltage level in the grid. There are two options, depending on operation and supplying voltage type: DC-converter or AC-inverter application.

2.3 Power electronic and wiring system

Inverter is the electronic circuit or electromechanical device that transforms direct current (DC) into alternating one (AC) and vice versa (can be used as a rectifier). It doesn’t produced any power and serve only for transformation one current type to another. It is used for suppling AC load in DC off-grid systems. Inverter input voltage varies from 12V DC to hundred thousands of volts with the PV array application. The output voltage is consumer-end voltage (typically 120 VAC and 240VAC) [35]. In order to increase output voltage rate, buck-boost converter (on the DC side) or transformer (on the AC side) is used in addition. Figure 2.6 illustrates full bridge single-phase inverter with four switches, which is one of the simplest implementation [36].

The output voltage of the bridge, depending on switches position, can be either +Vdc or−Vdc. Switches on one leg must not be closed at the same time because it leads to short current. For getting proper gating switch signal the Pulse Width Modulator (PWM) is applied. Low-pass filter is implemented in order to absorb high frequency harmonics and get clear output signal.

It is important to note that the higher switching frequency is the less filtering is required but the more switching losses is occurred [37, 38]. Nowadays the efficiency of inverters is very high and reaches 90–98% [39].

V

R

R

Load

C L

V

Figure 2.6: Single phase full-bridge inverter

DC/DC Converter is the electronic circuit or electromechanical device that transforms DC voltage from one level to another. Step down (buck); step up (boost) and inverting (buck- boost) converters are studied in this thesis (Figure 2.7).

Above-mentioned converters are based on switching regulator that use PWM, an inductor, a capacitor and a diode to transfer energy from input to output. The principle of operation of these devices is due to the relation between current and voltage of the inductor. In the further application it is basically used for transforming from generated/consumers voltage level to transmission line voltage level and vice versa [40].

The last part of the system is a connection between generation and consumer, transmission lines (TL) are needed for this purpose.

Vin Vout Vin Vout

Vⁱⁿ V^out

a) b)

c)

Figure 2.7: Converter scheme: a) Buck topology b) Boost topology c) Inverting topology Depending on surround conditions and options of laying (it could be overhead or underground), type of wire is chosen. However, despite various cable types, there is a common equation which allows to calculate wire resistance which is used in further calculations [41].

R_cable= p_Cu·L_cable Sline

, (2.20)

where p_Cu=0,17 ^Ω·mm_m ² is specific resistance of copper, L_cable [m] is a length of a line and S_line [mm²] is a cross-section of a wire.

Wiring could be organized in different ways. DC grid wire topology is possible to subdivide into unipolar and bipolar. Unipolar DC network configuration requires at least use two-core wiring, for example SWA BS5467 2 core 10 mm² SWA steel wire armoured cable. There are one positive or negative pole and natural one. Bipolar DC grid takes at least three-core wiring (SWA BS5467 3 core mm² SWA steel wire armoured cable). It consists of positive, negative and natural (zero voltage) poles. For example, there are +750 VDC, -750 VDC and 0 VDC

voltage conductor in±750 VDC system. Zero voltage on the neutral wire is obtained by sum of the voltages on the positive and negative pole, but it is only if poles are equally loaded. [42, 43]

The distribution DC low-voltage levels varies from 120 VDC to 1500 VDC [Low Voltage Di- rective (LVD) 2006/95/EC]. It is also possible to use extra low-voltage levels 24 VDC and 48 VDC but its transmission capacity and transmission distances are rather limited. Transmission capacity of overhead or cable line depends on such factors as load capacity, distance to the load, wire cross section, voltage level and system network type (bipolar/unipolar). The more detailed analysis of these dependences is presented in Chapter 3.

At the distribution level, AC grid wire topology could be either 1-phase or 3-phase according European standard [IEC 60038, IEC 50160, CENELEC HD 472 S1] and split-phase according to American one [ANSI C84.1-1989]. In one-phase topology, only two wires are used (the phase and the neutral). In three-phase systems three wires are reserved for phases and one or two for neutral and/or protective earth conductor. Therefore, for one and three phase systems, two and four core cable are used, respectively. For AC European distribution nominal low-voltages are from 220 VAC to 1000 VAC [Low Voltage Directive (LVD) 2006/95/EC]. Split-phase system is design in such way that one-phase 240 VAC divided into two phases (120 VAC). This topology provides not only efficiency but also the consumer safety. All types of above mentioned cables are applied in low voltage rate and complied with the requirements set out by the International Electrotechnical Commission standard IEC 60502.

