• No results found

Open Source Energy Model for the Electricity Sector of Sri Lanka

N/A
N/A
Protected

Academic year: 2021

Share "Open Source Energy Model for the Electricity Sector of Sri Lanka"

Copied!
60
0
0

Loading.... (view fulltext now)

Full text

(1)

Open Source Energy Model for the Electricity Sector of Sri Lanka

T.L.B. Attanayaka

Master of Science Thesis

KTH School of Industrial Engineering and Management Department of Energy Technology

Division of Energy System Analysis SE-100 44 STOCKHOLM TRITA-ITM-EX 2018:626

(2)

I

Master of Science Thesis TRITA-ITM-EX 2018:626

Open Source Energy Model for the Electricity Sector of Sri Lanka

T.L.B Attanayaka

Approved Examiner

Prof. Mark Howells

Supervisor

Ms. Agnese Beltramo

(3)

II

Abstract (English)

A long term generation expansion model for the electricity sector of Sri Lanka was developed in this thesis. The model provides the least cost development pathways to cater the future electricity demand within the user defined constraints that need to be adhered. Starting from the present electricity system of the south Asian island nation, the model spans for the period from 2018 to 2050.

Open Source Energy Modelling System (OSeMOSYS) was used to create the model. It utilises linear optimization and minimize the net present value of the modelled system in the entire period. Four electricity end user sectors were modelled namely, residential, industry, services and transport. Final electricity demand at present is around 13 TWh and it is projected to grow at a rate of 5.6% per annum for the next ten years to be around 24 TWh in year 2028 and to rise at a rate of 4.3% per annum there onwards to exceed 61 TWh in year 2050. Twelve fuel options were used by the existing and candidate technologies for electricity generation in the model, namely biomass, coal, diesel, furnace oil, hydro, liquefied natural gas, naphtha, natural gas, nuclear, residual oil, solar and wind. Electricity production in different levels such as transmission, distribution and end user locations were modelled in the system. Capital cost, fixed and variable operation and maintenance cost and salvage value of technologies were considered for the cost optimisation. Environmental emissions were included in the model and CO2 emission limit of 20% for the modelling period was included in the Base Scenario to represent the expected development pathway of the country in the future. Scenario analysis was conducted to examine the sensitivity of input variables such as electricity demand and hydro condition, and the impact of user defined constraints to the least cost solution. Renewable energy integration in to the system was studied and the impact of higher shares of renewable energy was examined. Capacity mix, energy mix, CO2 emission and LCOE of different scenarios were compared in the analysis.

(4)

III

Abstract (Svenska)

I denna avhandling utvecklades en långsiktig modell för utbyggnad av elproduktion för elsektorn i Sri Lanka. Inom de användardefinierade begränsningarna som måste följas, ger modellen den minsta kostnaden för att tillgodose det framtida elbehovet. Med utgångspunkt från det nuvarande elsystemet i den sydasiatiska önationen, sträcker sig modellen över perioden 2018-2050.

Open Source Energy Modeling System (OSeMOSYS) användes för att skapa modellen. Den utnyttjar linjär optimering och minimerar nuvärdet av det modellerade systemet under hela perioden. Fyra sektorer med slutanvändare av elektricitet modellerades, bostäder, industri, tjänster och transporter. Det slutliga elbehovet är för närvarande cirka 13 TWh och det förväntas växa med 5.6% per år under de närmaste tio åren till cirka 28 TWh år 2028 och stiga med 4.3% per år kommande år till att överstiga 61 TWh år 2050. Tolv bränslealternativ användes av befintliga och kandiderande teknologier för elproduktion i modellen, nämligen biomassa, kol, diesel, brännolja, vatten, flytande naturgas, nafta, naturgas, nukleär, restolja, sol och vind. Elproduktion på olika nivåer, såsom överföring, distribution och slutanvändares läge, modellerades i systemet. Kapitalkostnad, fast och rörlig drift och underhållskostnad och restvärde för teknik beaktades för kostnadsoptimering. Miljömässiga utsläpp inkluderades i modellen och koldioxidutsläppsgränsen på 20% för modellperioden inkluderades i basfallet för att representera landets förväntade utvecklingsväg i framtiden. Scenarioanalys genomfördes för att undersöka känsligheten hos inmatningsvariabler som efterfrågan på elektricitet och vattenförhållanden, och effekten av användardefinierade begränsningar till lösningen med minsta kostnaden. Integreringen av förnybar energi i systemet studerades och effekten av högre andelar av förnybar energi undersöktes. Kapacitetsmix, energimix, CO2-utsläpp och LCOE av olika scenarier jämfördes i analysen.

(5)

IV

Acknowledgement

I would like to thank all the people who supported me during the work on this thesis. First I would like to remind my family who always supported me in my life and in my studies. I also would like to thank my examiner Prof. Mark Howells and my supervisor Ms. Agnese Beltramo for giving me the opportunity to do my thesis on this interesting topic, and for the guidance and support provided. Furthermore, I would like to thank all the people in DESA who provided insights and helped me to improve my work.

I would like to express my gratitude to Dr. Jeevan Jayasuriya of InnoEnergy master’s school and Dr. Peter Hagström of SELECT master’s programme. And also I would like to express my thankfulness to Mr. Ruchira Abeyweera for coordinating the first year studies at OUSL. Further, I would like to sincerely remember the General Manager and the officers of Ceylon Electricity Board especially the Generation Planning Unit for providing me study leave to complete my master’s studies.

(6)

Contents

Abstract (English) ... II Abstract (Svenska) ... III Acknowledgement ... IV

List of Figures ... 3

List of Tables ... 4

List of Abbreviations ... 5

1. Introduction ... 6

1.1 Electricity Sector of Sri Lanka ... 6

1.2 Background for the Study ... 7

2. Aims and Objectives ... 9

3. Generation Planning Tools ... 10

4. Development of the Open Source Energy Model ... 11

4.1 OSeMOSYS ... 11

4.2 Methodology... 12

4.3 Reference Energy System ... 13

4.4 Data Collection ... 15

4.5 Assumptions... 15

4.6 Year Split ... 16

4.6.1 Load profile of Sri Lanka ... 16

4.6.2 Time Slices ... 17

4.7 Demand Projection ... 18

4.7.1 Specified Demand Profile ... 20

4.8 Technology Parameters ... 21

4.8.1 Implementation of Residual Capacities ... 22

4.8.2 Availability Factor... 22

4.8.3 Capacity Factor ... 22

4.8.4 Annual Maximum Capacity Investment... 22

4.8.5 Reserve Margin ... 23

4.8.6 Fuel Price ... 23

(7)

2 4.10 Storage ... 24 4.11 Emissions... 24 5. Results ... 25 5.1 Reference Scenario ... 25 5.2 Base Scenario ... 26 5.3 Environmental Emissions ... 28 5.4 Scenario Analysis... 29

5.4.1 Variation of Hydro Condition ... 29

5.4.2 Sensitivity to Electricity Demand ... 30

5.4.3 Sensitivity to Discount Rate ... 31

5.4.4 High Renewable Energy Integration Scenario ... 31

5.4.5 Summary of Scenario Analysis ... 32

6. Discussion ... 33

6.1 Levelized Cost of Electricity ... 33

6.2 Comparison with WASP IV Results ... 35

6.2.1 Capacity Additions... 35 6.2.2 Energy Mix ... 37 6.3 CO2 Emission Results ... 38 7. Conclusion ... 40 Bibliography... 41 Annex-1 ………. i Annex-2 ………. iii Annex-3 ………. iv Annex-4 ………. vi Annex-5 ………. vii Annex-6 ………. ix