DC/DC converter

DC/DC converter

DC/DC converter

DC/DC converter

DC/DC converter

DC/DC converter

DC/DC converter Rload

Rload

Rload

Rload

±350V

±350V

±350V

±120V

±120V

-120V

-48V +120V

+48V

+120V

+48V

-120V

-48V Main cables

Main cables

Distribution cables

Distribution cables

Consumer cables

Consumer cables

Consumer cables

Consumer cables DC source

Source voltage

+120

-120

+120

-120 +350

-350

(a)

DC/AC inverter

380V DC source

Source voltage

Rload

Rload

Rload Main cables

Distribution cables

A B C N

(b)

DC/AC inverter

1000V DC source

Source voltage

Main cables A

B C N

Distribution cables

Rload

Rload

Distribution cables

Rload

Rload

Distribution cables

Rload

Rload

240V

120V

120V

240V

120V

120V

240V

120V

120V

(c)

Figure 2.8: Configurations: (a) DC grid, (b) AC European grid. (c) AC American grid

In the further research of this thesis, the following configurations are applied:

• DC gridassumes usage DC bipolar system (Figure 3.1a). It is distinguished three voltage levels: main (up to 1500 VDC bipolar) - central transmission lines, distribution (from 120 to 750 VDC bipolar) - internal group voltage and client (48 VDC unipolar) - inner house end consumer voltage, it is chosen in a way for staying in low voltage rate. The grid parameters is calculated using below equations.

Transmission line DC resistance is calculated using Equation 2.20. Line transmission capacity is:

Sload.capacity = 2·dV_cable·V_DC²

R_cable (2.21)

where dV_cable = 13% is the maximum voltage drop for the intermediate DC grid point (for the end usersdV_cable = 3%),V_DC [V] is the nominal voltage, R_cable [Ω] is a resistance of a wire [44].

• AC European gridis a one or three phase AC voltage system (Figure 3.1b). It includes two voltage levels: main (up to 1000 VAC one or three phase) and distributed (220 VAC one or three phase). Network parameters is defined using undermentioned equations.

All load is estimated as an active one in order to equivalent DC and AC grid load size.

Therefore, transmission line AC resistance is defined using Equation 2.20.

Transmission line capacity for one and three phase systems are:

Sload.capacity.1ph = dV_cable·V_AC² 2·Rcable

(2.22)

Sload.capacity.3ph = 3· dVcable·V_AC²

R_cable (2.23)

wheredVcable= 5% is the maximum voltage drop for the AC grid, VAC [V] is the nominal voltage [45].

• AC American grid is similar to the previous approach but at the distributed level

splitted phase system is applied (Figure 2.8c).

Sload.capacity.m = 3· dV_cable·V_AC²

R_cable (2.24)

Sload.capacity.d = 2· dV_cable·V_AC²

R_cable (2.25)

In Chapter 4 it is presented additional technical-economical comparison all of these system and chosen the most suitable variant for the current application.

2.4 Chapter summary

Introduction and Chapter 2 cover the first and second stages of electrical network design methodology. It is defined that Namibia has a need and prescriptive for off-grid electricity + connectivity system. In addition in the part the surround conditions is determined, all components of system are described and common methodology of calculation is presented.

First steps

This chapter describes starting implementation of off-grid power system combining electricity and connectivity. Several stages and possible way of applications, including economical and technical calculation is presented here. In addition, the competitiveness of designed systems is shown.

3.1 Stage 1: Power supply for Kuha base station

Number and capacity of consumers could vary from case to case. System flexibility is an option that can deal with this issue. In order to design saleable system there is a need to define basic methodology of gradually system development. In this study and approach the first step of creation of the system that provides electricity and connectivity is to design the basic micro grid for supplying base station. Kuha base station provides connectivity via wireless 4G LTE network of which coverage is few kilometers and can supply six hundreds consumers simultaneously [46]. The network coverage depends on characteristic of Kuha base station, its dimensions, antenna sizes, signal transmission power, and the surrounding environment. The peak power of base station is 150 W and it should work 24 hours without any interruption in energy supply. Therefore the resulting required energy is 150·24 = 3600Wh.

The initial system is calculated for different equipment options. While designing this stage, the 28

battery lifetime and price comparisons is conducted. Moreover, economical difference between applying inverter and converter is shown. Equipment properties and prices, which is used in the calculations, is taken from websites selling solar off-grid systems or its separate parts. In the system the following equipment is used:

• JAP6K PV panels ( P_PV = 271 Wp,V_PV = 24 V, P rice_PV = 120 e) [47],

• Powercraft battery (E_ES= 230 Ah, V_ES= 12 V, P rice_ES = 270e) [48],

• GWL/Power CALB CA180FI battery (E_ES = 180 Ah, V_ES = 3.2 V, P rice_ES = 190 e) [49],

• Victron SinusMax 24/800 ( η_Inverter= 0.91, P_Inverter= 800V A,P rice_Inverter= 387 e) [50],

• Orion-Tr 24/48-8.5A (η_Converter = 0.89, P_Converter = 400W, P rice_Converter = 172 e) [51],

• Victron MPPT 150/35 (P rice_MPPT= 148 e) [52].