(8)

3

List of Figures

Figure 1 Energy Mix of Sri Lanka in 2017 ... 7

Figure 2 Capacity Mix of Sri Lanka (end of 2017) ... 7

Figure 3 Generation and Transmission Planning and Operation Functions by Timescale [8] ... 10

Figure 4 OSeMOSYS blocks and level of description [5] ... 11

Figure 5 Methodology of the study ... 12

Figure 6 Reference Energy System of Sri Lanka ... 14

Figure 7 Hourly load profile in peak demand day of recent years ... 16

Figure 8 Electricity Demand Projection ... 19

Figure 9 Sector-wise Electricity Demand Share ... 19

Figure 10 Hourly Electricity Demand ... 20

Figure 11 Demand Profile of the Residential Sector ... 20

Figure 12 Demand Profile of the Industrial Sector ... 21

Figure 13 Hourly wind speed and electricity generation profile in Puttalam ... 24

Figure 14 Reference Scenario Capacity Mix... 25

Figure 15 Reference Scenario Energy Mix ... 25

Figure 16 Base Scenario capacity mix ... 26

Figure 17 Base Scenario energy mix ... 27

Figure 18 Percentage share of energy mix in Base Scenario ... 27

Figure 19 CO2 Emission of Base Scenario... 28

Figure 20 Capacity addition comparison with the Base Scenario ... 29

Figure 21 Base and High Demand Projections ... 30

Figure 22 Energy mix comparison between (a) Base Scenario and (b) High RE Scenario ... 32

Figure 23 Scenario LCOE Comparison... 34

Figure 24 Installed Capacity Results of the Two Models ... 35

Figure 25 Energy Mix Results for Year 2020 (a) WASP-IV and (b) OSeMOSYS ... 37

Figure 26 Energy Mix Results for Year 2025 (a) WASP-IV and (b) OSeMOSYS ... 37

Figure 27 CO2 emission per unit of electricity generation in base scneario ... 38

(9)

4

List of Tables

Table 1 Time span of seasons in a year ... 17

Table 2 Day time split ... 17

Table 3 Year split for time slices ... 18

Table 4 Net present cost with discount rate variation (million US$) ... 31

Table 5 Results of scenario analysis ... 32

Table 6 Input Electricity Demand comparison of the two models ... 35

(10)

5

List of Abbreviations

CCY Combined Cycle Power Plant

CEB Ceylon Electricity Board

CIF Cost, Insurance and Freight

GDP Gross Domestic Product

GT Gas Turbine

IAEA International Atomic Energy Agency

IEA International Energy Agency

IPP Independent Power Producer

LCOE Levelized Cost of Electricity

LDC Load Duration Curve

LNG Liquefied Natural Gas

LTGEP Long Term Generation Expansion Plan

MAED Model for Analysis of Energy Demand

MMBTU Million British Thermal Units

NDC Nationally Determined Contributions

ORE Other Renewable Energy

O&M Operation and Maintenance

PF Plant Factor

PSPP Pumped Storage Power Plant

RES Reference Energy System

RM Reserve Margin

SAM System Advisor Model

US$ United States Dollars

VRE Variable Renewable Energy

(11)

6

1. Introduction

Affordable and clean energy for all is a primary sustainable development goal adopted by the United Nations in its mission to end poverty, protect the planet and ensure prosperity for all. Ensure access to affordable, reliable and modern energy services is its main target and providing electricity is one of the main objectives. Furthermore, electricity is closely related with the economic development of a country, while economic growth drives the demand for electricity (United Nations General Assembly, 2015) (Yılmaz, B.; Alp Özel, 2014).

1.1 Electricity Sector of Sri Lanka

Sri Lanka is an island nation located in the Indian Ocean with a population of around 21 million. GDP per capita of the country in 2017 was USD 4,065 and a GDP growth rate of 3.1% was recorded. Electricity is a basic requirement for the economic growth of any country. Around 99.3% from the total population had access to electricity from the national grid by end of 2016. Electricity demand grow at a higher rate contributing to the rapid development of the economy. The average per capita electricity consumption in 2017 was 626 kWh and it has been rising steadily during the past decade. To meet the increasing demand of electricity, new generation capacity has to be added to the system (Central Bank of Sri Lanka, 2017).

Ceylon Electricity Board (CEB) is the national electric utility responsible to develop and maintain an efficient and coordinated, economical system of electricity supply for the country. Present installed electricity generation capacity is around 4,050 MW with a total of 3,500 MW of dispatchable capacity. Major share of dispatchable capacity is owned by CEB, which includes 1,370 MW of hydro and 1,500 MW of thermal power plants. The balance share is owned by private sector developers. Installed capacity of other renewable energy (ORE) plants are around 550 MW and it consists of mini-hydro, biomass, wind and solar PV plants. Electricity transmission network is owned and operated by CEB at the voltages of 220 kV and 132 kV. Annual electricity generation recorded in 2017 was 14,620 GWh which is a 3.7 percent growth from the previous year (Ceylon Electricity Board, 2018b).

Electricity production in the country was dominant by hydropower until late 90’s. However, with the increase of electricity demand, after developing all potential major hydro power plants the country had to shift towards thermal power to accommodate the increase in demand. Oil fired power plants were added to the system with technologies of Reciprocating Engines, Simple Cycle and Combined Cycle Power Plants. The first coal power plant was added to the system in year 2012. The capacity mix at the end of year 2017 and energy mix for electricity generation in year 2017 is given in figures 1 and 2 (Ceylon Electricity Board, 2018b).

(12)

7 34% 20% 15% 17% 14% CEB Hydro CEB Thermal Coal CEB Thermal Oil IPP Thermal Oil Other RE 21% 35% 17% 17% 10%

Figure 2 Capacity Mix of Sri Lanka (end of 2017) Figure 1 Energy Mix of Sri Lanka in 2017

The drought conditions that prevailed during most part of 2017 has directly affected the output of hydro generation and increased the country’s reliance on thermal power generation. Consequently, hydropower generation, excluding mini hydro generation, has been reduced by 14% from previous year generation of 3,499 GWh to 3,014 GWh, whereas fuel oil-based power generation and coal power generation increased to 5,045 GWh and 5,071 GWh, respectively. With the rise in mini hydropower generation due to the rainfall received during the latter part of 2017, electricity generation through other renewable energy (ORE) sources increased by 28% to 1,489 GWh, compared to the previous year value of 1,160 GWh. Coal based electricity generation accounted for the largest share within the total generation with a percentage share of 35% followed by fuel oil, hydro and ORE power generation, with percentages 34%, 21% and 10% respectively (Central Bank of Sri Lanka, 2017).

In the past, Sri Lanka has been extensively depending on its hydro resources supplemented by gas turbines and diesel plants with small inputs from other renewable energy resources. However, with the growth of electricity demand exceeding the installed capacity of major hydro power plants, the country had to depend more on thermal power generation which are owned by both national utility and the private sector. This significantly increased the cost of electricity generation during last two decades (Asian Development Bank, 2007).

1.2 Background for the Study

Long term planning is needed for the electricity sector since many investment decisions have very long lead times, and when implemented they have long economic lives (Meier, Vagliasindi, & Imran, 2014). Long Term Generation Expansion Plan (LTGEP) 2018 – 2037 prepared by CEB is in draft stage and it projects the

(13)

8

electricity demand and proposes power plant additions for the future(Ceylon Electricity Board, 2018a). The LTGEP identifies coal power plants as the least-cost option for the country and it proposes to add several coal power plants in the future. All fossil fuel sources for power generation, coal and oil, are imported to the country at present. Liquefied natural gas (LNG) fired thermal plants are also proposed for the first time and they will open the island nation to the new fuel option, natural gas. There are discussions comparing LNG vs. coal and economic and environmental benefit of LNG in the power sector has to be studied in detail.

In addition, natural gas exploration studies have been going on in past few years in the Mannar basin, North West of the country. Preliminary studies indicate the availability of natural gas and the possibility of extracting it in the future with an estimated quantity of 0.5 MTPA (Ceylon Electricity Board, 2018a). Economic benefits of using this local natural gas for power generation is yet to be studied.

The daily load profile of the country records a surge in electricity demand in the evening time period around 19.00 to 21.00 (Public Utilities Commission of Sri Lanka, 2012). CEB has to invest on adequate generation capacity to meet this daily short term peak demand in order to maintain the required level of power supply capacity. LTGEP also proposes a pumped storage power plant (PSPP) that could utilize excess electricity in off-peak hours and supply energy in the peak hours (Ceylon Electricity Board, 2018a). This has to be further studied considering benefits of long term operation of PSPP plants.

Integration of variable renewable energy (VRE) plants, mainly wind and solar is another main problem that has to be analyzed in the island’s electricity system. Optimum level of VRE penetration has to be identified where the power system can operate with required level of reliability considering the intermittency issues of VRE. Electricity and heat generation sector is the largest contributor for the global CO2 emission from fuel combustion and it accounts for a share of 42%. Transport sector which accounts for a share of 24% is the second largest contributor according to the IEA statistics (International Energy Agency, 2017). Even though electricity sector is the major contributor for CO2 emissions in the world, the transport sector contributes to the majority of emissions in Sri Lanka. Total CO2 emission from fuel combustion in Sri Lanka is 19.5 million tons in year 2015 and the emission recorded from the electricity sector is 6.8 million tons which is about 35% share from the total (International Energy Agency, 2017).

(14)

9

2. Aims and Objectives

The aim of this thesis is to develop an open source energy model for the electricity sector of Sri Lanka using OSeMOSYS for the time horizon 2018 – 2050. In addition, this thesis intends to compare the results with model outcomes of existing generation planning tools of the country.

Following results and deliverables are expected from this research work.

 Optimized long term expansion plan for the electricity sector of Sri Lanka for the time horizon 2018 – 2050

 OSeMOSYS model for the electricity sector of Sri Lanka

 Scenario analysis to investigate the impacts of uncertainties in the electricity sector and to analyze different investment strategies

 CO2 emission analysis for the scenarios

Following chapters describe the work covered in the thesis. Present applications of generation expansion planning tools are explained in chapter 3. Model development and adopted methodology is explained in chapter 4 and the model output results are elaborated in chapter 5. In chapter 6, results are compared with published outputs of the existing plan and significant deviations are discussed and the conclusion is given in Chapter 7.

(15)

10

3. Generation Planning Tools

One main difference of electricity compared to other fuels is that it is not economical to store in large quantities. Therefore, system operator has to ensure that supply meets demand at all time to maintain the security of supply. Adjustment of generation need to be made in different time windows, in order to ensure the system to operate economically and securely from years to milliseconds with different combinations of generation technologies. The generation capacity expansion planning usually performed for long time horizons and it identifies new investments and time periods they need to be added to fulfil exogenously derived demand. Figure 3 shows the relationship of generation expansion planning with medium and short term planning and typical time horizons associated with it (Yuan, 2013). Due consideration on operational issues has to be given in generation expansion planning to ensure an economical and reliable electricity system.

Several energy planning models were used in recent past to analyse the electricity sector of Sri Lanka with assistance of International Atomic Energy Agency (IAEA). Model for Analysis of Energy Demand (MAED) is an end user demand projection tool which can be used to forecast sector-wise final energy demand (IAEA, 2006). Electricity sector demand projection for this study was obtained from MAED model and explained in Chapter 4.

The Wein Automatic System Planning Package (WASP-IV) of IAEA has been used for long term generation planning from the national electric utility. It is an optimisation tool within user defined constraints and utilizes probabilistic estimation of system production costs, unserved energy cost and reliability and applies linear programming (IAEA, 2001). Results of Long Term Generation Expansion Plan 2018 – 2037 prepared using WASP-IV simulations are compared with OSeMOSYS results in chapter 6.

Model for Energy Supply System Alternatives and their General Environmental Impacts (MESSAGE) which is also an IAEA tool was used by the national electric utility to study future scenarios of the electricity sector. However, all these tools are not available in open source and hence limit the use within the interested research community. Generation Expansion Planning Generation System’s Unit Commitment Economic Dispatch

Years Days / Hours Hours / Minutes

Generator AGC

Generator Excitation Control

Seconds Milliseconds

(16)

11

4. Development of the Open Source Energy Model

The methodology adopted to develop the model is described in this chapter. More specifically, the structure of the model, main assumption, step-wise approach to develop the model and main parameters are explained in detail.

4.1 OSeMOSYS

Open source energy modelling system, OSeMOSYS is a deterministic model which utilizes linear optimization. The energy models created in OSeMOSYS are dynamic, bottom-up, multi-year models. It assumes a perfect market with perfect competition and foresight (Henke, 2017).

The model consists of seven ‘blocks’ of functionality and has three levels of abstraction. The blocks are interrelated and connected, however one advantage is that a single block can be updated or replaced by a new block with new functionality. The seven blocks of the current OSeMOSYS version are the objective (1), costs (2), storage (3), capacity adequacy (4), energy balance (5), constraints (6), and emissions (7) which are given in figure 4 (Howells et al., 2011).

One main aim/objective for the development of OSeMOSYS was to enable students, business analysts, researchers and developing country governments to investigate and work on energy models (Howells et al., 2011). Therefore, the model is open source, and it is easy to understand with the plain English description and mathematical analogy. This promotes wide use of this tool and with the code available for modifications it facilitate researchers to add more features to the model. For example, Incorporating flexibility requirements into long term energy model of Ireland (Welsch et al., 2014) and analysis on storage capacity limit impact to solar power plant installation (Niet, Pitt, Rowe, & Wild, 2016) are studies done by modifying the OSeMOSYS code. Active online community gather interested users and help to solve problems and promote improvements in OSeMOSYS.

(17)

12

4.2 Methodology

At the beginning a background research and a literature survey was conducted on long term generation expansion planning problem. Published documents on OSeMOSYS were studied to understand the modelling process. Previous work by KTH dESA (division of Energy System Analysis) was used as a guideline for this thesis (Welsch et al., 2014),(Henke, 2017).

The reference energy system (RES) was defined to identify all fuel and technologies related to electricity sector. Then, relevant data were collected and open source energy model was developed for Sri Lanka. The diagram in figure 5 shows the complete methodology adopted in the study.

Figure 5 Methodology of the study

Variable Renewable Generation Profiles

 Hourly profiles for wind & solar

 Seasonal profiles for Mini-hydro Electricity Demand Projection

 End user method demand projection for Household, Industry, Services and Transport sector

Open Source Energy Model -Optimization-

Results Extraction

- Capacity mix, Energy mix, etc.