• Kuha base station [53].

Based on equations presented in the first chapter, parameters for four systems varying bat- tery type (Lead-acid or Li-ion) and end-consumer voltage type (DC or AC) are calculated (Firure 3.1). Table 3.1 illustrates four different system implementations changing battery type and power electronic equipment. Each column presents number and power of PV panels, num- ber, capacity and lifetime of batteries, and total system price. It is seen that systems with lead-acid batteries have less initial price than the other. Four solar panel is enough for sup- plying Kuha base station and battery storage charging. The life time of lead-acid batteries is almost three times less than Li-ion one. The additional analysis is realized further in this chapter.

Battery Charger

Battery

48V DC 24V DC

PV array

Kuha base station

5V DC

USB

DC/DC converter DC/DC

converter

(a)

Battery Charger

Battery

PV array

Kuha base station

5V DC

USB

AC/DC converter DC/AC

converter

220VAC

220VAC

Sockets

(b)

Figure 3.1: (a) DC configuration, (b) AC configuration

Table 3.1: Kuha power supply systems.

System component Inv.+L.A. Inv.+Li-ion Conv.+L.A. Conv.+Li-ion

N_PV 4 4 4 4

P_PV^kW 1.08 1.08 1.08 1.08

N_ES 9 24 9 25

E_ES^{kW h} 24.8 13.8 24.8 14.4

t^yearsBattery life 8.17 20.6 8.07 21.13

Total price 3626 5852 3584 6004

Lead Acid Li-ion

0 1000 2000 3000 4000 5000 6000

Euro

PV material costs Battery material costs Converter material costs

Figure 3.2: Total prices of stage 1 system (inverter implementation)

From Figure 3.2 it is seen that system with lead acid batteries and inverter requires 38% less initial investment than li-ion. Using converter (Figure 3.3), power supply part of off-grid has almost the same cost difference (1.75 times). Therefore, usage of either inverter or converter depends on only voltage type that consumers need. Total system price include only initial installation costs of electric equipment.

With designed energy part is possible to supply communication component of the system for 24 hours for two days without interruption in energy supply. The electricity equipment meet assigned requirement even in the worst sun hours conditions. Such kind of system provide only communication to the consumers. The possible implementation is depicted in Figure 3.4a and Figure 3.4b. For the further usage of specified battery type there is a need to compare not only

Lead Acid Li-ion 0

1000 2000 3000 4000 5000 6000

Euro

PV material costs Battery material costs Converter material costs

Figure 3.3: Total prices of stage 1 system (converter implementation)

(a) (b)

Figure 3.4: Implementation of stage 1 system: (a) Totem design, (b) Bus stop design

initial investment but also assume battery life time and its replacement during operation. The set life time of PV panel is 20 years. Therefore, for the detailed analysis of different battery type implementation the 20 years period is considered [54]. Figure 3.5 shows that despite low initial investment of system with lead-acid batteries, in a life time comparison its price is hirer than with Li-ion one. Because of small operation period of lead-acid batteries there is a need to replace it two times during operation period when li-ion batteries’ lifespan exceeds of considered time period. It is worth mentioning that due to complexity of prediction and as a

kind of simplification time cost changing is neglected in this calculation, consequently constant battery price is assumed.

Lead Acid Li-ion

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Euro

PV material costs Battery material costs Power Electronic material costs Power Electronic replacement costs Battery replacement costs

Figure 3.5: System costs comparison assuming equipment lifetime

3.2 Analysis of wire parameters

Before moving to the next stage of development it is necessary to study how one wire parameter influence on another. Figure 3.6 illustrates various cable parameter dependencies in DC grid.

Curves of AC wiring has similar view (because of Equation 2.20) that is why it isn’t considered in this section. From the first two rows it is seen hyperbolic decrement of possible power transfer with distance to the load rising, assuming constant cable cross section (one per plot).

Different solid lines represent various voltage types and distribution system types. More power is possible to transfer with higher voltage. Moving from plot with constant 2.5mm² cable cross section to the plot with 35mm² one, it is possible to define also rise of possible power transfer.

Therefore, higher power transfer capabilities can be achieved with larger wire cross section.

Figure 3.6 subplot at the left bottom illustrates dependence of required voltage from length to the consumer with constant load and various cable cross sections. Figure 3.6 plot depicts that higher voltage is possible to transfer power for farther distances. As it is also seen from the