Comparison with Published Long Term Generation Planning Results

LTGEP results (WASP IV)

Discussion

- LCOE

- Environmental emissions Scenario Analysis

- Hydro condition variation - High penetration of VRE - Sensitivity to demand variation - Discount rate variation Data Collection

Present Electricity System (technical data, cost data, etc.) Candidate Power Plants

(18)

13

RES, data collection and demand projection are explained in sections 4.3, 4.4 and 4.7 respectively. Main assumptions used are given in section 4.5 and main parameters in the model are explained in following sections. After the model is developed and optimized, the results were compared with the available results of the draft LTGEP until year 2037 since it was prepared only for 20 year time horizon. Detailed analysis was done for the milestone year 2025.

Scenarios were developed to investigate the sensitivity of input parameters to output results and to define development pathways with different investment strategies according to policy alternatives. Final results were analysed and further model development was carried out to identify the impact of integration of higher share of VRE.

List of all existing and candidate power plants are given in Annex 1. Candidate power plants include thermal power plants with fuel options natural gas, coal, oil and nuclear, and also renewable power plants such as wind, solar PV, biomass and hydro. CO2 emissions from thermal plants were integrated to each technology in the model.

4.3 Reference Energy System

The first step in developing an electricity system model is to create a Reference Energy System (RES) (Howells et al., 2011). RES is a qualitative representation indicating what the demands are, what technologies are used and how to meet the demands. The process of generating a RES is helpful to get an overall picture on the energy system, and identify the salient aspects need to be included in the model.

Four distinctive levels of fuel categories are identified namely, Primary, Secondary, Tertiary energy levels and End user electricity. Fuel produced within the country like biomass, and imported to the country such as coal, crude oil and refined oil products, belong to primary energy level. Natural gas which has an identified production potential as well as expected in the form of LNG import is also categorized under primary energy level.

Electricity which is produced through an energy conversion from a technology such as hydro or thermal power plants, belong to secondary energy level. All existing, committed and candidate technologies of electricity generation are included in the RES. Electricity transmission technology reflects the losses associate in transmitting electricity from secondary level to final level and four electricity distribution technologies are defined to deliver electricity from final level to four end user electricity levels namely, residential (household), industry, services and transport. The diagram in figure 6 depicts the RES for the electricity sector of Sri Lanka.

(19)

14 Biomass Production C o a l C ru d e O il B io m as s N at u ra l G as

Primary Energy Secondary Energy Final Energy

Coal Import

Crude Oil Import NG Production

Hydropower Plants

Coal Power Plant Thermal Plants (Utility owned) Gas Turbine PP Combined Cycle PP Reciprocating Engines O il P ro d u ct s Oil Import E le ct ric it y E le ct ric it y O il P ro d u ct s N at u ra l G as

Thermal Plants (IPP) Reciprocating Engines Combined Cycle PS Mini-hydro PP Wind PP Solar PV Biomass PP Candidate Coal PP Candidate Oil PP Candidate Gas PP Candidate Hydro PP Candidate RE PP Mini-hydro, Wind, Solar PV (Utility scale,

Rooftop), Biomass Pumped Storage PP Oil Distribution Gas Distribution Refinery Electricity Transmission

End Use Electricity

El . R e sid e n tia l El . I n d u st ry El . Se rv ic e s El . T ra n sp o rt Electricity Distribution LNG Import Solar PV Residential

(20)

15

4.4 Data Collection

In order to build the model in OSeMOSYS, it is necessary to identify required data for the model and collect those data. The list given below indicates the main data needed to model the electricity sector of Sri Lanka.

I. Technical data of existing and committed power plants (Maximum/minimum operating limits, heat

rates (efficiency), etc.)

II. Cost data of existing and committed power plants (Fixed and variable o&m cost, fuel cost) III. Technical and cost data of candidate power plants

IV. Annual electricity demand profiles of identified sectors (Residential, Industry and Services) V. Electricity demand projections for the modelled period

VI. Large hydro generation constraints (Seasonal and annual limits)

VII. VRE plant data

VIII. Wind and solar resource availability data for identified high potential regions

IX. Wind, Solar, mini-hydro and Biomass project development plans

X. Proposed pumped storage power plant data

XI. Policy directives related to electricity sector

Most of the data were gathered from the national electric utility of Sri Lanka (Generation Planning Unit, Ceylon Electricity Board).

4.5 Assumptions

Main assumptions made during the model development are explained in this section. It is important to consider these assumptions when analysing the output results of the model.

 Planning period is 2018 to 2050, however short term solutions may become attractive at the end years. Therefore the model period is extended until year 2055 to avoid this effect.

 Default unit for energy in OSeMOSYS is petajoule (PJ), therefore all input energy quantities such as demands are converted to PJ.

 Capacity is measured in gigawatt (GW).

 The monetary unit for the model is US Dollar (US$).

 All cost data are in the unit million US$ per GW (MUS$/GW) which is equivalent to US$/kW except variable operation cost which is in MUS$/PJ.

 To integrate emissions from fossil fuel fired power plants, EmissionActivityRatio is included to the model and it is measured in million tons per PJ (Mton/PJ).

(21)

16 0 500 1000 1500 2000 2500 3000 0 2 4 6 8 10 12 14 16 18 20 22 24 D e m a n d (M W ) Time (h) 2014 2015 2016 2017

Figure 7 Hourly load profile in peak demand day of recent years

 Discount rate of 10% is used in the model which is applied for long term planning in the country.  Capital investment is assumed to occur at the beginning of the year. Fixed and variable operation and

maintenance costs are assumed to occur at the middle of the year which the costs are incurred. All costs are discounted to the base year of the planning period (Howells et al., 2011).

 Construction periods of different power plants vary depending on the amount of infrastructure development required. Construction periods considered for candidate thermal power plants are given in Annex1.

 Sinking fund depreciation method is applied for salvage value calculation in the model (IAEA, 1984).

4.6 Year Split

4.6.1 Load profile of Sri Lanka

Hourly load profiles of the country is given in figure 7 and it consists with demand profiles of highest demand days for the past four years from 2014 to 2017. Peak demand is recorded in the evening from 19.00 to 21.00. And the minimum demand record at the night time from 22.00 to 05.00 in the morning. Daily electricity demand has been growing while maintaining the same demand profile. Annual electricity demand growth in past four years is around 5% per annum while peak demand rise is about 4.7% per annum. Peak electricity demand recorded in the country is 2,523 MW for year 2017 (Ceylon Electricity Board, 2018a).

(22)

17

4.6.2 Time Slices

A year is divided in to three seasons namely, Dry season, High wind (Intermediate) season and Wet season. Seasons are defined based on weather patterns prevailing in the country and hence cause considerable demand and supply changes. For example, low rainfall in dry season will limit hydro generation and high temperatures in this season increases the electricity demand for cooling. Typically the peak electricity consumption is occurred during the dry season. South-west monsoon is active in intermediate/ high wind season and therefore favourable wind speeds are observed for wind power generation throughout this season. Table below shows the seasons with months associated with it.

Table 1 Time span of seasons in a year

Season Months

Dry January to April

Intermediate/ High wind May to August

Wet September to December

A week is divided into week days and weekend days to reflect the difference of electricity consumption in working days to non-working days. Considering the daily demand profile, three time slices within a day are defined namely, Day, Peak and Off-peak as given in the table below.

Table 2 Day time split

Daily Time Slices Time

Day 06.00 – 18.00

Peak 19.00 – 21.00

Off-peak 22.00 – 05.00

Therefore, a year is divided into 18 time slices in the model. These time slices reflect a fraction of the year with specific demand and supply characteristics. Following equation explains the number of time slices.

(23)

18

Time slices defined in the model and year split calculated from the share of the time of a year is given in table below.

Table 3 Year split for time slices

Season Day Type Day Split Year Split

Dry Weekday Day 0.1276

Dry Weekday Peak 0.0295

Dry Weekday Off-peak 0.0785

Dry Weekend day Day 0.0505

Dry Weekend day Peak 0.0116

Dry Weekend day Off-peak 0.0311

Intermediate/ High-wind Weekday Day 0.1261

Intermediate/ High-wind Weekday Peak 0.0291

Intermediate/ High-wind Weekday Off-peak 0.0776

Intermediate/ High-wind Weekend day Day 0.0505

Intermediate/ High-wind Weekend day Peak 0.0116

Intermediate/ High-wind Weekend day Off-peak 0.0311

Wet Weekday Day 0.1336

Wet Weekday Peak 0.0308

Wet Weekday Off-peak 0.0822

Wet Weekend day Day 0.0534

Wet Weekend day Peak 0.0123

Wet Weekend day Off-peak 0.0329

4.7 Demand Projection

Electricity demand of four sectors Residential, Industry, Services and Transport are projected for the period 2018 – 2055 based on the MAED results available from the LTGEP studies (Ceylon Electricity Board, 2018a). Figure 8 shows the projected annual final level electricity demand for the planning period.

(24)

19 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 D e m and (G W h) Year

Residential Industry Service Transport

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 De m and (G W h)

Residential Industry Service Transport Figure 8 Electricity Demand Projection

Figure 9 Sector-wise Electricity Demand Share

Residential and Industry sectors contribute to the largest share of 38% each for the final level electricity demand of 13,929 GWh in year 2018. Service sector accounts for 24% share while electricity demand for transport is only 0.2% with no other main consumption than for pipeline oil transport at present. Electricity demand projection results are given in Annex 2 and percentage share of its sector-wise composition is given in figure 9.

(25)

20 0 100 200 300 400 500 600 700 800 900 1000 D em an d (M W ) Time Slices

Figure 10 Hourly Electricity Demand

Figure 11 Demand Profile of the Residential Sector 0 200 400 600 800 1000 1200 1400 1600 1800 2000 D ema n d (M W)

4.7.1 Specified Demand Profile

Hourly electricity demand profiles collected from past operational data are used to calculate demand fractions. Averaged demand profile of the whole country is used to represent residential sector and services sector since industrial demand has more constant demand profile compared to other two sectors. Hourly electricity demand profile of the country is shown in figure 10.

Calculated demand profile for residential sector is depicted in figure 11 and calculated demand profile tables are given in Annex 3.

(26)

21 0 100 200 300 400 500 600 700 D em an d (MW ) Time Slices

Figure 12 Demand Profile of the Industrial Sector

Industrial sector demand profile is collected from a large industrial consumer and demand profiles are calculated. Figure 12 shows the calculated demand profile.

Industrial sector electricity consumption profile is more constant in all time slices and hence justifies the assumption of using the country load profile to other two sectors. Constant electricity demand profile is used for transport sector since a justifiable demand profile could not be found and also its demand is minimal at present.

It is assumed that the demand profiles of Residential, Industry and Service sectors do not change during the planning horizon. There for the overall country demand profile is assumed to remain the same along the years in this study. However, recent analysis predict a higher growth in day demand and improvement of load factor (Ceylon Electricity Board, 2018a). In future work this demand profile change can be integrated in to the model.

4.8 Technology Parameters

All the processes in the model such as fuel import or production, electricity generation, transmission and distribution are represented by separate technologies. A technology is defined by a set of parameters. Every technology do not require all parameters to be defined. For example, wind and solar PV plants do not have a fuel cost and their operation & maintenance cost is defined as a fixed cost. Therefore variable cost is not applied to those plants.

Technical and cost parameters of operational power plants were collected from the electric utility data base. Candidate power plant parameters were taken from LTGEP published values (Ceylon Electricity Board, 2018a)

(27)

22

since they were based on feasibility studies done for Sri Lanka or the parameters were adjusted to south Asian region. Sources like IEA, World Bank and IRENA data bases were examined to verify the accuracy of these data (International Renewable Energy Agency (IRENA), 2017).

4.8.1 Implementation of Residual Capacities

Existing power plant data were collected and categorized by technology type and fuel. Capacities of these power plants were put in the base year to allow the operation of them from the beginning. Operational lifetimes were used to model the retirement of power plants at the end of their planned lifetime.

4.8.2 Availability Factor

Planned outages for maintenance activities and forced outages due to breakdowns can make the power plant unavailable for power generation. Availability factors were used to represent this loss of availability of a power plants. For example, Coal power plant availability factor in a year was set to 85% considering 45 days of average scheduled maintenance which accounts for 12% of the time and another 3% margin for forced outages.

4.8.3 Capacity Factor

Capacity output of a technology can be defined by the capacity factor and it can be set for each time slice. This was used to model seasonal variations of variable renewable energy which is explained in section 4.9. Capacity factor was also applied to represent net capacity output of a thermal power plant reducing its auxiliary consumption. Capacity factor of major hydropower plants were set to 70% in dry season to reflect the low capacity output in that season (Ceylon Electricity Board, 2018a).

4.8.4 Annual Maximum Capacity Investment

Most of the favourable sites for wind and solar PV are located away from the national grid and transmission network need to be develop to connect them. And also additional infrastructure need to be developed to strengthen the national grid to overcome stability and reliability issues when these variable plants. These developments require more time in addition to power plant construction. Considering these factors, 200MW upper limit was set as the maximum annual capacity development from wind and solar PV plants in specific locations. For instance, wind plants in northern region can be developed up to 200MW in a single year.

(28)

23

4.8.5 Reserve Margin

Reserve margin is an important parameter to ensure capacity adequacy of the electricity system. It indicates the reserve level of installed capacity maintained over the peak demand. In the model, a reserve margin of 30% was kept ensure that adequate capacity of power generation is available to cater the demand.

4.8.6 Fuel Price

Fuel price forecast published by the World Bank was used to estimate the price of LNG, crude oil and other fuel oils (World Bank, 2018). LNG imported price forecast for Japan is more relevant for Asia since the bulk consumption of Japan sets the prices in the region. Year 2030 price forecast of 10 US$/MMBtu was used as the CIF (cost insurance and freight) price of LNG. Long term average crude oil price forecast estimate a price of 70 US$/bbl for year 2030 was used for the study. Present ratio between crude oil price and other fuel oil used for power generation were assumed to remain constant and used to estimate other oil prices. World Bank fuel price forecast is given in Annex 4. Latest available coal import price for Sri Lanka was used as the coal fuel price which is 76 US$/ton.

Coal and LNG prices were compared to examine relative price variations in past years. When converted to the same energy unit, LNG price is more than two times high compared to coal price and this ratio varies between 2 to 6. Coal price used for the study when converted to the same energy unit is 3.04 US$/MMBtu and the LNG/coal ratio becomes 3.3 which is in the middle range when compared to past fuel price ratios. However, LNG prices have shown much fluctuations, rapid increase and decrease both in past ten years compared to coal prices.

4.9 Variable Renewable Generation

Renewable energy integration is an important objective in the generation expansion plan. Variability and non-dispatchable nature of renewable power plants such as wind and solar PV has to be modelled and it becomes more important when the share of these power plants increase. In OSeMOSYS, capacity factor of power generation technology can be altered from year to time slice, to reflect the variation of power generation. Therefore capacity factors were calculated for both wind and solar PV plants from collected hourly resource data which cover a complete year. Seasonal capacity factor was calculated for wind power plants, and solar PV had capacity factor only during the day. Figure 13 shows hourly wind speeds measured in Puttalam region (North western part of the country) and the calculated hourly energy outputs for a 20MW wind park from the System Advisor Model (SAM) of National Renewable Energy Laboratory (NREL) (Ceylon Electricity Board, 2018a).

(29)

24 0 5 10 15 20 25 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 W ind Speed (m /s ) H o ur ly Ener gy (MW h) Months

Hourly Energy (kWh) Hourly wind Speed (m/s)

Figure 13 Hourly wind speed and electricity generation profile in Puttalam

Capacity factors calculated for the above wind profile for the three seasons Dry, High wind and Wet are 20%, 55% and 25% respectively. In Mannar region (Northern part) which is the most promising wind site in the country, capacity factors are 19%, 71% and 37% in the three seasons. Solar PV power plants have 35% capacity factor during day time.

4.10 Storage

As explained in chapter 1, there are two main cascade hydro systems in the country, namely Mahaweli and Laxapana complex. They contribute to the major share of hydro generation. The storage capacity available in these two hydro cascades are represented in the model by two storages. Mahaweli complex have a storage capacity of around 2.5 PJ and Laxapana complex have a capacity of about 1.3 PJ storage.

Pumped storage power plant is also modelled in the system according to the available feasibility study data. A unit capacity of 200 MW is introduced in year 2025 and the capacity is increased up to 600 MW in two years. Electricity generation and consumption profile of PSPP is given in chapter 5.

4.11 Emissions

It is important to integrate emissions related with thermal power plants into the model to reflect the overall environmental impact of those plants. CO2 emission associated with thermal power plants were calculated from fuel characteristics and plant efficiencies and modelled linking to activity output of those technologies. For example, Coal power plant has CO2 emission of 0.94 kg/kWh which is equivalent to 0.26 million ton per PJ.

(30)

25 0 2000 4000 6000 8000 10000 12000 14000 16000 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 MW

Major Hydro Coal ORE Combined Cycle - LNG

Combined Cycle - NG Combined Cycle - Oil Reciprocating Engines (Old) Reciprocating Engines (New)

Gas Turbines Pumped Hydro

Figure 14 Reference Scenario Capacity Mix

0 10000 20000 30000 40000 50000 60000 70000 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 G W h

Major Hydro Coal ORE Combined Cycle - LNG

Combined Cycle - NG Combined Cycle - Oil Reciprocating Engines (Old) Reciprocating Engines (New)

Gas Turbines Pumped Hydro

Figure 15 Reference Scenario Energy Mix

5. Results

5.1 Reference Scenario

Reference scenario was developed by optimising the model without applying any constraint on CO2 emissions. This gives the lowest economic cost to the country for the planning horizon. Net present cost recorded was 16,485 million US$ and 784 million ton of CO2 emission was projected from the electricity sector for the time horizon 2018 to 2050. Figure 14 and 15 shows the capacity mix and energy mix of the reference scenario.

(31)

26 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 MW

Major Hydro Coal ORE Combined Cycle - LNG

Combined Cycle - NG Combined Cycle - Oil Reciprocating Engines (Old) Reciprocating Engines (New)

Gas Turbines Pumped Hydro

Figure 16 Base Scenario capacity mix

Installed capacity of new coal power plants increased after they are opened for selection in 2023. This implies that coal fired steam plants are the lowest cost option compared with other candidate technologies and fuel options like LNG, oil and natural gas. Capacity share of coal plants rise from 18% in year 2018 to 66% in year 2050. Energy mix illustrates that the coal electricity generation share of 35% in year 2018, exceed 90% in year 2050. Depending on this type of large energy share from one fuel option is not acceptable in the view point of energy security of the country. And also higher emissions of CO2 from coal plants compared to other technologies contribute largely to the energy sector emissions. This need to be limited according to the nationally determined contributions (NDCs) submitted by Sri Lanka.

5.2 Base Scenario

New case was developed by limiting overall CO2 emissions in the planning horizon by 20% from the emissions of reference scenario. Results are more realistic for the country and therefore the case was named as the Base Scenario. Net present cost of the base scenario was 17,226 million US$ and it was a 4% increase from reference scenario. CO2 emission was 627 million ton in the base scenario and it was a 158 million ton equivalent to 20% reduction compared to the reference scenario. Capacity mix of the base scenario is given in figure 16.

Total installed capacity in the base scenario is higher compared to the reference scenario in the long term since more renewable plants are added to the system which are having low capacity factors. For example, a 100MW wind power park in Mannar can contribute only 19MW capacity in dry season since its capacity factor is 19%.

(32)

27 0 10000 20000 30000 40000 50000 60000 70000 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 G W h

Major Hydro Coal ORE Combined Cycle - LNG

Combined Cycle - NG Combined Cycle - Oil Reciprocating Engines (Old) Reciprocating Engines (New)

Gas Turbines Pumped Hydro

Figure 17 Base Scenario energy mix

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 Ener gy Mi x (%)

Major Hydro Coal ORE Combined Cycle - LNG

Combined Cycle - NG Combined Cycle - Oil Reciprocating Engines (Old) Reciprocating Engines (New)

Gas Turbines Pumped Hydro

Figure 18 Percentage share of energy mix in Base Scenario

Likewise capacity factors in high wind and wet seasons are 71% and 37% respectively. Therefore, higher installed capacity of ORE like wind and solar need to be added to replace the firm generation capacity of coal plants. Energy mix of the base scenario is given in figures 17.

Figure 18 shows the percentage share contribution of different technologies and fuel options throughout the planning horizon.

(33)

28

Figure 19 CO2 Emission of Base Scenario

0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 Mi lli o n To n

Detailed results of energy mix is given in Annex 5. Hydro power plants contribute to over 20% energy share in initial years and then the percentage contribution declines when the electricity demand rises. Coal plants continue to supply the major share of electricity to the national grid with 35% share in 2018 and 52% share in 2025 while reducing again to 35% share in 2050. ORE contribution is 7% in 2018 and it increases to 22% in 2025 and maintain a 20% share until 2050. LNG fired combined cycle power plants supplies electricity in medium term horizon starting from 2020 and its share reduced when coal power plants are added to the system. Natural gas fired combined cycle plants picked in the model in year 2042 and its energy production is increasing steadily afterwards. It can be viewed as the local NG production requires a significant lead time for development and once developed, it can contribute to a significant share of electricity production.

5.3 Environmental Emissions

Figure 19 shows the annual emission of CO2 over the period for the base scenario. In year 2018, 6.1 million ton of CO2 is emitted and it is expected to grow up to 11.5 million ton in 2025 and to 31.8 million ton in 2050.

(34)

29 -300 -200 -100 0 100 200 300 400 500 600 700 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 C ap aci ty (MW )

Major Hydro Reciprocating Engines (New) Coal Combined Cycle - NG Combined Cycle - Oil Other RE

5.4 Scenario Analysis

Scenario analysis has two main purposes, to identify the sensitivity to main input parameter variations and to look in to different investment strategies with policy alternatives. Following scenarios were modelled and results were compared with the base scenario.

 Hydro condition variation

 Demand sensitivity

 Discount rate sensitivity

 High renewable energy integration

5.4.1 Variation of Hydro Condition

Average annual energy output from major hydro power plants of the country is around 4,100GWh. Dry weather condition can lead to low energy yield from hydro power plants. This can adversely effect on the electricity supply of the country. This scenario was done to examine the effect of dry hydro condition on the electricity system.

The dry hydro condition was modelled to have 20% less energy output from major hydro plants in a year compared to the base scenario. Availability factor of major hydro plants (70%) was reduced to 56% to reflect the energy output and capacity factor in the dry season was reduced to 50% to reflect the capacity output of hydro power plants. Resulting annual energy reduction from major hydro plants was around 1,000GWh. The main difference of capacity addition compared to the base scenario was the addition of new reciprocating engines of capacity 180MW in year 2018. Electricity generation from these reciprocating engines was around 1,200GWh in initial years with an average capacity factor of around 76%. Figure 20 depicts the capacity addition comparison between dry hydro condition scenario and base scenario. Net present cost of dry hydro condition is 18,727 million US$.

(35)

30 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 D em an d (G W h) Year Figure 21 Base and High Demand Projections

Hydro power plant additions in the base scenario in 2019 are delayed in this scenario and reciprocating engines are added in the base year mainly to meet the capacity demand. Additional capacity of about 300 MW of ORE are added in years 2022 and 2023 and around 100 MW of coal plants are added in year 2028. Thermal power plant capacity added in dry hydro condition scenario exceeds delayed hydro power plant capacity.

5.4.2 Sensitivity to Electricity Demand

Uncertainty of electricity demand projection was considered in this scenario. Electricity demand growth can vary from the projected values and a separate scenario was conducted to analyse the future when a higher demand growth occur. Residential, services and transport sectors were assumed to have a 1% higher growth rate while industry sector has a 0.5% higher growth rate compared to the base demand projection. Figure 21 shows the comparison of base and high demand projections.

Annual electricity demand growth rate for the ten year period from 2018 to 2028 is 6.4% and for the planning period of 2018 to 2050, the annual growth rate is 5.5%. This is about 0.8% increase compared to the base demand projection and it requires to install more capacity in the system. The net present cost of the high demand scenario is 20,578 million US$ in the planning period.

(36)

31

5.4.3 Sensitivity to Discount Rate

Discount rate is used to analyse economic cost and benefits occur in different time periods. It reflects the time value of money and other factors like purchasing power and utility of money. Base discount rate used for the study was 10%, which was taken from national planning studies in the country. Two scenarios with low and high discount rates, 5% and 15% respectively were conducted to analyse the sensitivity.

Low discount rate scenario results show that plants with high capital cost and low operational cost are selected in the final capacity mix. For instance, 22% energy share of ORE in the base scenario in year 2025 is increased to 29% share in the low discount scenario. This indicates that low discount rate is favourable for renewable power plants which have high capital cost and low operation cost.

On the other hand, high discount rate scenario results indicate the opposite of this and only have an energy share of 18% from ORE in year 2025. Gas turbines which require low capital investment and have relatively high operation cost are preferred with the high discount rate.

Since base scenario and reference scenario are two investment strategies, discount rate sensitivity was done for both cases to investigate the robustness of those cases and the results are given in the table below.

Table 4 Net present cost with discount rate variation (million US$) Low Discount Rate

5%

Base Discount Rate 10%

High Discount Rate 15%

Base Scenario 32,832 17,226 11,131

Reference Scenario 29,906 16,485 10,899

5.4.4 High Renewable Energy Integration Scenario

Increasing renewable energy contribution in the energy mix is an important requirement for sustainable development of the country. Renewable energy share at present is around 30% from total generated electricity. Average annual growth rate of electricity generation from renewables in the base scenario is around 4%, and it contributes to around 30% share for the total electricity throughout the planning period. High renewable energy integration scenario was done to investigate the model outputs when renewable energy share increases over 50% from total electricity demand. To achieve this target, ORE share which include wind, solar PV, Biomass

(37)

32

Figure 22 Energy mix comparison between (a) Base Scenario and (b) High RE Scenario 0 10 20 30 40 50 60 70 80 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2050 TW h 0 10 20 30 40 50 60 70 80 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2050 TW h

and mini-hydro was set to 40% and remaining energy share was generated by major hydro plants. Comparison of energy mix between the base scenario and high RE integration scenario is given in figure 22.

5.4.5 Summary of Scenario Analysis

Net present cost and CO2 emissions of scenarios are summarised in the table below. Table 5 Results of scenario analysis

Net Present Cost PV Cost Difference CO2 Emission

Million US$ Million US$ Million ton

Base Scenario 17,226 - 627

Reference Scenario 16,485 (741) 784

Hydro Condition Variation

Dry Hydro Condition 18,727 1,502 637

Demand Variation

High Demand 20,578 3,353 684

Discount Rate Variation

High Discount Rate 11,131 - 629

Low Discount Rate 32,832 - 610

(38)

33

6. Discussion

6.1 Levelized Cost of Electricity

Levelized cost of electricity (LCOE) was calculated to examine the overall economic performance. Related cost outputs, namely capital investment, fixed operation cost, variable operation cost and salvage value were used for the calculation. Annual costs were discounted to the base year 2018 using the discount rate of 10% and yearly electricity generated was also discounted to the base year. The equation applied to calculate LCOE in yearly basis is given below (IAEA, 1984).

𝐿𝐶𝑂𝐸𝑡 =

∑𝐼𝑡+ 𝐹𝑂𝑀𝑡+ 𝑉𝑂𝑀𝑡− 𝑆𝑉𝑡 (1 + 𝑟)(𝑡−2018)

∑ 𝐸𝑡

(1 + 𝑟)(𝑡−2018) t = year, takes values from 2018 to 2050

r = discount rate (10%)

It = Capital investment in year ‘t’

FOMt = Fixed operation and maintenance cost in year ‘t’

VOMt = Variable operation and maintenance cost including the fuel cost in year ‘t’ SVt = Salvage value in year ‘t’

Et = Energy Production in year ‘t’

Figure 23 shows the LCOE of scenarios including base and reference scenarios. The LCOE of base scenario is 7 UScts/kWh in 2018 and it follows a downward trend. Fluctuations can be seen in initial years where increase occur due to large capital investments in years 2019-2020, 2024-2025 and 2028, and the maximum LCOE of 9.73 UScts/kWh is recorded in year 2020. Most of the capacity additions in 2019 and 2020 are committed hydropower plants, and remaining capacity additions are from residential solar PV and wind plants. LNG import terminal cost added in 2020 drives the LCOE up in 2020 while high LCOE in 2025 and 2028 is due to coal plant additions in those years. However, initial investment on these power plants leads to a lower LCOE in long run since their operational costs are lower compared to other thermal power plants. ORE plant additions are also contributed to the high LCOE in initial years with majority share of capacity additions from wind and solar PV during the period 2018 – 2022 before major thermal power plants introduced in to the system considering their construction periods.

(39)

34 0.00 2.00 4.00 6.00 8.00 10.00 12.00 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 LC O E (US c ts /kW h)

Base Scenario Reference Scenario

High ORE Int. Scenario Dry Hydro Cond. Scenario High Demand Scenario

Figure 23 Scenario LCOE Comparison

LCOE of the two investment strategies, reference scenario and base scenario, deviate in latter part of the planning period after 2034, and higher LCOE of the base scenario is due to the constraint on total period CO2 emissions and hence the delay of low cost electricity generation with high level of CO2 emissions such as coal. Dry hydro condition scenario has a higher LCOE compared to other cases in initial years mainly due to its increased operation cost with low availability of hydroelectricity. High ORE integration scenario has the highest LCOE during the period in between 2034 to 2044 due to high capital cost of ORE plants. A capacity of 7.7 GW of new ORE plants are added to the system in this period. LCOE declined in the latter part of the planning period since the ORE plants have low operation cost compared to other thermal plants. Tabulated results of LCOE are given in Annex 6.

(40)

35

Figure 24 Installed Capacity Results of the Two Models 0 2000 4000 6000 8000 10000 12000 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 MW OSeMOSYS WASP-IV

6.2 Comparison with WASP IV Results

6.2.1 Capacity Additions

As explained in section 2.1, WASP-IV was used to develop the twenty year generation expansion plan, LTGEP 2018 by the national electric utility of Sri Lanka. To compare the study results, electricity demand projection inputs of the two models are compared at first and WASP-IV input demand is slightly higher compared to the OSeMOSYS Demand. The table below contains the electricity demands for milestone years and the percentage difference compared with OSeMOSYS demand.

Table 6 Input Electricity Demand comparison of the two models

Year OSeMOSYS Demand (GWh) WASP-IV Demand (GWh) Percentage Difference

2018 13,929 14,588 5%

2020 15,343 16,646 8%

2022 17,287 18,353 6%

2025 20,266 21,260 5%

2037 36,198 36,613 1%

Demand difference increase initially and record the highest deviation in year 2020 and gradually reduce along the time horizon. All cost data are inline in the two models. Output installed capacities of the two models are compared and figure 24 shows the total installed capacities.

(41)

36

During the initial years up to 2024, a significant difference in installed capacities can be observed while after that two models show a similar result during the years. The table below gives the percentage reduction of installed capacity in OSeMOSYS compared to the WASP-IV output.

Table 7 Percentage Reduction of Installed Capacity in OSeMOSYS compared to WASP-IV

Year Percentage Reduction of OSeMOSYS Installed Capacity

2018 6%

2020 20%

2022 17%

2025 6%

2037 -2%

Installed capacity in WASP-IV results are 6% to 20% higher than OSeMOSYS results in that period 2018-2025. Around 5% to 8% difference in annual electricity demand is present during this period which can contribute to this capacity deviation. Year to year comparison with input electricity demand and installed capacities in milestone years given in above tables show that major difference is in year 2020 and 2022 while the difference in other years vary only 1% to 3%. In year 2037 OSeMOSYS has a slightly higher installed capacity.

The averaging effect in demand profile modelling in OSeMOSYS can be a reason for this difference in years 2020 and 2022. During the demand profile modelling, peak period is taken as 3 hours from 19.00 to 21.00. However, actual peak demand is recorded in a specific time of the day. Resultant peak demand represented in the model becomes lower since the demand is averaged over 3 hours. Comparison between the actual peak demand and modelled peak demand in year 2016 shows a difference of 13% in this OSeMOSYS model. However, WASP-IV uses twelve monthly load duration curves to model the demand in a year and more accurately represent the peak demand since it represent the demand in every hour.

It is important to reflect the peak demand accurately in the model to ensure the capacity adequacy. The difference between actual peak and model peak can be minimized by taking a smaller time period to model the peak demand. However, this may require to increase the number of time slices in the OSeMOSYS model. The risk of inadequacy of supply capacity due to this averaging effect can be mitigated by maintaining a higher reserve margin in the model to ensure adequate capacity available to serve the demand.

(42)

37 23% 25% 17% 18% 17% 0% 25% 32% 24% 8% 11% 0% Major Hydro Coal

Combined Cycle (LNG/Oil) Other Thermal

Other RE Pumped Storage

Figure 25 Energy Mix Results for Year 2020 (a) WASP-IV and (b) OSeMOSYS

18% 44% 18% 0% 19% 1% 18% 52% 7% 0% 22% 1% Major Hydro Coal

Combined Cycle (LNG/Oil) Other Thermal

Other RE Pumped Storage

Figure 26 Energy Mix Results for Year 2025 (a) WASP-IV and (b) OSeMOSYS

6.2.2 Energy Mix

Energy mix results of WASP-IV and OSeMOSYS for the two years 2020 and 2025 are given in figure 25 and 26 respectively. Fuel options selected by both models are the same while energy shares deviate from one another.

(a) (b)

(43)

38 0 100 200 300 400 500 600 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2050 CO 2 em is si o n (g /kW h)

Figure 27 CO2 emission per unit of electricity generation in base scneario

Electricity generation from other thermal plants include the reciprocating engines which are retired from operation during the period 2020 to 2025. In year 2020, OSeMOSYS results show a low share of other RE compared to WASP-IV output while LNG/oil fuelled combined cycle generation is high. Electricity generation from coal plants are higher in OSeMOSYS for both years. Installed capacity of coal plants is almost equal in the two models around 825MW in year 2020 and 1,700MW in year 2025. Coal power plants in OSeMOSYS results operate at a higher plant factor around 78% compared to WASP-IV output, in which operate about 72% plant factor. This is due to the high forced outage rate given for existing coal power plants in WASP-IV limiting their availability. In OSeMOSYS 85% availability factor has been used for existing and candidate coal power plants.

Installed capacity of LNG/oil fuelled combined cycle plants is low in OSeMOSYS by around 250MW in year 2025 which leads to the low energy share. In contrast other RE capacity is high by around 200MW in OSeMOSYS in that year.

6.3 CO

2

Emission Results

Emission of CO2 per unit of electricity generation in base scenario is given in figure 27 which is calculated from OSeMOSYS results. It shows how the emissions are limited with the total period CO2 emission limit. CO2 emissions peaked to around 570 g/kWh in 2034 and reduced to a value around 460 g/kWh at the end of the planning period.

(44)

39

Figure 28 PV Cost and CO2 emission comparison 0 100 200 300 400 500 600 700 800 900 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 Reference Scenario

Base Scenario High RE Scenario

C O 2 em is si o n (m ill io n to n) P V c o st (m ill io n US D)

PV Cost Total CO2 Emissions

It is important to look into the cost impact when limiting emissions to evaluate the feasibility of these scenarios. Figure 28 depicts PV cost and CO2 emissions of reference, base and high RE scenarios.

In the base scenario, the PV cost is increased by 4% while CO2 emission is reduced by 20% compared to the reference scenario. High RE integration scenario which provides 50% energy share from renewable energy has a reduction of 27% CO2 emissions while the PV cost increases by 5% compared to the reference scenario. These insights are important when setting future policy directives of the country.

References

Related documents

The reasons behind this result of intraday reductions in price volatility are, according to Mauritzen (2010), due to the supply shift caused by the low marginal cost of

efficiency of electricity usage and therefore reduce the demand. It is also difficult to predict the future, from what has been, not looking at other changes in society.

To investigate if a relationship exists between price of electricity and the number of workers in the manufacturing industry, a Vector Autoregressive model test will be

Key-words: Energy investment, investment valuation, renewable energy production, electricity price modeling, long-term, combined heat and power, CHP, balancing power,

Considering this ambitious move in the Chinese electricity sector combined with the high potential of renewable energy resources in the country, assumption is

The data input to the model is the number of time steps for the optimization problem (n, the time period under study), the current water level in the reservoirs (Reservoirs 0 ,

The reduced regression model is a model where, in this case, the supply- and demand variables are used to approximate an equation which will later be used in the study to observe

The main formulas used were two: one to evaluate the value in a certain year based on the annual growth/degrowth coefficients (Eq. 1), and one to extrapolate the annual variation